University of Southern California Dissertations and Theses

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

(USC Thesis Other)

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

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Copyright 2023 Fatimata Sanogo

INVESTIGATING A PHYSIOLOGICAL PATHWAY FOR THE EFFECT OF GUIDED
IMAGERY ON INSULIN RESISTANCE

By

Fatimata Sanogo

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
EPIDEMIOLOGY

August 2023

ii

Epigraph

"Until the lion learns how to write, every story will glorify the hunter."

-African proverb

This proverb is a poignant reminder that who tells the story matters, and that
marginalized groups and perspectives may be overlooked or misrepresented if they
don't have a voice in shaping the narrative. As epidemiologists, may we use our
vocation as an opportunity to shed light on health disparities and inequalities affecting
vulnerable populations and bring attention to their experiences and needs. May this
work be a step towards a more equitable and just world.

iii
Dedication

To my ancestors, who for generations have passed down the gifts of scholarship,
mysticism, and sacred fellowship in service of healing, self-realization, and
unconditional love for humanity. Your resilience and wisdom continue to guide me.
To my parents, for your unwavering support, sacrifice, and belief in my
greatness. You instilled in me the values of integrity, faith, generosity, joy, freedom,
hard work, resilience, and perseverance.
To my grandparents, who paved the way for me to pursue education beyond my
native land, and who taught me the importance of intergenerational connections and
storytelling.
To my great grandmother, whose love liberated me to life and whose spirit lives
on in me.
To my American family, whose unconditional love, support, and acceptance,
have enriched my life in immeasurable ways.
And to Jordan Bigler, my life and soul partner, for your unwavering love,
encouragement, and partnership in this journey of growth and transformation.
May this work honor your legacies and inspire the future generations of young
Africans rising fully emancipated, empowered, and connected to their roots.

iv
Acknowledgements

Just as the African proverb goes, "It takes a village to raise a child," the same
could be said for obtaining a PhD. I am deeply grateful to my family for their unwavering
support and encouragement throughout my PhD journey. I thank my parents for
instilling in me the values of integrity, faith, hard work, and sacrifice, which have been
instrumental in shaping my character and work ethic. I am also grateful to my American
family for giving me a home away from home, and for teaching me the English language
in fun and creative ways. I deeply appreciate the immense impact their love has had on
me and my growth as a person.
I extend my heartfelt thanks to my friends and colleagues who have been an
integral part of my life and have contributed significantly to my holistic growth. I thank
my cohort members David Bogumil, Charlie Zhong, and peer mentors like Ugonna
Iheanacho, Keren Xu, Kaili Ding, Senkei Feng, Rashi Arora, Kruthika Swaminathan,
Pranav Kulkarni for their constant support, camaraderie, and friendship. I also thank the
Graduate Society for Biostatistics and Epidemiology (GSBE) board members, Eleanor
Zhang, Anmol Anand Pardeshi, Emily Beglarian, Vahan Aslanyan, and Raymond
Hughley for making our work so fulfilling and enjoyable.
I am deeply grateful to my dissertation committee for their guidance, support, and
mentorship throughout my research. I would like to express my special gratitude to my
dissertation chair, Dr. Richard Watanabe, for his invaluable insights and willingness to
explore the field of integrative medicine in the context of type 2 diabetes. Thank you for
your exceptional teaching, mentorship, generosity, and for modeling integrity and hard
work. My deepest thanks to Dr. Victoria Cortessis and Dr. Roberta Mckean-Cowdin for
v
the rigorous training in epidemiology, for teaching in fun ways, and inspiring me to learn
more. I am very grateful to Dr. Wendy Mack for invaluable teachings in biostatistics and
for giving me an opportunity to work as a graduate ‘consultant’, which has opened a
new passion for me. I would like to express my sincere gratitude to Marc Weigensberg
for heartfelt compassion, for all the teachings and for letting me use Imagine HEALTH
Data to complete my research.
Finally, a huge thank you to my partner, Jordan, for making me laugh every day,
being my constant support, source of strength, and love. Thank you for always having
my back and being an integral part of my life.
I am forever grateful to everyone who has played a part in making my PhD
journey a success. Thank you all!

