University of Southern California Dissertations and Theses

Exploring the interplay of birth order and birth weight on leukemia risk

(USC Thesis Other)

Exploring the interplay of birth order and birth weight on leukemia risk

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Copyright 2024
WENJIA LU
Exploring the Interplay of Birth Order and Birth Weight
on Leukemia Risk
Wenjia Lu
A Thesis Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF SCIENCE
(BIOSTATISTICS)
May 2024

ii
ACKNOWLEDGEMENTS
I would like to express my appreciation to my advisor Dr. Joseph Leo Wiemels, for
graciously welcoming me into his research and providing invaluable guidance throughout my
master's thesis journey. His unwavering support and insightful mentorship have been
instrumental in shaping this work. I would also like to thank my committee members Dr. William Gauderman and Dr. Sandrah
Proctor Eckel for their expertise and guidance. Their constructive feedback and contributions
have greatly enriched the quality of this thesis. Special thanks to Dr. Scott C. Kogan, from the Department of Laboratory Medicine at UCSF, who was invited by Dr. Wiemels to provide invaluable insights to this thesis. His expertise
has significantly contributed to the refinement of this work.

iii
TABLE OF CONTENTS
ACKNOWLEDGEMENTS .......................................................ii
LIST OF TABLES .....................................................................iv
LIST OF FIGURES ....................................................................v
ABSTRACT ...............................................................................vi
CHAPTER ONE: INTRODUCTION .........................................1
CHAPTER TWO: METHODS ...................................................4
CHAPTER THREE: RESULTS .................................................7
CHAPTER FOUR: DISCUSSION ............................................13
CHAPTER FIVE: CONCLUSION ............................................15
REFERENCES ...........................................................................16

iv
LIST OF TABLES
Table 1: Baseline characteristics of leukemia cases and corresponding
controls ...............................................................................................7
Table 2: Association between birth order, other characteristics, and
risk of Leukemia in the California CCRLP ........................................9
Table 3: Causal Mediation Analysis Results of Birth Order, Birth
Weight and Leukemia Risk ...............................................................10

v
LIST OF FIGURES
Figure 1: Proposed mediation models of birth weight in the interplay
of birth order and leukemia ...............................................................10
Figure 2: Causal Mediation Analysis and Interplay of Birth Order, Birth Weight, and Leukemia Risk (ACME: Average Causal
Mediation Effect, ADE: Average Direct Effect) ...............................11

vi
ABSTRACT
Background: The interplay between birth order, birth weight, and leukemia risk is a complex
and intriguing area of investigation within the field of epidemiology. Existing literature
reveals a spectrum of studies examining individual factors such as birth order or birth weight
in relation to leukemia. However, a comprehensive exploration of their combined effects
remains limited. This thesis focused on the interrelationship of birth order and birth weight, seeking to understand how these factors jointly contribute to Leukemia risk. Methods: The data used for this analysis come from the California Cancer Records Linkage
Project – a merge between California vital statistics and cancer registries. A total of 7538
subjects (3939 controls and 3599 leukemia cases) were included in this study. Statistical
assessments for significant associations between birth order and birth weight with leukemia
risk were conducted employing a logistic regression model, which was adjusted for
gestational age and sex of children to account for potential confounding factors. Analysis of
the dataset was facilitated using the mediation function, evaluating the mediated pathway
from birth order to leukemia risk through birth weight. The results are summarized to provide
insights into the magnitude and significance of the mediating effect. Results: In all logit models, birth weight exhibited a significant impact on Leukemia risk, while birth order displayed a slight effect, indicating a lower risk for higher birth order, although not reaching statistical significance. Among the predictors (including birth weight, birth order, gestational age, and gender) in the logistic regression model, only birth weight
exhibited a statistically positive association with log-odds of having leukemia cases
(p<0.0001). The causal mediation analysis, conducted through quasi-Bayesian confidence
intervals with 1000 simulations, indicated a positive average causal mediation effect (ACME), signifying a mediating role for birth weight on Leukemia risk through birth order. The overall

vii
total effect was statistically significant (p<0.0001), emphasizing the combined impact of both
direct and mediated pathways. Conclusion: This analysis delves into the intricate interplay of birth order and birth weight in
the context of leukemia risk. The identified mediating role of birth weight underscores its
significance in shaping Leukemia outcomes. These findings have implications for
understanding the complex etiology of leukemia, shedding light on potential avenues for
further research and intervention strategies. Further research will be needed to understand the
complexities of mediating pathways and their associated mechanisms. A deeper
understanding of these intricate pathways is paramount for the prevention and reduction of
leukemia risk in the vulnerable newborn population.

