University of Southern California Dissertations and Theses

The environmental and genetic determinants of cleft lip and palate in the global setting

(USC Thesis Other)

The environmental and genetic determinants of cleft lip and palate in the global setting

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content The Environmental and Genetic Determinants of Cleft
Lip and Palate in the Global Setting
By
Allyn Auslander
A Dissertation Presented to the
FACULTY OF THE USC KECK SCHOOL OF MEDICINE
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(EPIDEMIOLOGY )
August 2021
ii
Acknowledgments

I would first like to acknowledge the scientific contributions and mentorship of my
committee members including: David Conti, PhD for encouraging me to embark on this journey
and guidance on biostatistics and genetic methodology; Lourdes Baezconde-Garbanati, PhD for
insight into health behavior; Jessica Barrington- Trimis, PhD for support in gene- environment
interaction studies; Roberta Mckean- Cowdin, PhD for her mentorship in epidemiological
methods and career development as well as endless hours editing and discussing study design;
and William Magee III, MD, DDS for his guidance into global surgery, nonprofits, and both my
personal and professional growth. Many others were critical throughout this process including
Bruce Burkemper, PhD; Andre Everson Kim, PhD; Rita Burke, PhD; Zarko Manojlovic, PhD;
Pedro A Sanchez-Lara, MD; Stephanie Ly, PhD; and Rijuta Kapoor, MS. I would also like to
thank Children’s Hospital Los Angeles and the USC Graduate School for funding my work.
Thank you to Mary Trujillo, Sherri Fagan, and Renee Stanley Rapanot for assisting with
administrative requirements.
Secondarily, I would like to thank my Operation Smile family who have been my constant
inspiration and encouragement. Specifically, this would not have been possible without the core
members of the International Family Study team including Kathy Magee, Melissa DiBona, Lili
Arakaki, Frederick Brindopke, and Devin Feigelson. In addition, I would like to thank the Global
Surgery Fellows who have supported me as sounding boards and medical advisors. This
research would not be successful without our in-country teams, volunteer data collectors, and
patients and their families. I would also like to thank the variety of funders who have supported
this work over the years including Operation Smile donors, The Moore Family, The Marguerite
Foundation, The Rallis Foundation, and The Sorenson Legacy Foundation, and many others.

Finally, I would like to thank my husband, friends and family for their endless support.
iii
TABLE OF CONTENTS
Acknowledgments ...................................................................................................................... ii
List of Tables ............................................................................................................................... v
List of Figures ........................................................................................................................... vii
Abbreviations ............................................................................................................................. ix
Abstract ....................................................................................................................................... x
Background ................................................................................................................................. 1
Global Surgery .......................................................................................................................... 1
Cleft Lip and Palate ................................................................................................................... 3
Environmental Risk Factors ...................................................................................................... 5
Genetic Risk Factors ................................................................................................................. 9
Gene x Environment Interaction ............................................................................................. 12
Understanding the Barriers to Surgical Care .......................................................................... 15
The Role of Smoke from Cooking Indoors Over an Open Flame and Parental Smoking on
the Risk of Cleft Lip and Palate: A Case- Control Study in 7 Low-Resource Countries. ... 17
Introduction ............................................................................................................................. 18
Methods .................................................................................................................................. 20
Results .................................................................................................................................... 25
Discussion ............................................................................................................................... 28
Tables and Figures ................................................................................................................. 32
Supplemental Material ............................................................................................................ 38
The International Family Study of Nonsyndromic Orofacial Clefts: Design and Methods 46
Introduction ............................................................................................................................. 47
Methods .................................................................................................................................. 49
Results .................................................................................................................................... 60
Discussion ............................................................................................................................... 61
Tables and Figures ................................................................................................................. 64
Supplemental Material ............................................................................................................ 69
A Genome-Wide Association Study and Genome-Wide Interaction Scan with exposure to
smoke from cooking and the risk of nonsyndromic orofacial cleft in a Vietnamese
population. ................................................................................................................................ 73
Introduction ............................................................................................................................. 75
Methods .................................................................................................................................. 78
Results .................................................................................................................................... 84
Discussion ............................................................................................................................... 86
Tables and Figures ................................................................................................................. 91
Supplemental Material ............................................................................................................ 95
Understanding the Patient- Centered Barriers to NGO Based Cleft Surgical Care through
the Integrated Health Behavior Model .................................................................................. 103
Introduction ........................................................................................................................... 104
Methods ................................................................................................................................ 106
Results .................................................................................................................................. 111
iv
Discussion ............................................................................................................................. 114
Tables and Figures ............................................................................................................... 119
Conclusion .............................................................................................................................. 128
References .............................................................................................................................. 131

v
List of Tables

Table 1. Child Characteristics of Case and Control from all Countries (N=4151) ...................... 34
Table 2. Parental Characteristics of Cases and Controls from all Countries (N=4151) ............. 34
Table 3. Lifestyle Factors of Cases and Controls from all Countries (N=4151) ......................... 35
Table 4. Adjusted Odd's Ratios (OR) of Smoke Related Factors and Cleft lip and/or palate in all
countries (N= 4151) .................................................................................................................... 36
Table 5. Adjusted Odd's Ratios (OR) of Smoke Related Factors and Cleft lip with or without cleft
palate (excluding iCP) in all countries (N= 3765) ........................................................................ 37
Table S1. All Sites Used for Case and Control Collection .......................................................... 38
Table S2. Case and Control Breakdown by Year for All Countries ............................................ 39
Table S3. Adjusted Odd's Ratios (OR) of Smoke Related Factors and Isolated Cleft Palate in all
countries (N= 2320) .................................................................................................................... 41
Table 6. Descriptive Characteristics of the Study by Case/ Control Status through December
2017 (N= 5729) ........................................................................................................................... 68
Table S4. Partners by Country and Site Type ............................................................................ 69
Table S5. Descriptive Characteristics of the Study through December 2017 by Case/ Control
Status and Country (N= 5729) .................................................................................................... 72
Table 7. Demographic Characteristics of the Cases and Controls Used in the GWAS and GWIS
Analyses (N= 1304) .................................................................................................................... 91
Table 8. Top 10 SNPs looking at the GxCookSmoke Interaction from the 1- Step Methods (N=
1067) ........................................................................................................................................... 92
Table 9. Characteristics of the 10 most significant SNPs identified in the non-Imputed GWAS
analysis as associated with NSCLP adjusted for PCs 1-5 and sex. ........................................... 93
Table 10. Characteristics of the 10 most significant SNPs identified in the Imputed GWAS
analysis as associated with cleft (N= 1067) ................................................................................ 94
vi
Table 11. Descriptive Characteristics of the patients and their families (N=901) ..................... 119
Table 12. Health Care Accessibility for patients and their families (N=901) ............................. 121
Table 13. Fit Indexes for Integrated Health Behavior Model Latent Constructs ....................... 122
Table 14. Attitude, perceived norm, and personal agency variables by timeliness of care (N=
292) ........................................................................................................................................... 123
Table 15. Knowledge and skills needed to perform the behavior and environmental constraints
variables by timeliness of care (N= 292) ................................................................................... 125

vii
List of Figures

Figure 1. Cooking Indoors Over an Open Flame- Odds Ratio and 95% CI Excluding Each
Country (All Cleft Types Combined) ........................................................................................... 32
Figure 2. Cooking Indoors Over an Open Flame- Odds Ratio and 95% CI By Country (All Cleft
Types Combined) ....................................................................................................................... 33
Figure S1. Cooking Indoors Over an Open Flame- Odds Ratio and 95% CI Excluding each
Country (CL+/-P ONLY) .............................................................................................................. 42
Figure S2. Cooking Indoors Over an Open Flame- Odds Ratio and 95% CI by Country (CL+/-P
ONLY) ......................................................................................................................................... 43
Figure S3. Cooking Indoors Over an Open Flame- Odds Ratio and 95% CI Excluding each
Country (iCP ONLY) ................................................................................................................... 44
Figure S4. Cooking Indoors Over an Open Flame- Odds Ratio and 95% CI by Country (iCP
ONLY) ......................................................................................................................................... 45
Figure 3. Key Definitions ............................................................................................................ 64
Figure 4. IFS Environmental and Genetic Study Framework ..................................................... 65
Figure 5. IFS Global Structural Organization Chart ................................................................... 66
Figure 6. OS Mission Flow ......................................................................................................... 67
Figure 7. Manhattan plot of 3df test assessing GxCookSmoke Interaction and the risk of CLP
(N=1067) ..................................................................................................................................... 92
Figure 8. Manhattan Plot of non-Imputed SNPs adjusted for PCs 1-5 and sex (n= 1,288,712) 93
Figure 9. Manhattan Plot of Imputed GWAS Results adjusted for PCs 1-5 and sex (N= 1067) 94
Figure S5. SNP and Individual Exclusion Criteria ...................................................................... 95
Figure S6. Visual analysis of Top 5 Principal Components to Understand Population
Substructure in final sample data (N= 1304) .............................................................................. 96
viii
Figure S7. Manhattan plot of 2df test assessing GxCook-Smoke Interaction and the risk of CLP
(N=1067) ..................................................................................................................................... 97
Figure S8. Manhattan plot of Case ONLY test assessing GxCook-Smoke Interaction and the
risk of CLP (N=1067) .................................................................................................................. 98
Figure S9. Manhattan plot of Traditional GxE test assessing GxCook-Smoke Interaction and the
risk of CLP (N=1067) .................................................................................................................. 99
Figure S10. G|E 2- Step procedure results (N=1067, Bins 1-10) ............................................. 100
Figure S11. D|G 2- Step procedure results (N=1067, Bins 1-10) ............................................ 101
Figure S12. EDGE 2- Step procedure results (N=1067, Bins 1-10) ......................................... 102
Figure 10. Integrated Health Behavior Model for Barriers to Cleft Surgery ............................. 108
Figure 11. Intercorrelation between Latent Factors ................................................................. 127

ix
Abbreviations

Term Abbreviation
Nonsyndromic orofacial cleft NSOFC
Low- and middle- income countries LMICs
Cleft lip and palate CLP
Isolated Cleft lip iCL
Isolated cleft palate iCP
Cleft lip with or without Palate (CLP or iCL) CL+/-P
Operation Smile OS
International Family Study IFS

x
Abstract

In 2015, the Lancet commission published a report quantifying what Paul Farmer
deemed the “neglected stepchild of global public health”—surgery. Although developed health
systems have clearly established the necessity of surgery for a wide variety of health issues,
disease maintenance, and patient quality of life, this subject has been essentially ignored in the
otherwise passionate and multifaceted global health community. With an approximate 5 billion
individuals lacking access to safe, affordable surgical care, simply providing care for all those in
need is a distant and unlikely solution. Cleft lip and/ or palate is one of the most common birth
defects (approximately 1 in 700 live births globally) and one of the many diseases that is
inhibiting quality of life for patients who cannot receive surgical care. I propose to address this
problem by first looking at the genetic and environmental determinants of cleft in low- and
middle- income countries (LMICs), a birth defect that has traditionally only been addressed by
surgery, but preliminary evidence suggests could be targeted for prevention. Second, I will work
to understand the patient perspective through studying the barriers that patients face when
trying to receive specialized surgical care in low-resource settings.
The initial hypothesis I am pursuing is that smoke inhalation from cooking in the home
over an open flame is a risk factor for cleft in combination with parental and household smoke
exposure, which have been previously established in the literature. Smoke exposure from
cooking occurs infrequently in developed countries but represents a high-proportion of smoke
exposure in less-developed regions thus this exposure has been minimally studied with respect
to cleft lip and palate development. The data is from a population-sampled case-control study of
children with cleft lip and/or palate and healthy newborns from Vietnam, Philippines, Honduras,
Nicaragua, Morocco, Congo, and Madagascar. Multivariable regression models were used to
assess associations between maternal cooking during pregnancy, parental smoking, and
household tobacco smoke with cleft. 2,137 cases and 2,014 controls recruited between 2012-
xi
2017 were included. While maternal smoking was uncommon (<1%), 58.3% case and 36.1%
control mothers cooked over an open fire inside. Children whose mothers reported cook smoke
exposure were 49% (CI:1.2–1.8) more likely to have a child with a cleft. This was consistent in
five of seven countries. No significant associations were found for any other smoke exposure.
Our finding of maternal cook smoke and cleft in low-resource countries, similar to maternal
tobacco smoke in high-resource countries, may reflect a common etiology. This relationship
was present across geographically diverse countries with variable socioeconomic statuses and
access to care. Exposures specific to low-resource settings must be considered to develop
public health strategies that address the populations at increased risk of living with cleft and
inform the mechanisms leading to cleft development.
To further understand this finding, I evaluated the hypothesis that the main effect that the
first paper explored may miss susceptible genetic subpopulations of the cook-smoke and cleft
association in Vietnam (where our findings showed the strongest effect). To test this, a genome-
wide interaction scan (GWIS) using traditional GxE analyses, two-step methods, and statistically
efficient one- step tests and genome-wide association study (GWAS) were conducted.
Exposure was defined as exposure to smoke from cooking indoors over an open flame. 589
cases and 715 controls were available for analysis post QC measures (N= 1304). No SNPs
reached genome-wide significance from the GWAS or GWIS analyses. No variants reached
genome-wide significance, however 71 were of interest and reached a statistical threshold for
exploration of p < 2E-6. Of the top 10 (excluding those grouped in the same gene/ region), 2
were identified from 3df test, 5 from the 2df, 1 from case- only and 2 from the traditional GxE
test. No interesting hits were identified from the 2-step methods. Our top GWIS hits,
rs1459270985 and rs1243867285, were identified by the unified model (3df test) and reside in
the IQ motif containing H (IQCH), which is involved in testis development and spermatogenesis
however a theoretical link to GxCookSmoke interaction and/ or NSOFC is unclear. One SNP
identified via the case-only analysis, rs1254772630, is flanked by GRM8- a protein coding gene
xii
that has been linked to maternal plasma folate levels during pregnancy, which also plays a role
in cleft development. The continued use of novel, statistically efficient methods to understand
these relationships will be essential in the study of rare diseases, such as NSOFC, especially
for individuals that reside in under-represented regions of the world and are at highest risk of
living with disease.
The first two hypotheses were explored using data from the International Family Study
(IFS). IFS is an ongoing case-control study with supplemental parental trio data designed to
examine genetic, environmental, lifestyle and sociodemographic risk factors for NSOFCs in
eight LMICs (through August 2020). Interview and biological samples are collected for each
family. The interview includes demographics, family history of cleft, diet and water sources,
maternal pregnancy history, and other lifestyle and environmental factors. Seven of eight
countries are currently summarized (2012-2017) for a total of 2,955 case and 2,774 control
families with 11,946 unique biological samples from Vietnam, Philippines, Honduras,
Madagascar, Morocco, DRC, and Nicaragua. IFS is the largest case set of NSOFCs with an
associated biobank in LMICs currently assembled. The biobank, family, and case-control study
now include samples from 8 LMICs where local health care infrastructure cannot address the
surgical burden of cleft or investigate causal mechanisms. IFS can be a source of information
and may collaborate with local public health institutions regarding education and interventions to
potentially prevent NSOFC.
Although prevention is undoubtedly the best long-term solution to decrease the burden
of disease associated with cleft, there are believed to be upwards of five- million individuals
currently living with cleft. The final portion of my dissertation will shed light on this issue using
data collected with Operation Smile in five different countries assessing the patients structural,
financial and cultural barriers to care. We conducted a cross-sectional study of children with
cleft in Vietnam, Honduras, Madagascar, Mexico and Nicaragua between 2014 and 2018. An
interview was conducted with the primary caregiver of each patient covering demographic,
xiii
clinical, socioeconomic, geographic, and treatment characteristics. The Integrated Health
Behavior Model (IBM) was used to conceptualize behavior. Descriptive statistics and
confirmatory factor analysis were used to assess relationships between constructs and
timeliness of care. A total of 901 patients and their families were surveyed from the five
countries. Five latent constructs were included in the final framework: (personal agency-
structure, personal agency- financial, perceived norm, environmental constraints, and
knowledge and skills to perform the behavior). Of the patients seeking care for the first-time cleft
phenotype, opinion of the family, perceived quality of available treatment, mother’s employment
type, and father’s employment were all significantly related to reaching timely care before and
after adjusting for country. This study maps the cultural, financial and structural barriers
experienced by patients and their families to model care-seeking behavior across five diverse
countries. As public health continues to increase investment into global surgery initiatives, it is
necessary to understand these relationships and how they may differ by country to create
effective and comprehensive programs that mitigate these barriers and enable patients to reach
timely care.

1
Background
Global Surgery
Global Surgery 2030
Care for infectious disease, in both low and high resource settings, has improved
drastically in recent history through vaccines and other large-scale global health initiatives.
Surgically treatable conditions account for approximately one-third of all global disease, but
remain a lower priority for public health research and prevention because under the best
circumstances they can be treated, despite the fact that many impacted children and adults will
not have adequate access to surgical care.
1
In 2015, the Lancet Commission put out “Global
Surgery 2030”, which was the first report that quantified the surgical need globally and the
findings were staggering.
The main takeaways from the report were:
• “5 billion people lack access to safe, affordable surgical and anesthesia care when
needed
• 143 million additional surgical procedures are needed each year to save lives and
prevent disability
• 33 million individuals face catastrophic health expenditure due to payment for surgery
and anesthesia each year
• 28-32% of the global burden of disease can be attributed to surgically treatable
conditions
• The projected GDP loss from five major categories of surgical conditions between 2015
and 2030 in low- and middle-income countries (LMICs) is $12.3 trillion”
In short, this report established surgery as an indispensable part of health care that is not only
contributing largely to the global burden of disease but is cost-effective for low-resource
countries. This report was pivotal for the now booming field of Global Surgery.
2
Operation Smile
Operation Smile (OS) is an international non-profit that has been providing cleft lip and
palate care since 1982. What began with a singular medical mission to the Philippines, has now
grown into one of the most well-known international NGO’s that is currently active in over 30
low-resource countries. As OS has changed over the years, they have gone from taking
international teams into these settings to deliver surgery to delivering care that is made up of
80% local volunteers and includes surgeons, anesthesiologists, nurses, pediatricians, speech
therapists, dentists, nutritionists and many others. They have local foundations in many of their
countries and certain countries, such as Panama, have gone from being a country receiving
Operation Smile services to one who does their own fundraising and serves other countries.
In the past 10 years, Operation Smile has taken their mission one step further and gone
from providing care to enhancing it. These initiatives have focused on formally educating local
providers to increase their capacity, providing surgical rotations for countries without plastic
surgery residency programs, global conferences to ensure the highest standard of care, and
using technology to improve outcome monitoring and patient expectations among many others.
Research has also taken flight at the organization during this time and has led to a partnership
between OS and USC that has allowed them to work together and push the organization even
further. This partnership has created a unique model where an NGO, university can leverage
local community empowerment to both reach populations that have not previously been studied
and do this work with the academic rigor of a leading US institution. It is innovative groups, such
as this one, that will be able to use their work in order to move the needle forward and push the
boundaries on the overwhelming task of improving surgery for those most in need. This
partnership has led to the International Family Study (IFS), which will be used throughout my
dissertation. It is a case-control study, that also uses parent-child trios, to study the risk factors
for cleft in 7,000 families) from a vast array of ethnicities in 8 countries.
3
Cleft Lip and Palate

Phenotypes and Development

Cleft is characterized as an embryologic failure of fusion of facial elements that normally
develop into the lip, palate, or both. Cleft lip and palate (CLP, Figure 1- C, D, and E) is the most
common phenotype, followed by isolated cleft lip (iCL, Figure 1- A), and isolated cleft palate
(iCP, Figure 1- B). CLP in our context will account for both bilateral CLP, where there is a failure
of fusion in both sides of the lip (Figure 1- E), and unilateral CLP, where it is only affecting one
side of the lip (Figure 1- D).

Figure 1. A Visual of the Different Cleft Phenotypes
2, 3

Cleft lip with or without palate (CL+/-P) is often used to denote the inclusion of CLP and
iCL. In the literature, cleft is often divided into two categories: iCP and CL+/-P, which has been
validated by the finding that they do not segregate in the same family.
4
Neural crest cells
migrate to the oral cavity as early as the 4
th
week of embryonic development to begin the
primitive formation of the lip and palate. The upper lip and primary palate will be formed by the
end of the 6
th
week and this peak of cell division occurring creates susceptibility to teratogenic
insults and growth disturbances leading to a failure of fusion in the region.
5
The formulation of
both the hard and soft palate continues (dividing the oral and nasal cavities) and is complete by
the 10
th
week of development.
4
Cleft Incidence

Cleft lip and/ or palate is the most common birth defect globally. It has a global incidence
of 1.7 per 1,000 live births, but this varies by ethnicity and phenotype.
2
There is a large
variance in the reporting of incidence by subtype with CL+/-P reports ranging from .34- 2.29 per
1,000 births and iCP ranging from .13- 2.53 per 1,000 births.
6
The incidence of CL+/-P is more
frequent in males, while incidence of iCP is observed more frequently in females. When this
failure of fusion occurs without any other birth defect or complication, the patient is considered
to be non-syndromic.
2
Approximately 70% of CL+/-P cases and 50% of iCP are non-
syndromic.
7-9

The International Perinatal Database of Typical Oral Clefts consists of data from 54
registries in 30 countries over a minimum of one year between 2000-2005. Of the 7.5 million
births, a prevalence of CL+/-P was 0.99 per 1000 births with Mexico, Japan, Western Europe
and Canada having a higher prevalence than the average (iCP was not commented on or
included in their findings).
10
Literature reports an incidence range of 1.28-1.69 clefts (all
phenotypes) per 1000 live births in Asians, 0.96-1.87 per 1000 in Hispanics, 1 per 1000 in non-
Hispanic Whites, and 0.57-1.22 per 1000 in Blacks or populations of African descent.
2, 11-14
The
following cleft prevalence rates have been reported in the countries IFS includes (all per 1000
births with total sample size noted): 2.01 in Vietnam (n=13,954 births)
15
, 0.48 in Madagascar (n=
150,973)
16
, 0.84 in the Congo (n=203,836)
17
and 1.94 in the Philippines (n=47,969)
18
. Country
specific prevalence could not be found for Honduras, Nicaragua, and Morocco.
A study conducted in the National Birth Defects Prevention Study between 1997- 2004
found a prevalence of CL+/-P in 0.8 per 1000 births for US non-Hispanic whites and Hispanics,
0.5 per 1000 in non-Hispanic Blacks and 0.9 per 1000 in the other category.
19
The prevalence
of iCP was 0.5 per 1000 births for non-Hispanic whites and the other category and 0.3 per 1000
for non-Hispanic blacks and Hispanics. In data from the same source, but from the years 2007-
2011, similar trends are observed, however Asian/ Pacific islander and Native American are
5
represented. The prevalence reported by this data source on individuals of Asian descent is 0.8
per 1000 for CL+/-P, which is lower than what is reported by Asian countries and comparable to
non-Hispanic whites.
19
For example registry data from Japan between 2000- 2004, a prevalence
of 2 per 1000 births of CL+/-P was observed and is one of the highest reported.
10
In a
population-based study of cleft rates in Taiwan between 2002- 2009, they found a rate of CL+/-
P of 1 in 1000 births and iCP in .4 per 1000 births
20
, which is more similar to the numbers found
in the US. There are notable differences in the methodology for calculating prevalence and the
possibility of differential quality of the data may be due to: hospital v. population-based
registries, types of cleft (iCP is not outwardly visible and may be harder to detect), cultural
beliefs surrounding cleft, and other common data quality errors.

Environmental Risk Factors
Smoke Related Risk Factors
Maternal Tobacco Smoke
An extensive amount of literature exists exploring the relationship between maternal
tobacco smoking during pregnancy with respect to cleft. In a meta-analysis by Xuan et al., of the
9 studies that reported on all cleft types combined and the risk and smoking: 4 studies found a
significant positive association while the other five saw positive, but insignificant effects.
21
Of
the 11 studies that looked at non-syndromic CL+/-P, which is most relevant to IFS, all found a
positive main effect of which 5 were statistically significant. Overall, maternal smoking was
associated with a 30% increase in CL+/-P (OR=1.3; 95% CI: 1.2, 1.4). The majority of the
studies in the meta-analysis (27 of 29 included for syndromic and non-syndromic clefts) were
conducted in populations of European decent (excluding 1 study in Brazil and 1 in China). A
study out of the 2005 Natality Data Set conducted a study looking at differences in the effect of
smoking on cleft by race/ ethnicity and found that smoking increased risk of cleft regardless of
race/ ethnicity (OR= 1.66 (1.3, 2.1)), but interestingly they did not report differences by race for
this exposure.
22
A study in China found a much stronger effect than those in mainly European
6
populations with a 3.3 times increased risk in iCL and a 3.12 times increased risk in CLP if
women reported smoking between 1-10 cigarettes per day during pregnancy.
23
Generally, the
effect of smoking has been observed to be weaker or insignificant for iCP although a meta-
analysis reported a 1.24(1.1, 1.4) times increased risk of iCP if the mother smoked during
pregnancy.
21, 23-25

Less information exists on the dose response observed from smoking during pregnancy
and the effects of smoking prior to pregnancy. In the same meta-analysis mentioned above, 11
studies looked at risk by dose. Only four of those could categorize smoking into three groups
(low, medium, high) with only one showing a significant, positive dose response.
21
The other
studies dichotomized into low and medium, with 4 showing a significant, positive risk increase
between low and medium smoking levels.
21
All studies had slightly different cutoffs for the
levels, making the meanings of the categories slightly unclear. In a study by Honein et al., an
increased risk of CLP for women who smoked pre-conception but not post (OR= 1.4 (0.9, 2.1))
was observed; however this did not extend to iCP.
24
In the same analysis, when they assessed
dose of smoking (never v. light (< .5 pack/ day) v. medium (1 pack/day) v. heavy (> 1 pack/ day)
for CL+/-P cases: smoking (highest level from the month prior to pregnancy through the first
trimester) was associated with a 1.4 (1.1-1.8) times increased risk if light, 1.1 (0.8-1.6) if
medium, and 1.6 (0.9, 3.0) compared to never smokers. Numbers in both the medium (n= 47)
and high (n=15) categories were low, which could account for the wide confidence intervals.
24

Paternal Tobacco Smoke
There have been less consistent findings with respect to paternal tobacco smoking. A
study completed in a subset of Operation Smile data in the DRC, Philippines, Honduras and
Vietnam (n= 430 cases, 754 controls), found paternal smoking to be associated with a 1.5 times
increased risk of having a child with a cleft (95% CI: 1.1, 1.9).
26
One study in China and another
in the Netherlands found a similar effect, but it did not account for whether or not the mother
7
smoked when reporting on paternal habits.
27, 28
However, multiple studies that looked at paternal
smoking, specifically among women who do not smoke, found no effect on the child’s risk of
cleft.
29, 30

Environmental Tobacco Smoke
Environmental tobacco smoke, which may but does not always include the father
smoking by definition, has been well studied and is often reported in combination with maternal
active smoking. A meta-analysis of all “maternal passive smoking” (14 studies total) found an
overall 2.11 times increased risk of all clefts when mothers reported passive smoke exposure.
31

Interestingly, this was stronger among iCP patients (OR=2.11) than CL+/-P patients (OR=2.05),
which is inconsistent with all other environmental cleft risk factors. In the sensitivity analyses,
even after removing the studies with the most extreme effects, they observed a 1.7 times
increased risk of cleft. The most conservative estimate, using adjusted models, reports 1.5
times increase in risk of all cleft with passive smoke exposure, which is higher than the 1.3
times risk increase generally agreed upon in active smoking literature.
21
The meta-analysis
offered multiple theories for why this may be: increased duration of passive smoke compared to
active smoke from occupational/ lifetime (childhood) exposure and the potential for under-
reporting of true active smoking habits due to stigma.
One specific study based in a large, international cohort found a much more moderately
increased risk of cleft among non-smoking mothers who reported exposure to environmental
tobacco (OR= 1.14 (1.0, 1.3))
32
and others using population based methods have found it to
have no effect.
24
A case-control study in China found second hand smoke in any setting
(including the workplace) to be associated with an increased risk of all clefts (OR=1.5 (1.2,2.2))
and that higher exposure levels showed an increase trend in risk.
33

8
Cooking Practices
Only two studies have explored exposure to smoke through cooking with respect to cleft
risk. One case-control study in China restricted to individuals who cooked with coal and found
no effect on cleft risk.
34
They did find that there was a significant increase in risk of cleft if the
house was not ventilated (OR=4.5, 95% CI: 1.6–12.9) and attributed this to heating with coal in
the home. It is possible that where smoke from heating may be the major form of smoke in
China, where temperatures are more extreme, cooking may be the main source in more
moderate climates. This theory is supported by a second study in the Congo, one of the
countries used in Operation Smile research, that found cooking with charcoal indoors to have a
large effect on cleft risk (OR=6.5 (1.2, 34.5)).
17
This was a relatively small case control study
(n=162 cases, 162 controls) and all models reported were unadjusted although controls were
matched by maternity ward of birth and neighborhood. Interestingly, both studies only looked at
charcoal although wood is also a common fuel that has been studied with respect to other
diseases.
Cooking practices in low-resource countries are dramatically different from those in the
more developed world. Globally, household smoke resulting from heating and cooking fires is
considered a leading environmental cause of morbidity and mortality, with a substantial impact
on women, who disproportionately spend more time in the home than other family members.
35-
37
According to the WHO, more than 3 billion people are currently cooking with wood or
charcoal and it is responsible for killing 4.3 million people every year, including 500,000 children
under the age of 5, more than HIV/ AIDS, tuberculosis and malaria combined.
36, 38
In India
alone, biomass cooking fuel is used by approximately 80% of rural households and has been
consistently associated with the above diseases.
39
These cooking habits have been linked to
lower respiratory infections, chronic respiratory disease, and pneumonia. It is also believed to
exacerbate both infectious diseases (malaria, TB, HIV/ AIDS) and chronic diseases (chronic
respiratory infections and cardiovascular disease) as well as adverse pregnancy outcomes.
40-42