vi
Table of Contents
Epigraph ........................................................................................................................... ii
Dedication ........................................................................................................................ iii
Acknowledgements ......................................................................................................... iv
List of Tables .................................................................................................................. viii
List of Figures .................................................................................................................. ix
Abstract ............................................................................................................................ x
CHAPTER ONE: Background: Type 2 diabetes, Stress, Mind and Body Practices ........ 1
1. Type 2 Diabetes (T2D) .............................................................................................. 1
1.1 History ............................................................................................................ 1
1.2 Pathogenesis ................................................................................................. 2
1.3 Epidemiology .................................................................................................. 3
2. Relationship between Stress, Cortisol, Obesity, and Insulin Resistance .................. 7
3. Mind and Body Practices ........................................................................................... 9
3.1 The Increasing Prevalence of Use of Mind and Body Practices .................... 9
3.2 Guided Imagery: An increasingly Used Mind and Body Technique ............. 10
4. References .............................................................................................................. 12
CHAPTER TWO: Mind and Body-based Interventions Improve Glycemic Control in
Patients with Type 2 Diabetes: A Systematic Review and Meta-Analysis ..................... 19
1. Abstract ................................................................................................................... 19
2. Introduction .............................................................................................................. 20
3. Materials and Methods ............................................................................................ 22
3.1 Data Sources and Searches ........................................................................ 22
3.2 Inclusion Criteria .......................................................................................... 23
3.3 Study Selection ............................................................................................ 23
3.4 Data Extraction and Quality Assessment ..................................................... 24
3.5 Data Synthesis and Analysis ........................................................................ 24
3.6 Results ......................................................................................................... 25
3.7 Limitations .................................................................................................... 29
3.8 Discussion .................................................................................................... 30
4. Conclusions ............................................................................................................. 33
5. References .............................................................................................................. 34
CHAPTER THREE: Cortisol Biomarkers Are Not Associated with Perceived Stress in
Predominantly Latino Adolescents ................................................................................. 45
1. Abstract ................................................................................................................... 45
2. Introduction .............................................................................................................. 46
3. Methods ................................................................................................................... 48
vii
3.1 Study Design and Participants ..................................................................... 48
3.2 Outcome Measures ...................................................................................... 49
4. Results ..................................................................................................................... 52
5. Discussion ............................................................................................................... 53
6. Conclusion: .............................................................................................................. 57
7. References .............................................................................................................. 58
CHAPTER FOUR: Insulin Resistance is Associated with Serum Cortisol, but not
Perceived Stress Score in Latino Adolescents .............................................................. 69
1. Abstract ................................................................................................................... 69
2. Introduction .............................................................................................................. 71
3. Methods ................................................................................................................... 72
3.1 Study Design and Participants ..................................................................... 72
3.2 Measurement Visits ...................................................................................... 73
3.3 Diurnal Salivary Cortisol Pattern Measurement Visits .................................. 74
3.4 Assays .......................................................................................................... 75
3.5 Statistical Analyses ...................................................................................... 76
4. Results ..................................................................................................................... 77
5. Discussion ............................................................................................................... 78
6. Conclusion ............................................................................................................... 82
7. References .............................................................................................................. 83
CHAPTER FIVE: Public Health Implications, and Future Directions ............................. 96
1. Clinical and Public Health Implications .................................................................... 96
2. Future Research .................................................................................................... 100
3. Conclusion ............................................................................................................. 102
4. References ............................................................................................................ 103

viii

List of Tables

Table 1 Characteristics of Studies Included in the Meta-Analysis……………………39

Table 2 Participant Demographics and Clinical Characteristics………………………63

Table 3 Association between PSS and biomarkers of stress………………………….65

Table 4 Association between PSS, cortisol biomarkers, and FBG……………………66

s.Table
1
1 Participant Demographics and Clinical Characteristics by Treatment……..67

Table 5 Participant Demographics and Clinical Characteristics……………………….87

Table 6 Baseline Associations of log HOMA-IR with PSS and Cortisol Biomarkers….87

Table 7 Associations Between Fasting Insulin, PSS and biomarkers of stress……….90

Table 8 Associations Between BMI, PSS and biomarkers of stress…………………….91

Table 9 Associations Between BIA, PSS and Biomarkers of Stress……………………92

S.Table 2 Participant Demographics and Clinical Characteristics by Treatment……....93

1
S.Table = Supplemental Table
ix
List of Figures

Figure I Proposed Mechanism of Action Before Investigation……………………..XV

Figure II Proposed Mechanism of Action After Investigation………………………XVI

Figure 1 Stress, Cortisol and Metabolic Effect Relationships with HPA Activity…..8