1
CHAPTER ONE: INTRODUCTION
The influence of birth order, denoting the ordinal position of a child within a family, on
various health outcomes has been a subject of growing interest, with discernible alterations in
disease risks associated with higher birth order (Li et al, 2024). Importantly, first-borns
experience different gestational environments than their later-born siblings, as indexed using
a variety of different biomarkers. These environments may impact later disease risk and
support a biological basis for prenatal environmental conditions related to birth order. First- borns experience less sufficient placentation, higher estrogen levels, and lower insulin
sensitivity, which could all contribute to subsequent post- birth disease risks(Gluckman, P. D. et al, 2004)(Ayyavoo, A., et al, 2013)(Bernstein, L. et al, 1986)(Panagiotopoulou, K. et al, 1990). Despite these observations, the intricate mechanisms through which these factors, tied
to birth order, influence childhood outcomes remain elusive, highlighting the imperative for a
more profound comprehension of disease etiology and prevention. Leukemia, a malignant clonal disease of hematopoietic stem cells, involves uncontrolled
proliferation and development of leukocytes, leading to infiltration of internal
organs(Kampen KR, 2012)(Jakobsen NA et al, 2018)(Licht JD et al, 2005). As the most
common cancer in children, childhood leukemia is thought to be related to environmental
exposures during the prenatal, postnatal, and possibly the pre-conception periods, yet its
pathogenesis remains unclear. Typically, exposure early in life may lead to dramatic health
consequences, including the risk of cancer in the childhood and throughout the life of an
individual (Baird J et al, 2017). There is limited direct evidence for reduced exposure to
specific infections among children who develop leukemia(Lin JN, 2015) as early childhood
exposure to multiple pathogens is difficult to measure precisely, particularly in retrospective
studies relying on maternal recall. Therefore, proxy measures, such as birth order, have been

2
evaluated, revealing that later-born children exhibit increased serological evidence of
common infections(Law GR, 2008). Birth order, traditionally employed as a proxy for both prenatal and postnatal exposures
in childhood cancer studies, offers an easily measured characteristic associated with data
challenging to collect retrospectively—specifically, the timing and frequency of infectious
exposure in early life (Law GR, 2008). The association between family size and childhood cancer was first raised in 1962 by
MacMahon and Newill, who showed a monotonic decrease in childhood cancer mortality for
birth order 1–5(MacMahon B, 1962). Since then, studies from many countries have explored
the relationship between birth order and childhood leukemia, revealing inconsistent results in
terms of increased or decreased risks associated with birth order(Law GR, 2008)(Little J, 1999)(Ross JA et al, 1997)(Westergaard T et al, 1997)(Reynolds P et al, 2002)(Roman E et al, 2005)(Dockerty JD et al, 2001)(Hjalgrim LL et al, 2004)(Ma X, Metayer C et al, 2005)(Cnattingius S et al, 1995). These inconsistencies may stem from differences in study
design or heterogeneity in sociodemographic or perinatal characteristics among the studied
populations. International data demonstrate secular trends of increasing birth weight(Johnson w, 2012), as well as the proportion of first-born children due to decreasing birth rates
worldwide(Roser, M 2014). The current and future implications of these developments on
childhood leukemia risk have garnered limited attention. Birth weight has emerged as a clear
and consistent risk factor for childhood leukemia(Caughey RW, 2009). Birth weight, a key indicator of fetal nutrition and degree of development, is another
perinatal factor that has been examined with regard to childhood leukemia risk. Although
several previous studies have suggested that high birth weight increases the risk of
subsequent acute lymphoid leukemia(ALL), the results are inconsistent (Koifman S et al,

3
2008)(Ma X, Buffler PA, et al, 2005)(Okcu MF et al, 2002)(Oksuzyan S et al, 2012). In most
epidemiologic studies, birth weight has primarily been considered a potential confounder in
the association between birth order and leukemia, with limited exploration of its role as a
modifier. Investigating the interrelationship between birth order and birth weight may
uncover specific subgroups where biological or environmental determinants differentially
impact outcomes, thereby enhancing our understanding of the etiologic pathways leading to
childhood leukemia. Given the rarity of childhood leukemia and the limited explanatory power of identified
risk factors, large collaborative efforts are essential to accumulate sufficient cases for
epidemiologic studies. Most evidence regarding the etiology of childhood leukemia is derived
from retrospective studies or birth registries. We analyzed the data from a large California
state data merge between vital statistics (birth) and cancer registries (Wiemels JL et al, 2018)(Wang R, Wiemels JL et al, 2017). Our aim was to systematically explore the interplay
between birth order and birth weight on childhood leukemia, with a specific focus on
determining whether birth weight modifies the effect of birth order on leukemia risk. Leveraging the ample sample size at our disposal, this investigation aims to provide detailed
insights into these intricate relationships.

4
CHAPTER TWO: METHODS
Participants and Study design
The dataset utilized in this thesis was derived from the California Childhood Cancer
Record Linkage Project (CCRLP), a comprehensive linkage-based study of childhood
cancers conducted in California, United States. Leukemia cases diagnosed in children aged 0
to 14 years between 1988 and 2011 were identified through the California Cancer Registry
and linked to corresponding birth records (1982-2009) from the Office of Vital Records at the
California Department of Public Health (CDPH). Out of the initial 8,178 subjects considered
for this study, 640 participants were excluded due to missing values in either birth order or
birth weight. Thus, the final dataset comprised 7,538 participants (3,939 controls and 3,599
leukemia cases). Controls were meticulously selected from birth records, matching leukemia
cases based on year and month of birth, sex, and race/ethnicity (Hispanic, non-Hispanic
White, non-Hispanic Black, Asian/Pacific Islander, Other). Additional information gleaned
from birth records included gestational age, birth weight, and delivery mode. Definition of birth related varibles
Birth order, denoting the number of prior deliveries the mother had at the time of the
subject’s birth, was coded as an ordinal variable (1, 2, 3, 4+). Only singletons with older
siblings who were also singletons were included in the study, ensuring the inclusion of
relevant information. In cases where multiple participants within a sample set hailed from the
same family, only one was randomly chosen to maintain subject independence. Miscarriages
and abortions were not considered as delivery events. Statistical Analysis
The primary data source, CCRLP, was chosen for its extensive sample size and
population-based subject enrollment, minimizing selection bias as no consent was required. Despite the large number of cases and controls, sample sizes for some analyses were reduced