9
Genetic Risk Factors

There are 200+ syndromes that CL+/- P is currently associated with and more than 400 that
iCP is considered a feature of.
43
This project will focus entirely on nonysndromic cleft, which
encompasses any type cleft that occurs with no other documented genetic syndrome. Literature
has consistently shown, through twin and family studies, that reoccurrence risk can be up to 30
times greater than the population prevalence, supporting that genetics play a strong role in
nonsyndromic clefting.
44-46

Candidate Gene and Association Studies
Initial focus, and the most well studied genetic region, in cleft genetics was in the IRF6
region. This gene was initially implicated to play a role in Van der Woude syndrome, which is a
genetic syndrome often mistaken for nonsyndromic cleft.
47
Van der Woude case’s only
distinguishable difference between nonsyndromic cleft cases is pits in the lower lip, which only
occurs in 85% of cases.
48
Zucchero et al. tested the role of specific SNP’s in IRF6 using a
cohort of nonsyndromic cleft cases from East Asia, South America and Europe as well as
diverse controls (N=8003 individuals, 1968 families).
49
There was a significant over
transmission of the V allele in cases of Vietnam and Filipino cases as well as the “All Asian” and
“All South American” groups, but this was not true for the European group. The authors
concluded that “IRF6 was responsible for 12 percent of the genetic contribution to cleft lip or
palate and tripled the risk of recurrence in families that had already had one affected child”.
Their findings have since been consistently supported in the literature.
50, 51
FOXE1 has also
been implicated in cleft risk using association studies, and has been further explored in the
GWAS era.
52, 53
There are a wide variety of family, candidate genes and association studies
indicating other genes in cleft risk that have been further explored in the GWAS era.
GWAS Findings
There are currently more than 39 risk loci associated with cleft development. IRF6 are
8q24 are two of the main genes that have been consistently associated with cleft. A recent
10
meta-analysis looked at 31 case-control studies to assess the relationship between cleft and
IRF6 (3 main SNP’s) or 8q24 (1 main SNP reported).
54
The analysis highlighted that although
the findings for both of these regions is consistent, there is a large level of heterogeneity within
the findings for these genes and CL+/-P. There was only sufficient data to report on 3 SNP’s
from IRF6 (15,11 and 9 papers found) and 1 SNP for 8q24 (12 papers found), although there
had been 9 SNP’s found in the literature for IRF6 and 6 for 8q24 at that time.
The SNP’s associated with IRF6 were found to have a combined effect of (all compared to GG
genotype):
• rs2235371: 0.51 (95% CI: 0.37- 0.61) for AA and 0.42 (95% CI: 0.32- 0.50) for GA
decrease risk of CL+/-P in Asians, but the result was not found for Caucasians
• rs2013162: 0.65 (95% CI: 0.52- 0.82) for AA decrease in risk for CL+/-P in Caucasians,
but not Asian populations
• rs642961: 2.47 (95% CI: 1.4- 4.4) for AA and 1.40 (95% CI: 1.1-1.8) times increased risk
in Asian populations; 2.03 (95% CI: 1.5-2.7) for AA and 1.58 (95% CI: 1.4- 1.8) times
increased risk in Caucasian populations
The SNP associated with 8q24, rs987525, was found to have a combined effect of (all
compared to CC genotype):
• Asian ethnicity: 2.27 (95% CI: 1.4- 3.6) for AA and 1.34 (95% CI: 1.0-1.8) for CA
increase in risk of CL+/-P
• Caucasian ethnicity: 5.25 (95% CI: 4.0- 6.9) for AA and 2.13 (95% CI: 1.8- 2.5) for CA
increase in risk of CL+/-P
• Mixed ethnicities (includes African and Central/ South American): 1.42 (95% CI: 1.1- 1.8)
for AA and 1.28 (95% CI: 1.1- 1.5) increase in risk of CL+/-P
The first cleft GWAS was done by Birnbaum et al. and found a strong relationship
between CL+/-P and 8q24.
55
Their findings did not extend to iCP and the associated allele was
11
strongest among those of African ancestry. This relationship, including the findings by
phenotype, has been well replicated by other studies in European, African and Hispanic
populations
54, 56-58
, however has not been found to be significant in those of Asian decent.
59, 60
A
recent study in an African cohort, found that although 8q24 is the most significant locus for
CL+/-P, the SNP they found to be significant is not the one that has been reported among
European’s.
61
The PAX7 region has also been confirmed to be associated with cleft
development. This has been shown in human studies of Europeans, Asians and those of African
descent
56, 61, 62
, as well as confirmed in multiple animal studies.
63-65

One paper has been published using GWAS data from our study cohort.
66
The study
included case-parent trios as well as controls from the DRC, the Philippines and Vietnam. The
study found SNP’s significantly associated with CL+/-P in the region of VAX1 (10q25.3; Vietnam
only), IRF6 (1q32.2; Vietnamese and Filipino), and NOG (17q22; Vietnamese and Filipino only).
The sample from the DRC did not show any significant hits in the known cleft regions in this
cohort. None of the populations showed signs of association with 8q24 in this cohort.
Outside of this, very little of the literature reflects populations that overlap with the IFS
study. A cohort of Filipinos that has been studied and found to have a rare missense mutation in
MAFB over-represented and it was confirmed with a mouse model, however this finding has not
been well-replicated.
67
IRF6, 8q24, FOXE1, GREM1, PAX7, 2p24.2, and 10q25 have been
implicated in Central and South American populations, but neither country in our study
(Nicaragua or Honduras) was included in any study.
53, 56, 58, 68, 69
One study has been published
exclusively on African populations and found VAX1, PAX7, 8q24, CTNNA2, and SULT2A1 to be
associated with CL+/-P risk.
61
The authors used a sample of the DRC data from our cohort to
replicate their findings, as well as mouse models to confirm the role of CTNNA2 and SULT2A1
in craniofacial development. Although many studies have been successful in identifying risk loci
associated with cleft, the diversity in the populations is still minimal and does not include many
of the individuals at highest risk of living with disease.
12
Gene x Environment Interaction

Due to the complexity of nonsyndromic cleft, gene environment interactions have been
discussed as a possible explanation for the lack of clarity from genetic and environmental
studies alone in understanding the full etiology of the disease. These studies have mainly
focused on maternal alcohol use (n= 14), maternal smoking (n= 21), maternal multivitamin
usage (n= 17), paternal smoking (n=5) and environmental tobacco smoke (n= 9).
Maternal Smoking
Maternal smoking is one of the best documented environmental risk factors for
nonsyndromic cleft with an approximate 30% increase in risk observed among smoking
mothers.
21
Multiple GxE studies have looked into the role it plays with the MTHFR, which has
been associated with neural tube defects such as spina bfida, but of the three studies (2 GWAS,
1 candidate gene study) that looked at this relationship- none found an interaction between the
two.
70-72
In a study of 550 iCP case-parent trios, interactions between maternal smoking and
multiple SNP’s in the TBK1 and ZNF236 genes were observed, however these have not been
replicated nor have they been implicated in cleft development previously.
73
The authors note
that iCP is rarely studied alone due to limited case volume, so this is compelling evidence that
these patients should continue to be studied separately. A cleft group from Iowa and Denmark
(n= 1244 cases, 4183 controls) found that children who carried the GSTT1- null (a detoxification
gene) genotype and had a mother who smoked (10-19 cigarettes per day for Danish women,
>15 cigarettes per day for Iowan women) had a 4.2 (95% CI: 1.4, 12.4) times increased risk of
cleft compared to controls for Danish children and a 17.1 times increased risk compared to
controls for Iowan children.
74
They calculated that in a population with a smoking prevalence of
25% and GSTT1-null prevalent in approximately 15% of the population, this interaction
attributes up to 6% of clefts. This finding has been replicated in a small study from the
Netherlands.
75
The TGFA (transforming growth factor) gene has also been implicated to interact
with maternal smoking in multiple candidate gene studies, however one used controls with other
13
birth defects
76
and the other only found the association among their small number if iCP
patients
77
. These studies, which took place in 1995 and 1999 respectively, have not been
replicated in recent literature.
Environmental Tobacco Smoke
The GxE interaction with environmental tobacco smoke (ETS) has been studied with
slightly more consistent findings, however it is largely based in Chinese populations with
relatively small sample sizes. A study of Asian case-parent trios found that no genes were
associated with iCP on their own, but when taken into account with ETS 15 SNP’s mapped to
SLC2A9 and 9 to WDR1.
78
Although this finding has not been replicated, it was found in a small
sample (n=259 trios) and highlights the importance of what can be missed without taking GxE
interaction into account. A study done in China, had similar findings with the RUNX2 gene in
that they found no association in the SNP’s alone, but found nine that showed a GxETS
interaction, two of which they were able to replicate in a European cohort.
79
Four other studies
in China found marginally signficant GxETS interactions in: the microRNA- 140 gene (iCP only,
n=169 cases, 306 controls)
80
, ZNF33 (n= 211 case-parent trios, 188 controls)
81
, BMP4 (n= 326
case- parent trios)
82
, and IRF6 (n= 77 cases, 284 controls)
83
.
Multivitamin Use and Maternal Alcohol Consumption
Although I am not focusing on it in my dissertation, the rest of the literature focuses
exclusively on maternal alcohol and multivitamin use. Multivitamin use prior to and during
pregnancy, more specifically folic acid intake, has been established as protective to cleft
development. It has recently been found to interact with the estrogen-related receptor gamma
(ESRRG) gene in European and Asian case-parent trios (although neither were significant after
stratification by race)
84
as well as IRF6
83
, the ABCB1 gene
85
, and the BAALC gene with iCP
only
73
. There was suggestive evidence found in one study of a GxMultivatimin interaction with
TGFA
86
and MTHFR
70
, but the findings were not conclusive and have not be replicated. Only
one study of 550 iCP trios found a notable interaction between alcohol and the MLLT3 as well
14
as the SMC2 gene
73
, where the other studies that assessed alcohol with any type of cleft found
no result.
84, 87

Known Cleft Genes
Interestingly, although 8q24 and IRF6 are both well documented as genes contributing
to cleft, there is only one paper assessing potential interactions with IRF6 and environmental
factors.
83
Wu et al. looked at the potential for interaction between 22 candidate SNP’s in the
IRF6 region and maternal multivitamin supplementation as well as environmental tobacco
smoke. Although the sample size was not large (326 case-parent trios), they found two SNP’s
with a marginally significant interaction with multivitamin usage and one with environmental
tobacco smoke.
Generally, the literature on gene environment interaction and cleft tend to have low
sample sizes and few findings have been replicated in subsequent work. However, the
understanding of these interactions will be critical in populations where treatment may be limited
and there may be an increased impact of promoting healthy behavior.
70

15
Understanding the Barriers to Surgical Care

Surgery is undoubtedly one of the complex parts of a health system due to the extensive
requirement of: both basic (consistent water and electricity) and surgical (operating rooms,
intensive care units, recovery rooms) infrastructure, highly- trained personnel, equipment, and
having a system in place that can support both the emergent and elective needs of a population.
This has led to the general neglect of surgery in the public health world up until recently. Now
that it has been brought into the light, understanding the structural, financial and cultural barriers
to receiving surgery is going to be imperative as money and resources are being invested into
surgical systems in low resource countries. As we are focused on cleft care, we will focus on
non-emergent surgical barriers.
The most straight-forward barrier, which has been consistently documented in the
literature is the lack of financial means to both pay for care and in many situations even to reach
care. A population-based study done in Nepal found that individuals who required motorized
transportation to reach a facility equipped to provide surgery were 66% (OR=0.44 (0.2, 0.9) less
likely to have accessed care than those who did not.
88
A study with similar aims in Malawi found
that 39% of males and 59% of females lacked the financial resources to travel the median
reported time of 1- 2.5 hours to a hospital that provided surgery.
89

A study done in collaboration with Operation Smile found in Vietnam studied potential
demographic characteristics that could predict timely cleft care (cleft care prior to 1.5 years of
age).
90
The study of the 453 patients included in the study the two main factors that predicted
timely care were fathers education and whether the child was male. This could suggest that the
fathers spearhead the search for care and their education enables more timely access. The
gender component suggests that families still may seek out care earlier for male children than
female. Interestingly, neither cost nor time to the nearest hospital was associated with timely
cleft care, but distance was. This could mean that those closer to any health care, even if the
16
facility cannot provide cleft care, are more likely to become aware of other care options (ie
NGO’s) in a timely manner. Time, distance or cost to the Operation Smile mission site was not
associated with timely care, but how the individual heard about the organization was- with those
hearing from family and friends significantly more likely to seek care earlier.
A second study of perceived patient barriers, using the same Operation Smile Vietnam
data, cited cost and mistrust of medical providers as the most reported barriers to care by
patients.
91
There was not a significant difference in hospital access or household income in the
group that had previously received surgery and 83% of those who had accessed previous
surgery was through charity. Understanding the patient perspective of surgical care and what
inhibits them from successfully reaching timely care will be critical to the provision of safe, timely
and effective surgical care for all.

17
The Role of Smoke from Cooking Indoors Over an Open Flame and
Parental Smoking on the Risk of Cleft Lip and Palate: A Case- Control
Study in 7 Low-Resource Countries.

Abstract:

Background: Cleft is one of the most common birth defects globally and the lack of access to
surgery means millions are living untreated. Smoke exposure from cooking occurs infrequently
in developed countries but represents a high-proportion of smoke exposure in less-developed
regions. We aimed to study if smoke exposure from cooking is associated with an increased risk
in cleft, while accounting for other smoke sources.
Methods: We conducted a population-sampled case-control study of children with cleft lip
and/or palate and healthy newborns from Vietnam, Philippines, Honduras, Nicaragua, Morocco,
Congo, and Madagascar. Multivariable regression models were used to assess associations
between maternal cooking during pregnancy, parental smoking, and household tobacco smoke
with cleft.
Results: 2,137 cases and 2,014 controls recruited between 2012-2017 were included. While
maternal smoking was uncommon (<1%), 58.3% case and 36.1% control mothers cooked over
an open fire inside. Children whose mothers reported cook smoke exposure were 49% (CI:1.2–
1.8) more likely to have a child with a cleft. This was consistent in five of seven countries. No
significant associations were found for any other smoke exposure.
Conclusions: Our finding of maternal cook smoke and cleft in low-resource countries, similar to
maternal tobacco smoke in high-resource countries, may reflect a common etiology. This
relationship was present across geographically diverse countries with variable socioeconomic
statuses and access to care. Exposures specific to low-resource settings must be considered to
develop public health strategies that address the populations at increased risk of living with cleft
and inform the mechanisms leading to cleft development.

18
Introduction

Surgically treatable conditions account for approximately one-third of all global disease
and an additional 2.2 million providers would be needed to treat the existing surgical need.
1

Although such conditions are treatable, many impacted children and adults do not have
adequate access to care. Prevention of surgically treatable birth defects is therefore a
necessary goal as provision of surgical treatment for all patients is unlikely - especially for
diseases such as orofacial clefts that require complex multidisciplinary care.
Cleft lip with or without palate is one of the most common birth defects worldwide. The
global incidence is approximately 1 in 700 live births
2
, however incidence ranges from 1.28-
1.90 per 1000 live births in Asians, 0.96-1.87 per 1000 in Hispanics, 1 per 1000 in non-Hispanic
Whites, and 0.73-1.22 per 1000 in populations of African descent.
2, 11-13
Although there is clear
variability by ethnicity, the majority of etiologic data comes from individuals of European
descent.
7
Disruptions in craniofacial development occur during the first trimester of pregnancy
during weeks 4 to 13 of development.
9
Cleft is characterized as an embryologic failure of fusion
of facial elements that normally develop into the lip and palate. In the absence of any other birth
defect, the patient is considered to be non-syndromic (approximately 70% of cleft lip with or
without palate and 50% of isolated cleft palate patients).
2, 7-9
While the origins of syndromic
clefts are considered largely genetic, the etiology of non-syndromic clefts remains unclear.
Parental smoking has been considered an important determinant of developmental
disorders, but the environmental impact of smoke exposure from cooking and cleft risk has only
been mentioned in two existing studies.
17, 34
In low-resource countries, biomass cooking fuel is
used by approximately 80% of rural households and has been associated with a wide variety of
diseases, including stillbirths.
39
While maternal smoking has been associated with risk of cleft
21, 92, 93
, the association between smoke from cooking remains uncharacterized. Beyond
different types of smoke exposure, the main established risk factors for cleft are low maternal
19
education
94, 95
, lack of folic acid supplementation
96, 97
, advanced maternal age
98, 99
, family
history of clefts
46
, and ethnicity
19
. Other factors that have been less consistently associated
with cleft are periconceptional alcohol use
100, 101
and diabetes (either pregestational or
gestational)
22, 102
.
In the current analysis, we used data from over 4,000 children and their mothers
collected on surgical missions conducted by Operation Smile. Specifically, we assessed the
relationship between smoke exposure from cooking and the risk of non-syndromic cleft. Other
sources of smoke investigated include maternal smoking, paternal smoking, and household
tobacco smoke. This study is the first to evaluate cooking practices as an environmental
determinant of cleft in a population-recruited sample of children from diverse, low-resource
countries. Data were collected from 7 countries (Vietnam, Philippines, Morocco, Madagascar,
Democratic Republic of Congo (DRC), Honduras, Nicaragua) to evaluate the association overall
and to explore factors that may influence heterogeneity of effects by country. Clarifying the role
of prenatal exposure to smoke from sources common to different populations and cleft risk may
help to improve our understanding of risk factors contributing to non-syndromic cleft and inform
preventive strategies.

20
Methods

Data for this study was collected from 2012-2017 as part of a coordinated series of
population-sampled case-control studies focusing on genetic, lifestyle and environmental
exposures and cleft in children 6 months to 4 years of age. This study was conducted with
Operation Smile (OS), an internationally recognized not-for-profit that has been providing free
cleft surgery and related care to patients for over 36 years. Data for the current analysis
represents children from 7 countries sampled over multiple missions (Table S1). Participation
rates in the study varied by site from 77%-96% for cases and 45%-100% for controls. The
methods of this study have been previously published in depth.
66, 102
All work was approved by
the Institutional Review Board at the University of Southern California including country-specific
authorizations.

Case Definition
This study includes non-syndromic cases of cleft lip and / or cleft palate (ICD10 35-
37).
103
Cleft lip and palate (CLP) is the most common phenotype, followed by isolated cleft lip
(iCL) and isolated cleft palate (iCP). Cleft lip with or without palate (CL+/-P) is used to denote
CLP and iCL. Cases were screened to confirm diagnosis and absence of any genetic syndrome
or other birth defect by medical practitioners at the mission site. This included pediatricians,
nurses, anesthesiologists and surgeons who are all formally licensed, trained and OS certified
to work with cleft patients.
Patients were included in the study if they were accompanied by their biological mother
(18 years or older), 6 months to 4 years of age, and presented for cleft treatment at the time of
the OS mission. Patients were excluded if the child was not the most recent pregnancy, a
multiple birth, had a genetic syndrome, or had another co-morbid condition.

21
Case Recruitment
Cases for the International Family Study (IFS) are recruited on site during OS missions.
IFS countries were selected from sites OS identified a priori as having adequate infrastructure to
support research and the specific hospital was chosen based on its ability to meet the
organization’s ‘Global Standards of Care’. Extensive regional recruitment and community
outreach efforts are conducted by OS prior to each mission to assure saturation of the
communities. All patients arrive to the mission site to be screened for care over the span of one
or two days with all costs covered by OS. The patients are registered and seen by general
practitioners, nurses, anesthesiologists, surgeons, and dentists to assess surgical eligibility.
Case recruitment for the study occurs at the end of the screening process. Study eligibility
criteria were identical for cases in all countries.

Control Definition
Controls were newborns identified from regional neighborhood, clinic, and hospital-
based birth centers around the mission site (Table S1) whose mother agreed to complete
informed consent and the study interview. Individuals were excluded if they had a cleft or any
other birth defect, were a multiple birth, or if the mother was younger than 18.

Control Recruitment
Multiple neighborhood, clinic, and hospital-based birth centers were identified prior to
each mission by in-country OS partners to represent the catchment area of the OS mission and
improve case-control comparability. All maternity wards selected were public to better match
demographics of the mission patients. The leadership at the birth center was approached and
debriefed on the study, and local authorization was obtained to recruit families along with IRB
approval prior to the mission. Each site was visited daily during the mission.

22
Data Collection
Local volunteers with medical training (i.e. nursing/ medical students) were identified by
OS and underwent training by study members. Local interviewers were used to assure high
recruitment and allow completion of the interview in the language of the families; however, the
study supervisor was present during all interviews for consistency and to maintain quality.
Informed consent was completed before the interview and parents were assured that
participation was not required for their child to receive care. Families were interviewed in a
private to semi-private area (depending on screening space). Questionnaires have been
translated and back translated by certified translators to ensure consistency across countries.
Mother’s interviews took approximately 40-minutes. The interview included questions on
family history of cleft, lifestyle and environmental exposures (smoke, alcohol, diet, water
source), medical history (parental medical history, use of prescription and nonprescription
drugs), demographics (age, pregnancy history, education, employment), and paternal factors
(smoke, alcohol, age, employment, education). When the father was present, a limited interview
is independently completed on medical history, environmental, and lifestyle exposures. Our
current analysis included data exclusively from the mothers’ interviews.

Statistical analysis
Descriptive statistics, including proportions for categorical variables and means for
continuous variables, were constructed for the child characteristics, parental characteristics, and
lifestyle factors. Tests of statistical significance included t-tests for continuous and chi-squared
tests for categorical variables. Maternal exposure to indoor cook smoke was categorized as a
dichotomous variable (yes/no). Maternal smoking was dichotomized (ever/never) for the three
months prior to pregnancy and during pregnancy. Smoking pre-pregnancy was not collected in
Vietnam, so they are not included in those analyses. Fathers were dichotomized into lifetime
smokers (ever) or never smokers and the definition included any type of tobacco product
23
(cigarettes, pipes, chewing tobacco, cigars, other). Household tobacco smoke is defined as any
member of the household smoking inside during the mother’s pregnancy with the child (yes/ no).
All education variables were harmonized as less than secondary school or secondary school or
higher for mothers and fathers separately. Family history of cleft was defined as any first or
second degree relative having any cleft. Number of children was classified into 1, 2, or 3 or
more.
Logistic regression was used to calculate odds ratios and 95% confidence intervals for
smoke exposure (indoor cook smoke, maternal smoking, paternal or household tobacco smoke)
and cleft overall and by subtype (iCP, CL+/-P). Models were specified as minimal (adjusted for
country, maternal age at the child’s birth, mother’s education, father’s education, and family
history of cleft), full (adjusted for minimal model and rural/urban residence and maternal alcohol
consumption during pregnancy), and a mutually adjusted model including all previous covariates
and mutual adjustment for all smoke variables. Additional adjustment for demographics of the
child and parent (e.g. child’s sex, maternal employment, paternal employment, paternal age at
child’s birth), environmental and lifestyle factors (water source, folic acid use, and prescription/
nonprescription drug use) were considered as potential confounders but were not included in
the final model as the measures of association did not meaningfully change (difference in effect
< 10%). Heterogeneity of effects by country was investigated by including interaction terms for
exposures of interest and by stratification. Missing values were handled by exclusion as they
were generally low (< 10%).
Secondary analyses were conducted to assess if the findings differed by cleft subtype
(iCP and CL+/-P) as they are often considered to have different etiologies. Heterogeneity of the
indoor cook smoke finding by country was evaluated using stratified analysis and by excluding
country data one at a time. To evaluate the independent effect of cook smoke on cleft, without
confounding by maternal smoking, we repeated the analysis restricted to never smoking
mothers. Additional sensitivity analyses were done by maternal education, paternal education,
24
income level, and age of cases (limited to less than one year). Income was only available for
Vietnam, the Philippines and Morocco due to cultural sensitivity and medical mission
considerations. When available, income quartile groups were defined by country based off of
the income level reported by controls. All analyses were completed using SAS 9.4 and the R
statistical language.

25
Results

A total of 4426 eligible children were identified from the 7 countries between 2012 and
2017. Of these: 58 participants were excluded due to missing case status and an additional 217
were excluded because they exceeded the newborn to four-year inclusion criteria. 4151
participants were included in the final dataset: 2137 cases (51%) and 2014 controls (49%) with
the majority coming from Vietnam (31.8%), followed by the Philippines (22.3%), Honduras
(22.1%), Congo (10.2%), Madagascar (5.1%), Morocco (4.3%), and Nicaragua (4.2%). Cases
and controls were recruited simultaneously in all years the study was active with the exception
of a delay in control collection in Vietnam in 2012 due to approval delays (Table S2). The case
phenotype distribution consisted of 1198 (56.1%) with CLP, 553 (25.9%) with iCL, and 306
(14.3%) with iCP (Table 1).
Characteristics of the study population are described in Tables 1-3. Case mothers were
on average six months older than controls (P=0.008) and less often employed (P =0.03).
Control mothers (81.1% vs. 66.4%) and fathers (78.2% vs. 64.2%; both P < 0.001) were more
likely to have a secondary education. No difference was observed in father’s age, father’s
employment status, or maternal smoking (prior to or during pregnancy). A higher proportion of
cases reported cooking indoors over an open flame (58.3% vs. 36.1%). Fewer case mothers
reported drinking alcohol pre-pregnancy (8.5% vs. 12.4%; P < 0.001); however, they were more
likely to report drinking during pregnancy (P =0.09), living in a rural area, smoking in the
household, and that the father of the child smoked (all P < 0.001). The distribution of all five
smoking variables significantly differed across countries (all P < 0.05).
The relationship between smoke exposures and the odds of all cleft is shown in Table 4.
A strong positive association was found between cooking indoors over an open flame and risk
of all cleft types. Mothers who reported cooking over an open flame indoors were 49% more
likely to have a child with a cleft after adjusting for country, maternal age, mother and father
26
education, family history of cleft, rural/ urban residence, alcohol consumption during pregnancy
and all smoke variables [Model 3]. With respect to maternal smoking during pregnancy, the OR
was elevated but the confidence interval included the null (OR=1.65 (0.5, 5.6)). The prevalence
of mothers who smoked prior to or during pregnancy was low (prior: 39 cases (1.8%) and 39
controls (1.9%); during: 15 cases (0.7%) and 18 controls (0.9%)). No association was found with
smoking the three months prior to pregnancy or exposure to household tobacco smoke. There
was a positive association with ever paternal smoking and cleft (OR=1.18 (0.96,1.5)) however
the confidence interval included the null. The results of Table 4 were nearly identical when
restricted to women with no history of smoking (n= 4057). No evidence of interaction by country
was found (all P > 0.05).
The analyses restricting to CL+/-P (cases=1751; excluding iCP) are explored in Table 5.
Cooking indoors over an open flame was associated with a modest increase in risk of cleft
(OR=1.55 (1.3, 1.9)) compared to the full case set. For iCP (cases= 306), only cooking indoors
over an open flame in the minimally adjusted model showed elevated risk (OR=1.65 (1.2, 2.3))
(Table S3). No other smoking variables were associated with iCP in any model.
The impact of rural/ urban residence, parental education, and case age on risk of all cleft
types was explored in stratified analyses. Cooking indoors over an open flame was associated
with cleft in both rural and urban residence. However, the magnitude was higher for urban
(OR=1.71 (1.3,2.3)) than rural (OR=1.28 (1.0, 1.7)) (data not shown). Sensitivity analyses
showed that the results were as strong by education group for both parents (less than
secondary vs. secondary or higher) using stratification, by paternal smoking status (yes v. no)
using stratification, after adjusting for income, and when restricting to cases under one year of
age (group most comparable to newborn controls).
Effect estimates after exclusion of individual countries are shown in Figure 1. The overall
effect was not dramatically influenced by data from any single country with the exception of
Vietnam. The association was slightly reduced (OR=1.25(1.0,1.6)) with the removal of Vietnam,
27
which contributed 31.8% of the dataset. The individual association of cooking indoors over an
open flame by country is shown in Figure 2. Positive associations were found in every country
besides Nicaragua. The most extreme results were observed in Vietnam (OR=2.05(1.5, 2.8))
followed by the Congo (OR=1.82(1.0, 3.4)), and the Philippines (OR=1.47(1.0,2.2)).
Madagascar did not contribute to the figures due to minimal variation (93.9% reported cooking
indoors over an open flame). The results observed for CL+/-P were comparable and slightly
further from the null with Congo, Honduras and the Philippines reaching statistical significance
(Figure S1, S2). The iCP results showed a generally positive association with cooking indoors
over an open flame (Figure S3, S4).
28
Discussion
This is the first study to report on the association between cooking indoors over an open
flame and non-syndromic cleft using data from multiple low-resource countries. Mothers who
reported cooking indoor cook smoke exposure were approximately 50% more likely to have a
child with a cleft. This result was present after controlling for suspected confounders (including
other sources of smoke exposure) and in countries with different socioeconomic characteristics
and access to care, as well as being consistent for 5 of 7 countries (Vietnam, the Philippines,
Honduras, DRC, and Morocco). The negative association found in Nicaragua was
underpowered and will need to be explored further as the sample size increases. No association
was found with maternal smoking (prior to or during pregnancy), paternal smoking (lifetime),
environmental tobacco smoke, or folic acid supplementation.
Similar to our findings, a prior case-control study by Liu et al. in China found that indoor
air pollution was associated with an increased odds of cleft if the house was not ventilated
(OR=4.5(1.6–12.9)) and attributed this to coal-burning heating sources.
34
A second study in the
Congo found the odds of cleft was 6-times higher among mothers who reported cooking
indoors.
17
However, the sample size was small (n=162 cases; 162 matched controls) and they
didn’t adjust for additional confounding factors beyond matching. Notably, they also found no
maternal smoking effect but did find a positive association with paternal smoking.
Although we did not observe an effect of mother’s smoking on the risk of cleft, the
frequency of maternal smoking was very low, and smoking is most often not economically
accessible or culturally accepted for women in low-resource populations. According to 2016
World Bank data, 1% of women smoke in Vietnam, 7.8% in the Philippines, 2% in Honduras,
and 0.80% in Morocco.
104
Although the US Surgeon General report identified maternal smoking
during pregnancy as a risk factor for cleft
105
, it is unclear if these findings are generalizable to
low-resource settings. The literature consistently supports an association between maternal
smoking and risk of cleft. Specifically, a meta-analysis of 23 case-control and 6 cohort studies
29
found that mothers who smoked were 37% more likely to have a child with CL+/-P than never
smokers.
21, 24, 92
It is worth noting that 27 of 29 studies were conducted in populations of
European decent and developed countries, which may reflect a different risk profile than the
individuals in our study.
The effects of paternal smoking and ETS on cleft is less conclusive. Paternal smoking
was evaluated in a subset of the current data from Vietnam, the Philippines, Morocco, and
Honduras (n=626 father/child duos) and no association with cleft was found.
26
Studies in
Norway
25
, India
106
, and China
107
have found that exposure to ETS, defined as an exposure to
passive tobacco smoke during the first trimester at home or work, is associated with an
increased risk of cleft (OR=1.6 (1.0-2.5); 2.0 (1.2-3.4); 2.46 (0.99-6.08) respectively), however a
case-control study based in a large American birth defect registry found no effect.
24
Our study
did not see an effect of either paternal smoking or ETS, where ETS is defined as maternal
report of passive smoke exposure in the home during pregnancy.
The mechanism for a role of smoke on cleft formation in embryonic development has
been described in the current literature. An association between maternal periconceptional
exposure to secondhand tobacco smoke and cleft in offspring has been consistently found
31-33,
108-110
in epidemiological studies, while 2 studies have linked clefting with maternal exposure to
other indoor air pollutants and combustion byproducts.
34, 111
The incomplete combustion of
tobacco and other organic compounds, including fuels for cooking and heating, produces
numerous airborne chemicals. Secondhand smoke from tobacco and smoke from fuel
combustion contain known teratogens, including polycyclic aromatic hydrocarbons (PAHs),
carbon monoxide (CO), and heavy metals. Shum et al. demonstrated that periconceptional
maternal exposure to the PAH benzo[a]pyrene causes cleft in genetically “nonresponsive”
inbred mice, which were metabolically deficient.
112
Maternal exposure to low levels of CO has
been shown to cause tissue hypoxia in rat fetuses
113
, which diminishes cellular metabolism of
benzo[a]pyrene
114
, suggesting a potential role of CO in cleft development. Combustion of both
30
tobacco and biomass fuels emit heavy metals, including cadmium, which has been shown to
cause cleft development in rats.
115, 116
At the human population level, Langlois et al. showed
that maternal occupational exposure in work environments with greater levels of PAHs was
associated with greater odds of cleft.
111

Bias due to control selection is a concern common to all case-control studies. Selection
of more affluent controls could influence the representativeness of cooking methods in the
sample with respect to the underlying base population. The replication of the association across
sites (Figure 2) suggests that selection of non-comparable controls is unlikely to explain the
finding due to the variability of SES and access to medical care across countries. In support of
this, the association between indoor cook smoke exposure and cleft was present in countries
with a variety of surgical care and those where medical missions are the primary care source.
Further, differential recruitment by age of cases and controls did not explain the results as the
effect magnitude was not diminished when restricting to cases one-year of age and under. The
original design restricted to cases age 4 years and under to limit recall bias and in fact
approximately 70% of our cases were under one year. Correction for differential selection due to
SES was addressed by adjustments for household income and parental education in the
analysis. Both the adjusted and stratified models by SES were consistent with the original
findings.
A limitation of our study is that we cannot be certain cases are fully population-based.
However, regional recruitment efforts were extensive and conducted at least four months prior
to each mission. Control samples were collected from women at public neighborhood, clinic,
and hospital-based birth centers to limit the oversampling of higher income families. Another
concern may be underreporting of smoking or alcohol history by case mothers due to stigma
around these behaviors while pregnant. While we did ask information on amount of tobacco
products parents used weekly, the data was too sparse and variability too low among mothers
to conduct a detailed analysis. Similarly, we did not have data on potential changes in the
31
father’s smoking habits during the pregnancy, which would not adequately classify the fathers
smoking status by trimester.