Figure 2 Meta-analysis Flow Diagram…………………………………………………38

Figure 3 Results of the mean change in HbA1c………………………………………40

Figure 4 Results of the mean change in FBG…………………………………………41

Figure 5 Bubble plot of change in HbA1c and FBG…………………………………..42

Figure 6 Cumulative meta-analyses by study weight………………………………...43

S. Figure
2
1 Bubble plots for studies reporting on both HbA1c and FBG………………44

2
S.Figure is an abbreviation for Supplementary Figure

x

Abstract

The prevalence of type 2 diabetes and prediabetes is increasing globally, with
minority groups in both developed and developing countries being disproportionately
affected. This trend is particularly concerning because there is an increase in type 2
diabetes in young adults, adolescents, and children. Despite the availability of a range
of therapies to manage hyperglycemia, only 51% of patients with type 2 diabetes
achieve the recommended target hemoglobin A1C (HbA1c) level less than 7%. If left
uncontrolled, type 2 diabetes can lead to various complications, including heart disease,
stroke, and poor blood circulation, making it the ninth leading cause of death in the
world. Thus, there is an urgent need to find lifestyle interventions to control
hyperglycemia in patients with type 2 diabetes and prevent its development in high-risk
populations.
Insulin resistance is considered an underlying cause of impaired fasting glucose
and type 2 diabetes. Insulin resistance is associated with a range of stress-related
factors, including hypothalamic-pituitary adrenal (HPA) axis activation
1, 2
, sympathetic
nervous system overactivity
3
. HPA activation can increase cortisol secretion, resulting in
weight gain that may lead to insulin resistance and hyperglycemia
4
. Cortisol is a steroid
hormone involved in physiological processes including metabolism and stress
response
5, 6
. Normal diurnal variations in cortisol concentration occur in the human
body. Measures of cortisol include serum cortisol, the cortisol awakening response
(CAR), and the diurnal cortisol slope (DCS)
7
. CAR is the increase in salivary cortisol
levels within 30-60 min after awakening in the morning and DCS is the decrease in
xi
salivary cortisol from morning to evening over the waking day
8
. Factors that influence
diurnal cortisol patterns include age, sex, menstrual cycle phase, time of awakening,
sleep duration and quality, ambient light levels, weekdays vs. weekend, prior day
experiences, anticipation of day ahead/prospective memory load, diet and nutrition,
physical activity, medications that can directly impact cortisol levels, health conditions
such as Cushing syndrome, and genetics
8
.
In the United States, it is estimated that 66% of patients with T2D use mind and
body practices
9
, of whom 6-20% are using it specifically to treat their diabetes
10
. While
mind and body practices have been increasingly used to control glycemia in patients
with type 2 diabetes, the effectiveness of these practices is inconsistent across studies,
and the physiological mechanisms underlying their effects on insulin resistance remain
unclear.
In the first project, our goal was to conduct a systematic review and meta-
analysis to determine the effects of mind and body practices on glycemic control in
patients with type 2 diabetes. We had hypothesized that commonly used practices such
as yoga, meditation, mindfulness-based stress reduction, and qigong could improve
glycemic control in patients with type 2 diabetes. We identified 28 intervention studies,
and after meta-analysis found that mind and body practices improved glycemic control,
resulting in a mean decrease of 0.84% in HbA1c and a mean decrease of 22.81 mg/dl
in fasting blood glucose. These results are both statistically significant and clinically
relevant. Notably, these reductions were achieved in addition to standard patient care.
Furthermore, the absolute mean reduction in mean HbA1c is comparable to metformin
monotherapy. Our findings suggest that mind and body practices may be an effective
xii
adjunct to standard pharmacological therapy in managing type 2 diabetes and a
potential preventative measure for those at risk. We also observed that the effects of
mind and body practices were similar across the different modalities, although there
was significant heterogeneity in the yoga intervention studies. Our study-level meta-
regression analysis revealed a significant association between the weekly frequency of
yoga practice and the mean change in HbA1c. We hypothesized that a physiological
mechanism is that mind and body practices lower psychological stress levels, leading to
decreased cortisol levels, which in turn reduces insulin resistance and improves
glycemic control.
In the second project, we tested part of our physiological hypothesis for the effect
of mind and body practices on insulin resistance, and subsequent glycemic control. We
used data from the Imagine HEALTH study to determine whether perceived stress,
known to be affected by mind and body practices, is associated with cortisol levels, also
known to be affected by mind and body practices. We conducted analyses to determine
the relationship between Perceived Stress Scale (PSS), cortisol biomarkers and fasting
blood glucose levels. Our hypothesis was that perceived stress would be positively
associated with cortisol levels and fasting blood glucose. However, we did not find
evidence of an association between perceived stress and cortisol biomarkers of stress,
nor fasting blood glucose. We did observe a significant association between serum
cortisol and fasting blood glucose, both cross-sectionally and longitudinally. These
findings suggest that mind and body practices may modulate cortisol levels, and
subsequent glycemic control through a pathway independent of perceived stress. In
other words, the effect of these practices on cortisol biomarkers and subsequent
xiii
glycemia may occur regardless of changes in perceived stress levels. Our findings also
suggest that PSS may be related to other physiological stress responses that do not
involve activation of the HPA axis. These findings suggest that mind and body practices
may affect glycemia, through a stress pathway, either psychological or physiological,
and mutually independent of each other.
In our third project, we investigated the baseline and longitudinal associations of
insulin resistance with perceived stress, and cortisol biomarker measures using data
from the Imagine HEALTH Study. We hypothesized a positive association between
insulin resistance and cortisol biomarkers, but not association between insulin
resistance and perceived stress. We used HOMA-IR, fasting insulin, BMI and BIA as
indices of insulin resistance, and found fasting insulin levels were associated with
changes in serum cortisol, but not with diurnal cortisol patterns or perceived stress.
These findings reinforce that the impact of mind and body practices on insulin
resistance, and subsequent glycemic control is achieved through a process that is not
influenced by perceived stress, which raises the question by which mechanism mind
and body practices effect serum cortisol levels. The mechanism behind this unknown
phenomenon requires further investigation in future studies. It further supports the
notion that serum cortisol levels may be more predictive of metabolic outcomes, like
insulin resistance, than diurnal cortisol patterns or self-reported perceived stress in this
population. Further research is necessary to fully understand how mind and body
practices enhance serum cortisol levels.