5
due to missing data in either birth order or birth weight, that was attributable largely to
differences in the information collected on birth certificates from year to year. No differences
in patterns of missingness were detected between cases and controls. The dependent variable for this investigation was leukemia, while the main independent
variable, birth order, was analyzed as ordinal categories. Covariates included sex of the child, gestational age by week, birth weight, and maternal smoking. The criteria for covariate
inclusion in the models were based on their significance in existing literature or their
statistical association with the birth order-leukemia relationship. The analytical framework commenced with generalized linear regression, where birth
order served as the predictor and birth weight as the outcome. This initial step laid the
groundwork for understanding the intricate relationship between birth order and birth weight. Subsequently, an outcome model was developed using binomial logistic regression, predicting leukemia case status solely based on birth order. Odds ratios and 95% confidence
intervals (95% CI) were estimated, providing a fundamental insight into the direct impact of
birth order on leukemia risk. Unconditional multivariable models were employed to adjust for
sex, birth weight, gestational age, maternal smoking, and year of birth. Adjusted odds ratios
(ORs) and corresponding 95% confidence intervals (CIs) were computed to assess the
association between birth order and leukemia risk while accounting for these covariates. The
significance of these associations was evaluated using p values. Expanding upon the baseline model, we evaluated the moderated mediation models to
understand the extent to which the direct and indirect effects of birth order on leukemia risk
varied by birth weight. The process involved fitting the mediator model: Ybirth weight = β0 + β1 ×
Ybirth order + є, using birth weight as the outcome and birth order as the predictor. Subsequently, an outcome model without the mediator was developed using binomial logistic regression
with leukemia as the outcome and birth order as the predictor: Yleukemia = γ0 + γ1 × Ybirth order +

6
δ. Finally, the outcome model with the mediator birth weight was fitted: Yleukemia = α0 + α1 ×
Ybirth order+ α2 × Ybirth weight + є’, aiming to elucidate the indirect effects of birth order on
leukemia risk mediated through birth weight. Our analytical approach encompassed the computation of pivotal indices, including the
Average Causal Mediation Effect (ACME), which signifies the average impact of the
mediator (birth weight) on leukemia cases through the treatment variable (birth order). Additionally, we calculated the Average Direct Effect (ADE), reflecting if the average
influence of birth order on leukemia cases independent of birth weight. The Total Effect
amalgamated both direct and indirect effects of birth order on leukemia risk. The Proportion
Mediated represents the proportion of the total effect that is mediated by the mediator. We
also employed quasi-Bayesian confidence intervals for effect estimation, these indices
formed the cornerstone for quantifying mediation and direct effects in our analysis. Statistical significance for all analyses was set at p ≤ 0.05. R 4.0.0 served as the
analytical tool, ensuring a robust and comprehensive exploration of the dataset. This
methodological approach ensures a meticulous examination of the interplay between birth
order, birth weight, and leukemia risk, providing a foundation for nuanced insights into the
complex relationships within this epidemiology study.

7
CHAPTER THREE: RESULTS
Baseline Characteristics of Participant Population
Table 1 shows the baseline characteristics of the research subjects, providing insights
into the distribution of key variables among leukemia cases and corresponding controls. A
total of 8,178 subjects were initially considered for this study. However, due to missing
values in either birth order or birth weight, 640 participants were excluded from the analysis, resulting in a final dataset comprising 7,538 participants (3,939 controls and 3,599 leukemia
cases). Analyzing birth order distribution further, the majority of cases and controls were born
as follows: 1st order (39.2% cases, 38.6% controls), 2nd order (31.6% cases, 31.6% controls), 3rd order (16.6% cases, 17.0% controls), and 4th order or higher (12.5% cases, 12.9%
controls), indicating that over 70% of the children were first order or second order. These
findings are summarized in Table 1. In terms of birth characteristics, leukemia cases
exhibited slightly higher birth weight compared to controls, tobacco and gestational age for
both cases and controls were very similar. Tobacco was included due to its known effects on
lowering birthweight and its carcinogenicity. Overall, the baseline characteristics, including birth order, birth weight, tobacco
exposure, and gestational age, showed no significant differences between leukemia cases and
controls, laying the foundation for subsequent analyses exploring their interplay in the
context of leukemia risk. Table 1. Baseline characteristics of leukemia cases and corresponding controls Cases
(N, %)
Controls
(N, %)
birth weight(grams, mean ± SD) 3,433 ± 539 3,380 ± 552
Tobacco( mean ± SD)
* 26.6 ± 6.56 26.4 ± 6.84
Gestational age(weeks, mean ± SD) 39.1 ± 2.13 39.0 ± 2.24
Birth order 1
st 1,412 (39.2) 1,519 (38.6)
2
nd 1,138 (31.6) 1,243 (31.6)
3
rd 599 (16.6) 670 (17.0)