Conclusions
Exposure to smoke while cooking is already a well-established health risk in low-
resource countries for a wide variety of diseases but has been minimally studied with respect to
cleft. We found a 50% increase in cleft risk for mothers reporting cooking over an open flame
indoors compared to controls in a diverse group of LMICs. It is necessary to take risk factors
specific to low resource settings into account, as those individuals are at the highest risk of
being unable to access care and therefore live with the negative health consequences of
disease. This information can inform public health interventions and education to potentially
prevent disease in populations where care is sparse, and children are most likely to feel the
detrimental, lifelong medical and social effects of cleft. Modifiable, patient-centric solutions, such
as providing a clean-burning cookstove, will be critical for efforts to decrease the burden of cleft
globally and improve lives around the world.

32
Tables and Figures

Figure 1. Cooking Indoors Over an Open Flame- Odds Ratio and 95% CI Excluding Each
Country (All Cleft Types Combined)

33

Figure 2. Cooking Indoors Over an Open Flame- Odds Ratio and 95% CI By Country (All Cleft
Types Combined)

34

Table 1. Child Characteristics of Case and Control from all Countries (N=4151)

35
Table 2. Parental Characteristics of Cases and Controls from all Countries (N=4151)

36

Table 3. Lifestyle Factors of Cases and Controls from all Countries (N=4151)

37
Table 4. Adjusted Odd's Ratios (OR) of Smoke Related Factors and Cleft lip and/or
palate in all countries (N= 4151)

Table 5. Adjusted Odd's Ratios (OR) of Smoke Related Factors and Cleft lip with or
without cleft palate (excluding iCP) in all countries (N= 3765)

Model 1* Model 2** Model 3***

OR 95% CI p-value OR 95% CI p-value OR 95% CI p-value

Cooking
indoors over
open flame
1.93 (1.64, 2.27) < .0001 1.51 (1.26, 1.81) < .0001 1.49 (1.23, 1.79) < .0001

Smoking Pre-
pregnancy-
Mother
0.88 (0.49, 1.58) 0.67 0.92 (0.49, 1.70) 0.78 1.65 (0.50, 5.61) 0.52

Smoking during
pregnancy-
Mother
1.20 (0.45, 3.34) 0.72 1.39 (0.49, 4.05) 0.54 0.79 (0.38, 1.61) 0.41

Smoking-
Father

1.11 (0.95, 1.30) 0.19 1.10 (0.93, 1.31) 0.28 1.18 (0.96, 1.47) 0.12

Smoking in the
household
during
pregnancy

1.04 (0.89, 1.23) 0.61 0.98 (0.83, 1.17) 0.86 0.85 (0.68, 1.06) 0.14
*Model 1-- Adjusted for country, maternal age, mother and father education (primary or less/ secondary or more), family history of cleft
**Model 2-- Additionally adjusted for rural/ urban home and alcohol consumption during pregnancy
***Model 3-- Mutually adjusted for all smoke variables and Model 2 covariates

Model 1* Model 2** Model 3***

OR 95% CI p-value OR 95% CI p-value OR 95% CI p-value

Cooking
indoors over
open flame
2.02 (1.70, 2.41) < .0001 1.56 (1.29, 1.90) < .0001 1.55 (1.27, 1.89) < .0001

Smoking Pre-
pregnancy-
Mother
0.91 (0.50, 1.66) 0.76 0.95 (0.50, 1.80) 0.87 2.08 (0.61, 7.30) 0.43

Smoking
during
pregnancy-
Mother
1.43 (0.53, 4.02) 0.48 1.65 (0.57, 4.86) 0.35 0.74 (0.34, 1.57) 0.24

Smoking-
Father

1.08 (0.92, 1.28) 0.35 1.08 (0.90, 1.29) 0.42 1.14 (0.91, 1.43) 0.26

Smoking in the
household
during
pregnancy

1.06 (0.89, 1.25) 0.54 1.00 (0.83, 1.20) 0.98 0.87 (0.69, 1.10) 0.25
*Model 1-- Adjusted for country, maternal age, mother and father education (primary or less/ secondary or more), family history of cleft
**Model 2-- Additionally adjusted for rural/ urban home and alcohol consumption during pregnancy
***Model 3-- Mutually adjusted for all smoke variables and Model 2 covariates

38
Supplemental Material
Table S1. All Sites Used for Case and Control Collection

Country City Site
Democratic Republic of Congo Kinshasa
Operation Smile Democratic Republic of
Congo
Counseil National de l'Orde des Medecins
Clinique Ngaliema
Kinshasa General
Roi Baudouin
Maternite Kingasani
El Rapha Clinic
Kitambo
Centre De Sante Maternite Lisungi
Honduras
Choluteca Hospital Regional Del Sur
Comayagua Hospital Santa Teresa
Santa Rosa de Copan Hospital Regional de Occidente
Tegucigalpa
Operacion Sonrisa Honduras
Hospital San Felipe
Madagascar
Antananarivo
Operation Smile Madagascar
Joseph Ravoahangy Andrianavalona
Hospital
Antsirabe Vakinankaratra Regional Center Hospital
Morocco
Oujda Hospital Al Farabi
Marrakesh Ibn Tofail Hospital
Tiznit Hospital Hasan I
Nicaragua Managua
Operacion Sonrisa Nicaragua
Hospital Aleman Nicaraguanese
Hospital Berta Calderon
Philippines
Bacolod
Our Lady of Mercy Hospital
Adventist Miller Hospital
Cauayan City
Cauayan District Hospital
Isabela United Doctors Medical Center
Cavite
General Emilio Aguinaldo Memorial Hospital
Pakamutan ng Dasmarinas
St. Paul Hospital
Cebu
Mariquita Young Foundation
Paanakan Se Mandaue (Mandaue)
St. Anthony’s Birthing Clinic
Municipal Health Office of Consolacion
Daisy’s Birthing Clinic
University of Cebu Medical Center
Agnes Birthing Center
Davao City
Brokenshire Hospital
Mindanao Cleft Center
General Santos General Santos District Hospital
Iloilo
Qualimed Hospital
CFC Birthing Clinic
Pampanga
Ricardo Rodriguez Hospital
Diosdado Macapagal Hospital
Jesus A Datu Medical Center
Vietnam
An Giang An Giang General Hospital
Dak Lak Dak Lak District Hospital
Hanoi
Vietnam Cuba Friendship Hospital
Operation Smile Vietnam
Phu San HN (Hanoi Maternity Clinic)
39
Ho Chi Minh City
HCMC University Medical Center
Thu Duc District Hospital
Hue Hue University Medical Center
Quang Ngai Quang Ngai Hospital

40

2012 2013 2014 2015 2016 2017

Control Case Control Case Control Case Control Case Control Case Control Case
(n=210) (n=136) (n=248) (n=106) (n=419) (n=532) (n=553) (n=493) (n=412) (n=530) (n=172) (n=340)
Country

Congo 117
(55.7%)
50
(36.8%)
91
(36.7%)
33
(31.1%)
0 (0%) 0 (0%)
74
(13.4%)
59
(12.0%)
0 (0%) 0 (0%) 0 (0%) 0 (0%)
Honduras
0 (0%) 0 (0%) 0 (0%) 0 (0%)
169
(40.3%)
132
(24.8%)
145
(26.2%)
74
(15.0%)
193
(46.8%)
129
(24.3%)
21
(12.2%)
54
(15.9%)
Madagascar
0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%)
85
(20.6%)
128
(24.2%)
0 (0%) 0 (0%)
Morocco
0 (0%) 0 (0%) 0 (0%) 0 (0%)
23
(5.5%)
45
(8.5%)
30
(5.4%)
39
(7.9%)
16
(3.9%)
27
(5.1%)
0 (0%) 0 (0%)
Nicaragua
0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%)
47
(27.3%)
124
(36.5%)
Philippines 93
(44.3%)
50
(36.8%)
0 (0%) 0 (0%)
206
(49.2%)
181
(34.0%)
38
(6.9%)
157
(31.8%)
13
(3.2%)
110
(20.8%)
3 (1.7%)
75
(22.1%)
Vietnam
0 (0%)
36
(26.5%)
157
(63.3%)
73
(68.9%)
21
(5.0%)
174
(32.7%)
266
(48.1%)
164
(33.3%)
105
(25.5%)
136
(25.7%)
101
(58.7%)
87
(25.6%)

Table S2. Case and Control Breakdown by Year for All Countries

41
Table S3. Adjusted Odd's Ratios (OR) of Smoke Related Factors and Isolated Cleft
Palate in all countries (N= 2320)

Model 1* Model 2** Model 3***

OR 95% CI p-value OR 95% CI p-value OR 95% CI p-value

Cooking indoors
over open flame
1.65 (1.22, 2.23) 0.001 1.33 (0.95, 1.85) 0.10 1.28 (0.91, 1.79) 0.16

Smoking Pre-
pregnancy- Mother
0.75 (0.21, 2.10) 0.61 0.82 (0.19, 2.52) 0.76 NA NA NA

Smoking during
pregnancy- Mother
NA NA NA NA NA NA 1.15 (0.26, 3.75) 0.98

Smoking- Father

1.20 (0.90, 1.60) 0.22 1.17 (0.86, 1.59) 0.31 1.35 (0.99, 2.00) 0.14

Smoking in the
household during
pregnancy

0.99 (0.74, 1.34) 0.97 0.97 (0.71, 1.32) 0.84 0.80 (0.53, 1.19) 0.27
*Model 1-- Adjusted for country, maternal age, mother and father education (primary or less/ secondary or more), family history of cleft
**Model 2-- Additionally adjusted for rural/ urban home and alcohol consumption during pregnancy
***Model 3-- Mutually adjusted for all smoke variables and Model 2 covariates

42
Figure S1. Cooking Indoors Over an Open Flame- Odds Ratio and 95% CI Excluding each
Country (CL+/-P ONLY)

43
Figure S2. Cooking Indoors Over an Open Flame- Odds Ratio and 95% CI by Country (CL+/-P
ONLY)

44
Figure S3. Cooking Indoors Over an Open Flame- Odds Ratio and 95% CI Excluding each
Country (iCP ONLY)

45
Figure S4. Cooking Indoors Over an Open Flame- Odds Ratio and 95% CI by Country (iCP
ONLY)

46
The International Family Study of Nonsyndromic Orofacial Clefts:
Design and Methods

Abstract

Background: The majority of research to understand the risk factors of nonsyndromic orofacial
clefts (NSOFCs) has been conducted in high- income populations. Although patients with
NSOFCs in low- and middle-income countries (LMICs) are at the highest risk of not receiving
care, global health infrastructure allows innovative partnerships to explore the etiologic
mechanisms of cleft and targets for prevention unique to these populations.
Methods: The International Family Study (IFS) is an ongoing case-control study with
supplemental parental trio data designed to examine genetic, environmental, lifestyle and
sociodemographic risk factors for NSOFCs in eight LMICs (through August 2020). Interview and
biological samples are collected for each family. The interview includes demographics, family
history of cleft, diet and water sources, maternal pregnancy history, and other lifestyle and
environmental factors.
Results: Seven of eight countries are currently summarized (2012-2017) for a total of 2,955
case and 2,774 control families with 11,946 unique biological samples from Vietnam,
Philippines, Honduras, Madagascar, Morocco, DRC, and Nicaragua. The phenotype distribution
was 1,641(55.5%) cases with cleft lip and palate (CLP), 782(26.5%) with isolated cleft lip, and
432(14.6%) with isolated cleft palate.
Discussion: IFS is the largest case set of NSOFCs with an associated biobank in LMICs
currently assembled. The biobank, family, and case-control study now include samples from 8
LMICs where local health care infrastructure cannot address the surgical burden of cleft or
investigate causal mechanisms. IFS can be a source of information and may collaborate with
local public health institutions regarding education and interventions to potentially prevent
NSOFCs.
47
Introduction
Cleft lip with or without cleft palate is one of the most common birth defects worldwide
with a global incidence of 1 in 700 live births
2
. Patients with cleft face significant health risks
from birth due to feeding difficulties leading to malnutrition and associated health risks
117-119
.
Throughout their lifetime, when adequate care is available, a child born with a cleft lip and
palate will require multiple orofacial surgeries, specialized dental and orthodontic care, and
speech therapy at a minimum. Some individuals will experience lifelong difficulties from the
condition and may experience lasting psychosocial effects
120
. Current estimates are that over 5
billion people globally lack access to safe and affordable surgical care for any surgically
treatable condition
1
. Given the complexity of cleft surgery and the necessity of ancillary services
with the large number of impacted children and adults awaiting care, it is unlikely that a global
solution dependent solely on surgical intervention will ease the burden of cleft disease.
Non-syndromic orofacial cleft lip and/ or cleft palate (NSOFCs) is defined as those cases
with no other major malformations or syndromes, and these represent approximately 70% of all
orofacial clefts
7-9
. The risk factors for NSOFCs are multifactorial and remain largely
uncharacterized. The incidence of cleft varies widely by race/ethnicity. The highest rates of cleft
are found in Asian populations with an incidence as high as 1 in 500 births
12
. The lowest rates
are in African populations (1 in 2000 births) with populations of European ancestry falling
between 1 in 1000 births
2, 11, 13
. The most consistently identified risk factors for non-syndromic
cleft are family history, race/ethnicity, maternal smoking, and environmental tobacco smoke
19,
31, 46, 92, 93
. Other environmental factors that have been found to be associated with non-
syndromic cleft in at least two studies include: periconceptional alcohol use
100, 101
, lack of folic
acid supplementation
96, 97
, low maternal education
94, 95
, advanced maternal age
98, 99
, and
diabetes (either pregestational or gestational)
22, 102
.
The International Family Study (IFS) was designed to address the lack of
comprehensive data available on the genetic, environmental, lifestyle and sociodemographic
48
factors influencing the risk of NSOFCs in low- and middle-income countries (LMICs). Through
an ongoing partnership with Operation Smile (OS), a well-established international cleft
organization providing free cleft care (https://www.operationsmile.org), we conducted a
“population- sampled” case-control study (see discussion for limitations related to design-
hereinafter referred to as case-control) to explore risk factors for non-syndromic cleft in eight
multiethnic LMICs- collecting both questionnaire data and biological samples. While the etiologic
mechanisms resulting in controlling risk of NSOFCs may be similar globally, the specific factors
contributing to these mechanisms, the prevalence of exposures, and the numbers of cases
attributable to specific mechanisms may vary by region of the world. Our effort to collect and
document the experience of a large, regionally diverse set of children attending OS missions
may inform risk mechanisms through novel factors or serve to replicate or elucidate findings
from populations of European ancestry. The study will add to our understanding of determinants
of NSOFCs in multiethnic populations (Asian, African, Latino) and will evaluate environmental,
genetic, and lifestyle factors included in the current body of cleft research. Using the growing
volume of patients served by OS, this study will be statistically powered to test these risk factors
by cleft phenotype, country and region.

49
Methods
Study Design
IFS is an ongoing case-control study with supplemental parental trio data led by
investigators at the University of Southern California (USC) and Children’s Hospital Los Angeles
(CHLA) in collaboration with OS. It is designed to examine genetic, environmental, lifestyle and
sociodemographic risk factors for NSOFCs in eight LMICs (as of August 2020). Many of these
cases are treated only through not-for-profit organizations and are therefore difficult to locate,
document and characterize through other data sources. Because of the lack of centralized
information on cleft cases in LMICs, less information is currently available in the medical
literature on environmental and genetic risk factors for cleft compared to Europe, the US and
China. The study is currently active in eight countries: Vietnam, the Philippines, Honduras,
Nicaragua, Guatemala, Madagascar, Morocco and the Democratic Republic of the Congo
(DRC). The data snapshot presented summarizes collections through 2017 and does not
include Guatemala. All work was approved by the Institutional Review Board at the university
including country-specific authorizations (IRB #s: HS-11-00602, HS-12-00682, HS-13-00616,
HS-16-00790, HS-18-00325, HS-16-00138, HS-14-00253, HS-11-00233). Funding for the study
has come from a variety of donors and foundations through the children’s hospital and OS. Key
definitions for the study are provided in Figure 1.

Specific Aims
The primary aims of IFS were designed to address the leading suspected risk factors for
environmental, lifestyle and genetic causes of NSOFCs: 1) to evaluate the relationship between
personal and environmental exposure to paternal and maternal smoke sources, both prior to
and during pregnancy, and NSOFCs including maternal and paternal use of tobacco products,
exposure to tobacco smoke in the home environment, and use of an open flame for cooking in
the home; 2) to evaluate the relationship between modifiable maternal prenatal lifestyle factors
(diet, water source, vitamin, drug or medication, and alcohol use) with risk to NSOFCs; 3) to
50
evaluate the relationship between current and past paternal occupation and exposure to agents
known to be mutagenic and NSOFCs; and 4) to explore both rare and common genetic variants
and their relationship to risk of NSOFCs using genome wide association and tests for gene-
environment interaction methodology in multiethnic populations. The study framework is shown
in further detail in Figure 2.

Operation Smile Partnership
IFS was designed and is conducted by collaborators at USC, CHLA, and OS. An
organizational chart for the IFS and OS program is shown in Figure 3. The epidemiology,
biostatistics, laboratory, and other preventive medicine resources and personnel are based at
the university’s school of medicine. Clinical study personnel and additional genetic laboratory
resources are based at CHLA, which is an affiliated pediatric partner hospital of USC. These
individuals make up the scientific portion of the core IFS study team, which is responsible for
study design, data storage, genetic storage, IRB requirements, data analysis, and all other post-
collection study needs. Multiple OS staff members dedicate time to work on the study,
participate with the scientific team on field coordination and manage related logistics. The OS
teams (both international and local) are responsible for communicating and working with the
local teams and partners.
In many OS mission countries, there is a local foundation functioning semi-
autonomously from the global organization. In these cases, there is often a large, dynamic in-
country team with long-standing relationships and infrastructure that can support research. This
is true for seven of the eight study countries (all but the DRC). Study countries are chosen
based on both scientific and practical considerations by all stakeholders. The study team has
chosen to work in what has been deemed a feasible number of the 27 currently active OS
program countries, determined by staff, resources including available time, local infrastructure,
presence of strong leadership, a participating volunteer base of research team members as well
51
as a representation of a regionally diverse group of LMICs. Allocating study resources to a
smaller group of LMICs also strategically allows sample sizes to accumulate faster to enable
country level analysis. The project has both a full-time project coordinator and manager to
ensure the teams work effectively and communicate weekly to meet research and quality control
standards.

Case Definition
IFS defines cases as any patients with non-syndromic cleft lip and/or cleft palate (ICD10
35-37) presenting to an OS mission from ages 6 months to 4 years
103
. Nonsyndromic clefts are
broken into three unique phenotypes: isolated cleft lip (iCL), isolated cleft palate (iCP) and cleft
lip and palate (CLP). NSOFCs have a wide range of severity within phenotypes, ranging from a
slight notch in the lip to the lip and palate being completely open. Cases were deemed
nonsyndromic (as defined in Figure 1) by credentialed OS clinical volunteers across a wide
variety of specialties formally trained to work with patients with cleft.
Further eligibility criteria include that the biological mother or biological father of the child
seeking medical care for cleft must be present to conduct the interview and provide informed
consent for participation at the time of the OS mission. The consenting parent must be 18 years
of age or older. A child is considered ineligible for IFS participation if the mother is pregnant, a
biological sibling was born after the child seeking treatment, the child has a clinically
recognizable syndrome, the child is one of a multiple birth, or the child has a medical condition
other than cleft. Criteria were selected to maximize the ability of the biological parent to recall
the exposure window relevant to the most recent pregnancy with the child proband. If biological
samples and interview data are only available for the biological father and the child (no maternal
samples), the case-father dyads are used for genetic analyses and studies looking at paternal
exposures, but they are excluded in studies based on maternal responses. Beginning in 2017,
52
saliva samples were collected for OS mission children up to 7 years of age for inclusion in the
genetic portion of the study only.

Case Ascertainment
Cases for IFS are recruited on selected OS mission sites. The specific hospital used for
each OS mission, and by default this study, is chosen based on its ability to meet the
organization’s ‘Global Standards of Care’. Cases are considered representative of the region
due to extensive OS community outreach. Recruitment methods differ slightly by country based
on knowledge of geographical features and community structure established by local teams.
The catchment area for study recruitment differs by country, and sometimes by region or
province, based on the availability of specialized, affordable cleft care. Eligibility criteria were
identical for case recruitment in all countries.

Control Definition
Control children are identified as any child born during an approximate one- to two-week
window encompassing the mission date among all children identified to be a healthy newborn
by local medical practitioners. Control child eligibility requires the child’s biological parent(s) to
agree to participate in the study, complete the study interview (same questionnaire as case
families), be 18 years of age or older and provide informed consent for participation in the study.
Consistent with case exclusion criteria, a child is considered ineligible for IFS participation if: the
mother is pregnant, the child has a syndrome or major birth defect (including cleft), or the child
is one of a multiple birth.

Control Ascertainment
Multiple regional neighborhood, clinic, and hospital-based birth centers are identified
prior to each mission by in-country OS partners to represent the catchment area of the
Operation Smile mission and improve case-control comparability. All selected birth centers are
public to better match demographics of the mission patients. Leadership is approached and
53
debriefed on the study aims and procedures, and permissions are secured to recruit control
families. Mothers are approached at the center by our study team members. Each site is visited
daily or every other day during the mission to screen and recruit control families of healthy
eligible newborns.

Study Area
The study area includes 8 different OS countries located across Asia, Africa and Central
America. The countries (number of unique sites, year of initial collection) in order that they were
included are: Vietnam (9 sites in 8 cities, 2012), Philippines (25 sites in 13 cities, 2012), DRC (7
sites in 1 city, 2012), Morocco (8 sites in 8 cities, 2014), Honduras (5 sites in 5 cities, 2013),
Madagascar (5 sites in 4 cities, 2016), Nicaragua (2 sites in 1 city, 2016), and Guatemala (2
sites in 2 cities, 2018). These countries were chosen for the following reasons 1) geographic
diversity; 2) strong relationships with local stakeholders (both OS and medical volunteers) who
support the research; 3) strong university, hospital, or government partners to obtain all
necessary approvals; and 4) inclusion of OS countries underrepresented in current NSOFC
literature. Additional detail including the collection cities, hospital and university partners, and
organizations utilized during these collections can be found in Supplemental Table 1.

Pre-Mission Control Site Recruitment
Beginning one year to two- months prior to the mission, contact is made with the
potential control sites (regional neighborhood, clinic, and hospital-based birth centers in the OS
mission region) to assess the feasibility and appropriateness of each site for control recruitment,
as well as to gauge interest in participating. Leadership from the selected sites are then invited
to participate and be debriefed on the study aims and procedures. Authorization from officials
representing the local OS organization, center leadership, and/or Ministry of Health and related
health agencies within the OS host country is obtained prior to recruitment.

54
OS Mission Recruitment Techniques
Operation Smile advertises its patient outreach extensively throughout local communities
and often recruits hundreds of patients for its medical missions. This is done through radio
broadcasting, television ads, flyers, billboards, and utilization of the local health system, among
many other methods. Innovative techniques are being used by the organization to ensure that
even remote regions are completely saturated with information and patients have the highest
probability of accessing care. In Honduras, meter readers from the utility company, who visit
every home in the country, were trained to recognize cleft conditions and provide contact
information for OS follow-up when patients were identified. The reach of the recruitment
initiatives depends on the country’s geography, the number of missions scheduled per year, and
mission locations. In countries where medical missions happen throughout the year in multiple
geographic sites, such as Vietnam, patients will be contacted by the organization about the next
mission in their region. In Madagascar, where missions are less frequent, patients will often
travel long distances to the mission because it is unknown if one will happen in their region in a
timely window. Case recruitment initiatives often occur one year to at least six- months prior to
the mission to ensure regional saturation and prepare logistically. There are many countries
where a patient tracking system has been established so patients are contacted about
upcoming missions when they can most conveniently receive timely care.

Interviewer Recruitment and Training
IFS interviewers are selected from a previously identified volunteer pool wishing to
support an OS mission. Specifically, they are chosen if they a) can be present for the majority of
the mission to volunteer, b) have work experience with the study, c) are proficient in English as
well as their local language, and/or d) have a medical or research background (such as medical
or nursing students). Once they have been selected, they participate in a training conducted by
the core-study team member present on the mission. Interviewers are trained on how to assess
each family for their eligibility, complete the consenting process and explain the study to
55
families, and how to record eligibility information for each family visiting the research station.
The interviewers are then given copies of the questionnaire to familiarize themselves with the
interview flow and the questions. The core-study team member reviews questions with
interviewers and procedures to ensure that each interviewer is able to conduct interviews
appropriately with study families. Interviewers are trained in English and/or their native
language.
Interviewers are also trained to collect saliva from the parents and study child. Practice
saliva kits are used to demonstrate how to properly collect saliva and the different strategies to
ensure quality samples without cross contamination for parents (spit kit) or children (swab).
Interviewers are then trained on how to handle the collected saliva samples to ensure that the
samples and their corresponding questionnaires are properly linked. The core-study team
member is always present for both the saliva collection and questionnaire portions of the
interview process to address any questions or issues that arise.

Interview Flow
The child and family proceed through clinical screening and recruitment as shown in
Figure 4, which is the typical medical and comprehensive care screening for all OS surgical
candidates in a mission setting. Clinical screening consists of creating a medical record, surgical
evaluation, examination by anesthesiologists and pediatricians, vital signs taken by nurses,
consultation from a speech language pathologist and child psychologist, and patient navigators
who help guide the patients through the screening process. The research station follows clinical
recruitment, but is located prior to the phlebotomy station, which is only required for a subset of
patients.
Case eligibility criteria are assessed by the core-study team member. Eligible families
are then assigned to an interviewer to complete the questionnaire. It is highly emphasized that
their participation in the research study has no impact on the selection process for surgery with
56
OS. Once families have agreed to participate in the study and are formally consented, they are
brought in to a private room or area whenever possible with an interviewer who can administer
the questionnaire in the family’s native language. If both the mother and father of the patient are
present, the interviewer tries to spend time with each parent individually. After the questionnaire
is completed, the interviewer collects saliva samples from the mother, father and proband. The
barcodes from the saliva samples are added to the master data sheet and parental
questionnaires to ensure samples and questionnaire data can be linked without the use of
identifiers.