xiv
References

1. Andrews RC, Walker BR. Glucocorticoids and insulin resistance: old hormones,
new targets. Clin Sci (Lond). May 1999;96(5):513-523.

2. Antuna-Puente B, Feve B, Fellahi S, Bastard JP. Adipokines: the missing link
between insulin resistance and obesity. Diabetes Metab. Feb 2008;34(1):2-11.

3. Lambert GW, Straznicky NE, Lambert EA, et al. Sympathetic nervous activation
in obesity and the metabolic syndrome--causes, consequences and therapeutic
implications. Pharmacol Ther. May 2010;126(2):159-172.

4. Peckett AJ, Wright DC, Riddell MC. The effects of glucocorticoids on adipose
tissue lipid metabolism. Metabolism. Nov 2011;60(11):1500-1510.

5. Lightman SL, Birnie MT, Conway-Campbell BL. Dynamics of ACTH and Cortisol
Secretion and Implications for Disease. Endocr Rev. Jun 1 2020;41(3).

6. Hewagalamulage SD, Lee TK, Clarke IJ, Henry BA. Stress, cortisol, and obesity:
a role for cortisol responsiveness in identifying individuals prone to obesity.
Domest Anim Endocrinol. Jul 2016;56 Suppl:S112-120.

7. Adam EK, Quinn ME, Tavernier R, et al. Diurnal cortisol slopes and mental and
physical health outcomes: A systematic review and meta-analysis.
Psychoneuroendocrinology. Sep 2017;83:25-41.

8. Stalder T, Kirschbaum C, Kudielka BM, et al. Assessment of the cortisol
awakening response: Expert consensus guidelines. Psychoneuroendocrinology.
Jan 2016;63:414-432.

9. Bell RA, Suerken CK, Grzywacz JG, et al. Complementary and alternative
medicine use among adults with diabetes in the United States. Altern Ther Health
Med. Sep-Oct 2006;12(5):16-22.

10. Nahin RL, Byrd-Clark D, Stussman BJ, Kalyanaraman N. Disease severity is
associated with the use of complementary medicine to treat or manage type-2
diabetes: data from the 2002 and 2007 National Health Interview Survey. BMC
Complement Altern Med. Oct 22 2012;12:193.

Copyright 2023 Fatimata Sanogo

FIGURE I: Proposed Mechanism of Action Before Investigation

Mind and Body Practice
Decrease Perceived Stress
HPA activity
Change in cortisol Biomarkers
Decrease Insulin Resistance
What We Already Know
Thesis Project 2
Thesis Project 3
Decrease HbA1c, FBG
ii

FIGURE II: Proposed Mechanism of Action After Investigation
Mind and Body Practice
Decrease Perceived Stress
HPA activity
Change in cortisol Biomarkers
Decrease Insulin Resistance Decrease HbA1c, FBG
1
CHAPTER ONE: Background: Type 2 diabetes, Stress, Mind and Body Practices