8
4
th or higher 450 (12.5) 507 (12.9)
SexMale 2,011 (55.9) 2,224 (56.5)
Female 1,588 (44.1) 1,715 (43.5)
*
tobacco expressed as inverse DNA methylation status of the indicator gene AHRR (Zhong et al., 2023)
Association Analysis
Table 2 shows the unadjusted and multivariable-adjusted relationships between
leukemia and birth order, along with other relevant characteristics, utilizing logistic
regression. Notably, increasing birth order was associated with a decreased risk of leukemia. In the unadjusted analyses, birth weight (100g increments) displayed a statistically
significant association with leukemia risk (OR = 1.018, 95% CI: 1.010-1.027, P = 2.48×10
-5), indicating that a higher risk associated with higher birth weight. Conversely, gestational age
also demonstrated an association (OR = 1.018, 95% CI: 0.997-1.039, P = 0.09), hinting at a
higher risk with longer gestational periods, though this result did not reach statistical
significance. Notably, maternal smoking showed no significant association with leukemia
risk in the unadjusted model (OR = 1.002, 95% CI: 0.996-1.009, P = 0.50). With respect to
sex, males showed a non-significant association in the unadjusted model (OR = 0.977, 95%
CI: 0.892-1.070, P = 0.61), using females as the reference category. Additionally, the multivariable-adjusted model revealed that birth weight remained
significantly associated with leukemia risk (OR = 1.020, 95% CI: 1.011-1.030, P = 3.55×10
- 5), supporting the observed trend in the unadjusted analysis. When compared to first-borns, second-borns demonstrated a slight decrease in leukemia risk (OR: 0.985, 95% CI: 0.884- 1.097, P = 0.78). However, this association did not reach statistical significance, and after
adjustment, the effect size slightly decreased (Adjusted OR: 0.968, 95% CI: 0.868-1.080, P =
0.56). Similarly, third-born or higher individuals also exhibited reduced unadjusted leukemia
risk (OR:0.962, 95% CI:0.843-1.097, P = 0.56 and OR: 0.955, 95% CI: 0.824-1.105, P = 0.54, respectively), yet these associations remained non-significant after adjustment (Adjusted OR:

9
0.937, 95% CI: 0.820-1.070, P = 0.34 and Adjusted OR: 0.926, 95% CI: 0.799-1.074, P =
0.31, respectively). Overall, we found no statistically significant inverse relationship between
leukemia risk and increasing birth order. In summary, while the initial unadjusted analyses suggested a potential association
between birth order and leukemia risk, these associations did not demonstrate significant after
adjusting for other characteristics, as did gestational age, maternal and sex. Conversely, birth
weight consistently maintained a significant positive association with leukemia risk across
both unadjusted and adjusted analyses. Birth weight, however, showed a consistently
significant positive association with leukemia risk across both unadjusted and adjusted
models. It is important to note that all p-values for statistically significant associations are
denoted as p < 0.05. The unadjusted odds ratios were derived from logistic regression models
with one independent variable at a time, while the adjusted odds ratios were derived from a
single multivariable logistic regression model incorporating all variables simultaneously. Table 2. Association between birth order, other characteristics, and risk of Leukemia in the California CCRLP
Unadjusted
a OR (95% CI) P value Adjusted
a OR (95% CI) P value birth weight(100grams) 1.018 (1.010-1.027)* 2.48×10
-5 1.020 (1.011-1.030)* 3.55×10
-5 Tobacco
* 1.002 (0.996-1.009) 0.50 1.004 (0.997-1.010) 0.27
Gestational age(weeks) 1.018 (0.997-1.039) 0.09 0.997 (0.974-1.020) 0.78
Birth order 1
st Reference Reference 2
nd 0.985( 0.884-1.097) 0.78 0.968 (0.868-1.080) 0.56
3
rd 0.962 (0.843-1.097) 0.56 0.937 (0.820-1.070) 0.34
4
th or higher 0.955 (0.825-1.105) 0.54 0.926 (0.799-1.074) 0.31
SexFemale Reference Reference Male 0.977 (0.892-1.070) 0.61 0.953 (0.869-1.044) 0.30
aUnadjusted odds ratios were derived from logistical regression models that only included one independent variable at a time. Adjusted odds ratios were derived from a single multivariable logistic regression model that simultaneously included all variables. *tobacco expressed as inverse DNA methylation status of the indicator gene AHRR (Zhong et al., 2023) *p < 0.05, statistically significant
The Mediating Role of Birth Weight

10
Figure 1 illustrates the conceptual framework of our proposed mediation model. This model
delves into the nuanced pathways through which birth order influences leukemia, with birth
weight serving as the mediator.
Figure 1. Proposed mediation models of birth weight in the interplay of birth order and leukemia As shown in Table 3 and Figure 2, birth weight mediated the effect of birth order on
leukemia, which provides the results of the causal mediation analysis. Table 3. Causal Mediation Analysis Results of Birth Order, Birth Weight and Leukemia Risk
β (95% CI) P value ACME
a(control) 0.00349 (0.00154-0.01)* 0.002
ACME(treated) 0.00348 (0.00153-0.01)* 0.002
ADE
b(control) -0.00752 (-0.03410-0.02) 0.576
ADE(treated) -0.00752 (-0.03412-0.02) 0.576
Prop.Mediated
c (control) -0.15446 (-3.58747-6.11) 0.776
Prop.Mediated (treated) -0.15406 (-3.58559-6.10) 0.776
ACME(average) 0.00349 (0.00154-0.01)* 0.002
ADE(average) -0.00752 (-0.03411-0.02) 0.576
Prop.Mediated (average) 0.15426 (-3.58653-6.10) 0.776
Total Effectd -0.00403 (-0.03034-0.02) 0.768
aACME: Average Causal Mediation Effect, average impact of mediator on outcome through treatment variable
bADE: Average Direct Effect, average effect of treatment variable on outcome without mediator
cProp.Mediated: proportion of total effect that is mediated by mediator
dTotal Effect: total effect of treatment variable on outcome, combining both direct and indirect effects *p < 0.05, statistically significant