Questionnaire Protocols and Procedures
Mothers are asked to complete a detailed interview, which takes approximately 40
minutes. Variables collected through the interview relevant to the etiology of cleft are
summarized in Figure 2. All questionnaires, informed consents, and study materials have been
translated and back translated before being used in country. Extensive data are collected from
the maternal questionnaire on family history of cleft (in the immediate and extended family on
both sides), parental exposures prior to and during pregnancy (smoking, alcohol, diet, medical
issues, medication, water usage, cooking behavior), child and parent demographic
characteristics (age the child was born, pregnancy history, education, employment), and
information about the father (smoking, alcohol, age the child was born, employment history and
exposures, education). Questions are kept simple to decrease any potential mistranslation or
confusion among both data collectors and study participants. Any individuals interested in
obtaining a copy of the questionnaire or informed consent may contact the authors for a copy.
If the father is present, which is less common, he is asked to complete a short, 15-20-
minute interview, which includes limited information on his medical history as well as
environmental and behavioral characteristics. These characteristics include family history of
cleft, smoking behaviors, alcohol intake, occupation, area of residence, education, and chemical
57
exposures. The father is interviewed separately from the mother to increase the reliability of the
response. The questions on the father questionnaire are written to a level of detail reasonable
for the fathers to adequately respond.

Saliva Sample Protocols and Procedures
All biological samples are collected in the form of saliva using Mawi (Mawi DNA
Technologies, ISWAB-DNA-250) and DNA Genotek (DNA Genotek, Inc., OGR-250, OGR-525,
and OGR-575) collection kits. Adults are shown how to use the saliva tubes for collection and
children are swabbed by a study member with gloves to ensure no contamination. Adults and
older children most often (approximately 3+) have no issues providing samples. It is often most
difficult in small children and babies due to contamination (if they were recently eating or
drinking) as well as the realities of the setting; most cases have spent an entire day being
screened for potential surgical care and are often agitated. Families are informed that if they are
uncomfortable with providing a saliva sample, they are welcome to decline participation in this
portion of the study and participate only in the questionnaire portion. The consent document
ensures participants “To prevent this [loss of privacy], we will not put your name or other
personal information on the surveys and saliva samples. Instead, they will be given a code
number to protect your identity.”

Quality Control
There are multiple quality control measures taken throughout the interview,
transportation, storage, and data entry processes of the study. A core-study team member
monitors all data collections.

Data Collection
Master data sheets are used to track study ID numbers, saliva kit numbers, cleft
diagnosis information, and study eligibility criteria kept for each individual approaching the
research station. This allows the study team to cross-check completed questionnaires with the
58
master data sheets at the end of each day and again at the close of each mission to ensure all
questionnaires are present and no duplicates have been collected. The study team also
calculates participation rates and tracks the number of families who approach the research
station but do not meet eligibility criteria.
Questionnaire Entry and Data Aggregation
All questionnaire data are entered by study staff trained in data management and
computer programming. Data entry personnel are assigned to the study as interns for at least
one year, ensuring familiarity with the questionnaire, acceptable responses, and other details
that could be missed by untrained data entry personnel. Data are entered into Qualtrics, which
has electronic quality control, including duplicate checks, built-in throughout the entry platform.
Genetic Data Storage and Processing
Saliva samples are shipped directly back to the university lab via courier from the
respective mission sites to avoid issues with research personnel transporting biological samples
out of the mission country or into the United States. Upon arrival, the saliva samples are
inventoried, checked for the proper match against questionnaire data, and stored in the lab
labeled with mission site and collection date. The location of each sample is kept in the study
master file and database. All samples are stored in a temperature-controlled room at the CHLA
laboratory prior to extraction to increase the life of the stored saliva. Once DNA is extracted, the
manifests are updated, and samples are stored in freezers monitored by the USC Molecular
Genetics laboratory.

Statistical Analysis
For this paper, descriptive statistics were calculated to summarize cleft phenotype,
country, year, maternal education, and paternal education separately for case or control. Chi-
squared tests for categorical data were used to assess differences between cases and controls
using a conventional p=0.05 for significance.
59
The most appropriate analysis for these aims will be logistic regression to analyze case-
control data and estimate odds ratios (OR) with their 95% confidence intervals (95% CI). Our
proposed analyses to address the specific aims will include adjusted models based on
knowledge of the literature, known confounders and interactions with respect to the main
exposure of interest. Genetic analyses will be either case- and population-controls or case-trio
data depending on the most appropriate comparison for the null hypothesis of independence
between a genetic marker and case-status. This will allow genome-wide association study
(GWAS) analyses to look for common genetic variants in samples as well as exome or whole
genome sequencing to test for association between aggregated rare variants and de novo
mutations. The established environmental database, in combination with our genetic findings,
will enable data from the IFS to be used to study a wide array of gene by environment
interactions through traditional interaction analysis as well as two and three step approaches to
preserve statistical power.

60
Results
From here forward we reference only the 7 countries with data available, which excludes
Guatemala. Table 1 briefly describes the complete IFS dataset through 2017. Between 2012
and 2017, there were 2,955 case families (51.6%) and 2,774 control families (48.4%) with a
total of 11,946 individual saliva samples collected. The participation overall for cases was 90%
and 85% for controls. The country specific participations rates are (country (case %; control %):
Vietnam (86%, 93%), Philippines (92%, 82%), Congo (97%, 100%), Madagascar (93%, 90%),
Morocco (90%, 72%), Nicaragua (77%, 85%), and Honduras (94%, 65%). The majority of the
data were collected in 2016 (25.6%), followed by 2017 (23.7%), 2015 (21.5%), 2014 (15.1%),
2012 (7.3%), and 2013 (6.8%). The highest yield countries were Vietnam (36.6%), the
Philippines (21.4%), and Honduras (18.7%). Congo represents 8.7% of the IFS dataset,
Madagascar 6.2%, Morocco 3.6%, and Nicaragua 4.7%. Most cases had the CLP phenotype
(55.5%), followed by iCL (26.5%) and iCP (14.6%). We have collected 1020 (34.5%) case-trios
compared to 467 (16.8%) control trios. Education was generally higher in the control group with
82.2% of control mothers compared to 65.3% of case mothers having a secondary education or
higher (p < 0.01). This was also true among fathers with 78.7% of control fathers having a
secondary education or more compared to 62.6% of case fathers (p <0.01). A further
breakdown of this information by country is provided in Supplementary Table 2.

61
Discussion
The International Family Study is the largest case-control study with a supplementary
biobank, regionally sampled controls, and child-parent trios for exploration of etiologic questions
around NSOFCs in a diverse group of LMICs with extensive population-wide recruitment. It was
designed to have the unique ability to study genetic and environmental risk factors among
racial/ethnic groups that have been under-represented in the current literature, using the
established network of a large, volunteer-based international non-profit whose goal is to serve
untreated children. Our study will be able to both explore populations that are not part of the
existing literature, as well as strengthen and validate findings from other researchers who are
working in similar populations
17, 34, 49, 61, 67
. The strengths of the study are the ability to reach
LMICs that are underrepresented in the current literature, an unconventional partnership
between USC, CHLA, and OS, the large and diverse group of cases seen by OS, and the ability
to study risk factors of NSOFCs that are highly prevalent in low-resource settings.
The key innovations include:
• Utilizing academic resources in collaboration with the networks and operational capacity
of a large, international non-profit organization allowing engagement of in-country
stakeholders to ensure the success of the project.
• Providing a pathway for interviewers to join the core-study team when appropriate to
improve local engagement, collect controls beyond the mission timeframe if needed, and
decrease study costs due to less travel being required from the US.
• Creating a multi-ethnic database that can be used by other research teams to validate
their findings and make comparisons by country, region or cleft phenotype.
The data from this study can be used to identify unique environmental risk factors for cleft
prevalent in low-resource areas and investigate genetic risk factors in multiple racial/ethnic
groups underrepresented in cleft research. A recent paper by Sirugo et al. found that 78% of all
GWAS studies are of European ancestry with only 2% African and 1% Hispanic or Latin
62
American representation
121
. Without studying populations like those included in IFS, not only
will we be unable to fully understand cleft genetics, but health inequalities will be further
exacerbated as the era of personalized medicine continues to grow. This holds true for
environmental exposures as well, which may not have been studied due to negligible
prevalence in high-resource countries. For example, while maternal smoking may be an
important contributor to cleft risk in westernized populations, in low resource countries where
few women smoke, the primary source of exposure to smoke-related agents and therefore
attributable risk of NSOFCs, may be from biomass cooking
17, 122
.
A potential limitation of this study is uncertainty about the representativeness of controls with
respect to the base population for all environmental and lifestyle factors of interest. Controls are
recruited from local birthing centers of similar socioeconomic status, restricted to similar age
range and sampled at the same time frame as case recruitment to improve comparability. Data
were collected to control for other sources of selection bias in the analysis including household
income, parental education, parental employment, and child’s place of birth (clinic, hospital or
home birth). Case-parent trios also extend our ability to explore genetic risk factors using a
family-based design and help to better account for this possibility.
We used extensive, regional recruitment efforts to identify all eligible cases of cleft within the
geographically targeted region. We recognize that the designation of population-based is a
difficult standard to meet and is likely to be achieved only for diseases or conditions identified in
registries such as cancer or nested in population-based cohorts or clinical trials. We use the
descriptor of “population-sampled” case-control study to distinguish our design from exclusively
clinic-based or samples of convenience. While we do not have adequate data to validate the
representativeness of the final cases, the recruitment efforts of each mission were extensive
through coordination with local government agencies, hospitals, and clinics to reach children
and infants with unaddressed NSOFCs. Additionally, we saturated the region in the months
leading up to each mission by distributing recruitment materials through radio announcements,
63
roadside banners, fliers, and direct conversations with local stakeholders. The organization’s
38-year history, well-recognized brand, and strong in-country relationships have made them a
large part of the cleft landscape in all IFS countries- ensuring a high likelihood of interaction with
untreated patients with NSOFCs.
The potential of generalizing to the region or countries with a similar gross domestic product
(GDP) will continue to be explored further and has high potential to increase the impact of the
IFS findings. There is also the possibility that subtle signs and symptoms of malformation
syndromes may be missed, and these patients could be included in our study population.
Through the screening process, patients were reviewed by multiple medical providers with
expertise in cleft, increasing the likelihood for a syndrome to be identified.
In summary, IFS aims to fill a sizable gap in our current understanding of cleft risk factors
because we cannot assume that the risk profile for clefts is comparable between LMICs and
high-resource regions. It is imperative to take risk factors specific to LMIC settings into account,
as those individuals are at the highest risk of not being able to access care and live with the
negative health consequences of this common craniofacial malformation. This information can
inform public health interventions and education to potentially prevent disease in populations
where care is sparse, and where children are most likely to feel the detrimental, lifelong medical
and social effects of cleft. Modifiable, patient-centric solutions, such as providing a clean-
burning cookstove, will be critical for efforts to decrease the burden of this malformation globally
and improve lives around the world.

64
Tables and Figures
Figure 3. Key Definitions
The following definitions are utilized in the study to standardize methods and patient
selection:

Term Definition
Nonsyndromic orofacial cleft
(NSOFC)
A defect ranging from a small notch in the lip to a
complete cleft through the lip, alveolus, and palate
(ICD10 35- 27) with no other syndrome or other birth
defect present (this does not include submucous
clefts or bifid uvula without cleft palate).
Cleft lip and palate (CLP) A NSOFC where both the lip and palate are affected
Isolated Cleft Lip (iCL) A NSOFC where the lip with or without the alveolus is
affected, but the palate is phenotypically normal
Isolated Cleft Palate (iCP) A NSOFC where only the palate is affected, but the lip
is phenotypically normal (does not include submucous
clefts or bifid uvula without cleft palate)
Cleft lip with or without Palate
(CL+/-P)
The denotation when patients with iCL and CLP are
combined, as they are considered to have a common
etiology
Low- and middle- income
countries (LMICs)
A country whose national income per person is less
than $12,375 as designated by the World Bank
(definition as of 2019).

65
Figure 4. IFS Environmental and Genetic Study Framework

66
Figure 5. IFS Global Structural Organization Chart

67
Figure 6. OS Mission Flow

68
Table 6. Descriptive Characteristics of the Study by Case/ Control Status through
December 2017 (N= 5729)

Case
(n=2955)
Control
(n=2774)
Total
(n=5729)
Year

2012 208 (7.0%) 212 (7.6%) 420 (7.3%)
2013 90 (3.0%) 301 (10.9%) 391 (6.8%)
2014 508 (17.2%) 357 (12.9%) 865 (15.1%)
2015 530 (17.9%) 700 (25.2%) 1230 (21.5%)
2016 815 (27.6%) 653 (23.5%) 1468 (25.6%)
2017 804 (27.2%) 551 (19.9%) 1355 (23.7%)
Country

Congo 217 (7.3%) 284 (10.2%) 501 (8.7%)
Honduras 485 (16.4%) 588 (21.2%) 1073 (18.7%)
Madagascar 238 (8.1%) 120 (4.3%) 358 (6.2%)
Morocco 121 (4.1%) 88 (3.2%) 209 (3.6%)
Nicaragua 214 (7.2%) 55 (2.0%) 269 (4.7%)
Philippines 772 (26.1%) 453 (16.3%) 1225 (21.4%)
Vietnam 908 (30.7%) 1186 (42.8%) 2094 (36.6%)
Cleft Type
1

Cleft Lip and Palate 1641 (55.5%) NA 1641 (55.5%)
Cleft Lip Only 782 (26.5%) NA 782 (26.5%)
Cleft Palate Only 432 (14.6%) NA 432 (14.6%)
Complete Trio
2

Yes 1020 (34.5%) 467 (16.8%) 1487 (26.0%)
Mother Education
3

None 187 (6.3%) 57 (2.1%) 244 (4.3%)
Primary 805 (27.2%) 408 (14.7%) 1213 (21.2%)
Secondary 1424 (48.2%) 1567 (56.5%) 2991 (52.2%)
More than secondary 506 (17.1%) 713 (25.7%) 1219 (21.3%)
Father Education
4

None 38 (1.3%) 8 (0.3%) 46 (0.8%)
Primary 815 (27.6%) 420 (15.1%) 1235 (21.6%)
Secondary 1330 (45.0%) 1363 (49.1%) 2693 (47.0%)
More than secondary 521 (17.6%) 822 (29.6%) 1343 (23.4%)

1
Missing (cases, controls): 100, NA;
2
Missing (cases, controls): (13,16);
3
Missing (cases, controls):
33, 29;
4
Missing (cases, controls): 251, 161

69
Supplemental Material

Table S4. Partners by Country and Site Type

PARTNERS
COUNTRY National Partners University
Partners
Case Site Partners* Control Site Partners*
Vietnam Operation Smile
Vietnam
Vietnam Ministry of
Health

Vietnam Cuba Friendship
Hospital, Hanoi
Quang Ngai District
Hospital, Quang Ngai
Dak Lak District Hospital,
Dak Lak
Nghe An General
Hospital, Vinh City
An Giang General
Hospital, An Giang
HCMC University Medical
Center, Ho Chi Minh City
Thu Duc District Hospital,
Ho Chi Minh City
Hue University Medical
Center, Hue
Phu San HN (Hanoi
Maternity Clinic), Hanoi
Vietnam Cuba
Friendship Hospital,
Hanoi
Quang Ngai District
Hospital, Quang Ngai
Dak Lak District
Hospital, Dak Lak
Nghe An General
Hospital, Vinh City
An Giang General
Hospital, An Giang
HCMC University
Medical Center, Ho Chi
Minh City
Thu Duc District
Hospital, Ho Chi Minh
City
Hue University Medical
Center, Hue
Phu San HN (Hanoi
Maternity Clinic), Hanoi
Philippines Operation Smile
Philippines
Kapampangan
Development
Foundation
Mariquita S. Yeung
Foundation
Hope Foundation

University of
Santo Tomas
Iloilo Doctors
College
Santa Ana Hospital,
Manila
Ricardo Rodriquez
Hospital, Pampanga
University of Cebu
Medical Center, Cebu
Isabela United Doctors
Medical Center, Cauayan
City/Isabela
Diosdado Macapagal
Hospital, Pampanga
Jesus A Datu Medical
Center, Pampanga
Our Lady of Mercy
Hospital, Bacolod
Silay City Health Center,
Silay City
Teresita L. Jalandoni
Provincial Hospital, Silay
City
Brokenshire Hospital,
Davao City
Mindanao Cleft Center,
Davao City
General Emilio Aguinaldo
Memorial Hospital, Cavite
St. Paul Hospital,
Cavite/Dasmarinas
Eastern Samar Provincial
Hospital, Borongan/Samar
RHU – Maternity Clinic,
Municipal Health Office
of Consolacion
Pakamutan ng,
Dasmarinas
St. Anthony's Birthing
Clinic, Cebu
Daisy's Birthing Clinic,
Cebu
Agnes Birthing Center,
Cebu
Cauayan District
Hospital, Cauayan
City/Isabela
Grengia Maternity
Home, Lapu-Lapu City
CFC Birthing Clinic,
Iloilo
Teresita L. Jalandoni
Provincial Hospital,
Silay City
Paanakan Sa Mandaue,
Cebu
Saint Anthony Mother
and Child Hospital,
Cebu
Grengia Maternity
House, Cebu
RHU – Maternity Clinic,
Quezon
Our Lady of Mercy
Hospital, Bacolod
70
Quezon
General Santos District
Hospital, General Santos
Adventist Miller Hospital,
Bacolod
Qualimed Hospital, Iloilo
Negros Oriental Provincial
Hospital, Dumaguete
General Emilio
Aguinaldo Memorial
Hospital, Cavite
Eastern Samar
Provincial Hospital,
Borongan/Samar
Negros Oriental
Provincial Hospital,
Dumaguete
University of Cebu
Medical Center, Cebu

Honduras Operacion Sonrisa
Honduras

Hospital Leonardo
Martinez, San Pedro Sula
Hospital San Felipe,
Tegucigalpa
Hospital Regional Del Sur,
Choluteca
Hospital Santa Teresa,
Comayagua
Hospital Regional de
Occidente, Santa Rosa de
Copan
Hospital Leonardo
Martinez, San Pedro
Sula
Hospital San Felipe,
Tegucigalpa
Hospital Regional Del
Sur, Choluteca
Hospital Santa Teresa,
Comayagua
Hospital Regional de
Occidente, Santa Rosa
de Copan
Nicaragua Operacion Sonrisa
Nicaragua

Hospital Aleman
Nicaraguanese, Managua
Hospital Berta Calderon,
Managua
Hospital Aleman
Nicaraguanese,
Managua
Hospital Berta
Calderon, Managua
DRC Operation Smile
Democratic Republic
of Congo
Counseil National de
l'Orde des Medecins
Operation Smile
South Africa

Clinique Ngaliema,
Kinshasa
Kinshasa General,
Kinshasa
Roi Baudouin, Kinshasa
Maternite Kingasani,
Kinshasa
El Rapha Clinic, Kinshasa
Kitambo Clinic, Kinshasa

Centre de Sante
Maternite Lisungi,
Kinshasa
Clinique Ngaliema,
Kinshasa
Kinshasa General,
Kinshasa
Roi Baudouin, Kinshasa
Maternite Kingasani,
Kinshasa
El Rapha Clinic,
Kinshasa
Kitambo Clinic,
Kinshasa
Madagascar Operation Smile
Madagascar

Joseph Ravoahangy
Andrianavalona Hospital,
Antananarivo
Vakinankaratra Regional
Center Hospital, Antsirabe
University Hospital Center
Androva, Mahajanga
University Hospital Center
Tamatave, Tamatave
Befelatanana University
Hospital Center of
Obstetric Gynecology,
Antananarivo
Vakinankaratra
Regional Center
Hospital, Antsirabe

Morocco Operation Smile
Morocco

Hospital Hasan I, Tiznit
Hospital Hasan II, Dakhla
Ibn Tofail Hospital,
Marrakesh
Hospital Sidi Hssian
Bennaceur, Ouarzazate
Hospital Al Farabi, Oujda
Hospital Hassan II,
Laayoune
Hospital Mohammed V,
Safi
Hospital Al Farabi,
Oujda
Hospital Hasan I, Tiznit
Hospital Hasan II,
Dakhla
Hospital Sidi Hssian
Bennaceur, Ouarzazate
71
Hospital Moulay Youssef,
Rabat

Guatemala Operacion Sonrisa
Guatemala

Hospital Hilario Galndo,
Retalhuleu
Hospital Juan Pablo II,
Guatemala City
Hospital Juan Pablo II,
Guatemala City
*Site name, City
72

Table S5. Descriptive Characteristics of the Study through December 2017 by Case/ Control
Status and Country (N= 5729)

73
A Genome-Wide Association Study and Genome-Wide Interaction
Scan with exposure to smoke from cooking and the risk of
nonsyndromic orofacial cleft in a Vietnamese population.

Abstract
Background: Cleft is one of the most common birth defects globally and gene- environment
interactions have been discussed as a possible explanation for the lack of understanding
around risk factors for nonsyndromic occurrence of the disease. In this analysis, we explore
potential genetic variants associated with nonsyndromic orofacial clefts (NSOFC) as well as
gene- cook smoke interactions in an entirely Vietnamese population.
Methods: We conducted a population-sampled case-control study of children with cleft lip
and/or palate and healthy newborns from Vietnam. Saliva samples were collected from the child
and pertinent environmental information was self-reported by the mother. Exposure was defined
as exposure to smoke from cooking indoors over an open flame. A genome-wide association
study (GWAS) using both imputed and non-imputed data as well as a genome- wide interaction
scan (GWIS) using traditional GxE analyses, two-step methods, and statistically efficient one-
step tests were conducted.
Results: 589 cases and 715 controls were available for analysis post QC measures (N= 1304).
No SNPs reached genome-wide significance from the GWAS or GWIS analyses. The tests
produced multiple suggestive findings that warrant further exploration including: rs871570
(OR=1.47, p=6.7E-6), rs113296466 (OR=3.29, p=1.7E-5), rs10834414 (OR= 2.36, p=7.6E-8),
and rs576853071 (OR= 0.57, p=1.3E-6) from the GWAS; and rs1459270985 (OR=0.43,
p=2.8E-7) and rs1254772630 (OR= 0.60, p=1.5E-6) from the GWIS.
Conclusions: This study is a promising exploratory and hypothesis generating analysis to
understand the genetic risk factors for NSOFC in a novel Vietnamese population, as well as the
potential interaction effects with a minimally studied, but highly prevalent, environmental risk
factor- cook smoke exposure. The continued use of novel, statistically efficient methods to
understand these relationships will be essential in the study of rare diseases, such as NSOFC,
74
especially for individuals that reside in under-represented regions of the world and are at
highest risk of living with disease.
75
Introduction
Cleft lip and/ or palate is one of the most common birth defects globally with an
incidence of 1 in 700 live births. A wide incidence range is found by race/ethnicity with the
highest incidence found among Asian populations (1.28-1.90 per 1000 births)
2, 9, 12
. Patients with
cleft are considered nonsyndromic in the absence of any other birth defect, which accounts for
approximately 70% of patients with cleft lip with or without palate (CL+/-P) and 50% of patients
with isolated cleft palate (iCP)
2, 7-9
. While the origins of syndromic clefts are considered largely
genetic, the etiology of nonsyndromic orofacial clefts (NSOFC) remains unclear. Gene-
environment interactions have been discussed as a possible explanation for the lack of clarity
from genetic and environmental studies alone.
With respect to environmental exposures, prenatal exposure to maternal smoking is the
most widely cited risk factor for NSOFC. A recent meta-analysis of 23 case-control and 6 cohort
studies found that mothers who were ever-smokers were 37% more likely to have a child with
CL+/-P than never smokers.
21, 24, 92
Although maternal smoking has been established as a risk
factor for NSOFC, the environmental impact of smoke exposure from cooking has only been
mentioned in two existing studies, but is an important source of prenatal exposure as it is a
leading cause of morbidity and mortality in low- and middle- income countries (LMICs)
17, 34, 123
. A
case-control study by Liu et al. in China found that indoor air pollution was associated with an
increased odds of NSOFC if the house was not ventilated (OR=4.5(1.6–12.9)) and attributed
this to coal-burning heating sources
34
. A second study in the Democratic Republic of the Congo
found the odds of NSOFC was 6-times higher among mothers who reported smoke exposure
from cooking indoors
17
. While indoor cook smoke is not an important source of prenatal smoke
in high- income countries, indoor cooking practices are used by approximately 80% of rural
households in LMICs and have been associated with a wide variety of diseases, including
stillbirths
39
. Smoke exposure from cooking indoors represents the primary source of prenatal
76
exposure to smoke related agents for the majority of women in these countries as maternal
smoking rates are extremely low in most LMICs
39
.
With respect to genetic risk factors, genome-wide association studies (GWAS) of
NSOFC have identified more than 39 risk loci associated with cleft development including IRF6
54
, MTHFR
124
, TGFA
125
, 8q24
55
, PAX7
126
, FOXE1
52
, VAX1
61
, and MSX1
127
among others. A
recent meta-analysis looked at 31 case-control studies to assess the relationship between
NSOFC and IRF6 (3 main SNP’s) or 8q24 (1 main SNP reported)
54
. The analysis highlighted
that although these two loci have been consistently identified as key regions of interest, there is
a large level of heterogeneity within the findings by race/ethnicity. This is continually reflected by
the literature surrounding all of the known risk loci as there is still a lack of understanding of
what percentage of disease risk they account for and differences by race/ ethnicity.
The potential interaction between maternal smoking and genetic factors has been
explored in a small number of studies. Three gene- environment interaction (GxE) studies (2
GWAS, 1 candidate gene study) examined the potential interaction between prenatal, maternal
smoking with MTHFR, a gene associated with neural tube defects, with no statistically
significant findings
70-72
. Three small studies of prenatal, maternal smoking and candidate genes
have found statistically significant interactions with TBK1
73
, ZNF236
73
, and GSTT1
74
, but only
the finding in GSTT1 has been replicated. Environmental tobacco smoke (ETS) has also been
studied in GxE analyses with slightly more consistent findings. A study of Asian case-parent
trios found no marginal associations with genetic variants and risk of iCP, but 15 SNP’s that
mapped to SLC2A9 and 9 to WDR1 were significantly associated with iCP when considering
history of ETS
78
. A study in China also found no direct genetic variant associations with iCP but
found nine variants associated with iCP in RUNX2 when considering ETS history
79
. Two of
these GxETS interaction findings were replicated in a European cohort. Four other studies in
China found marginally, statistically significant GxETS interactions in: the microRNA- 140
gene
80
, ZNF33
81
, BMP4
82
, and IRF6
83
, but none have been replicated.
77
Since the completion of these early GxE studies, new efficient approaches to evaluate
genomic data have been developed for gene-environment interactions of birth defects or other
childhood diseases, where sample size and power are often an issue. In the current study, we
1) perform a genome-wide association study (GWAS) using both imputed and non-imputed data
to explore genetic variants associated with NSOFC and 2) perform a genome-wide interaction
scan (GWIS) to assess for gene- cook smoke exposure (GxCookSmoke) interactions using a
variety of traditional, two-step, and statistically efficient one- step tests. We conducted this study
in a set of Vietnamese cases and controls (n= 1496) collected as part of a population-sampled
case-control study in partnership with Operation Smile, a large international nonprofit
organization. The understanding of these interactions will be critical in populations, such as
Vietnam, where treatment may be limited and there may be an increased impact of promoting
healthy behavior.
70

78
Methods
Subjects
Data for this study was collected from 2012-2018 as part of a coordinated series of
population-sampled case-control studies focusing on genetic, lifestyle and environmental
exposures and NSOFC in children 6 months to 4 years of age. This study was conducted with
Operation Smile (OS), an internationally recognized not-for-profit that has been providing free
cleft surgery and related care to patients for over 36 years. Data for the current analysis
represents children from the collections that took place in 8 cities (25 unique case site, 16
control sites) in Vietnam. Participation rates in the study varied by site from 40.5%-100% for
eligible cases and 66.1%-100% for controls. The methods of this study have been previously
published in depth
66, 102
. All work was approved by the Institutional Review Board at the
University of Southern California including site-specific authorizations.

Case Definition
This study includes non-syndromic cases of cleft lip and / or cleft palate (ICD10 35-
37)
103
. Cleft phenotype is classified as either cleft lip and palate (CLP), isolated cleft lip (iCL) or
isolated cleft palate (iCP). Cleft lip with or without palate (CL+/-P) is used to denote CLP and
iCL. Cases were screened to confirm diagnosis and absence of any genetic syndrome or birth
defect by medical practitioners at the mission site. This included pediatricians, nurses,
anesthesiologists and surgeons who are all formally licensed, trained and OS certified to work
with cleft patients.
Patients were eligible for the study if they were accompanied by their biological mother
(18 years or older), 6 months to 4 years of age, and presented for cleft treatment at the time of
the OS mission. Patients were excluded if the child was not the most recent pregnancy, a
multiple birth, had a genetic syndrome, or had another co-morbid condition.

79
Case Recruitment
Cases for the International Family Study (IFS) are recruited on site during OS missions.
Extensive, on-going regional recruitment and community outreach efforts are conducted by OS
prior to each mission to assure saturation of the event information in the communities. All
patients arrive to the mission site to be screened for care over the span of one or two days with
all costs covered by OS. The patients are registered and seen by general practitioners, nurses,
anesthesiologists, surgeons, and dentists to assess surgical eligibility. Case recruitment for the
study occurs at the end of the screening process. Study eligibility criteria were identical for
cases in all case sites.

Control Definition
Controls were newborns identified from regional neighborhood, clinic, and hospital-
based birth centers around the mission site whose mother agreed to complete informed consent
and the study interview. Individuals were excluded if they had a cleft or any other birth defect,
were a multiple birth, or if the mother was younger than 18.

Control Recruitment
Multiple neighborhood, clinic, and hospital-based birth centers were identified prior to
each mission by the OS Vietnam team to represent the catchment area of the OS mission and
improve case-control comparability. All maternity wards selected were public to better match
demographics of the mission patients. The leadership at the birth center was approached and
debriefed on the study, and local authorization was obtained to recruit families along with IRB
approval prior to the mission. Each site was visited daily during the mission.