1. Type 2 Diabetes (T2D)

1.1 History

As far back as 1500 BC, a disease characterized by excessive urinary output had
been described in Egyptian manuscripts
1
. Indian physicians called it madhumeha
(‘honey urine’) because it attracted ants
2
. Aretaeus the Cappadocian coined the word
diabetes, derived from the Greek word diabeinein, to go to excess
3
. The first use of the
word appears around the second century AD.
Aretaeus (81-138 AD) provided the first detailed description of the symptoms
associated with the disease: intolerable thirst, excessive urination, a “burning in the
intestines”, and the two stages of the disease: chronic and acutely fatal
4
. Avicenna
(980-1037 AD), the great Persian physician, in The Canon of Medicine, concocted a
mixture of seeds (lupin, fenugreek, zedoary) as a panacea, used to treat the condition
2
.
To date, fenugreek is still used in the treatment of type 2 diabetes and a recent meta-
analysis showed its use resulted in significant reduction in HbA1c and fasting blood
glucose
5
. Early remedies for diabetes included diverse and interesting prescriptions like
oil of roses, dates, raw quinces and gruel, jelly of viper’s flesh, broken red coral, sweet
almonds, fresh flowers of blind nettles
6
. Prescriptions before the insulin era also
included calorie restrictions. Diet and exercise advocacy was the hallmark of treatment
for diabetes by 19
th
century physicians and still remains an important component of
diabetes management today
7
.

2
The word mellitus, Latin for ‘sweet like honey’ was coined by the British surgeon-
general, John Rollo in 1798 to distinguish this diabetes (that produce sweet urine) from
the other diabetes in which the urine is tasteless
1
. Hence, the term diabetes (Greek)
mellitus (Latin) to describe the condition known today as T2D. During the 18th and 19th
centuries glycosuria was used as a diagnostic characteristic of diabetes mellitus;
however, the pathogenic factors resulting in the disease were not understood.

1.2 Pathogenesis

In 1869, a 22-year-old medical student, Paul Langerhans identified the cells
known as the ‘islets of Langerhans’ in the pancreas. In 1889, Von Mering and
Minkowski, when experimenting on dogs, found that removal of the pancreas led to
diabetes
8
. These studies marked the first association between the pancreas and
diabetes. The name insulin for the ‘juice’ secreted by the islets, which could reduce
blood glucose levels, was coined in 1909 and 1910, individually by De Meyer and
Schaefer, respectively
9, 10
. In 1922, Banting, Best and Collip first reported the isolation
of insulin from pancreatic tissue and received the Nobel Prize for medicine and
physiology for the discovery of insulin
11
. In their work, they conclusively established that
the deficiency of insulin was the cause of diabetes.
12
. Following the isolation of insulin,
in 1951, Sanger and colleagues reported the first in a series of papers detailing the
amino acid sequence of insulin and earned the Nobel Prize in 1959
13
. In 1960, Yalow
and Berson
14
developed the radioimmunoassay (RIA) for insulin, which enabled the
measurent of insulin in the blood. Yalow also received the Nobel Prize for her work.
3
With the ability to isolate and measure insulin, scientists gained new insights into the
disease.
The existence of two prevalent forms of diabetes has now been established:
insulin-dependent diabetes mellitus (IDDM) and insulin-independent diabetes mellitus
(NIDDM). Both forms of the disease are characterized by insulin secretory dysfunction;
however, the pathogenesis of these secretory defects are distinctly different. IDDM is
characterized by autoimmune destruction of the islets of Langerhans that secrete
insulin, whereas NIDDM or T2D is a complex metabolic disorder characterized by
persistent hyperglycemia. The core defect leading to chronic hyperglycemia in T2D are
insulin resistance and impaired insulin secretion
15
. Additional pathways/mechanism
known to contribute to hyperglycemia are inappropriate release of glucagon from
pancreatic alpha cells (especially in postprandial state), decreased incretin effect,
hypothalamic (central nervous system) insulin resistance, and increased lipolysis
15, 16
.
Routine blood sugar tests at prescribed intervals were used for the diagnosis of
diabetes until the introduction of the glycosylated hemoglobin (HbA1c) estimation
2
. To
date, both HbA1c and FBG are used for the diagnosis of T2D and the assessment of
glycemic control. HbA1c reflects a 3-month average of glycemia while FBG reflects an
acute measure of glycemia. Hence, HbA1c is a more stable measure of glycemic
control than FBG.