11
Figure 2. Causal Mediation Analysis and Interplay of Birth Order, Birth Weight, and Leukemia Risk (ACME: Average Causal Mediation Effect, ADE: Average Direct Effect)
The Average Causal Mediation Effect (ACME)(0.00349)(95% CI: 0.00154-0.01, P =
0.002) indicates a positive mediation effect, suggesting that birth order influences leukemia
risk partially through its impact on birth weight. The Average Direct Effect (ADE)(- 0.00752)(95% CI: -0.03411-0.02, P = 0.576) suggests a negative direct effect, which implies
that birth order has a direct effect on leukemia risk independent of its influence on birth
weight, although this effect is not statistically significant. The total effect combines both the
direct and indirect effects of the treatment variable on the outcome. In this case, the total
effect is close to zero(-0.00403)(95% CI: -0.03034-0.02, P = 0.768) and not statistically
significant, suggesting that birth order may not have a substantial overall impact on leukemia
risk. The proportion mediated indicates the proportion of the total effect that is mediated by
the mediator, birth weight. The negative values and wide confidence intervals suggest
uncertainty(-0.15426)(95% CI: -3.58653-6.10, P = 0.766), and the proportion mediated is not

12
statistically significant. This implies that the mediation effect through birth weight is not
conclusively established. In summary, the results indicate a positive average causal mediation effect, suggesting
that birth order partially influences leukemia risk through its impact on birth weight. However, the overall total effect is not statistically significant, highlighting the need for
further exploration and consideration of other factors in understanding the complex interplay
between birth order, birth weight, and leukemia risk. The average direct effect is negative but
lacks statistical significance, emphasizing the nuanced nature of these relationships. All p-values indicating statistically significant associations are represented as p < 0.05.

13
CHAPTER FOUR: DISCUSSION
Our case-control study indicated a potential inverse association between birth order and
childhood leukemia risk, although this association did not reach statistical significance after
adjusting for potential confounders, including gestational age, child's sex, and maternal
smoking. Further investigation is warranted to elucidate this relationship. The intriguing
findings indicate that later-born children have a reduced leukemia risk, aligning with results
from a comprehensive pooled analysis across five U.S. states (Von Behren J et al, 2011). Birth order, extending beyond familial dynamics, emerges as a potential indicator of
infectious exposures. The prevailing perspective suggests that later-born children likely
encounter pathogens earlier through interactions with older siblings, although the impact
could vary based on factors like birth interval and external sources such as daycare (Perrillat
F et al, 2002; Ma X, Buffler PA, et al, 2005; Gilham C et al, 2005). Early hormonal
exposures may influence future cancer development, with a hypothesis that firstborn children
may face higher estrogen exposures than their later-born counterparts (Bernstein L et al, 1986;
Panagiotopoulou K et al, 1990; Maccoby EE et al, 1979). However, an intriguing shift in
focus emerges when considering birth order's potential influence on in utero development, as
highlighted by Shaobo Li's research on DNA methylation impacts associated with birth order. Changes in placentation and exposures to in utero growth factors with successive pregnancies
may impact later life disease risk via persistent DNA methylation alterations (Li et al, 2024). In contrast to the conventional focus on high birth weight as a leukemia risk factor, our
study introduces a novel perspective by examining the mediating effect of birth weight on the
birth order-leukemia relationship. While previous research often links high birth weight to
leukemia risk, our findings suggest a potential protective effect associated with lower birth
weight. This observation aligns with discussions highlighting the scarcity of childhood acute
lymphoid leukemia (ALL) cases among children with low birth weight (Roman E, 2013). The

14
protective effect in older siblings with lower birth weight prompts thought-provoking
questions about potential underlying mechanisms, such as the role of neonatal infections and
their enduring impact throughout childhood (Miller JE, 2016). This complexity warrants
further exploration into the intricate interplay of birth weight, birth order, and leukemia risk. The focus on causal mediation effect in our study reveals a positive association between
birth weight and the birth order-leukemia relationship. Birth weight plays a partial mediating
role, indicating that prenatal factors significantly contribute to the intricate connection
between birth order and leukemia risk. However, the lack of statistical significance in the
overall total effect, encompassing both direct and indirect influences. This nuanced finding
suggests the presence of other potential factors influencing leukemia risk beyond birth weight
and birth order. Our study's strength lies in its reliance on population registries with nearly complete
birth and cancer registrations in California, minimizing selection bias associated with
volunteer recruitment. With over 7,538 children with leukemia, our analysis possesses robust
statistical power, allowing for a thorough examination of the mediating role of birth weight. Controlling for factors like maternal age and smoking enhances the study's reliability. However, acknowledging certain limitations is essential. The examination of unrelated
individuals instead of same-family siblings introduces potential noise and reduces the
findings' reliability. Genetic, socio-economic, and cultural differences across study
participants and cohorts might have influenced the results. Furthermore, the impact of
additional perinatal variables and maternal characteristics on the findings requires further
confirmation. Although our odds ratios were adjusted for multiple variables, the potential
effects of residual confounding variables remain uncertain. Future research should delve into
these limitations to refine our understanding of the intricate relationships between birth order, birth weight, and childhood leukemia risk.