80
Smoke Exposure
Smoke exposure was collected by in-person interview with the mothers of cases and
controls for all participants in the study. Local volunteers with medical training (i.e. nursing/
medical students) were identified by OS and underwent training by study members to conduct
the interviews. Local interviewers were used to assure high recruitment and allow completion of
the interview in the language of the families; however, the study supervisor was present during
all interviews for consistency and to maintain quality. Informed consent was completed before
the interview and parents were assured that participation was not required for their child to
receive care. Families were interviewed in a private to semi-private area (depending on
screening space). Questionnaires have been translated and back translated by certified
translators to ensure consistency across countries.
Mother’s interviews took approximately 40-minutes. Smoke exposure history was
collected by five different categories. Maternal smoking (yes/ no) prior to pregnancy, maternal
smoking during pregnancy (yes/ no), ever paternal smoking (yes/ no and quantity), smoking in
the household during pregnancy, and smoke from cooking indoors over an open flame. The
environmental exposure for the GxE portion of this analysis is defined as maternal self-reported
cooking indoors over an open flame as the primary method of cooking (time frame was not
specified).

Genotyping and Quality Control (QC)
Genome- wide genotyping was conducted using DNA from saliva collected from the
case and control children. All DNA is collected in the form of saliva using Mawi (Mawi DNA
Technologies, ISWAB-DNA-250) and DNA Genotek (DNA Genotek, Inc., OGR-525, and OGR-
575) collection kits. Children under 3 are swabbed by a study member with gloves to ensure no
contamination. Families are instructed that their family genetic material will be used in a de-
identified (group level) manner. Families may agree to participate in only the questionnaire
81
portion of the study or both the questionnaire and genetic portions of the study (however less
than 2% refused to participate in both). Only those families who agreed to both are included in
the GxE analyses. Genome-wide SNP data was generated using the Illumina MEGAChip.
Initially, 1496 individuals were sequenced, and 1,705,969 SNPs were available for
analysis. A primary cut off of a call rate less than 80% was used to exclude bad markers/
samples (206,866 SNPs and 31 individuals were excluded) followed by a secondary exclusion
at 98%, which lead to an additional 189,779 SNPs and 51 individuals excluded. 43 individuals
were excluded because they had missing data on phenotype and 67 individuals were removed
to unexplained relatedness. Another 2,204 SNPs were excluded due to not meeting Hardy-
Weinberg equilibrium (HWE) and 18,408 were excluded due to a minor allele frequency less
than 1%. Counting all remaining SNPs (not indels), a total of 1304 individuals and 1,288,712
variants were available for the analyses (Supplemental Figure 1). All QC was done using Plink
1.9. The Michigan Imputation Server (https://imputationserver.sph.umich.edu/index.html#!) was
then utilized with the 1,000 Genomes East Asian reference panel (which includes a Vietnamese
subpopulation) to impute SNPs on all 22 chromosomes. A cutoff of R- squared > 0.3 was used
for all imputed SNPs to ensure quality imputation.

Statistical Analysis
A principal component (PC) analysis was done with the final dataset to adjust for population
substructure among our sample. Areas of high linkage disequilibrium (as defined by
https://genome.sph.umich.edu/wiki/Regions_of_high_linkage_disequilibrium_(LD)) were
excluded for the PC analysis. The scree plot and PCs were visually assessed to better
understand the clustering and population structure of the data (Supplementary Table 1). PCs 1-
5 were retained to account for genetic variance observed among our sample. Sex was the only
other adjustment covariate used for these analyses. A standard GWAS was conducted to look
for the marginal associations between genetic variants (G) and cleft (null hypothesis βG=0) for
82
all 1,288,712 non-imputed SNPs included in the analysis as well as using the imputed data set.
A Bonferroni multiple-testing correction was applied and all SNPs passing the threshold of p <=
1e-8 were considered statistically significant.
Data was evaluated for GxE interaction using both traditional methods (1 degree- of-
freedom GxE test and 2df G and GxE joint test) as well as case- only analyses, two-step
approaches, and a novel unified model
128, 129
. QQ plots for GxCookSmoke of all tests
suggested a sufficient lack of correlation between the Cleft- G effect, G-CookSmoke effect, and
GxCookSmoke correlation in the base population (plots not shown). The following null
hypothesis and model were used for the single step tests:
• Traditional interaction test- H
0
: βGxE=0
o Logit(Pr(D=1|G) = β 0 + β GG + β EE + β GxEGxE
• Joint G, GxE test- H
0
: b
G
=b
GxE
=0
o Logit(Pr(D=1|G) = β 0 + β GG + β EE + β GxEGxE
• Case only analysis- H
0
: βE = 0
o Logit(Pr(G=g|E, D=1) = β 0 + β 1 + β EE
• Unified model- H
0
: β
D

= β
E

= β
D×E
= 0
o Logit[Pr(G = 1 | D, E)] = β
0
+ β
D
D + β
E
E + β
D×E
D × E
The unified model incorporates the marginal effect of G, GxE, and the case only analysis into a
single test using the understanding that either D or G can serve as the outcome variable. Thus,
a 3 degree-of-freedom Wald test to look at the D- G (marginal test), G- E (case-only), and GxE
(interaction- through DxE) is possible while preserving power.
Three two- step approaches were conducted. They have been shown to improve power
while preserving the Type 1 error rate- assuming independence between the two steps
129-132
. A
threshold of significance is established (alpha= 0.05) to prioritize a group of SNPs in Step 1 that
83
will then be used in Step 2- effectively limiting the number of tests being done. Step-2 then bins
the most significant SNPs from Step-1 using decreasing α2 values with an initial bin size of 5 (for
Bin 1= (α2/2)/B; the next 2B most significant SNPs from step 1 are tested at (α2/4)/2B etc.).
SNPs are then considered significant at a genome- wide level if p < α2. Step-1 of the three
methods are outlined below:
• DG| GxE- Step 1 screens SNPs on the marginal DG at alpha= 0.05.
• GE| GxE- Step 1 screens SNPs on the marginal GE association at alpha= 0.05.
• EDGE- Step 1 screens SNPs combining the DG and GE associations (2df test) at
alpha= 0.05.
All analyses included adjustment for gender and the first five PCs. QC, principal component
analyses, G, and GxE analyses were done in a combination of PLINK
(http://zzz.bwh.harvard.edu/plink/) and R (http://www.r-project.org/) using the GxEScanR
package.

84
Results

Table 1 summarizes the descriptive characteristics of the 589 cases and 715 controls
available for analysis following QC exclusions (N= 1304). Gender distribution was similar
between cases and controls (p= 0.19). The cleft phenotype of patients included 271 cases with
cleft lip and palate (46%), 116 with cleft lip only (19.7%), and 101 with cleft palate only (17.1%).
Cases were more likely to report cooking over an open flame indoors than controls (42.1% vs
27%, p < 0.001). Cases were also more often from a rural environment (58.9% vs 42.0%, p <
0.001) and less likely to report a paternal education of secondary school or higher (56.9% cs
69.7%).
The 10 SNPS with the smallest p-values from all one-step GWIS analyses are
summarized in Table 2. None reached genome- wide significance, but suggestive (p < 5E-6)
hits were identified through the different testing methods. SNPs on chromosomes 1, 3, 7, 11,
14, and 15 were identified by the combined set of tests. For all results, when SNPs from the
same gene were clustered into the top hits, only the two SNPs with the highest p-values were
included in the table. The top 3 SNPs with the smallest p-values were identified by the 3df test
and lie in IQCH (chromosome 15) followed by one SNP in RGS7 (chromosome 1) identified by
the traditional GxE test. The Manhattan plot of the 3DF test, which produced the most
suggestive group of SNPs, is shown in Figure 2. All Manhattan plots for the other 1-step testing
methods are included in the supplementary material (Figure S3- S5). Two-step methods did not
identify any promising results, however the top 10 bins for each method are included visually in
supplemental figures (S6- S8).
The 10 SNPs with the smallest, suggestive p-values (preliminary cutoff p <1E-5) from
our primary, non-imputed GWAS analysis are summarized in Table 3 with the corresponding
Manhattan plot represented in Figure 3. Suggestive peaks were seen on chromosomes 1, 3,
12, 13, 15, 18 and 19. The two strongest p-values were found for variants located on
85
chromosome 3 and chromosome 12. We found an approximate 50% increase in risk of NSOFC
among carriers of minor allele A on rs871570 (chromosome 3; OR= 1.466, p= 6.72E-06), which
is located near XXYLT1, a gene of potential interest. A second variant, minor allele A at
rs113296466 on chromosome 12 also was associated with elevated risk of NSOFC (OR= 3.285,
p= 1.68E-05).
The 10 SNPs with the smallest p-values from our primary, imputed GWAS analysis are
summarized in Table 4 with the corresponding Manhattan plot represented in Figure 4. We did
not find any SNPs that reached statistical significance in this analysis, but suggestive peaks
were visually observed on chromosomes 1, 3, 6, 8, 11, 13, and 18. The imputed and non-
imputed GWAS identified different chromosomes of interest- with 6, 8 and 15 being added by
the imputed results while the peaks on 12 and 19 was no longer observed (1, 3, 13, and 18
were consistent). The 8 SNPs with the lowest p-values (only top two are shown in the table)
were clustered on chromosome 11 in the LOC101927709 gene. SNPs within FHOD3
(chromosome 18), LINC02815 (chromosome 1), TFDP2 (chromosome 13), and CSMD1
(chromosome 8) were also top hits in this analysis.

86
Discussion

This study was the first GWIS of NSOFC to be conducted in a large population-sampled
set of Vietnamese children; a population who experience one of the highest incidences of cleft
globally
133
. This is also the first study to look at GxCookSmoke interaction in any population.
While other studies have examined GxSmoke from prenatal, maternal cigarette smoking,
maternal tobacco smoking is less common in LMICs and exposure to smoke from cooking is a
highly prevalent, leading cause of morbidity and mortality among women in these countries
36, 37
.
We recently published an analysis of data from 7 LMICs, which included this subset of children,
showing a consistent association of maternal cook smoke exposure with NSOFC while no
association was found for maternal smoking. In the current analysis, we conducted a genome-
wide interaction study (GWIS) considering maternal exposure to cook smoke with risk of
NSOFC in a Vietnamese population. Using statistically efficient and traditional 1-step interaction
tests (GxE interaction, case only analysis, two-degree of freedom test, and the unified model),
we identified 3 SNPs in IQHC (smallest p = 2.83E-7) and 1 near GRM8 (p = 1.51E-6) that were
of interest, however they did not reach the pre-set threshold for genome-wide significance (p <
1E-8). None of the two step methods identified variants that reached statistical significance.
Our top GWIS hits, rs1459270985 and rs1243867285, were identified by the unified
model (3df test) and reside in the IQ motif containing H (IQCH), which is a testis-specific gene
whose expression was observed to be confined in the male germ line cells and is involved in
testis development and spermatogenesis
134
. Past studies have identified SNPs in the IQCH
gene that were associated with adult height
135
, age at menarche
136
, and cardiac structure/
function
137
. However, the exact mechanisms by which IQCH contributes to these processes
remain unknown and a theoretical link to GxCookSmoke interaction and/ or NSOFC is unclear.
One SNP that was included the top 10 hits (after excluding SNPs in repetitive genes) via the
case-only analysis, rs1254772630, is flanked by GRM8- a protein coding gene involved in
87
modulating neural functions and development
138
. Joubert et al. showed an epigenetic link
between GRM8 and maternal plasma folate levels during pregnancy
139
. Higher maternal folate
has been suggested to decrease the incidence of clefts and has been studied in depth with
respect to other birth defects such as spina bfida
140
. This link supports the importance of
furthering our understanding of the epigenetic relationship between GRM8 and environmental
insults during the first trimester, especially one as prevalent as exposure to cook smoke.
While the primary purpose of this analysis was the GWIS, we also evaluated both the
non-imputed and imputed GWAS. The direct genotyping data (non-imputed) GWAS produced
two areas of interest, although they did not reach genome-wide significance (p < 1e-8).
rs871570 (p=6.7E-6) is flanked by XXYLT1 (xyloside xylosyltransferase 1) on chromosome 3,
which plays a role in the epidermal growth factor (EGF) pathway and functions in Notch
signaling transduction
141
. A case report of a patient in China found a deletion in XXYLT1 as one
of two candidate genes to explain a patients Sprengel’s deformity, a congenital defect effecting
early neural and skeletal development, which has been previously linked with isolated cleft
palate
142, 143
. Disfunction in XXYLT1 could impair notch signaling, which is known to be critical
in development as well as malignant transformation
144
. rs113296466 (p=1.68E-5) on
chromosome 12 is located in USP15 and presents another area of potential interest for NSOFC.
USP15 encodes the deubquitinase (DUB) ubiquitin-specific protease 15 which functions to
catalytically remove ubiquitin from substrate proteins
145
. One of the cellular events that USP15
is involved in is the TGF-β signaling pathway by deubiquitinating the TGF-β receptor itself in
addition to other proteins in the pathway
146, 147
. Mutations various components of the
transforming growth factor β (TGF- β) signaling pathway, such as TGFBR1, TGFBR2 and
TGFB3 have been shown to be associated with the formation of NSOFC (specifically isolated
cleft palate)
148-150
. While there currently is no literature on the correlation of USP15 and NSOFC,
interaction of USP15 with key proteins in the TGF- β pathway combined with previous findings
on the role of TGF-β suggests a promising finding for future research.
88
The imputed GWAS identified two other potential genes of interest that were not
highlighted through the direct GWAS results including one SNP (rs576853071, p=1.3E-6)
residing in CSMD1 and two SNPs (rs10834414, p=7.64E-8; rs1266850282, p=1.39E-7) in
LOC101927708. Cub and Sushi Multiple Domain 1 (CSMD1) is a protein encoder gene involved
in the innate immune system response
151
. Studies have shown the association of CSMD1 and
development of head and neck cancers
152, 153
, inflammation in the central nervous system
151
,
and schizophrenia
154
. While the direct relevance of the functions of CSMD1 in NSOFC is
unclear, one previous study observed a recessive locus in the CSMD1 gene that appeared to be
implicated in NSOFC
155
. LOC101927708 is an RNA gene that is affiliated with long non-coding
RNA (lncRNA) class. IncRNAs are involved in chromatin remodeling in addition to
transcriptional and post-transcriptional regulation to silence or downregulate target genes
156
.
Disruptions to lncRNAs usually result in excessive cell proliferations and resistance to
apoptosis, thus leading to various types of cancers
157, 158
. Although there are no studies linking
LOC101927708 specifically to known diseases, there have been studies on mouse models
158-160

as well as humans
161, 162
on possible associations generally between lncRNAs and NSOFC.
While no other studies have specifically evaluated prenatal exposure to cook smoke with
genetic risk, a study by Wang et al. evaluated the interaction between total indoor air pollution
(cook smoke is globally considered the main contributor of this) with neural tube defects
(NTDs)
163
. They found an association between a SNP in the CYP1B1 gene and increased
exposure to indoor air pollution, leading to a dose-dependent risk for developing NTDs. NTDs
and orofacial clefts are both birth defects occurring in the first trimester of pregnancy resulting in
a failure of fusion. The two have been linked previously through genetic mutations, specifically
in IRF6, as well as environmental exposures such as vitamin deficiencies
164, 165
. Similarly,
multiple studies linking Gx Polycyclic aromatic hydrocarbons, which cook smoke exposure can
also be considered a large contributor of, to other congenital disorders such as congenital heart
disease
166, 167
which has also been linked to NSOFC
119
.
89
As mentioned in the introduction, 21 studies previously evaluated the role of GxE
interaction in middle to high income countries with smoke exposure from maternal tobacco use,
however few of the findings have been validated. There was no overlap in the genes identified
in the current analysis to those of the GxSmoke from tobacco. However, a cleft group from Iowa
and Denmark (n= 1244 cases, 4183 controls) found that children who carried the GSTT1- null (a
detoxification gene) genotype and had a mother who smoked (10-19 cigarettes per day for
Danish women, >15 cigarettes per day for Iowan women) were at 4.2 (95% CI: 1.4, 12.4) times
higher risk of cleft compared to unexposed Danish children and a 17.1 times higher risk
compared to unexposed Iowan children
74
. They calculated that in a population with a smoking
prevalence of 25% and GSTT1-null prevalent in approximately 15% of the population, this
interaction could attribute up to 6% of clefts. This finding has been replicated in a small study
from the Netherlands
75
. When taken into account with the global prevalence of cook smoke
exposure, which is believed to effect 3 billion women and children according to the WHO
36, 38
,
these findings strengthen the importance of further studying any potential GxCookSmoke
relationship.
In this analysis we used seven different methods to analyze the GxCookSmoke
interaction and we found that the unified model (3 df test) identified the highest number of
variants approaching statistically significant results- whereas nothing suggestive was identified
by the two- step methods. Gauderman et al. demonstrated that the unified model has equivalent
if not better power to detect a GxE hit regardless of the marginal G effects; however it is more
powerful than all other tests when there is no strong marginal G association, which was the
case in our anlayses
128
. The lack of findings from the two-step methods is something that needs
to be explored further. A future direction of this study will be further limiting the number of SNPs
that pass step-1 through eliminating SNPs clustered in the same region and/ or gene. We also
intend to re-do these analyses by cleft type to explore the potential of differing genetic and
interaction mechanisms.
90
The main limitation of this study was the sample size available. Although it was a large
case-control sample for existing cleft research, it still had limited statistical power to detect at a
genome-wide significance level. In the future, we intend to use these findings as a basis while
we collect and analyze a larger Vietnamese case and control population. The expanded sample
size will also allow us to further study other types of smoke exposure, such as maternal smoking
and environmental tobacco smoking as well as stratify by cleft phenotype. Another limitation is
that, as this is the first GWAS and GWIS done in a Vietnamese population, we don’t have the
ability to validate or further explore our findings in a Vietnamese replication cohort. One future
direction of this work will be to study the existing, available Asian NSOFC populations (such as
those from China and the Philippines) to replicate and better understand our findings. It is
notable that our study is either a similar size or larger than what has been done previously-
highlighting the difficulty of adequately powering GxE studies in cleft- especially in countries
where any type of genetics research is minimal.
This study is a promising exploratory and hypothesis generating analysis to better
understand the potential interaction effects with a minimally studied, highly prevalent
environmental risk factor, smoke exposure from cooking, and NSOFC. The use of novel,
statistically efficient methods to understand these relationships will be essential in the study of
rare diseases, such as NSOFC, especially for individuals that reside in under-represented
regions of the world. Even when care is available, NSOFC care is costly, resource intensive,
and requires lifelong maintenance to adequately minimize the effects of living with the condition.
As we continue to explore and elucidate GxE relationships and mechanisms- specifically those
relevant to individuals living in LMICs- it may be possible to mitigate environmental exposures
and prevent disease for the individuals who are at the highest risk of living with the lifelong
health complications of being born with a cleft.
91
Tables and Figures

Table 7. Demographic Characteristics of the Cases and Controls Used in the GWAS
and GWIS Analyses (N= 1304)

Control Case Overall p-
value
(N=715) (N=589) (N=1304)
Patient gender

Female 385 (53.8%) 338 (57.4%) 723 (55.4%) 0.19
Male 330 (46.2%) 250 (42.4%) 580 (44.5%)
Missing 0 (0%) 1 (0.2%) 1 (0.1%)
Patient cleft type

Lip and palate NA 271 (46.0%) 271 (20.8%) NA
Lip only NA 116 (19.7%) 116 (8.9%)
Palate only NA 101 (17.1%) 101 (7.7%)
Missing 715 (100%) 101 (17.1%) 816 (62.6%)

Cooking over an open flame
indoors

No 385 (53.8%) 242 (41.1%) 627 (48.1%) < 0.001
Yes 193 (27.0%) 248 (42.1%) 441 (33.8%)
Missing 137 (19.2%) 99 (16.8%) 236 (18.1%)

Home environment

Urban 300 (42.0%) 347 (58.9%) 647 (49.6%) < 0.001
Rural 229 (32.0%) 111 (18.8%) 340 (26.1%)
Missing 186 (26.0%) 131 (22.2%) 317 (24.3%)

Paternal Education Level

Less than secondary 54 (7.6%) 133 (22.6%) 187 (14.3%) < 0.001
Secondary or more 498 (69.7%) 335 (56.9%) 833 (63.9%)
Missing 163 (22.8%) 121 (20.5%) 284 (21.8%)

92
Table 8. Top 10 SNPs looking at the GxCookSmoke Interaction from the 1- Step
Methods (N= 1067)

Rank SNP CHR Location Gene Minor
Allele
OR Test P
1 rs1459270985 15 67499477 IQCH A 0.434 3df 2.82975E-
07
2 rs1243867285 15 67478645 IQCH CA 0.442 3df 4.03519E-
07
3 rs1243034462 1 241273417 RGS7 T 0.034 GxE 5.88213E-
07
4 rs1358607090 1 241271215 RGS7 G 0.067 GxE 8.07281E-
07
5 rs1281892226 11 3578671 LOC107987159 G 0.724 2df 9.13119E-
07
6 rs1038650214 3 16520231 RFTN1,
LINC00690
T 0.584 2df 9.55434E-
07
7 rs1195831671 3 16531527 RFTN1,
LINC00690
A 0.538 2df 1.2311E-
06
8 rs1266850282 11 3555780 LOC107987159 T 0.786 2df 1.38273E-
06
9 rs1254772630 7 126426376 LOC105375488,
GRM8
T 0.601 Case
only
1.50679E-
06
10 rs1594933663 14 56725866 RPL36AP1,
OTX2
G 0.412 2df 1.87491E-
06

Figure 7. Manhattan plot of 3df test assessing GxCookSmoke Interaction and the risk of CLP
(N=1067)

93
Table 9. Characteristics of the 10 most significant SNPs identified in the non-Imputed
GWAS analysis as associated with NSCLP adjusted for PCs 1-5 and sex.

SNP CHR Location Gene Minor
Allele
N OR P
1 rs4128463 18 36431536 RPL12P40,
RN7SKP182
A 1303 0.6694 1.84E-
06
2 rs4128465 18 36431644 RPL12P40,
RN7SKP182
A 1303 0.6694 1.84E-
06
3 rs654654 1 82368812 ADGRL2 A 1303 1.511 2.75E-
06
4 rs2026596 1 82364673 ADGRL2 A 1303 1.479 5.23E-
06
5 rs7249334 19 9158204 TRQ-TTG8-1,
OR1M4P
A 1295 0.6862 6.50E-
06
6 rs871570 3 194558537 LINC01968,
XXYLT1
A 1302 1.466 6.72E-
06
7 rs7183789 15 98442159 LINC00923,
ARRDC4
G 1303 3.819 8.76E-
06
8 rs9515055 13 110044136 MYO16,
LINC00399
A 1303 2.255 1.31E-
05
9 rs2591605 19 9196725 OR1M4P,
OR1M1
A 1302 1.436 1.51E-
05
10 rs113296466 12 62687161 USP15 A 1303 3.285 1.68E-
05

Figure 8. Manhattan Plot of non-Imputed SNPs adjusted for PCs 1-5 and sex (n= 1,288,712)

94
Table 10. Characteristics of the 10 most significant SNPs identified in the Imputed
GWAS analysis as associated with cleft (N= 1067)

Rank SNP CHR Location Gene Minor
Allele
OR P
1 rs10834414 11 3578671 LOC101927708 G 2.335 7.64E-08
2 rs1266850282 11 3555780 LOC101927708 T 2.117 1.39E-07
3 18:36460716 18 36460716 FHOD3 T 1.635 5.37E-07
4 rs1241132153 1 229055951 LINC02815 A 1.714 8.25E-07
5 3:16520231 3 16520231 RFTN1,
LINC00690
T 0.505 8.49E-07
6 rs1231230361 3 16527930 RFTN1,
LINC00690
T 0.527 8.58E-07
7 rs1043819069 18 36431643 FHOD3 T 0.632 9.70E-07
8 rs1242267867 13 113638709 TFDP1 T 1.757 1.00E-06
9 6:32965703 6 32965703 HLA-DMA,
BRD2
A 0.263 1.13E-06
10 rs576853071 8 3247693 CSMD1 C 0.568 1.33E-06

Figure 9. Manhattan Plot of Imputed GWAS Results adjusted for PCs 1-5 and sex (N= 1067)

95
Supplemental Material

Figure S5. SNP and Individual Exclusion Criteria

Initial
Individuals
Available
1496
Call rate < 80%
31
removed
Secondary call
rate < 98%
51
removed
Missing
phenotype
43
removed
Sex Check
0
removed
Duplicates and
unexplained
relatedness
67
removed
Final sample
size
1304
Initial Variants
available
1705969
Call rate < 80%
206, 866
removed
Secondary call
rate < 98%
189, 779
removed
Filtered on HWE
(p < 1e-5)
2, 204
removed
Filtered on
Minor allele
frequency < 1%
18, 408
removed
Final pre
imputation
variant count
1,288,712
96
Figure S6. Visual analysis of Top 5 Principal Components to Understand Population
Substructure in final sample data (N= 1304)

97
Figure S7. Manhattan plot of 2df test assessing GxCook-Smoke Interaction and the risk of CLP
(N=1067)

98
Figure S8. Manhattan plot of Case ONLY test assessing GxCook-Smoke Interaction and the risk
of CLP (N=1067)

99
Figure S9. Manhattan plot of Traditional GxE test assessing GxCook-Smoke Interaction and the
risk of CLP (N=1067)

100
Figure S10. G|E 2- Step procedure results (N=1067, Bins 1-10)

101
Figure S11. D|G 2- Step procedure results (N=1067, Bins 1-10)

102
Figure S12. EDGE 2- Step procedure results (N=1067, Bins 1-10)

103
Understanding the Patient- Centered Barriers to NGO Based Cleft
Surgical Care through the Integrated Health Behavior Model

Abstract
Background: Orofacial clefts are one of the most common congenital anomalies globally,
however substantial barriers exist to seeking, reaching, and receiving care. Although barriers to
surgical care have been previously described, past studies did not assess barriers across
cultural and geographic lines. In the current analysis, we evaluated behavioral constructs across
five countries to understand barriers to care with the aim of improving service delivery.
Methods: We conducted a cross-sectional study of children with cleft in Vietnam, Honduras,
Madagascar, Mexico and Nicaragua between 2014 and 2018. An interview was conducted with
the primary caregiver of each patient covering demographic, clinical, socioeconomic,
geographic, and treatment characteristics. The Integrated Health Behavior Model (IBM) was
used to conceptualize behavior. Descriptive statistics and confirmatory factor analysis were
used to assess relationships between constructs and timeliness of care.
Results: A total of 901 patients and their families were surveyed from the five countries. Five
latent constructs were included in the final framework (personal agency- structure, personal
agency- financial, perceived norm, environmental constraints, and knowledge and skills to
perform the behavior) all of which had minimal correlation (R < .3). Of the patients seeking care
for the first-time cleft phenotype, opinion of the family, perceived quality of available treatment,
mother’s employment type, and father’s employment were all significantly related to reaching
timely care before and after adjusting for country.
Conclusions: This study maps the cultural, financial and structural barriers experienced by
patients and their families to model care-seeking behavior across five diverse countries. As
public health continues to increase investment into global surgery initiatives, it is necessary to
understand these relationships and how they may differ by country to create effective and
comprehensive programs that mitigate these barriers and enable patients to reach timely care.
104
Introduction
Congenital anomalies treatable through surgery, such as cleft lip and palate, represent
1% of the global disability-adjusted life years (DALYs) burden
168
. Orofacial clefts are some of
the most common congenital anomalies in the world, with one cleft occurring per 500-700 live
births
169
. The incidence ranges widely by ethnicity (as high as 1 in 500 for Southeast Asia and
as low as 1 in 2000 for Africa). However, even where incidence is low the lack of access to
surgery often means the prevalence of unrepaired cleft remains high
12, 13
. Delayed surgical
interventions may result in mortality or lifelong health and developmental complications due to
malnutrition as well as higher incidences of misarticulation, poorer overall speech outcomes,
lasting negative psychosocial effects, and dental problems
120, 170, 171
. More information is needed
to understand the barriers that impact not only the lack of access to surgical care, but timely
access to surgical care in order to maximize care delivery and reach patients at the highest risk
of experiencing the lifelong implications of cleft disease.
Of the 5 billion people lacking access to surgical care worldwide the majority reside in
low- and- middle- income countries (LMICs) and the burden of inaccessible surgery falls most
heavily on poor and marginalized people
172, 173
. Surgery is resource intensive due to the need
for physicians with a high-level training and skills as well as access to operating suites with
electricity, running water, and expensive surgical equipment. Improved understanding of the
barriers to pediatric surgery in LMICs is specifically needed because that is where 1) the
majority of children globally reside
174
and 2) birth defects (such as cleft lip and palate) are
leading causes of mortality and morbidity for children under-five
175
. Previous literature has
shown that there is a complex relationship between the economic, cultural, and structural
barriers to providing and receiving pediatric surgical care in LMICs
174
. Barriers to pediatric
surgery have been deemed a combination of country-level factors including disease distribution,
health spending, income level and cultural practices
174, 176
as well as patient-centered barriers
including low socioeconomic status
90
, the cost of treatment
177
, lack and cost of transportation
89,
105
178, 179
, fear of surgery
178, 180
, and a lack of supplies and surgical personnel
91, 181, 182
. More region-
specific barriers have been identified through work in Sub-Saharan Africa including family
circumstances
178
and high levels of illiteracy
174
. In Central America barriers include excessive
wait times, community distrust of surgeons, and geographically inaccessibility
183
.
Although barriers to surgical care have been described in the literature, past studies that
have examined specific barriers to pediatric cleft surgery in LMICs were not designed to assess
barriers across cultural and geographic lines
90, 91
. There remains a limited understanding of
whether a clear set of unifying factors exist across LMICs or if barriers are specific to a country
or region. In the current analysis, we used existing Operation Smile data from 900 cleft lip and
palate patients and their families from five countries to evaluate constructs that explain the
barriers to surgical care with the aim of informing methods for improvement of service delivery.
This study will identify barriers to surgical care that enable providers, NGOs, and researchers
alike to develop specific and appropriate solutions improving patients’ access to timely, effective
surgical care.