1.3 Epidemiology

1.3.1 Trends in type 2 diabetes and the Prediabetes Epidemic

4
With an estimated 537 million affected adults in 2021
17
and projected 642 million
by 2040, T2D is the ninth major underlying cause of death worldwide.
18
. The
prevalence of T2D has been rapidly rising worldwide over the past three decades,
particularly in developing countries. T2D was relatively rare in developing countries
some decades ago; for example, the prevalence of the disease was <1% in China in
1980
19
. Today, China has the highest prevalence of T2D in the world, followed by India
and the United States
20
. Eighty percent of cases of T2D worldwide live in less
developed countries and areas
21
. In addition to Asia, the prevalence and burden of T2D
are rising quickly in the gulf region of the Middle East
21
and sub-Saharan Africa
22, 23
.
Compared with developed countries, the proportion of young to middle-aged individuals
with T2D is higher in developing countries
19
and the rural-urban difference in prevalence
is predicted to narrow with urbanization.
24
. In the United States, there is a racial and
ethnic disparity in the prevalence of T2D. The prevalence of diagnosed T2D is 13.2% in
African Americans, 12.8% in Hispanic, 9.0% in Asians and 7.5% in non-Hispanic
Whites. In Native Americans, the prevalence ranges from 6.0% in Alaskan Natives to
24.1% in Southern Arizona Native American groups
25
.
Parallel to T2D, prediabetes has also been rapidly increasing. According to the
American Diabetes Association, prediabetes is a condition characterized by either
impaired fasting glucose (IFG) or impaired glucose tolerance (IGT). Normal fasting
glucose is estimated at 99mg/dl, and impaired fasting glucose between 110-124mg/dl, a
higher fasting blood glucose than normal but not high enough to be diagnosed as T2D.
IGT is a condition characterized by 2-hour post load glucose of 140-199mg/dl, higher
than normal (less than 140 mg/dl) but less than the cut off for T2D diagnosis (200 mg/dl
5
or higher)
26
. It is reported that more than 96 million (38% of the Unites States
population) Americans have prediabetes. According to the NIDDK, this highly prevalent
condition affects disproportionally ethnic minorities who are at a higher risk of
developing insulin resistance
27
. Similarly, the rise in prediabetes is also noticeable
among children and adolescents.

1.3.2 Type 2 Diabetes in Youth

In addition to the early onset of T2D in young adults, an increasing trend of T2D
and prediabetes is noticeable among children and adolescents
28
. In 2001, the crude
prevalence of T2D among North American youth aged 10-19 years was estimated to be
42 cases per 100,0000 youth
29
. Since 2001, the incidence has also been on the rise,
from 8 per 100,000 in 2003 to nearly 15 per 100,000 in 2015
30
. In the United States, the
highest incidence and prevalence are reported among adolescents of high-risk ethnic
groups, such as American Indians, Asian/Pacific Islanders, Hispanic and African
American populations. In the SEARCH for Diabetes in Youth Study in the United States,
the incidence rate in racial and ethnic minority youth 15-19 years old was higher (17.0 to
49.4 per 100,000 person-years) compared to their non-Hispanic white counterparts (5.6
per 100,000 person-years)
31, 32
. The same trend is observed in Australia, where
incidence of T2D in youth is six times higher in Australian indigenous individuals than in
the general population
33
.
Parallel to the rising prevalence of T2D in youth, the trend for prediabetes and
pediatric obesity is also increasing among adolescents. Asian countries such as China
and India have seen an increase in the prevalence of obesity in youth
34
. A similar trend
6
is observed in the United States
35

36
. According to the United States’ National Health
and Nutrition Examination Survey (NHANES), the prevalence of impaired fasting
glucose (IFG) increased from 7% in 2000 to 13.1% in 2005 among U.S. adolescents
aged 12-19 years
37
. In 2020, using survey data from 2005-2016, NHANES reported
that 18% of adolescents and 24% of young adults are prediabetic. IFG constituted the
largest proportion of prediabetes, with prevalence 11.1% in adolescents and 15.8% in
young adults
38
. IFG and impaired glucose tolerance (IGT) form an intermediate state in
the natural history of T2D
39
. Insulin resistance has been considered an underlying
cause of IFG and IGT
40
.

1.3.3 Risk Factors for Type 2 Diabetes

Overweight and obesity are the single most important predictors of T2D
41
. The
effect of obesity on lifetime risk of T2D is stronger in younger adults compared to those
65 years and older
42
. Visceral adiposity (abnormally high deposition of visceral
adipose tissue)
43
and nonalcoholic fatty liver disease (increased buildup of fat in the
liver)
44
have also been shown to be risk factors for T2D. Other reported modifiable risk
factors of T2D include sedentary behavior (sitting or lying positions requiring very low
expenditure), physical inactivity (not meeting physical activity guidelines)
45
, dietary
factors (such as high fat, calories and cholesterol diet), and smoking, previously
identified glucose tolerance (IFG or IGT), abnormal lipids (elevated triglycerides, low
HDL, cholesterol levels), hypertension, and inflammation
24
. Another modifiable risk
factor of T2D is intrauterine environment: adverse intrauterine environment such as
undernutrition in utero may lead to epigenetic modifications during fetal development
46,
7
47
. Low birth weight
48
, fetal malnutrition
49
, and fetal undernutrition
50, 51
have been
shown to increase T2D risk. In addition, women with gestational diabetes mellitus are
7.43 times as likely as those who had normoglycemic pregnancy to develop T2D
52
.
Disparity in the risk of T2D between different ethnic groups after controlling for
diverse environmental attributes indicates a genetic predisposition in the development
of T2D. The common variants of the TCF7L2 gene are significantly associated with risk
of T2D, with a pooled odds ratio of 1.46 for the rs7903146 variant
53
. Other major non-
modifiable risk factors for T2D include age, sex, ethnicity, family history of T2D,and
polycystic ovary syndrome
24
. Moreover, a number of novel factors have been identified
to be independently associated with the development of T2D such as sleeping
disorder
54
, depression
55
, antidepressant medication
56
, environmental toxins such as
endocrine disrupting chemicals (EDCs)
57
and particulate matter
58
.