15
CHAPTER FIVE: CONCLUSION
Our study sheds light on the intricate interplay between birth order, birth weight, and
childhood leukemia risk, challenging prevailing assumptions. Contrary to recent large-scale
registry studies reporting weak or null associations (Schuz J et al, 2015; Crump C et al, 2015), our analysis suggests that birth weight, when examined not just as covariates but as potential
effect mediators, may reveal previously undetected mediating effects. This shift in focus, influenced by the insights from Shaobo Li's paper on DNA methylation impacts, prompts a
reevaluation of the nuanced influences of birth order on in utero development. In conclusion, our study injects a fresh perspective into the discourse on childhood
leukemia risk by delving into the modifying role of birth weight within the birth orderleukemia relationship. The observed protective effect, particularly among older siblings and
lower birth weight, challenges existing paradigms, prompting a reevaluation of established
notions. While birth weight partially mediates the relationship, the overall impact of birth
order on leukemia risk remains nuanced and warrants sustained exploration. Future research
should navigate the intricate web of influences on childhood leukemia risk, considering
additional factors to deepen our understanding of this multifaceted phenomenon.

16
REFERENCES
Ayyavoo, A., Savage, T., Derraik, J. G. B., Hofman, P. L. & Cutfield, W. S. First-born
children have reduced insulin sensitivity and higher daytime blood pressure
compared to later-born children. J. Clin. Endocrinol. Metab. 98, 1248–1253 (2013). Baird J, Jacob C, Barker M, Fall CHD, Hanson M, Harvey NC, et al. Developmental origins
of health and disease: a life course approach to the prevention of non- communicable diseases. Healthcare 2017;5:14. doi: 10.3390/healthcare5010014
Bernstein L, Depue RH, Ross RK, Judd HL, Pike MC, Henderson BE. Higher maternal levels
of free estradiol in first compared to second pregnancy: early gestational
differences. J Natl Cancer Inst 1986;76:1035–9. Caughey RW, Michels KB. Birth weightand childhood leukemia: a meta-analysis andreview
of the current evidence. Int Cancer 2009;124(11):2658-70. doi: 10.1002/ijc24225
Cnattingius S, Zack M, Ekbom A, Gunnarskog J, Linet M, Adami HO. Prenatal and neonatal
risk factors for childhood myeloid leukemia. Cancer Epidemiol Biomarkers Prev
1995;4: 441–5. Crump C, Sundquist J, Sieh W, et al. Perinataland familial risk factors for acute
lymphoblasticleukemia in a Swedish national cohort. Cancer 2015;121(7):1040- 7.https://doi.org/10.1002/cncr.29172. Gilham C, Peto J, Simpson J, Roman E, Eden TO, Greaves MF, Alexander FE. Day care in
infancy and risk of childhood acute lymphoblastic leukaemia: findings from UK
case-control study. BMJ 2005;330:1294. Gluckman, P. D. & Hanson, M. A. Maternal constraint of fetal growth and its consequences. Semin. Fetal Neonatal Med. 9, 419–425 (2004). Dockerty JD, Draper G, Vincent T, Rowan SD, Bunch KJ. Case-control study of parental age, parity and socioeconomic level in relation to childhood cancers. Int J Epidemiol
2001;30:1428–37. Ekbom A. Growing evidence that several human cancers may originate in utero. Semin
Cancer Biol 1998;8:237–44. Hjalgrim LL, Rostgaard K, Hjalgrim H, Westergaard T, Thomassen H, Forestier E, Gustafsson G, Kristinsson J, Melbye M, Schmiegelow K. Birth weight and risk for

17
childhood leukemia in Denmark, Sweden, Norway, and Iceland. J Natl Cancer Inst
2004;96:1549–56. Jakobsen NA, Vyas P. From genomics to targeted treatment in haematological malignancies:
a focus on acute myeloid leukaemia. Clin. Med 2018;18:47. doi:
10.7861/clinmedicine.18-2-s47
Johnson w, Choh AC, Soloway LE, et al. Eightyyear trends in infant weight and length
growth:the Fels longitudinal study.J Pediatr 2012;160(5):762-8. doi:
10.1016/jjpeds.2011.11.002. Kampen KR. The discovery and early understanding of leukemia. Leuk. Res 2012; 36:6–13. doi: 10.1016/j.leukres.2011. 09.028
Koifman S, Pombo-de-Oliveira MS, Brazilian Collaborative Study Group of Infant Acute
Leukemia. High birth weight as an important risk factor for infant leukemia. Br J
Cancer 2008;98:664–7. doi: 10.1038/sj.bjc.6604202
Law GR. Host, family and community proxies forinfections potentially associated with
leukaemia. Radiat Prot Dosimetry 2008;132(2):267 72. doi: 10.1093/rpd/ncn263. Licht JD, Sternberg DW. The molecular pathology of acute myeloid leukemia. Hematology
2005:137–42. doi: 10.1182/asheducation-2005.1.137
Lin JN, Lin CL, Lin MC, et al, Risk ofleukaemiain children infected with enterovirus: a
nation-wide, retrospective, population-based, Taiwaneseregistry, cohort study. Lancet Oncol 2015;16(13)1335-43. doi: 10.1016/S1470-2045(15)00060-1. Little J. Epidemiology of Childhood Cancered, vol.149. Lyon, France: International Agency
for Research on Cancer, 1999. Li, S., Spitz, N., Ghantous, A., ... Wiemels, J. L. A Pregnancy and Childhood Epigenetics
Consortium (PACE) meta analysis highlights potential relationships between birth
order and neonatal blood DNA methylation. Communications Biology 2024;7:66. doi: 10.1038/s42003-023-05698-x. Maccoby EE, Doering CH, Jacklin CN, Kraemer H. Concentrations of sex hormones in
umbilical-cord blood: their relation to sex and birth order of infants. Child Dev
1979;50:632–42. MacMahon B, Newill VA. Birth characteristics ofchildren dying of malignant neoplasms. J
NatlCancer Inst 1962;28(1):231-44. doi: 10.1093/jnci/28.1.231.