106
Methods
Setting and Study Sources
This cross-sectional study was conducted with Operation Smile (OS), an internationally
recognized not-for-profit that has been providing free cleft surgery and related care to patients
for over 38 years, to better understand the barriers their patients face prior to receiving care.
Data for the current analysis represents all surveyed patients and their caregivers attending
eleven medical missions from five countries between 2014 and 2018. The Vietnamese cohort
was collected during a multi-site mission in 2014 including participants from the cities of Hanoi,
Nghe An, Hue, Ho Chi Minh, and An Giang. Similarly, data was collected from one 2016 mission
in Madagascar (Antsirabe), one 2015 and one 2016 mission in Honduras (Tegucigalpa), two
2017 missions in Nicaragua (Managua and Esteli), and one 2018 mission in Mexico (Chiapas).
All patients who presented to Operation Smile with a cleft lip and/ or palate (ICD10 35-37) were
included in the study
103
. The methods for this study have been published previously with respect
to the Vietnamese portion of the patients
90, 91
. All data collection and use was approved by
Operation Smile and the IRB at the University of Southern California.

Questionnaire Design
An in-person interview was conducted with the primary caregiver of each patient who
presented to the OS mission on demographic and clinical characteristics, family
socioeconomics, geographic factors, and cleft treatment site features that may be barriers to
accessing both cleft and general surgical care. The survey tool was developed using previously
validated questions from the Multi Country Survey Study and World Health Survey from the
health system responsiveness questionnaire from the World Health Organization (WHO)
184
.
Questions on medical and surgical history were included from the International Family Study,
which is aimed at identifying environmental and genetic determinants of cleft in LMICs
122
.
Minor modifications deemed necessary by in-country partners to ensure culture
sensitivity were identified and addressed. Barriers collected include directly measurable (cleft
107
phenotype, patient gender, income, education, employment, family characteristics, general
health and surgical care accessibility) as well as perceived barriers (opinions of the caregiver on
cultural, financial and structural elements).

Data Collection
Interviews were conducted by local, bilingual volunteers who were identified and trained
by the study team supervisor to ensure consistent data collection, professionalism, and cultural
sensitivity. Further, the study supervisor reviewed each survey for consistency and
completeness as they were completed during the mission. Interviews were conducted in a
confidential setting to the child’s primary care giver, aged 17 years or older who was
knowledgeable on the child’s medical history and household characteristics (e.g., a parent,
grandparent, or aunt/uncle). The interview with the adult was done in a private to semi-private
area depending on the confines of the mission setting. Consent was obtained prior to each
interview and patients/ families were assured that their care would not be affected by their
choice to participate. The interviews took approximately 30 minutes.

Integrated Health Behavior Model
The Integrated Health Behavior Model (IBM) was used to conceptualize behavior and
behavior change constructs in our population. The IBM was developed based on several
behavioral theories to model behavior and behavior change including the theory of reasoned
action and planned behavior
185, 186
, social cognitive theory
187, 188
, and the Health Belief Model
189
.
The major constructs of the IBM conceptual model include attitude (experiential and
instrumental), perceived norm (injunctive and descriptive), personal agency (perceived control
and self-efficacy), knowledge and skills to perform the behavior, and environmental constraints.
Latent constructs from the Operation Smile data were constructed from variables selected from
patients perceived barriers to care (yes or no). The specific data variables contributing to each
108
of the IBM constructs are shown in figure 1; no information was collected on direct intent to
receive follow-up care following the mission attendance.

Figure 10. Integrated Health Behavior Model for Barriers to Cleft Surgery

Attitude: Experiential attitude was defined as the emotional response to the idea of cleft
surgery (1 variable- fear of surgery) and instrumental attitude as beliefs about the outcome of
cleft surgery (1 variable- religious concerns).
Perceived Norm: Injunctive norm is the social pressure one feels to (not) receive cleft surgery
and was represented by 4 variables (knowledge about existing treatment, preference for
traditional healing, understanding of need by cleft phenotype, and gender of the child).
Descriptive norm is the perception of what others will think about (not) receiving cleft surgery
and was measured through 2 variables (the perceived opinion of the family and opinion of the
community).
Personal Agency: Perceived control was defined as amount of control over receiving cleft
surgery felt by the caregiver. This construct included 6 variables about financial control (lack of
109
saved money or income, loss of income while seeking care, cost of transportation to the
hospital, living expenses while seeking care, cost of surgical care, cost of food/ lodging while
seeking care) and 7 barriers about structural control (distance to a hospital that can provide cleft
surgery, no transportation available to the hospital, no doctor who can provide the treatment,
unfriendly health workers, waiting time, hospitals opening/ closing hours, and a paperwork/
administrative delay). There were no variables available for self-efficacy, which is defined as the
individual’s belief in their ability to obtain cleft surgery.
Knowledge and skills to perform the behavior: This construct includes variables that pertain
to the knowledge and skills of the caregivers to obtain cleft surgery for the patient. It was
measured through 6 variables (mothers’ education, fathers’ education, mother age at birth,
father age at birth, trust in the medical system, and perceived quality of available treatment).
Environmental constraints: This construct was defined as external constraints that may limit
the ability to obtain cleft surgery and included 8 variables (distance to the closest to healthcare
facility, time to the closest healthcare facility, distance to the OS mission site, time to the OS
mission site, income, mothers’ job, fathers’ job, and country).

Data Harmonization
Data was harmonized based on country-specific definitions and guidance by local
partners from each country to ensure consistency across the dataset. All annual income was
converted to USD based on the exchange rate at the time of interview. Mother and father
education were grouped to more or less than “secondary school”, where secondary school was
defined as their equivalent to high school in the US. Job categories for mothers and fathers
were clearly defined during the interviewer training to ensure consistency. Distance was
assessed in kilometers for all countries. All perceived barrier questions were dichotomized into
either “yes” or “no” to ensure clarity in the responses.

110
Statistical Analysis
Descriptive statistics, including proportions for categorical variables and means for
continuous variables, were constructed for the patient characteristics, socioeconomic variables,
access to care variables, and perceived barriers to care. Tests of statistical significance
included t-tests for continuous variables and chi-squared tests for categorical variables. ANOVA
and pairwise chi-squared tests were used to look for country-specific effects and were corrected
for multiple testing. Of patients presenting to a mission for the first time, timely care was defined
as <= 1 year old for cleft lip with or without cleft palate and 1.5 years old for patients with
isolated cleft palate. All data analysis used a two-sided alpha level of 0.05 (all CIs reported at
95%) and was done using RStudio.
Confirmatory factor analysis (CFA) was performed for 5 of the 6 latent constructs to
assess the appropriateness of the variables chosen. The attitude construct could not be
included in these analyses as there were only 2 variables available, which is not enough to
support factor analysis. A total of 33 variables initially contributed to the 5 constructs (Figure 1).
Each factor or construct was then assessed to identify the individual variables that loaded the
best model fit. All variables that were proposed in Figure 1 based on the framework for
perceived norm and personal agency (structural and financial) were used in the final latent
construct. Father’s age at the child’s birth was not influential in knowledge and skills needed to
perform the behavior and was removed. Inclusion of the distance to mission variable and
distance to health care facility did not improve the fit of the environmental constraint construct
and therefore were excluded. We report the Comparative Fit Index (CFI) and the Tucker- Lewis
Index (TLI) to assess the fit of the constructs. Values approaching 0.95 were considered good
as they indicate that 95% or more of the covariation in the data can be reproduced by the latent
construct. The root mean- square error of approximation (RMSEA) was also used as a measure
of fit that controls for sample size with RMSEA < 0.05 indicating a well- fit model. CFA was done
using M-Plus software.
111
Results
A total of 901 patients and their families were surveyed from the five countries: 119
(13.2%) from Honduras, 64 (7.1%) from Madagascar, 60 (6.6%) from Mexico, 225 (25.0%) from
Nicaragua, and 433 (48.1%) from Vietnam. Table 1 summarizes patient and socioeconomic
characteristics of the families. Patients were more likely to be male (54.6% vs 44.3%) and have
a cleft lip and palate (51.9%) followed by a cleft lip only (23.2%) and cleft palate only (21.1%).
The majority of mothers had less than a secondary education (52.8%) as did the fathers
(47.5%). Mothers most often reported being housewives or unemployed (35.5%) and farm or
labor workers (36.1%). Fathers most common employment was farm or labor worker (55.4%).
Across all countries, the mean annual income was $2,380 (STD $3,020) and the mean number
of individuals living in the household was 5.21 with a mean of 2.06 of those individuals
employed. Patient gender, cleft phenotype, mother’s education level, father’s education level,
annual income, and number of individuals living in the household all differed significantly by
country (p-value < .001). Only number of individuals employed in the household did not exhibit a
significant country level effect (p= 0.205).
In Table 2, a summary of the time and distance to both the closest health care facility
and the Operation Smile mission are provided. The mean distance to a health facility was 17.9
km (STD 60.1) and patients reported travel time to the closest health facility by any means of
transportation at slightly under an hour (Mean= 0.94 hour (STD 2.02)). The mission was on
average 149 km (STD 168) and 6.1 hours (STD 22.4) from the patient’s home. There was no
significant difference observed between countries for distance to the nearest healthcare facility
in KM, but there was by time to the facility (p= 0.02). The country-level difference in travel time
was largely due to the contrast between Vietnam (mean= 0.74 hrs.) and Nicaragua (mean=1.22
hrs.). Both distance and time to the mission was significantly different by country (p <0.001 and
p= 0.02, respectively). Madagascar compared to Vietnam, Honduras and Nicaragua differed by
distance (all p<0.0001), whereas only Vietnam and Nicaragua differed by time (p= 0.01).
112
The fit of the variables into each latent construct is summarized in Table 3. Personal
agency was split into two latent variables: structural and financial, which both had strong fit
statistics (CFI: 0.96, 0.99; TLI: 0.94, 0.99 for structural and financial respectively; p <0.0001 for
both). Perceived norm (CFI= 0.96; TLI= 0.93; p=0.15) and environmental constraints (CFI=
0.93; TLI= 0.88; p=0.01) had acceptable fit statistics with the consideration that the variables
were well mapped theoretically. Only knowledge and skills to perform the behavior did not reach
acceptable fit initially, but after variance adjustment was well mapped to the construct (CFI=
0.99; TLI= 0.99; p=0.05).
The correlation coefficients between the latent constructs are shown in Figure 2. The
correlation coefficients for each latent construct were small (R < .3) indicating minimal
correlation. Personal agency (structural) was weakly correlated with knowledge/ skills (R=0.33)
and personal agency financial (R=0.39). The statistics indicate that the constructs were
independent.
For the 292 patients that were seeking care for the first time, the characteristics used in
the IBM are shown in Tables 4 and 5 dichotomized by timeliness. As the question of if they had
attended a previous mission was not collected in Madagascar, they were excluded from this
analysis. Table 4 includes variables for attitude, perceived norm, and personal agency. Cleft
phenotype, opinion of the family, and waiting time to receive care were all significantly related to
timeliness of care (p <0.05). To understand if country could explain these relationships, we
further adjusted the significant findings for country. After adjustment for country, patients with an
isolated cleft palate were 64% less likely to present for timely care than those with a cleft lip
(OR= 0.36 (0.18, 0.70)). Those who perceived their families as a barrier to care were
approximately 50% less likely to present for timely care than those who didn’t after adjustment
for country (OR= 0.49 (0.28, 0.87)). A difference was no longer observed for perceived wait time
or cleft lip compared to cleft lip with palate after adjustment. Fear of treatment, patient gender,
113
cost of surgical care, availability of transportation to the hospital, and a paperwork/
administrative delay approached statistical significance (p <0.1).
Differences in knowledge and skills needed to perform the behavior and environmental
constraints are shown in Table 5. Father’s education, perceived quality of available treatment,
mother’s employment type, father’s employment type, and country were all significantly
associated with timeliness of care (p <0.05). After adjustment for country, those who perceived
quality of locally available treatment as a barrier were 48% less likely to present for timely care.
Mothers who we’re a housewife or unemployed were 2.6 times as likely to present their children
for timely care compared to mothers who were farmers or labor workers (OR= 2.61 (1.25, 5.56))
after adjustment. Similarly, fathers who were grouped in the employed- other category (not
professional/ service/ public employees) were 2.66 times as likely to have their child present for
timely care than those who were farm or labor workers regardless of country (OR= 2.66 (1.37,
5.23)).
114
Discussion
Surgery is one of the most complex services to provide to patients due to the extensive
requirements for delivery, which are all necessary and include basic and surgical infrastructure
(water, electricity, operating rooms, intensive care units, recovery rooms), highly- trained
personnel, medical equipment, and a health system that can support both the emergent and
elective needs of a population. This is the first study to map the wide variety of cultural, financial
and structural barriers that patients seeking cleft lip and/ or palate surgery experienced via the
IBM framework, which has been previously used to model care-seeking behavior
190
. Using this
structure, we found that cleft phenotype (perceived norm), opinion of family (perceived norm),
perceived quality of available treatment (knowledge and skills to perform the behavior), and
parental employment (environmental constraints) were all significantly associated with
presenting for timely care after adjusting for country of mission. As public health continues to
increase investment into global surgery initiatives, it is necessary to understand these
relationships and how they may differ by country to create effective and comprehensive
programs.
A key finding of this study is that perceived quality of available treatment was significantly
related to timeliness. Interestingly, we found that self-reported trust in the medical system and
knowledge of existing treatment availability were not strong predictors of timeliness of care. This
demonstrated that while the majority of patients reported having some general knowledge of
cleft treatment and did not have trust in the medical systems of their own countries, those who
reported that quality of care was not a barrier for them were likely to seek care sooner. Thus,
patients’ level of knowledge about the quality of cleft care- not that care itself in exists- may be
driving the timeliness in which they seek out care from NGOs. This finding is notable due to the
drastic differences in the existing, non-NGO health systems of the countries included. For
example, in Vietnam there is a wide range of available cleft care that is covered by insurance
which all citizens under 5 have at no cost. Operation Smile Vietnam missions are done almost
115
exclusively with Vietnamese healthcare teams, meaning individuals are waiting to get care from
the same providers in both situations, potentially due to their lack of comfort in the perceived
quality of the care. In Madagascar there is minimal (if any) non-NGO cleft surgery provided and
medical costs are for surgery are catastrophic for the majority of the population. Mexico is more
similar to Vietnam, whereas Honduras and Nicaragua may be closer to Madagascar, except for
the smaller size of those two countries. In a previous study in Sierra Leone, the investigators
also found a complex relationship between the perceived quality of hospitals and surgeons, and
that participants judged the quality of healthcare from word of mouth within their social
networks
191
. The relationship between perceived quality of care and timely presentation may be
closely tied to a patients’ social network and the previous experiences of those people with the
local surgical system. Another possible influence is that Operation Smile has a large presence
and good reputation within these countries, demonstrating that patients may perceive a higher
quality of care in an NGO setting based on the experience of parents reported within the
community. This barrier—public perceptions of the quality cleft care—represents something that
is actionable, relatively low-cost to address, and may be generalizable across a wide variety of
countries regardless of socioeconomic status.
Our findings also indicate that the employment status of the mother plays a significant role in
the timeliness of cleft surgery. Mothers working outside the home as farmers or laborers were
less likely to seek timely care for their child with cleft based on those assessed for timely
presentation. This is likely related to a lower economic status of the family, since the mother is
working to supplement the father’s earnings. More specifically, this may be an example that
working mothers often have less flexibility to take their children to the mission itself, which will
take them away from home for up to a week and lead to a loss of income, greatly effecting the
entire family. Education surrounding cleft care and timely surgical interventions may have more
of an impact if it is targeted at mothers, especially working mothers whose schedules are less
flexible. Increasing transportation to surgical facilities and planning missions during lulls in crop
116
harvesting or outside of typically working hours may help children with working mothers reach
care at the appropriate age. Other financial interventions, such as compensating for lost income,
that may put the family at risk of not eating, will also be a key aspect of overcoming this barrier
and ensuring those of the lowest SES receive care.
Similarly, our finding that patients with cleft palate only were more likely to present outside
the optimal window for care is consistent with past literature as cleft palate is not outwardly
visible and involves a more complex operation
90
. More education around the importance of
timely cleft palate care should be created to promote this behavior. A delay in care can lead to
lifelong speech impediments
171
, therefore we believe community messaging should stress that
palate surgery be prioritized as timeliness is even more important than corrective lip surgery. A
potential solution to maximize the number of patients with cleft palate only receiving timely care
in LMICs includes increased training of community health workers to identify cleft palate as well
as education of patients’ families on the reasons why it is so important to seek timely cleft
services.
It is important to note that while only select variables used in the IBM framework significantly
contributed to predicting timeliness of cleft care, other variables were reported as a barrier to
receiving care by at least half of the patients in the subgroup. All financial variables were
reported as a barrier by the majority of the individuals studied for timeliness, supporting the
literature that it is a universally relevant concern. It is also important to note that even though
there is a wide range of income levels within the LMICs included, these results were consistent
across income levels. For example, Vietnam has a GDP per capita of $2,566.60 whereas
Madagascar has a GDP per capita of $527.50. Although primary medical costs are provided by
the government in some countries (and are covered through NGOs) other ancillary costs still
need to be taken into account. A population-based study done in Nepal found that individuals
who required motorized transportation to reach a facility equipped to provide surgery were 66%
(OR=0.44 (0.2, 0.9)) less likely to have accessed care than those who did not
88
. A study with
117
similar aims in Malawi found that 39% of males and 59% of females lacked the financial
resources to travel the median reported time of 1- 2.5 hours to a hospital that provided
surgery
89
. These are consistent with our findings that cost of food, living expenses, wages lost,
and transportation were all barriers to our population.
Generally, the other findings in our study were similar to past studies done in other
settings (for different types of surgery) and countries, suggesting some level of generalizability
when considering the barriers to surgical care. Fear of surgery
178, 180
, the gender of the patient
90
,
family’s opinion, cost of treatment
177
, lack of transportation
89, 178, 179
, and waiting time have all
been cited in other studies assessing barriers to surgical care and were also found as self-
reported barriers in our work. We also found either parent being a farmer (a common job in
mission countries included in our survey) compared to a professional to be directly related to
timeliness, which supports past findings that socioeconomic status
90
continues to be a barrier to
surgical care, even when the care is provided by NGOs. All of these findings affirm that
actionable education and outreach could make a large impact reducing cleft barriers globally,
especially with respect to NGOs who are already covering the cost of care once patients reach
it.
The main limitation of this study and thus the use of care-seeking behavior frameworks is
that we are only sampling patients who have successfully reached care. Although this is a
limitation in most barriers to care studies, the clear guidelines around timing of cleft surgery
gives us the ability to dichotomize a portion of our patients into delayed vs timely presentation-
thus effectively exploring care-seeking behavior. A larger sample of diverse, first-time mission
patients will help confirm and expand upon these findings. We also recognize that self-reported
distance to both the mission and healthcare site may be misclassified to some extent compared
to self-reported time of travel, as the patients don’t often know the exact kilometers to the
mission site or hospital. However, the fact that time variables better mapped to the CFA
suggests that it may be a more appropriate measurement regardless. Another limitation of our
118
study is that the definition of employment in LMIC’s is much different than traditional industry
and occupation coding is completed in high income countries. Many individuals in LMICs have
informal employment or are self-employed. Therefore, those who are self-reporting unemployed
in LMICs may represent participants who are actually unable to work. This group is of special
interest and should be further explored, although very few male participants described
themselves as unemployed in our study population (0- 4.2% per country).

Conclusion
In this study, we attempted to better understand caregiver’s perspective of the barriers to
receiving timely surgical care for their children with cleft lip and/ or palate in five LMICs. We
found that the caregiver’s perception of quality of care and the mother’s employment status
represent key barriers across a diverse set of LMICs that, if intervened upon could, improve
patients’ access to timely care. NGOs should work with local health systems to improve health
education and messaging around cleft surgery to ensure that caregivers understand the life-long
benefits their child will receive through timely intervention. Innovative ways to do this, such as
through the parents of past patients acting as community advocates, will continue to improve
the likelihood that all patients can reach and receive timely care. Cleft NGOs specifically, in
combination with their local partners, are in a unique position to lead the way in these efforts
due to their long histories and presence around the world. Efforts such as these, in both
identifying and then intervening upon factors that prevent patients from successfully reaching
timely care, can serve as an example for other surgically treatable conditions with similar
obstacles to overcome. We believe this is a key example of one of many initiatives needed to
continue understanding and simultaneously addressing a wide variety of barriers to surgery for
the 5 billion people who still lack access to care.

119
Tables and Figures

Table 11. Descriptive Characteristics of the patients and their families (N=901)

Honduras

a

Madagascar
b

Mexico
c
Nicaragu
a
d

Vietna
m
e

Overall P-
value
*
(N=119) (N=64) (N=60) (N=225) (N=433
)
(N=901
)

Patient gender

Female 39
(32.8%)
28 (43.8%) 40
(66.7%)
109
(48.4%)
183
(42.3%)
399
(44.3%)
<
0.001
Male 78
(65.5%)
35 (54.7%) 18
(30.0%)
113
(50.2%)
248
(57.3%)
492
(54.6%)

Patient cleft
phenotype**

Lip and
Palate
62
(52.1%)
41 (64.1%) 26
(43.3%)
126
(56.0%)
213
(49.2%)
468
(51.9%)
<
0.001
Lip Only 25
(21.0%)
12 (18.8%) 11
(18.3%)
66
(29.3%)
95
(21.9%)
209
(23.2%)

Palate
Only
17
(14.3%)
8 (12.5%) 19
(31.7%)
29
(12.9%)
117
(27.0%)
190
(21.1%)

Mothers
education**

None 18
(15.1%)
21 (32.8%) 7 (11.7%) 37
(16.4%)
37
(8.5%)
120
(13.3%)
<
0.001
Less than
Secondary
58
(48.7%)
22 (34.4%) 33
(55.0%)
96
(42.7%)
147
(33.9%)
356
(39.5%)

Some
secondary or
higher
24
(20.2%)
7 (10.9%) 16
(26.7%)
43
(19.1%)
211
(48.7%)
301
(33.4%)

Fathers
education

None 14
(11.8%)
19 (29.7%) 4 (6.7%) 41
(18.2%)
33
(7.6%)
111
(12.3%)
<
0.001
Less than
Secondary
56
(47.1%)
19 (29.7%) 29
(48.3%)
73
(32.4%)
140
(32.3%)
317
(35.2%)

Some
secondary or
higher
18
(15.1%)
7 (10.9%) 13
(21.7%)
35
(15.6%)
201
(46.4%)
274
(30.4%)

Mothers
employment

Farm or
Labor Worker
7 (5.9%) 30 (46.9%) 5 (8.3%) 2 (0.9%) 281
(64.9%)
325
(36.1%)
<
0.001
Housewife/
unemployed
87
(73.1%)
7 (10.9%) 0 (0%) 172
(76.4%)
54
(12.5%)
320
(35.5%)

Profession
al/ Service/
Public
Employee
8 (6.7%) 8 (12.5%) 6 (10.0%) 17 (7.6%) 39
(9.0%)
78
(8.7%)

Other 15
(12.6%)
12 (18.8%) 3 (5.0%) 27
(12.0%)
52
(12.0%)
109
(12.1%)

120
Fathers
employment

Farm or
Labor Worker
52
(43.7%)
35 (54.7%) 25
(41.7%)
106
(47.1%)
281
(64.9%)
499
(55.4%)
<
0.001
Profession
al/ Service/
Public
Employee
18
(15.1%)
12 (18.8%) 17
(28.3%)
16 (7.1%) 57
(13.2%)
120
(13.3%)

Unemploye
d
5 (4.2%) 0 (0%) 1 (1.7%) 5 (2.2%) 4
(0.9%)
15
(1.7%)

Other 23
(19.3%)
8 (12.5%) 4 (6.7%) 61
(27.1%)
74
(17.1%)
170
(18.9%)

Annual
income (USD,
conversion
rate at time of
collection)

Mean (SD) $ 3339.31
(±
4475.89)
b

$ 624.12 (±
825.55)
a,c,d,e

$ 2394.89
(±2690.02
)
b

$
2748.22
(±
3381.48)
b

$
2418.8
4 (±
2438.4
2)
b

$
2520.2
9 (±
3033.0
7)
<
0.001
Number of
individuals
living in the
household

Mean (SD) 5.69 (±
1.99)
e

NA (± NA) 5.40 (±
2.64)
5.70 (±
2.41)
e

4.81 (±
1.43)
a,d

5.212
(± 1.96)
<
0.001
Number of
individuals
employed and
living in the
household

Mean (SD) 1.84 (±
1.62)
NA (± NA) 2.07 (±
1.57)
2.16 (±
1.48)
2.06 (±
0.97)
2.06 (±
1.27)
0.21
*P-value computed using chi-squared test or ANOVA
**Secondary education is defined as the schooling that comes directly post primary school
(equivalent to approximately 5
th
grade in the US)
For each row, means with different letters (a-e) across the countries are statistically significantly different
from one another using Tukey multiple comparison procedure (P<0.05).

121
Table 12. Health Care Accessibility for patients and their families (N=901)

Honduras
a

Madagascar
b

Mexico
c
Nicaragua
d
Vietnam
e
Overall P-
Value
(N=119) (N=64) (N=60) (N=225) (N=433) (N=901)

Distance
to any
healthcar
e facility
(km)

Mean
(SD)
3.23 (±
3.34)
NA (± NA) 8.86 (±
19.27)
9.04 (±
53.78)
21.17 (±
63.74)
17.89 (±
60.14)
0.19
Time to
any
healthcar
e facility
(hr)

Mean
(SD)
1.09 (±
1.48)
NA (± NA) 1.39 (±
3.97)
1.22 (± 2.83) 0.74 (±
1.21)
0.94 (±
2.02)
0.02
Distance
to the OS
mission
facility
(km)

Mean
(SD)
92.82
(± 124.04)
b

362.48
(± 309.63)
a,d,e

237.43
(±
281.32)
94.92
(± 80.52)
b

133.39
(± 136.55)
b

148.88
(±
168.48)
<
0.01
Time to
the OS
mission
facility
(hr)

Mean
(SD)
5.15
(± 7.00)
10.49
(± 15.84)
4.52
(± 9.06)
10.17
(± 43.12)
e

4.22
(± 4.67)
d

6.09
(± 22.45)
0.02
*P-value computed using chi-squared test or ANOVA
For each row, means with different letters (a-e) across the countries are statistically significantly different
from one another using Tukey multiple comparison procedure (P<0.05).