2. Relationship between Stress, Cortisol, Obesity, and Insulin Resistance

8

Figure 1: Stress, Cortisol, and Metabolic Effect Relationships in the Context of HPA Activity

Normal diurnal variations in cortisol concentration occur in the human body.
Fluctuations in cortisol also take place in response to both physiological and
psychological stressors that affect the HPA activity. Stress is defined as a state of
threatened homeostasis
59
. Stress-related HPA activation starts in the hypothalamus
where corticotrophin-releasing hormone (CRH) is released, which stimulates the release
of adrenocorticotropic hormone (ACTH) from the anterior pituitary gland. ACTH
circulates in the blood stream to the adrenal cortex, signaling adrenal glands to secrete
cortisol
59
. Cortisol, a glucocorticoid hormone, promotes the conversion of preadipocytes
Figure 1 shows the effect of stress on
HPA activity.

Stress stimulates the hypothalamus to
secrete corticotrophin-releasing hormone
(CRH).

CRH activates the pituitary gland which
secretes adrenocorticotropic hormone
(ACTH).

ACTH travels to the adrenal cortex through
the blood stream.

ACTH stimulates the adrenal gland to
produce cortisol. Cortisol leads to a series
of metabolic effects in the body which
leads to insulin resistance (IR)
IR
9
to mature adipocytes
60
, which leads to fat accumulation and weight gain. Cortisol action
has also been shown to contribute to the pathophysiology of insulin resistance
61
via
proliferation of adipokines and the secretion of proinflammatory cytokines (secreted by
adipose tissue).
62
.
Because psychological and physiological stressors can trigger cortisol secretion,
they both have implication for weight gain, obesity, insulin resistance, and metabolic
disorders associated with such states, like T2D.

3. Mind and Body Practices

3.1 The Increasing Prevalence of Use of Mind and Body Practices

According to the National Center for Complementary and Integrative Health
(NCCIH), mind and body practices are a large and diverse subset of complementary
and alternative medicine (CAM) procedures and techniques
63
. Yoga, meditation,
mindfulness-based stress reduction (MBSR), qigong, and guided imagery are common
types of mind and body practices
64
. A meta-analysis has shown association between
workplace based mindfulness-based intervention and improvement in physiological
indices of stress
65
. Other reviews also show effectiveness of mindfulness-based
intervention in reducing stress experienced by nurses
66
, as well as yoga and MBSR
effectiveness in reducing physiological stress
67
in the general population.
In the United States, it is estimated that 66% of patients with T2D use mind and
body practices
68
, of whom 6-20% are using it specifically to treat their diabetes
69
.
Studies have shown inconsistent results on the association between mind and body
practices and improvements in glycemic control, measured by HbA1c and FBG, in
patients with T2D
70, 71
. In addition, the biological mechanism(s) by which mind and body
10
practices improve glycemic control has largely been unexplored. A recent meta-analysis
on the effects of yoga interventions on glycemic control reported overall improvement
reflected by improved HbA1c and FBG
72
. However, the study also reports high
heterogeneity in the findings with no exploration of factors that could explain the
heterogeneity. This earlier meta-analysis does not include other common modes of
mind and body practices.