18
Marcotte EL, Ritz B, Cockburn M, et al. Exposureto infections and risk of leukemia in young
chil.dren. Cancer Epidemiol Biomarkers Prev 2014;23(7):1195-203. doi.org/10.1158/1055-9965.EPI-13-1330. Ma X, Buffler PA, Wiemels JL, Selvin S, Metayer C, Loh M, Does MB, Wiencke JK. Ethnic
difference in daycare attendance, early infections, and risk of childhood acute
lymphoblastic leukemia. Cancer Epidemiol Biomarkers Prev 2005;14: 1928–34. Ma X, Metayer C, Does MB, Buffler PA. Maternal pregnancy loss, birth characteristics, and
childhood leukemia (United States). Cancer Causes Control 2005;16:1075–83. doi:
10.1007/s10552-005-0356-9
Miller JE, Hammond GC, Strunk T, et al. Association of gestational age and growth measures
at birth with infection-related admissions to hospital throughout childhood: a
population-based, data- linkage study from Western Australia. Lancet Infect Dis
2016;16(8):952–61. doi.org/10. 1016/S1473-3099(16)00150-X. Okcu MF, Goodman KJ, Carozza SE, Weiss NS, Burau KD, Bleyer WA, et al. Birth weight, ethnicity, and occurrence of cancer in children: a population- based, incident case- control study in the State of Texas, USA. Cancer Causes Control 2002;13:595–602. doi: 10.1023/A:1019555912243
Oksuzyan S, Crespi CM, Cockburn M, Mezei G, Kheifets L. Birth weight and other perinatal
characteristics and childhood leukemia in California. Cancer Epidemiol
2012;36:e359–65. doi: 10.1016/j.canep.2012.08.00
Panagiotopoulou K, Katsouyanni K, Petridou E, Garas Y, Tzonou A,Trichopoulos D. Maternal age, parity, and pregnancy estrogens. Cancer Causes Control
1990;1:119–24. Perrillat F, Clavel J, Auclerc MF, Baruchel A, Leverger G, Nelken B, Philippe N, Schaison G, Sommelet D, Vilmer E, Hemon D. Day-care, early common infections and
childhood acute leukaemia: a multicentre French case-control study. Br J Cancer
2002;86:1064–9. Reynolds P, Von Behren J, Elkin EP. Birth characteristics and leukemia in young children. Am J Epidemiol 2002;155:603–13. Roman E, Lightfoot T, Smith AG, et al. Childhood acute lymphoblastic leukaemia and birth- weight: insights from a pooled analysis of case-control data from Germany, the
United King- dom and the United States. Eur J Cancer 2013; 49(6):1437–47. doi.org/10.1016/j.ejca. 2012.11.017.

19
Roman E, Simpson J, Ansell P, Lightfoot T, Mitchell C, Eden TO. Perinatal and reproductive
factors: a report on haematological malignancies from the UKCCS. Eur J Cancer
2005;41:749–59. Wang R, Wiemels JL, Metayer C, Morimoto L, Francis SS, Kadan-Lottick N, DeWan AT, Zhang Y, Ma X. Cesarean Section and Risk of Childhood Acute Lymphoblastic
Leukemia in a Population-Based, Record-Linkage Study in California. Am J
Epidemiol 2017 Jan 15;185(2):96-105. doi: 10.1093/aje/kww153. Roser, M. Fertility Rate 2014. Our World Data. Ross JA, Potter JD, Shu XO, Reaman GH, Lampkin B, Robison LL. Evaluating the
relationships among maternal reproductive history, birth characteristics, and infant
leukemia: a report from the Children’s Cancer Group. Ann Epidemiol 1997;7:172–9. Schuz J, Luta G, Erdmann F, et al. Birth orderand risk of childhood cancer in the Danish
birthcohort of 1973-2010. Cancer Causes Control 2015:26(11):1575-82. doi.org/10.1007/s10552-015-0651-z. Von Behren J, Spector LG, Mueller BA.et al. Birth order and risk of childhood cancer:
apooled analysis from five US States. Int J Cancer 2011;128(11):2709-16. doi.org/10.1002/ijc25593. Westergaard T, Andersen PK, Pedersen JB, Olsen JH, Frisch M, Sorensen HT, Wohlfahrt J, Melbye M. Birth characteristics, sibling patterns, and acute leukemia risk in
childhood: a population-based cohort study. J Natl Cancer Inst 1997;89:939–47. Wiemels JL, Walsh KM, de Smith AJ, Metayer C, Gonseth S, Hansen HM, Francis SS, Ojha
J, Smirnov I, Barcellos L, Xiao X, Morimoto L, McKean-Cowdin R, Wang R, Yu
H, Hoh J, DeWan AT, Ma X. GWAS in childhood acute lymphoblastic leukemia
reveals novel genetic associations at chromosomes 17q12 and 8q24.21. Nat
Commun 2018 Jan 18;9(1):286. doi: 10.1038/s41467-017-02596-9. Zhong C, Li S, Arroyo K, Morimoto LM, de Smith AJ, Metayer C, Ma X, Kogan SC, Gauderman WJ, Wiemels JL. Gene-Environment Analyses Reveal Novel Genetic
Candidates with Prenatal Tobacco Exposure in Relation to Risk for Childhood
Acute Lymphoblastic Leukemia. Cancer Epidemiol Biomarkers Prev 2023 Dec
1;32(12):1707-1715. doi: 10.1158/1055-9965.