122
Table 13. Fit Indexes for Integrated Health Behavior Model Latent Constructs

Latent Construct CFI TLI Prob RMSEA
<= 0.05
Perceived norm 0.95 0.93 0.15
Personal agency (structural) 0.96 0.94 < 0.01
Personal agency (financial) 0.99 0.99 0.01
Knowledge and skills to perform the
behavior
0.99 0.99 0.05
Environmental constraints 0.93 0.88 0.01

123
Table 14. Attitude, perceived norm, and personal agency variables by timeliness of care
(N= 292)

Delayed
presentation
Timely
Presentation
Overall P-value* P-
value**
(N=187) (N=98) (N=292)

Fear of treatment

0.076

Not a barrier 171 (91.4%) 95 (96.9%) 268 (91.8%)

0.073
Barrier 13 (7.0%) 2 (2.0%) 16 (5.5%)

Religious concerns

0.809

Not a barrier 108 (57.8%) 54 (55.1%) 164 (56.2%)

Barrier 0 (0%) 2 (2.0%) 5 (1.7%)

Male

0.059

Not a barrier 94 (50.3%) 38 (38.8%) 133 (45.5%)

0.093
Barrier 92 (49.2%) 60 (61.2%) 155 (53.1%)

Cleft phenotype

<.001

Lip and Palate 53 (28.3%) 36 (36.7%) 90 (30.8%)

Lip Only 46 (24.6%) 43 (43.9%) 89 (30.5%)

0.404
Palate Only 80 (42.8%) 19 (19.4%) 100 (34.2%)

0.003
Opinion of family

0.024

Not a barrier 106 (56.7%) 67 (68.4%) 176 (60.3%)

0.014
Barrier 81 (43.3%) 28 (28.6%) 110 (37.7%)

Missing 0 (0%) 3 (3.1%) 6 (2.1%)

Opinion of community

0.361

Not a barrier 149 (79.7%) 82 (83.7%) 234 (80.1%)

Barrier 37 (19.8%) 15 (15.3%) 53 (18.2%)

Knowledge of existing
treatment

0.116

Not a barrier 152 (81.3%) 87 (88.8%) 243 (83.2%)

Barrier 27 (14.4%) 8 (8.2%) 35 (12.0%)

Preference for Traditional
Healing

0.36

Not a barrier 75 (40.1%) 44 (44.9%) 121 (41.4%)

Barrier 107 (57.2%) 48 (49.0%) 157 (53.8%)

Lack of saved money or
income

0.234

Not a barrier 48 (25.7%) 32 (32.7%) 81 (27.7%)

Barrier 138 (73.8%) 64 (65.3%) 205 (70.2%)

Loss of income while seeking
care

0.246

Not a barrier 67 (35.8%) 42 (42.9%) 110 (37.7%)

Barrier 118 (63.1%) 55 (56.1%) 176 (60.3%)

Cost of transportation to the
hospital

0.525

Not a barrier 63 (33.7%) 36 (36.7%) 99 (33.9%)

Barrier 124 (66.3%) 60 (61.2%) 188 (64.4%)

124
Living expenses while seeking
care

0.738

Not a barrier 67 (35.8%) 33 (33.7%) 102 (34.9%)

Barrier 119 (63.6%) 64 (65.3%) 185 (63.4%)

Cost of surgical care

0.061

Not a barrier 60 (32.1%) 42 (42.9%) 105 (36.0%)

Barrier 125 (66.8%) 54 (55.1%) 180 (61.6%)

0.127
Cost of food/ lodging while
seeking care

0.426

Not a barrier 65 (34.8%) 38 (38.8%) 105 (36.0%)

Barrier 122 (65.2%) 58 (59.2%) 182 (62.3%)

Distance to hospital that can
provide cleft surgery

0.619

Not a barrier 68 (36.4%) 38 (38.8%) 108 (37.0%)

Barrier 118 (63.1%) 58 (59.2%) 178 (61.0%)

No transportation available to
the hospital

0.073

Not a barrier 68 (36.4%) 45 (45.9%) 115 (39.4%)

Barrier 115 (61.5%) 48 (49.0%) 165 (56.5%)

0.124
No doctor who can provide the
treatment

0.291

Not a barrier 35 (18.7%) 23 (23.5%) 60 (20.5%)

Barrier 151 (80.7%) 72 (73.5%) 225 (77.1%)

Unfriendly health workers

0.88

Not a barrier 70 (37.4%) 38 (38.8%) 109 (37.3%)

Barrier 113 (60.4%) 59 (60.2%) 175 (59.9%)

Waiting time

0.046

Not a barrier 80 (42.8%) 53 (54.1%) 135 (46.2%)

0.106
Barrier 105 (56.1%) 42 (42.9%) 149 (51.0%)

Hospitals opening/ closing
hours

0.176

Not a barrier 103 (55.1%) 60 (61.2%) 166 (56.8%)

Barrier 83 (44.4%) 34 (34.7%) 118 (40.4%)

A paperwork/ administrative
delay

0.097

Not a barrier 80 (42.8%) 52 (53.1%) 133 (45.5%)

0.137
Barrier 103 (55.1%) 44 (44.9%) 150 (51.4%)

*P-value from chi-squared or paired t-test
** P-value from logistic regression adjusted for country

125

Delayed
Presentation
Timely
Presentation
Overall

(N=187) (N=98) (N=292) P-
Value
P-
Value
Mothers age at patients
birth

Mean (SD) 26.1 (6.19) 26.1 (5.18) 26.1 (5.82) 0.99 0.757
Mothers education

0.12

None 20 (10.7%) 11 (11.2%) 32 (11.0%)

Less than Secondary 67 (35.8%) 27 (27.6%) 97 (33.2%)

Some secondary or
higher
74 (39.6%) 54 (55.1%) 128 (43.8%)

0.567
Fathers education

0.009

None 18 (9.6%) 10 (10.2%) 28 (9.6%)

Less than Secondary 65 (34.8%) 20 (20.4%) 86 (29.5%)

0.31
Some secondary or
higher
68 (36.4%) 54 (55.1%) 123 (42.1%)

0.223
Trust in the medical system

Not a barrier 72 (38.5%) 39 (39.8%) 113 (38.7%) 0.88

Barrier 111 (59.4%) 56 (57.1%) 169 (57.9%)

Perceived quality of
available treatment

Not a barrier 88 (47.1%) 60 (61.2%) 150 (51.4%) 0.017 0.017
Barrier 97 (51.9%) 36 (36.7%) 135 (46.2%)

Mothers employment

0.024

Farm or Labor Worker 87 (46.5%) 36 (36.7%) 125 (42.8%)

Housewife/ unemployed 39 (20.9%) 37 (37.8%) 78 (26.7%)

0.0115
Other 14 (7.5%) 12 (12.2%) 26 (8.9%)

0.061
Professional/ Service/
Public Employee
24 (12.8%) 10 (10.2%) 34 (11.6%)

0.882
Fathers employment

0.014

Farm or Labor Worker 111 (59.4%) 49 (50.0%) 162 (55.5%)

Professional/ Service/
Public Employee
32 (17.1%) 14 (14.3%) 46 (15.8%)

0.741
Other 22 (11.8%) 27 (27.6%) 49 (16.8%)

0.004
Unemployed 5 (2.7%) 2 (2.0%) 8 (2.7%)

0.934
Annual income (USD,
conversion rate at time of
collection)

Mean (SD) 2800 (3190) 2890 (4530) 2820 (3680) 0.869

Country

Madagascar 0 (0%) 0 (0%) 0 (0%) 0.046

Vietnam 126 (67.4%) 67 (68.4%) 195 (66.8%)

126
Table 15. Knowledge and skills needed to perform the behavior and environmental
constraints variables by timeliness of care (N= 292)

*P-value from chi-squared or paired t-test
** P-value from logistic regression adjusted for country

Honduras 23 (12.3%) 13 (13.3%) 37 (12.7%)

0.871
Nicaragua 12 (6.4%) 13 (13.3%) 28 (9.6%)

0.096
Mexico 26 (13.9%) 5 (5.1%) 32 (11.0%)

0.046
Hours to the closest health
care facility

Mean (SD) 0.933 (1.66) 0.993 (3.25) 0.956 (2.32) 0.506

Hours to the mission site

Mean (SD) 5.25 (7.59) 4.29 (9.76) 4.89 (8.35) 0.536

127
Figure 11. Intercorrelation between Latent Factors

128
Conclusion

Even when care is available, NSOFC care is costly, resource intensive, and requires
lifelong maintenance to adequately minimize the effects of living with the condition. This
highlights the importance of NGOs creating intentional programs to reach those who are least
likely to have access to care as well as strategically working towards the long-term goal of
prevention in these settings. With this in mind, the overarching question of my dissertation was
how to mitigate the burden of cleft disease from two angles- 1) what knowledge will help us to
prevent patients from being born with a cleft in the future and 2) how do we best reach and
serve those who are already living with disease.
The first paper demonstrated a 50% increase in cleft risk for mothers reporting cooking
over an open flame indoors compared to controls in a diverse group of LMICs. This exposure
has only been assessed in two other studies even though it is the main contributor of smoke
exposure in LMICs- affecting up to 80% of women. This information can inform public health
interventions and education to potentially prevent disease in populations where care is sparse,
and children are most likely to feel the detrimental, lifelong medical and social effects of cleft.
With this information, Operation Smile has already started to partner with clean burning
cookstove organizations and intends to continue to do so to improve the health of their patients
and families. Future directions of this work include both the expansion of education programs
and clean cook-stove interventions to mitigate this potential risk factor as well as further studies
to dive deeper into this finding (i.e. quantity of exposure, type of fuel being used, etc.). My next
steps also include conducting similar analyses into other environmental exposures, such as
pesticide exposure and maternal vitamin use, that we are able to study through the IFS
database.
This project led to my second question- can we take the environmental findings a step
further to understand the interaction with genetics and if there may be susceptible

129
subpopulations affected by exposure to cook smoke. This was done with the end goal that these
findings could help us prioritize the best places for our environmental interventions as well as
contribute new knowledge to the existing cleft genetic literature from underserved populations. I
successfully conducted both a genome-wide interaction scan (GWIS) and genome-wide
analysis study (GWAS) to explore potential gene-cook smoke interactions and genetic variants
of interest. The study identified multiple SNPs in genes of interest for further evaluation and was
is a promising exploratory and hypothesis generating analysis to better understand potential
interaction effects with a minimally studied, highly prevalent environmental risk factor, smoke
exposure from cooking, and NSOFC. The next steps are to increase our sample size in Vietnam
to gain better statistical power, sequence data from other IFS countries to test our hypothesis in
different populations, validate our findings with other existing Asian populations, and perform
functional studies to delve further into these findings as well as other potential mechanisms.
Finally, I had concerns about a related but distinct question to address the opposite side
of the coin- those currently living with cleft who are unable to reach care. Through my analysis
of the caregiver’s perspective of the barriers to receiving timely surgical care, I found that their
perception of quality of care and the mother’s employment status represent key barriers across
a diverse set of LMICs. If intervened upon, these could have the potential to directly improve
patients’ access to timely care. These findings support the idea that organizations, like
Operation Smile, should continue to create and utilize innovative methods to improve the
likelihood that all patients can reach and receive timely care. I plan to work with Operation Smile
to encourage both education initiatives to ensure families are comfortable with quality of
services as well as financial programs addressing lost wages and/ or timing missions around
growing seasons to overcome mother’s employment barriers. In the next phase of my career, I
also hope to expand the study to a larger group of individuals seeking care for the first time to
ask more specific questions of that subgroup and directly measure both intention and behavior.

130
In summary, by simultaneously studying populations with a risk set that has not been
explored by current research and better understanding barriers to receiving surgical care
organizations, such as Operation Smile, can work to make considerable progress towards
limiting the burden of disease. My dissertation addresses this concern from multiple angles and
future work will allow me to expand on my findings and impact the lives of patients living with
cleft lip and/ or palate in LMICs.

131
References
1. Meara JG, Leather AJ, Hagander L, et al. Global Surgery 2030: evidence and solutions
for achieving health, welfare, and economic development. Lancet. Aug 8 2015;386(9993):569-
624. doi:10.1016/S0140-6736(15)60160-X

2. Mossey PA, Little J, Munger RG, Dixon MJ, Shaw WC. Cleft lip and palate. Lancet. Nov
21 2009;374(9703):1773-85. doi:10.1016/S0140-6736(09)60695-4

3. Shaw WC. 'Orthodontics & occlusal management'. Br Dent J. Aug 20 1994;177(4):120-1.

4. Neilson DE, Brunger JW, Heeger S, Bamshad M, Robin NH. Mixed clefting type in
Rapp-Hodgkin syndrome. American journal of medical genetics. Apr 1 2002;108(4):281-4.

5. Sperber G. Formation of the primary and secondary palate. In: Wyszynski D, ed. Cleft lip
and palate: from origin to treatment. Oxford University Press; 2002:5-24.

6. Mossey PC, A; Eduardo, E. Global registry and database on craniofacial anomalies :
report of a WHO Registry Meeting on Craniofacial Anomalies. World Health Organization; 2003:

7. Calzolari E, Pierini A, Astolfi G, Bianchi F, Neville AJ, Rivieri F. Associated anomalies in
multi-malformed infants with cleft lip and palate: An epidemiologic study of nearly 6 million births
in 23 EUROCAT registries. Am J Med Genet A. Mar 15 2007;143A(6):528-37.
doi:10.1002/ajmg.a.31447

8. Rittler M, Lopez-Camelo J, Castilla EE. Sex ratio and associated risk factors for 50
congenital anomaly types: clues for causal heterogeneity. Birth Defects Res A Clin Mol Teratol.
Jan 2004;70(1):13-9. doi:10.1002/bdra.10131

9. Leslie EJ, Marazita ML. Genetics of cleft lip and cleft palate. Am J Med Genet C Semin
Med Genet. Nov 2013;163C(4):246-58. doi:10.1002/ajmg.c.31381

10. Group IW. Prevalence at birth of cleft lip with or without cleft palate: data from the
International Perinatal Database of Typical Oral Clefts (IPDTOC). Cleft Palate Craniofac J. Jan
2011;48(1):66-81. doi:10.1597/09-217

11. Gundlach KK, Maus C. Epidemiological studies on the frequency of clefts in Europe and
world-wide. J Craniomaxillofac Surg. Sep 2006;34 Suppl 2:1-2. doi:10.1016/S1010-
5182(06)60001-2

12. Cooper ME, Ratay JS, Marazita ML. Asian oral-facial cleft birth prevalence. Cleft Palate
Craniofac J. Sep 2006;43(5):580-9. doi:10.1597/05-167

13. Kadir A, Mossey PA, Blencowe H, et al. Systematic Review and Meta-Analysis of the
Birth Prevalence of Orofacial Clefts in Low- and Middle-Income Countries. Cleft Palate
Craniofac J. Sep 2017;54(5):571-581. doi:10.1597/15-221

14. Panamonta V, Pradubwong S, Panamonta M, Chowchuen B. Global Birth Prevalence of
Orofacial Clefts: A Systematic Review. J Med Assoc Thai. Aug 2015;98 Suppl 7:S11-21.

132
15. Hoang T, Nguyen DT, Nguyen PV, et al. External birth defects in Southern Vietnam: a
population-based study at the grassroots level of health care in Binh Thuan Province. BMC
Pediatr. Apr 30 2013;13:67. doi:10.1186/1471-2431-13-67

16. Rakotoarison RA, Rakotoarivony AE, Rabesandratana N, et al. Cleft lip and palate in
Madagascar 1998-2007. Br J Oral Maxillofac Surg. Jul 2012;50(5):430-4.
doi:10.1016/j.bjoms.2011.06.004

17. Mbuyi-Musanzayi S, Kayembe TJ, Kashal MK, et al. Non-syndromic cleft lip and/or cleft
palate: Epidemiology and risk factors in Lubumbashi (DR Congo), a case-control study. J
Craniomaxillofac Surg. Jul 2018;46(7):1051-1058. doi:10.1016/j.jcms.2018.05.006

18. Murray JC, Daack-Hirsch S, Buetow KH, et al. Clinical and epidemiologic studies of cleft
lip and palate in the Philippines. Cleft Palate Craniofac J. Jan 1997;34(1):7-10.
doi:10.1597/1545-1569_1997_034_0007_caesoc_2.3.co_2

19. Genisca AE, Frias JL, Broussard CS, et al. Orofacial clefts in the National Birth Defects
Prevention Study, 1997-2004. Am J Med Genet A. Jun 2009;149A(6):1149-58.
doi:10.1002/ajmg.a.32854

20. Lei RL, Chen HS, Huang BY, et al. Population-based study of birth prevalence and
factors associated with cleft lip and/or palate in Taiwan 2002-2009. PloS one.
2013;8(3):e58690. doi:10.1371/journal.pone.0058690

21. Xuan Z, Zhongpeng Y, Yanjun G, et al. Maternal active smoking and risk of oral clefts: a
meta-analysis. Oral Surg Oral Med Oral Pathol Oral Radiol. Dec 2016;122(6):680-690.
doi:10.1016/j.oooo.2016.08.007

22. Lebby KD, Tan F, Brown CP. Maternal factors and disparities associated with oral clefts.
Ethn Dis. Winter 2010;20(1 Suppl 1):S1-146-9.

23. Zhang B, Jiao X, Mao L, Xue J. Maternal cigarette smoking and the associated risk of
having a child with orofacial clefts in China: a case-control study. J Craniomaxillofac Surg. Jul
2011;39(5):313-8. doi:10.1016/j.jcms.2010.07.005

24. Honein MA, Rasmussen SA, Reefhuis J, et al. Maternal smoking and environmental
tobacco smoke exposure and the risk of orofacial clefts. Epidemiology (Cambridge, Mass). Mar
2007;18(2):226-33. doi:10.1097/01.ede.0000254430.61294.c0

25. Lie RT, Wilcox AJ, Taylor J, et al. Maternal smoking and oral clefts: the role of
detoxification pathway genes. Epidemiology (Cambridge, Mass). Jul 2008;19(4):606-15.
doi:10.1097/EDE.0b013e3181690731

26. Ly S, Burg ML, Ihenacho U, et al. Paternal Risk Factors for Oral Clefts in Northern
Africans, Southeast Asians, and Central Americans. Int J Environ Res Public Health. Jun 19
2017;14(6)doi:10.3390/ijerph14060657

27. Hao Y, Tian S, Jiao X, et al. Association of Parental Environmental Exposures and
Supplementation Intake with Risk of Nonsyndromic Orofacial Clefts: A Case-Control Study in
Heilongjiang Province, China. Nutrients. Aug 27 2015;7(9):7172-84. doi:10.3390/nu7095328

133
28. Krapels IP, Zielhuis GA, Vroom F, et al. Periconceptional health and lifestyle factors of
both parents affect the risk of live-born children with orofacial clefts. Birth Defects Res A Clin
Mol Teratol. Aug 2006;76(8):613-20. doi:10.1002/bdra.20285

29. Shaw GM, Wasserman CR, Lammer EJ, et al. Orofacial clefts, parental cigarette
smoking, and transforming growth factor-alpha gene variants. American journal of human
genetics. Mar 1996;58(3):551-61.

30. Ebadifar A, Hamedi R, KhorramKhorshid HR, Kamali K, Moghadam FA. Parental
cigarette smoking, transforming growth factor-alpha gene variant and the risk of orofacial cleft in
Iranian infants. Iran J Basic Med Sci. Apr 2016;19(4):366-73.

31. Sabbagh HJ, Hassan MH, Innes NP, Elkodary HM, Little J, Mossey PA. Passive
smoking in the etiology of non-syndromic orofacial clefts: a systematic review and meta-
analysis. PloS one. 2015;10(3):e0116963. doi:10.1371/journal.pone.0116963

32. Kummet CM, Moreno LM, Wilcox AJ, et al. Passive Smoke Exposure as a Risk Factor
for Oral Clefts-A Large International Population-Based Study. American journal of epidemiology.
May 1 2016;183(9):834-41. doi:10.1093/aje/kwv279

33. Pi X, Li Z, Jin L, et al. Secondhand smoke during the periconceptional period increases
the risk for orofacial clefts in offspring. Paediatric and perinatal epidemiology. Sep
2018;32(5):423-427. doi:10.1111/ppe.12497

34. Liu Y, Wang B, Li Z, Zhang L, Liu J, Ren A. Indoor air pollution and the risk of orofacial
clefts in a rural population in Shanxi province, China. Birth Defects Res A Clin Mol Teratol. Aug
2016;106(8):708-15. doi:10.1002/bdra.23522

35. Lim SS, Vos T, Flaxman AD, et al. A comparative risk assessment of burden of disease
and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990-2010: a
systematic analysis for the Global Burden of Disease Study 2010. Lancet. Dec 15
2012;380(9859):2224-60. doi:10.1016/S0140-6736(12)61766-8

36. Martin WJ, 2nd, Glass RI, Balbus JM, Collins FS. Public health. A major environmental
cause of death. Science. Oct 14 2011;334(6053):180-1. doi:10.1126/science.1213088

37. WHO. Quantifying Environmental Health Impacts: Global Estimates of Burden of
Disease Caused by Environmental Risks. WHO. 2018.
http://www.who.int/quantifying_ehimpacts/global/globalair2004/en/

38. Langbein J. Firewood, smoke and respiratory diseases in developing countries-The
neglected role of outdoor cooking. PloS one. 2017;12(6):e0178631.
doi:10.1371/journal.pone.0178631

39. Sehgal M, Rizwan SA, Krishnan A. Disease burden due to biomass cooking-fuel-related
household air pollution among women in India. Glob Health Action. 2014;7:25326.
doi:10.3402/gha.v7.25326

40. McCracken JP, Wellenius GA, Bloomfield GS, et al. Household Air Pollution from Solid
Fuel Use: Evidence for Links to CVD. Glob Heart. Sep 2012;7(3):223-34.
doi:10.1016/j.gheart.2012.06.010

134

41. Pope DP, Mishra V, Thompson L, et al. Risk of low birth weight and stillbirth associated
with indoor air pollution from solid fuel use in developing countries. Epidemiologic reviews.
2010;32:70-81. doi:10.1093/epirev/mxq005

42. Amegah AK, Quansah R, Jaakkola JJ. Household air pollution from solid fuel use and
risk of adverse pregnancy outcomes: a systematic review and meta-analysis of the empirical
evidence. PloS one. 2014;9(12):e113920. doi:10.1371/journal.pone.0113920

43. Wong FK, Hagg U. An update on the aetiology of orofacial clefts. Hong Kong Med J. Oct
2004;10(5):331-6.

44. Wyszynski DF, Zeiger J, Tilli MT, Bailey-Wilson JE, Beaty TH. Survey of genetic
counselors and clinical geneticists regarding recurrence risks for families with nonsyndromic
cleft lip with or without cleft palate. American journal of medical genetics. Sep 23
1998;79(3):184-90.

45. Grosen D, Chevrier C, Skytthe A, et al. A cohort study of recurrence patterns among
more than 54,000 relatives of oral cleft cases in Denmark: support for the multifactorial
threshold model of inheritance. J Med Genet. Mar 2010;47(3):162-8.
doi:10.1136/jmg.2009.069385

46. Sivertsen A, Wilcox AJ, Skjaerven R, et al. Familial risk of oral clefts by morphological
type and severity: population based cohort study of first degree relatives. BMJ. Feb 23
2008;336(7641):432-4. doi:10.1136/bmj.39458.563611.AE

47. Kondo S, Schutte BC, Richardson RJ, et al. Mutations in IRF6 cause Van der Woude
and popliteal pterygium syndromes. Nat Genet. Oct 2002;32(2):285-9. doi:10.1038/ng985

48. Burdick AB, Bixler D, Puckett CL. Genetic analysis in families with van der Woude
syndrome. J Craniofac Genet Dev Biol. 1985;5(2):181-208.

49. Zucchero TM, Cooper ME, Maher BS, et al. Interferon regulatory factor 6 (IRF6) gene
variants and the risk of isolated cleft lip or palate. N Engl J Med. Aug 19 2004;351(8):769-80.
doi:10.1056/NEJMoa032909

50. Rahimov F, Marazita ML, Visel A, et al. Disruption of an AP-2alpha binding site in an
IRF6 enhancer is associated with cleft lip. Nat Genet. Nov 2008;40(11):1341-7.
doi:10.1038/ng.242

51. Houdayer C, Bonaiti-Pellie C, Erguy C, et al. Possible relationship between the van der
Woude syndrome (vWS) locus and nonsyndromic cleft lip with or without cleft palate (NSCL/P).
American journal of medical genetics. Nov 15 2001;104(1):86-92.

52. Moreno LM, Mansilla MA, Bullard SA, et al. FOXE1 association with both isolated cleft
lip with or without cleft palate, and isolated cleft palate. Hum Mol Genet. Dec 15
2009;18(24):4879-96. doi:10.1093/hmg/ddp444

53. Ludwig KU, Bohmer AC, Rubini M, et al. Strong association of variants around FOXE1
and orofacial clefting. J Dent Res. Apr 2014;93(4):376-81. doi:10.1177/0022034514523987

135
54. Wattanawong K, Rattanasiri S, McEvoy M, Attia J, Thakkinstian A. Association between
IRF6 and 8q24 polymorphisms and nonsyndromic cleft lip with or without cleft palate:
Systematic review and meta-analysis. Birth Defects Res A Clin Mol Teratol. Sep
2016;106(9):773-88. doi:10.1002/bdra.23540

55. Birnbaum S, Ludwig KU, Reutter H, et al. Key susceptibility locus for nonsyndromic cleft
lip with or without cleft palate on chromosome 8q24. Nat Genet. Apr 2009;41(4):473-7.
doi:10.1038/ng.333

56. Leslie EJ, Carlson JC, Shaffer JR, et al. A multi-ethnic genome-wide association study
identifies novel loci for non-syndromic cleft lip with or without cleft palate on 2p24.2, 17q23 and
19q13. Hum Mol Genet. Jul 1 2016;25(13):2862-2872. doi:10.1093/hmg/ddw104

57. Grant SF, Wang K, Zhang H, et al. A genome-wide association study identifies a locus
for nonsyndromic cleft lip with or without cleft palate on 8q24. The Journal of pediatrics. Dec
2009;155(6):909-13. doi:10.1016/j.jpeds.2009.06.020

58. Ludwig KU, Wahle P, Reutter H, et al. Evaluating eight newly identified susceptibility loci
for nonsyndromic cleft lip with or without cleft palate in a Mesoamerican population. Birth
Defects Res A Clin Mol Teratol. Jan 2014;100(1):43-7. doi:10.1002/bdra.23209

59. Hikida M, Tsuda M, Watanabe A, et al. No evidence of association between 8q24 and
susceptibility to nonsyndromic cleft lip with or without palate in Japanese population. Cleft
Palate Craniofac J. Nov 2012;49(6):714-7. doi:10.1597/10-242

60. Xu MY, Deng XL, Tata LJ, et al. Case-control and family-based association studies of
novel susceptibility locus 8q24 in nonsyndromic cleft lip with or without cleft palate in a Southern
Han Chinese population located in Guangdong Province. DNA and cell biology. May
2012;31(5):700-5. doi:10.1089/dna.2011.1383

61. Butali A, Mossey PA, Adeyemo WL, et al. Genomic analyses in african populations
identify novel risk loci for cleft palate. Hum Mol Genet. Nov 19 2018;doi:10.1093/hmg/ddy402

62. Beaty TH, Taub MA, Scott AF, et al. Confirming genes influencing risk to cleft lip
with/without cleft palate in a case-parent trio study. Human genetics. Jul 2013;132(7):771-81.
doi:10.1007/s00439-013-1283-6

63. Sull JW, Liang KY, Hetmanski JB, et al. Maternal transmission effects of the PAX genes
among cleft case-parent trios from four populations. Eur J Hum Genet. Jun 2009;17(6):831-9.
doi:10.1038/ejhg.2008.250

64. Peters H, Neubuser A, Balling R. Pax genes and organogenesis: Pax9 meets tooth
development. European journal of oral sciences. Jan 1998;106 Suppl 1:38-43.

65. Basch ML, Bronner-Fraser M, Garcia-Castro MI. Specification of the neural crest occurs
during gastrulation and requires Pax7. Nature. May 11 2006;441(7090):218-22.
doi:10.1038/nature04684

66. Figueiredo JC, Ly S, Raimondi H, et al. Genetic risk factors for orofacial clefts in Central
Africans and Southeast Asians. Am J Med Genet A. Oct 2014;164A(10):2572-80.
doi:10.1002/ajmg.a.36693

136

67. Beaty TH, Murray JC, Marazita ML, et al. A genome-wide association study of cleft lip
with and without cleft palate identifies risk variants near MAFB and ABCA4. Nat Genet. Jun
2010;42(6):525-9. doi:10.1038/ng.580

68. Rojas-Martinez A, Reutter H, Chacon-Camacho O, et al. Genetic risk factors for
nonsyndromic cleft lip with or without cleft palate in a Mesoamerican population: Evidence for
IRF6 and variants at 8q24 and 10q25. Birth Defects Res A Clin Mol Teratol. Jul 2010;88(7):535-
7. doi:10.1002/bdra.20689

69. Ludwig KU, Ahmed ST, Bohmer AC, et al. Meta-analysis Reveals Genome-Wide
Significance at 15q13 for Nonsyndromic Clefting of Both the Lip and the Palate, and Functional
Analyses Implicate GREM1 As a Plausible Causative Gene. PLoS Genet. Mar
2016;12(3):e1005914. doi:10.1371/journal.pgen.1005914

70. Mossey PA, Little J, Steegers-Theunissen R, et al. Genetic Interactions in Nonsyndromic
Orofacial Clefts in Europe-EUROCRAN Study. Cleft Palate Craniofac J. Nov 2017;54(6):623-
630. doi:10.1597/16-037

71. Estandia-Ortega B, Velazquez-Aragon JA, Alcantara-Ortigoza MA, Reyna-Fabian ME,
Villagomez-Martinez S, Gonzalez-Del Angel A. 5,10-Methylenetetrahydrofolate reductase single
nucleotide polymorphisms and gene-environment interaction analysis in non-syndromic cleft
lip/palate. European journal of oral sciences. Apr 2014;122(2):109-13. doi:10.1111/eos.12114

72. Ibarra-Lopez JJ, Duarte P, Antonio-Vejar V, et al. Maternal C677T MTHFR
polymorphism and environmental factors are associated with cleft lip and palate in a Mexican
population. J Investig Med. Aug 2013;61(6):1030-5. doi:10.2310/JIM.0b013e31829a7e7e

73. Beaty TH, Ruczinski I, Murray JC, et al. Evidence for gene-environment interaction in a
genome wide study of nonsyndromic cleft palate. Genetic epidemiology. Sep 2011;35(6):469-
78. doi:10.1002/gepi.20595

74. Shi M, Christensen K, Weinberg CR, et al. Orofacial cleft risk is increased with maternal
smoking and specific detoxification-gene variants. American journal of human genetics. Jan
2007;80(1):76-90. doi:10.1086/510518

75. van Rooij IA, Wegerif MJ, Roelofs HM, et al. Smoking, genetic polymorphisms in
biotransformation enzymes, and nonsyndromic oral clefting: a gene-environment interaction.
Epidemiology (Cambridge, Mass). Sep 2001;12(5):502-7.

76. Hwang SJ, Beaty TH, Panny SR, et al. Association study of transforming growth factor
alpha (TGF alpha) TaqI polymorphism and oral clefts: indication of gene-environment interaction
in a population-based sample of infants with birth defects. American journal of epidemiology.
Apr 1 1995;141(7):629-36. doi:10.1093/oxfordjournals.aje.a117478

77. Romitti PA, Lidral AC, Munger RG, Daack-Hirsch S, Burns TL, Murray JC. Candidate
genes for nonsyndromic cleft lip and palate and maternal cigarette smoking and alcohol
consumption: evaluation of genotype-environment interactions from a population-based case-
control study of orofacial clefts. Teratology. Jan 1999;59(1):39-50. doi:10.1002/(SICI)1096-
9926(199901)59:1<39::AID-TERA9>3.0.CO;2-7

137
78. Wu T, Schwender H, Ruczinski I, et al. Evidence of gene-environment interaction for two
genes on chromosome 4 and environmental tobacco smoke in controlling the risk of
nonsyndromic cleft palate. PloS one. 2014;9(2):e88088. doi:10.1371/journal.pone.0088088

79. Wu T, Fallin MD, Shi M, et al. Evidence of gene-environment interaction for the RUNX2
gene and environmental tobacco smoke in controlling the risk of cleft lip with/without cleft palate.
Birth Defects Res A Clin Mol Teratol. Feb 2012;94(2):76-83. doi:10.1002/bdra.22885

80. Li L, Zhu GQ, Meng T, et al. Biological and epidemiological evidence of interaction of
infant genotypes at Rs7205289 and maternal passive smoking in cleft palate. Am J Med Genet
A. Dec 2011;155a(12):2940-8. doi:10.1002/ajmg.a.34254

81. Wu J, Zheng Q, Huang YQ, et al. Significant evidence of association between
polymorphisms in ZNF533, environmental factors, and nonsyndromic orofacial clefts in the
Western Han Chinese population. DNA and cell biology. Jan 2011;30(1):47-54.
doi:10.1089/dna.2010.1082

82. Jianyan L, Zeqiang G, Yongjuan C, Kaihong D, Bing D, Rongsheng L. Analysis of
interactions between genetic variants of BMP4 and environmental factors with nonsyndromic
cleft lip with or without cleft palate susceptibility. International journal of oral and maxillofacial
surgery. Jan 2010;39(1):50-6. doi:10.1016/j.ijom.2009.10.010

83. Wu T, Liang KY, Hetmanski JB, et al. Evidence of gene-environment interaction for the
IRF6 gene and maternal multivitamin supplementation in controlling the risk of cleft lip
with/without cleft palate. Human genetics. Oct 2010;128(4):401-10. doi:10.1007/s00439-010-
0863-y

84. Haaland OA, Lie RT, Romanowska J, Gjerdevik M, Gjessing HK, Jugessur A. A
Genome-Wide Search for Gene-Environment Effects in Isolated Cleft Lip with or without Cleft
Palate Triads Points to an Interaction between Maternal Periconceptional Vitamin Use and
Variants in ESRRG. Frontiers in genetics. 2018;9:60. doi:10.3389/fgene.2018.00060

85. Bliek BJ, van Schaik RH, van der Heiden IP, et al. Maternal medication use, carriership
of the ABCB1 3435C > T polymorphism and the risk of a child with cleft lip with or without cleft
palate. Am J Med Genet A. Oct 2009;149a(10):2088-92. doi:10.1002/ajmg.a.33036

86. Jugessur A, Lie RT, Wilcox AJ, et al. Cleft palate, transforming growth factor alpha gene
variants, and maternal exposures: assessing gene-environment interactions in case-parent
triads. Genetic epidemiology. Dec 2003;25(4):367-74. doi:10.1002/gepi.10268

87. Chen Q, Wang H, Schwender H, et al. Joint testing of genotypic and gene-environment
interaction identified novel association for BMP4 with non-syndromic CL/P in an Asian
population using data from an International Cleft Consortium. PloS one. 2014;9(10):e109038.
doi:10.1371/journal.pone.0109038

88. Boeck MA, Nagarajan N, Gupta S, et al. Assessing access to surgical care in Nepal via
a cross-sectional, countrywide survey. Surgery. Aug 2016;160(2):501-8.
doi:10.1016/j.surg.2016.03.012

89. Varela C, Young S, Mkandawire N, Groen RS, Banza L, Viste A. TRANSPORTATION
BARRIERS TO ACCESS HEALTH CARE FOR SURGICAL CONDITIONS IN MALAWI a cross

138
sectional nationwide household survey. BMC Public Health. Mar 5 2019;19(1):264.
doi:10.1186/s12889-019-6577-8

90. Swanson JW, Yao CA, Auslander A, et al. Patient Barriers to Accessing Surgical Cleft
Care in Vietnam: A Multi-site, Cross-Sectional Outcomes Study. World J Surg. Jun
2017;41(6):1435-1446. doi:10.1007/s00268-017-3896-8

91. Yao CA, Swanson J, Chanson D, et al. Barriers to Reconstructive Surgery in Low- and
Middle-Income Countries: A Cross-Sectional Study of 453 Cleft Lip and Cleft Palate Patients in
Vietnam. Plast Reconstr Surg. Nov 2016;138(5):887e-895e.
doi:10.1097/PRS.0000000000002656

92. Hackshaw A, Rodeck C, Boniface S. Maternal smoking in pregnancy and birth defects: a
systematic review based on 173 687 malformed cases and 11.7 million controls. Hum Reprod
Update. Sep-Oct 2011;17(5):589-604. doi:10.1093/humupd/dmr022

93. Wyszynski DF, Duffy DL, Beaty TH. Maternal cigarette smoking and oral clefts: a meta-
analysis. Cleft Palate Craniofac J. May 1997;34(3):206-10. doi:10.1597/1545-
1569_1997_034_0206_mcsaoc_2.3.co_2

94. Yang J, Carmichael SL, Canfield M, Song J, Shaw GM, National Birth Defects
Prevention S. Socioeconomic status in relation to selected birth defects in a large multicentered
US case-control study. American journal of epidemiology. Jan 15 2008;167(2):145-54.
doi:10.1093/aje/kwm283

95. Acuna-Gonzalez G, Medina-Solis CE, Maupome G, et al. Family history and
socioeconomic risk factors for non-syndromic cleft lip and palate: a matched case-control study
in a less developed country. Biomedica. Jul-Sep 2011;31(3):381-91. doi:10.1590/S0120-
41572011000300010

96. Jahanbin A, Shadkam E, Miri HH, Shirazi AS, Abtahi M. Maternal Folic Acid
Supplementation and the Risk of Oral Clefts in Offspring. J Craniofac Surg. Sep
2018;29(6):e534-e541. doi:10.1097/SCS.0000000000004488

97. Kelly D, O'Dowd T, Reulbach U. Use of folic acid supplements and risk of cleft lip and
palate in infants: a population-based cohort study. Br J Gen Pract. Jul 2012;62(600):e466-72.
doi:10.3399/bjgp12X652328

98. Berg E, Lie RT, Sivertsen A, Haaland OA. Parental age and the risk of isolated cleft lip: a
registry-based study. Ann Epidemiol. Dec 2015;25(12):942-7 e1.
doi:10.1016/j.annepidem.2015.05.003

99. Mai CT, Cassell CH, Meyer RE, et al. Birth defects data from population-based birth
defects surveillance programs in the United States, 2007 to 2011: highlighting orofacial clefts.
Birth Defects Res A Clin Mol Teratol. Nov 2014;100(11):895-904. doi:10.1002/bdra.23329

100. Bell JC, Raynes-Greenow C, Turner RM, Bower C, Nassar N, O'Leary CM. Maternal
alcohol consumption during pregnancy and the risk of orofacial clefts in infants: a systematic
review and meta-analysis. Paediatr Perinat Epidemiol. Jul 2014;28(4):322-32.
doi:10.1111/ppe.12131

139
101. Romitti PA, Sun L, Honein MA, Reefhuis J, Correa A, Rasmussen SA. Maternal
periconceptional alcohol consumption and risk of orofacial clefts. American journal of
epidemiology. Oct 1 2007;166(7):775-85. doi:10.1093/aje/kwm146

102. Figueiredo JC, Ly S, Magee KS, et al. Parental risk factors for oral clefts among Central
Africans, Southeast Asians, and Central Americans. Birth Defects Res A Clin Mol Teratol. Oct
2015;103(10):863-79. doi:10.1002/bdra.23417

103. ICD-10 Classifications of Mental and Behavioural Disorder: Clinical Descriptions and
Disgnostic Guidelines. 1992;Geneva. World Health Organisation.

104. The World Bank. Global Health Observatory Data Repository (2016), Smoking
prevalence, females (% of adults) https://data.worldbank.org/indicator/SH.PRV.SMOK.FE

105. The Health Consequences of Smoking-50 Years of Progress: A Report of the Surgeon
General. 2014. Reports of the Surgeon General.

106. Junaid M, Narayanan MBA, Jayanthi D, Kumar SGR, Selvamary AL. Association
between maternal exposure to tobacco, presence of TGFA gene, and the occurrence of oral
clefts. A case control study. Clin Oral Investig. Jan 2018;22(1):217-223. doi:10.1007/s00784-
017-2102-6

107. Jia ZL, Shi B, Chen CH, Shi JY, Wu J, Xu X. Maternal malnutrition, environmental
exposure during pregnancy and the risk of non-syndromic orofacial clefts. Oral Dis. Sep
2011;17(6):584-9. doi:10.1111/j.1601-0825.2011.01810.x

108. Hoyt AT, Canfield MA, Romitti PA, et al. Associations between maternal
periconceptional exposure to secondhand tobacco smoke and major birth defects. American
journal of obstetrics and gynecology. Nov 2016;215(5):613.e1-613.e11.
doi:10.1016/j.ajog.2016.07.022

109. McKinney CM, Pisek A, Chowchuen B, et al. Case-control study of nutritional and
environmental factors and the risk of oral clefts in Thailand. Birth Defects Res A Clin Mol
Teratol. Jul 2016;106(7):624-32. doi:10.1002/bdra.23505

110. Li Z, Liu J, Ye R, Zhang L, Zheng X, Ren A. Maternal passive smoking and risk of cleft
lip with or without cleft palate. Epidemiology (Cambridge, Mass). Mar 2010;21(2):240-2.
doi:10.1097/EDE.0b013e3181c9f941

111. Langlois PH, Hoyt AT, Lupo PJ, et al. Maternal occupational exposure to polycyclic
aromatic hydrocarbons and risk of oral cleft-affected pregnancies. The Cleft palate-craniofacial
journal : official publication of the American Cleft Palate-Craniofacial Association. May
2013;50(3):337-46. doi:10.1597/12-104

112. Shum S, Jensen NM, Nebert DW. The murine Ah locus: in utero toxicity and
teratogenesis associated with genetic differences in benzo[a]pyrene metabolism. Teratology.
Dec 1979;20(3):365-76. doi:10.1002/tera.1420200307

113. Garvey DJ, Longo LD. Chronic Low Level Maternal Carbon Monoxide Exposure and
Fetal Growth and Development. Biology of Reproduction. 1978;19(1):8-14.
doi:10.1095/biolreprod19.1.8

140

114. Schults MA, Sanen K, Godschalk RW, Theys J, van Schooten FJ, Chiu RK. Hypoxia
diminishes the detoxification of the environmental mutagen benzo[a]pyrene. Mutagenesis. Nov
2014;29(6):481-7. doi:10.1093/mutage/geu034

115. Chernoff N. Teratogenic effects of cadmium in rats. Teratology. 1973;8(1):29-32.

116. Ferm VH. Developmental malformations induced by cadmium. A study of timed
injections during embryogenesis. Biology of the neonate. 1971;19(1):101-7.
doi:10.1159/000240405

117. Tungotyo M, Atwine D, Nanjebe D, Hodges A, Situma M. The prevalence and factors
associated with malnutrition among infants with cleft palate and/or lip at a hospital in Uganda: a
cross-sectional study. BMC Pediatr. Jan 13 2017;17(1):17. doi:10.1186/s12887-016-0775-7

118. Masarei AG, Sell D, Habel A, Mars M, Sommerlad BC, Wade A. The nature of feeding in
infants with unrepaired cleft lip and/or palate compared with healthy noncleft infants. Cleft
Palate Craniofac J. May 2007;44(3):321-8. doi:10.1597/05-185

119. Munabi NCO, Swanson J, Auslander A, Sanchez-Lara PA, Davidson Ward SL, Magee
WP, 3rd. The Prevalence of Congenital Heart Disease in Nonsyndromic Cleft Lip and/or Palate:
A Systematic Review of the Literature. Ann Plast Surg. Aug 2017;79(2):214-220.
doi:10.1097/SAP.0000000000001069

120. Hunt O, Burden D, Hepper P, Stevenson M, Johnston C. Parent reports of the
psychosocial functioning of children with cleft lip and/or palate. Cleft Palate Craniofac J. May
2007;44(3):304-11. doi:10.1597/05-205

121. Sirugo G, Williams SM, Tishkoff SA. The Missing Diversity in Human Genetic Studies.
Cell. Mar 21 2019;177(1):26-31. doi:10.1016/j.cell.2019.02.048

122. Auslander A, McKean-Cowdin R, Brindopke F, et al. The role of smoke from cooking
indoors over an open flame and parental smoking on the risk of cleft lip and palate: A case-
control study in 7 low-resource countries. J Glob Health. Dec 2020;10(2):020410.
doi:10.7189/jogh.10.020410

123. Amegah AK, Jaakkola JJ. Household air pollution and the sustainable development
goals. Bull World Health Organ. Mar 1 2016;94(3):215-21. doi:10.2471/BLT.15.155812

124. Amooee A, Dastgheib SA, Niktabar SM, et al. Association of Fetal MTHFR 677C > T
Polymorphism with Non-Syndromic Cleft Lip with or without Palate Risk: A Systematic Review
and Meta-Analysis. Fetal Pediatr Pathol. Dec 27 2019:1-17.
doi:10.1080/15513815.2019.1707918

125. Sull JW, Liang KY, Hetmanski JB, et al. Evidence that TGFA influences risk to cleft lip
with/without cleft palate through unconventional genetic mechanisms. Hum Genet. Sep
2009;126(3):385-94. doi:10.1007/s00439-009-0680-3

126. Leslie EJ, Taub MA, Liu H, et al. Identification of functional variants for cleft lip with or
without cleft palate in or near PAX7, FGFR2, and NOG by targeted sequencing of GWAS loci.
Am J Hum Genet. Mar 5 2015;96(3):397-411. doi:10.1016/j.ajhg.2015.01.004

141

127. Cardoso ML, Bezerra JF, Oliveira GH, et al. MSX1 gene polymorphisms in non-
syndromic cleft lip and/or palate. Oral Dis. Jul 2013;19(5):507-12. doi:10.1111/odi.12033

128. Gauderman WJ, Kim A, Conti DV, et al. A Unified Model for the Analysis of Gene-
Environment Interaction. Am J Epidemiol. Apr 1 2019;188(4):760-767. doi:10.1093/aje/kwy278

129. Gauderman WJ, Zhang P, Morrison JL, Lewinger JP. Finding novel genes by testing G x
E interactions in a genome-wide association study. Genet Epidemiol. Sep 2013;37(6):603-13.
doi:10.1002/gepi.21748

130. Murcray CE, Lewinger JP, Gauderman WJ. Gene-environment interaction in genome-
wide association studies. Am J Epidemiol. Jan 15 2009;169(2):219-26. doi:10.1093/aje/kwn353

131. Gauderman WJ, Thomas DC, Murcray CE, Conti D, Li D, Lewinger JP. Efficient
genome-wide association testing of gene-environment interaction in case-parent trios. Am J
Epidemiol. Jul 1 2010;172(1):116-22. doi:10.1093/aje/kwq097

132. Li D, Conti DV. Detecting gene-environment interactions using a combined case-only
and case-control approach. Am J Epidemiol. Feb 15 2009;169(4):497-504.
doi:10.1093/aje/kwn339

133. Besseling S, Dubois L. The prevalence of caries in children with a cleft lip and/or palate
in Southern Vietnam. Cleft Palate Craniofac J. Nov 2004;41(6):629-32. doi:10.1597/03-008.1

134. Yin LL, Li JM, Zhou ZM, Sha JH. Identification of a novel testis-specific gene and its
potential roles in testis development/spermatogenesis. Asian J Androl. Jun 2005;7(2):127-37.
doi:10.1111/j.1745-7262.2005.00041.x

135. Kimura T, Kobayashi T, Munkhbat B, et al. Genome-wide association analysis with
selective genotyping identifies candidate loci for adult height at 8q21.13 and 15q22.33-q23 in
Mongolians. Hum Genet. Jul 2008;123(6):655-60. doi:10.1007/s00439-008-0512-x

136. Elks CE, Perry JR, Sulem P, et al. Thirty new loci for age at menarche identified by a
meta-analysis of genome-wide association studies. Nat Genet. Dec 2010;42(12):1077-85.
doi:10.1038/ng.714

137. Wild PS, Felix JF, Schillert A, et al. Large-scale genome-wide analysis identifies genetic
variants associated with cardiac structure and function. J Clin Invest. May 1 2017;127(5):1798-
1812. doi:10.1172/JCI84840

138. Stepulak A, Rola R, Polberg K, Ikonomidou C. Glutamate and its receptors in cancer. J
Neural Transm (Vienna). Aug 2014;121(8):933-44. doi:10.1007/s00702-014-1182-6

139. Joubert BR, den Dekker HT, Felix JF, et al. Maternal plasma folate impacts differential
DNA methylation in an epigenome-wide meta-analysis of newborns. Nat Commun. Feb 10
2016;7:10577. doi:10.1038/ncomms10577

140. Bortolus R, Blom F, Filippini F, et al. Prevention of congenital malformations and other
adverse pregnancy outcomes with 4.0 mg of folic acid: community-based randomized clinical

142
trial in Italy and the Netherlands. BMC Pregnancy Childbirth. May 13 2014;14:166.
doi:10.1186/1471-2393-14-166

141. Lee TV, Sethi MK, Leonardi J, et al. Negative regulation of notch signaling by xylose.
PLoS Genet. Jun 2013;9(6):e1003547. doi:10.1371/journal.pgen.1003547

142. Guo S, Fan XF, Jin JY, et al. A novel proximal 3q29 chromosome microdeletion in a
Chinese patient with Chiari malformation type II and Sprengel's deformity. Mol Cytogenet.
2018;11:8. doi:10.1186/s13039-018-0358-4

143. Hodgson SV, Chiu DC. Dominant transmission of Sprengel's shoulder and cleft palate. J
Med Genet. Aug 1981;18(4):263-5. doi:10.1136/jmg.18.4.263

144. Bray SJ. Notch signalling: a simple pathway becomes complex. Nat Rev Mol Cell Biol.
Sep 2006;7(9):678-89. doi:10.1038/nrm2009

145. Chou CK, Chang YT, Korinek M, et al. The Regulations of Deubiquitinase USP15 and Its
Pathophysiological Mechanisms in Diseases. Int J Mol Sci. Feb 24
2017;18(3)doi:10.3390/ijms18030483

146. Eichhorn PJ, Rodon L, Gonzalez-Junca A, et al. USP15 stabilizes TGF-beta receptor I
and promotes oncogenesis through the activation of TGF-beta signaling in glioblastoma. Nat
Med. Feb 19 2012;18(3):429-35. doi:10.1038/nm.2619

147. Iyengar PV, Jaynes P, Rodon L, et al. USP15 regulates SMURF2 kinetics through C-
lobe mediated deubiquitination. Sci Rep. Oct 5 2015;5:14733. doi:10.1038/srep14733

148. Loeys BL, Chen J, Neptune ER, et al. A syndrome of altered cardiovascular,
craniofacial, neurocognitive and skeletal development caused by mutations in TGFBR1 or
TGFBR2. Nat Genet. Mar 2005;37(3):275-81. doi:10.1038/ng1511

149. Proetzel G, Pawlowski SA, Wiles MV, et al. Transforming growth factor-beta 3 is
required for secondary palate fusion. Nat Genet. Dec 1995;11(4):409-14. doi:10.1038/ng1295-
409

150. Kaartinen V, Voncken JW, Shuler C, et al. Abnormal lung development and cleft palate
in mice lacking TGF-beta 3 indicates defects of epithelial-mesenchymal interaction. Nat Genet.
Dec 1995;11(4):415-21. doi:10.1038/ng1295-415

151. Kraus DM, Elliott GS, Chute H, et al. CSMD1 is a novel multiple domain complement-
regulatory protein highly expressed in the central nervous system and epithelial tissues. J
Immunol. Apr 1 2006;176(7):4419-30. doi:10.4049/jimmunol.176.7.4419

152. Scholnick SB, Richter TM. The role of CSMD1 in head and neck carcinogenesis. Genes
Chromosomes Cancer. Nov 2003;38(3):281-3. doi:10.1002/gcc.10279

153. Toomes C, Jackson A, Maguire K, et al. The presence of multiple regions of
homozygous deletion at the CSMD1 locus in oral squamous cell carcinoma question the role of
CSMD1 in head and neck carcinogenesis. Genes Chromosomes Cancer. Jun 2003;37(2):132-
40. doi:10.1002/gcc.10191

143
154. Havik B, Le Hellard S, Rietschel M, et al. The complement control-related genes CSMD1
and CSMD2 associate to schizophrenia. Biol Psychiatry. Jul 1 2011;70(1):35-42.
doi:10.1016/j.biopsych.2011.01.030

155. Camargo M, Rivera D, Moreno L, et al. GWAS reveals new recessive loci associated
with non-syndromic facial clefting. Eur J Med Genet. Oct 2012;55(10):510-4.
doi:10.1016/j.ejmg.2012.06.005

156. Fang Y, Fullwood MJ. Roles, Functions, and Mechanisms of Long Non-coding RNAs in
Cancer. Genomics Proteomics Bioinformatics. Feb 2016;14(1):42-54.
doi:10.1016/j.gpb.2015.09.006

157. Gutschner T, Diederichs S. The hallmarks of cancer: a long non-coding RNA point of
view. RNA Biol. Jun 2012;9(6):703-19. doi:10.4161/rna.20481

158. Brunner AL, Beck AH, Edris B, et al. Transcriptional profiling of long non-coding RNAs
and novel transcribed regions across a diverse panel of archived human cancers. Genome Biol.
Aug 28 2012;13(8):R75. doi:10.1186/gb-2012-13-8-r75

159. Gao LY, Zhang FQ, Zhao WH, et al. LncRNA H19 and Target Gene-mediated Cleft
Palate Induced by TCDD. Biomed Environ Sci. Sep 2017;30(9):676-680.
doi:10.3967/bes2017.090

160. Shu X, Shu S, Cheng H. A novel lncRNA-mediated trans-regulatory mechanism in the
development of cleft palate in mouse. Mol Genet Genomic Med. Feb 2019;7(2):e00522.
doi:10.1002/mgg3.522

161. Gao Y, Zang Q, Song H, et al. Comprehensive analysis of differentially expressed
profiles of noncoding RNAs in peripheral blood and ceRNA regulatory networks in
nonsyndromic orofacial clefts. Mol Med Rep. Jul 2019;20(1):513-528.
doi:10.3892/mmr.2019.10261

162. Yun L, Ma L, Wang M, et al. Rs2262251 in lncRNA RP11-462G12.2 is associated with
nonsyndromic cleft lip with/without cleft palate. Hum Mutat. Nov 2019;40(11):2057-2067.
doi:10.1002/humu.23859

163. Wang L, Li Z, Jin L, et al. Indoor air pollution and neural tube defects: effect modification
by maternal genes. Epidemiology. Sep 2014;25(5):658-65.
doi:10.1097/EDE.0000000000000129

164. Weingartner J, Lotz K, Fanghanel J, Gedrange T, Bienengraber V, Proff P. Induction and
prevention of cleft lip, alveolus and palate and neural tube defects with special consideration of
B vitamins and the methylation cycle. J Orofac Orthop. Jul 2007;68(4):266-77.
doi:10.1007/s00056-007-0701-6

165. Kousa YA, Zhu H, Fakhouri WD, et al. The TFAP2A-IRF6-GRHL3 genetic pathway is
conserved in neurulation. Hum Mol Genet. May 15 2019;28(10):1726-1737.
doi:10.1093/hmg/ddz010

144
166. Li N, Mu Y, Liu Z, et al. Assessment of interaction between maternal polycyclic aromatic
hydrocarbons exposure and genetic polymorphisms on the risk of congenital heart diseases. Sci
Rep. Feb 15 2018;8(1):3075. doi:10.1038/s41598-018-21380-3

167. Wang L, Jin L, Liu J, et al. Maternal genetic polymorphisms of phase II metabolic
enzymes and the risk of fetal neural tube defects. Birth Defects Res A Clin Mol Teratol. Jan
2014;100(1):13-21. doi:10.1002/bdra.23196

168. Jamison DT BJ, Measham AR, et al., editors. Disease Control Priorities in Developing
Countries. 2nd edition. The International Bank for Reconstruction and Development / The World
Bank; 2006.

169. Organization WH. International collaborative research on craniofacial anomalies project.
World Health Organization. Accessed September 18, 2020, 2020.
https://www.who.int/genomics/anomalies/en/

170. Clinic M. Cleft Lip and Cleft Palate. Mayo Clinic. Accessed Septermber 18, 2020, 2020.
https://www.mayoclinic.org/diseases-conditions/cleft-palate/symptoms-causes/syc-20370985

171. Pasick CM, Shay PL, Stransky CA, Solot CB, Cohen MA, Jackson OA. Long term
speech outcomes following late cleft palate repair using the modified Furlow technique. Int J
Pediatr Otorhinolaryngol. Dec 2014;78(12):2275-80. doi:10.1016/j.ijporl.2014.10.033

172. Meara JG, Leather AJ, Hagander L, et al. Global Surgery 2030: evidence and solutions
for achieving health, welfare, and economic development. Int J Obstet Anesth. Feb 2016;25:75-
8. doi:10.1016/j.ijoa.2015.09.006

173. Zafar SN, Fatmi Z, Iqbal A, Channa R, Haider AH. Disparities in access to surgical care
within a lower income country: an alarming inequity. World J Surg. Jul 2013;37(7):1470-7.
doi:10.1007/s00268-012-1732-8

174. Garber K, Cabrera CCR, Dinh QL, et al. The Heterogeneity of Global Pediatric Surgery:
Defining Needs and Opportunities Around the World. World J Surg. Jun 2019;43(6):1404-1415.
doi:10.1007/s00268-018-04884-x

175. Collaborators GBDCM. Global, regional, national, and selected subnational levels of
stillbirths, neonatal, infant, and under-5 mortality, 1980-2015: a systematic analysis for the
Global Burden of Disease Study 2015. Lancet. Oct 8 2016;388(10053):1725-1774.
doi:10.1016/S0140-6736(16)31575-6

176. Grimes CE, Bowman KG, Dodgion CM, Lavy CB. Systematic review of barriers to
surgical care in low-income and middle-income countries. World J Surg. May 2011;35(5):941-
50. doi:10.1007/s00268-011-1010-1

177. Lin BM, White M, Glover A, et al. Barriers to Surgical Care and Health Outcomes: A
Prospective Study on the Relation Between Wealth, Sex, and Postoperative Complications in
the Republic of Congo. World J Surg. Jan 2017;41(1):14-23. doi:10.1007/s00268-016-3676-x

178. Shrime MG, Hamer M, Mukhopadhyay S, et al. Effect of removing the barrier of
transportation costs on surgical utilisation in Guinea, Madagascar and the Republic of Congo.
BMJ Glob Health. 2017;2(Suppl 4):e000434. doi:10.1136/bmjgh-2017-000434

145

179. Shah R, Rehfuess EA, Paudel D, Maskey MK, Delius M. Barriers and facilitators to
institutional delivery in rural areas of Chitwan district, Nepal: a qualitative study. Reprod Health.
Jun 20 2018;15(1):110. doi:10.1186/s12978-018-0553-0

180. van Loenhout JA, Delbiso TD, Gupta S, et al. Barriers to surgical care in Nepal. BMC
Health Serv Res. Jan 23 2017;17(1):72. doi:10.1186/s12913-017-2024-7

181. Cairo S, Kakembo N, Kisa P, et al. Disparity in access and outcomes for emergency
neonatal surgery: intestinal atresia in Kampala, Uganda. Pediatr Surg Int. Aug 2017;33(8):907-
915. doi:10.1007/s00383-017-4120-5

182. Merchant A, Hendel S, Shockley R, Schlesinger J, Vansell H, McQueen K. Evaluating
Progress in the Global Surgical Crisis: Contrasting Access to Emergency and Essential Surgery
and Safe Anesthesia Around the World. World J Surg. Nov 2015;39(11):2630-5.
doi:10.1007/s00268-015-3179-1

183. Nguyen K, Bhattacharya SD, Maloney MJ, et al. Self-reported barriers to pediatric
surgical care in Guatemala. Am Surg. Sep 2013;79(9):885-8.

184. Organization WH. Multi-country survey study, World Health Survey questionnaires.
Health System Responsiveness Unit.

185. Ajzen I, Fishbein M. Understanding attitudes and predicting social behavior. Prentice-
Hall; 1980:x, 278 p.

186. Fishbein M, Ajzen I. Belief, attitude, intention, and behavior : an introduction to theory
and research. Addison-Wesley series in social psychology. Addison-Wesley Pub. Co.; 1975:xi,
578 p.

187. Bandura A. Social learning theory. Prentice Hall; 1977:viii, 247 p.

188. Bandura A. Self-efficacy : the exercise of control. W.H. Freeman; 1997:ix, 604 p.

189. Becker MH. The health belief model and personal health behavior. C. B. Slack;
1974:154 p.

190. Rhodes F, Stein JA, Fishbein M, Goldstein RB, Rotheram-Borus MJ. Using theory to
understand how interventions work: Project RESPECT, condom use, and the Integrative Model.
AIDS Behav. May 2007;11(3):393-407. doi:10.1007/s10461-007-9208-9

191. Groen RS, Sriram VM, Kamara TB, Kushner AL, Blok L. Individual and community
perceptions of surgical care in Sierra Leone. Trop Med Int Health. Jan 2014;19(1):107-16.
doi:10.1111/tmi.12215

Abstract (if available)

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Identifying genetic, environmental, and lifestyle determinants of ethnic variation in risk of pancreatic cancer

PDF

Genetic and environmental risk factors for childhood cancer

PDF

Barriers to surgery for in low- and middle-income countries: a cross-sectional study of cleft lip and cleft palate patients in Vietnam

PDF

Competency based education in low resource settings: design and implementation of a novel surgical training program

PDF

Genetic risk factors in multiple myeloma

PDF

Assessing the impact of air pollution on adverse birth outcomes in a low resource setting

PDF

Reprogramming the NiTi expander: an alternative to conventional rapid and slow maxillary expansion modalities for the treatment of patients with cleft lip and palate

PDF

Association of maternal and environmental factors with infant feeding behaviors in a birth cohort study

PDF

Genes and environment in prostate cancer risk and prognosis

PDF

Detrimental effects of dental encroachment on secondary alveolar bone graft outcomes in the treatment of patients with cleft lip and palate: a cone-beam computed tomography study

PDF

Vision epidemiology and the impact of vision loss on vision-specific quality of life

PDF

Predictive factors of breast cancer survival: a population-based study

PDF

Quality of life of patients with cleft lip and palate undergoing orthodontic treatment during early vs. late adolescence

PDF

Smoking in pregnancy: from effects to solutions

PDF

Investigating a physiological pathway for the effect of guided imagery on insulin resistance

PDF

Genes and hormonal factors involved in the development or recurrence of breast cancer

PDF

Genomic risk factors associated with Ewing Sarcoma susceptibility

PDF

Smoke-free housing policies and secondhand smoke exposure in low income multiunit housing in Los Angeles County

PDF

Prenatal environmental exposures and fetal growth in the MADRES cohort

PDF

Determinants of delay in care seeking for pelvic floor disorders among Latina women in Los Angeles

Asset Metadata

Creator Auslander, Allyn (author)

Core Title The environmental and genetic determinants of cleft lip and palate in the global setting

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Epidemiology

Degree Conferral Date 2021-08

Publication Date 07/31/2021

Defense Date 03/26/2021

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag cleft,cleft lip and palate,cook smoke,Epidemiology,genetics,genome-wide interaction scan (GWIS),global health,global surgery,low- and middle-income countries,maternal smoking,OAI-PMH Harvest

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Mckean-Cowdin, Roberta (committee chair), Baez Conde-Garbanati, Lourdes (committee member), Barrington-Trimis, Jessica (committee member), Conti, David (committee member), Magee, William III (committee member)

Creator Email aausland@usc.edu,allyn.auslander@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC15672112

Unique identifier UC15672112

Legacy Identifier etd-AuslanderA-9960

Document Type Dissertation

Format application/pdf (imt)

Rights Auslander, Allyn

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright. The original signature page accompanying the original submission of the work to the USC Libraries is retained by the USC Libraries and a copy of it may be obtained by authorized requesters contacting the repository e-mail address given.

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email cisadmin@lib.usc.edu