Abstract (if available)

Abstract Background: The interplay between birth order, birth weight, and leukemia risk is a complex and intriguing area of investigation within the field of epidemiology. Existing literature reveals a spectrum of studies examining individual factors such as birth order or birth weight in relation to leukemia. However, a comprehensive exploration of their combined effects remains limited. This thesis focused on the interrelationship of birth order and birth weight, seeking to understand how these factors jointly contribute to Leukemia risk.

Methods: The data used for this analysis come from the California Cancer Records Linkage Project – a merge between California vital statistics and cancer registries. A total of 7538 subjects (3939 controls and 3599 leukemia cases) were included in this study. Statistical assessments for significant associations between birth order and birth weight with leukemia risk were conducted employing a logistic regression model, which was adjusted for gestational age and sex of children to account for potential confounding factors. Analysis of the dataset was facilitated using the mediation function, evaluating the mediated pathway from birth order to leukemia risk through birth weight. The results are summarized to provide insights into the magnitude and significance of the mediating effect.

Results: In all logit models, birth weight exhibited a significant impact on Leukemia risk, while birth order displayed a slight effect, indicating a lower risk for higher birth order, although not reaching statistical significance. Among the predictors (including birth weight, birth order, gestational age, and gender) in the logistic regression model, only birth weight exhibited a statistically positive association with log-odds of having leukemia cases (p<0.0001). The causal mediation analysis, conducted through quasi-Bayesian confidence intervals with 1000 simulations, indicated a positive average causal mediation effect (ACME), signifying a mediating role for birth weight on Leukemia risk through birth order. The overall total effect was statistically significant (p<0.0001), emphasizing the combined impact of both direct and mediated pathways.

Conclusion: This analysis delves into the intricate interplay of birth order and birth weight in the context of leukemia risk. The identified mediating role of birth weight underscores its significance in shaping Leukemia outcomes. These findings have implications for understanding the complex etiology of leukemia, shedding light on potential avenues for further research and intervention strategies. Further research will be needed to understand the complexities of mediating pathways and their associated mechanisms. A deeper understanding of these intricate pathways is paramount for the prevention and reduction of leukemia risk in the vulnerable newborn population.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The influence of DNA repair genes and prenatal tobacco exposure on childhood acute lymphoblastic leukemia risk: a gene-environment interaction study

PDF

The interplay between tobacco exposure and polygenic risk score for growth on birthweight and childhood acute lymphoblastic leukemia

PDF

Associations of cumulative pollution burden and environmental health vulnerabilities with gestational weight gain in a cohort of predominantly low-income Hispanic women

PDF

The effect of cytomegalovirus on gene expression of pediatric acute lymphoblastic leukemia

PDF

Understanding acute lymphoblastic leukemia in different ethnic groups in the United States

PDF

Insulin’s effect on lactate levels in extremely low birth weight neonates. a multi-center, observational study

PDF

Air pollution and breast cancer survival in California teachers: using address histories and individual-level data

PDF

Prenatal sleep health, cortisol, and gestational weight gain

PDF

The role of genetic ancestry in estimation of the risk of age-related degeneration (AMD) in the Los Angeles Latino population

PDF

Detecting joint interactions between sets of variables in the context of studies with a dichotomous phenotype, with applications to asthma susceptibility involving epigenetics and epistasis

PDF

Statistical methods for causal inference and densely dependent random sums

PDF

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

PDF

Nonlinear modeling and machine learning methods for environmental epidemiology

PDF

Genetic and environmental risk factors for childhood cancer

PDF

Fish consumption and risk of colorectal cancer

PDF

The risk estimates of pneumoconiosis and its relevant complications: a systematic review and meta-analysis

PDF

Randomized clinical trial generalizability and outcomes for children and adolescents with high-risk acute lymphoblastic leukemia

PDF

The longitudinal risk factors of diabetic retinopathy: the Los Angeles Latino Eye Study

PDF

Red and processed meat consumption and colorectal cancer risk: meta-analysis of case-control studies

PDF

Genetic epidemiological approaches in the study of risk factors for hematologic malignancies

Asset Metadata

Creator Lu, Wenjia (author)

Core Title Exploring the interplay of birth order and birth weight on leukemia risk

School Keck School of Medicine

Degree Master of Science

Degree Program Biostatistics

Degree Conferral Date 2024-05

Publication Date 05/17/2024

Defense Date 05/17/2024

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

Tag birth order,birth weight,interplay,leukemia,OAI-PMH Harvest

Format theses (aat)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Wiemels, Joseph Leo (committee chair), Eckel, Sandrah Proctor (committee member), Gauderman, William (committee member)

Creator Email luwenjia@usc.edu,wennylu0110@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC113940229

Unique identifier UC113940229

Identifier etd-LuWenjia-12955.pdf (filename)

Legacy Identifier etd-LuWenjia-12955

Document Type Thesis

Format theses (aat)

Rights Lu, Wenjia

Internet Media Type application/pdf

Type texts

Source 20240517-usctheses-batch-1155 (batch), 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.

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