3.2 Guided Imagery: An increasingly Used Mind and Body Technique

Imagery is a technique that has been used in a variety of forms, across different
cultures for centuries and as far back as ancient Greek times
73
. The technique of
guided imagery is a well-established therapeutic approach in Chinese medicine and
American Indian traditions. In modern application, guided imagery is a mind and body
technique that is used as a therapeutic tool in a wide range of clinical applications that
include, but are not limited to acute and chronic pain relief, preparation for surgery and
medical procedures, cancer treatment, smoking cessation, weight control, fertility,
birthing and delivery, medication compliance and adherence issues, relaxation training,
and stress reduction and management
74
.
Guided imagery involves a series of relaxation techniques followed by the
generation of mental images in order to evoke a state of relaxation (reduced stress) or
achieve a specific health outcome (e.g. mobilize immune system response)
75

Guided imagery has profound physiological consequences given the two higher orders
of communication that are used by the nervous system: The first higher-order
communication system is a most recent evolutionary phenomenon which uses words
11
and verbal thought. It is sequential processing language that is rational, logical, and
analytical called the conscious mind and controls voluntary movements of the skeletal
muscle (conscious activities)
76
. The second higher-order communication system is an
older phenomenon characterized by simultaneous processing language. It is intuitive,
creative, synthetic and regulates the psycho-neuro-immune (PNI) system. It is called the
unconscious mind
76
. Because this more ancient higher order of communication
responds to imagery as it would to a genuine external experience, imagery is used to
induce physiological changes in the body. For example, if you vividly imagine slowly
sucking on a sour, tart slice of a fresh, juicy lemon, you will soon begin to salivate.
Imagery has been shown to affect almost all major physiological systems of the
body, including respiration, heart rate, blood pressure
77
, metabolic rates in cells
74
,
gastrointestinal mobility and secretion
74
, sexual function
78
, blood lipids, immune
responsiveness
79
and even cortisol levels
80
.
With respect to producing specific physiological changes that can promote
healing, guided imagery represents an important alternative to pharmacotherapy with
much greater safety and far fewer complications, precautions, and contra-indications.

12
4. References

1. MacCracken J, Hoel D. From ants to analogues. Puzzles and promises in
diabetes management. Postgrad Med. Apr 1997;101(4):138-140, 143-135, 149-
150.

2. Lakhtakia R. The history of diabetes mellitus. Sultan Qaboos Univ Med J. Aug
2013;13(3):368-370.

3. Ahmed AM. History of diabetes mellitus. Saudi Med J. Apr 2002;23(4):373-378.

4. Engelhardt DV. Diabetes Its Medical and Cultural History 1989.

5. Gong J, Fang K, Dong H, et al. Effect of fenugreek on hyperglycaemia and
hyperlipidemia in diabetes and prediabetes: A meta-analysis. J Ethnopharmacol.
Dec 24 2016;194:260-268.

6. Al-Samarrie N. The History of Diabetes. Accessed November 2021.

7. Allan FN. The writings of Thomas Willis, M.D.; diabetes three hundred years ago.
Diabetes. Jan-Feb 1953;2(1):74-77.

8. Minkowski O. Historical development of the theory of pancreatic diabetes by
Oscar Minkowski, 1929: introduction and translation by Rachmiel Levine.
Diabetes. Jan 1989;38(1):1-6.

9. E A Sharpey-Schäfer S. An introduction to the study of the endocrine glands and
internal secretions: Lane medical lectures, 1913. 1914.

10. Meyer D. Action de la sécrétion interne du pancéras sur différents organes et en
particulier sur la sécrétion rénale. 1909.

11. Banting FG, Best CH. The internal secretion of the pancreas. 1922. Indian J Med
Res. Mar 2007;125(3):251-266.

12. Banting FG, Best CH, Collip JB, et al. Pancreatic Extracts in the Treatment of
Diabetes Mellitus. Can Med Assoc J. Mar 1922;12(3):141-146.

13. Sanger F, Tuppy H. The amino-acid sequence in the phenylalanyl chain of
insulin. 2. The investigation of peptides from enzymic hydrolysates. Biochem J.
Sep 1951;49(4):481-490.

14. Yalow RS, Berson SA. Plasma insulin concentrations in nondiabetic and early
diabetic subjects. Determinations by a new sensitive immuno-assay technic.
Diabetes. Jul-Aug 1960;9:254-260.
13
15. Defronzo RA. Banting Lecture. From the triumvirate to the ominous octet: a new
paradigm for the treatment of type 2 diabetes mellitus. Diabetes. Apr
2009;58(4):773-795.

16. Solis-Herrera C, Triplitt C, Cersosimo E, DeFronzo RA. Pathogenesis of Type 2
Diabetes Mellitus. In: Feingold KR, Anawalt B, Boyce A, et al., eds. Endotext.
South Dartmouth (MA): MDText.com, Inc. Copyright © 2000-2022, MDText.com,
Inc.; 2000.

17. International Diabetes Federation. IDF Atlas. Diabetes Around the World in 2021.
2021.

18. Zheng Y, Ley SH, Hu FB. Global aetiology and epidemiology of type 2 diabetes
mellitus and its complications. Nat Rev Endocrinol. Feb 2018;14(2):88-98.

19. Chan JC, Malik V, Jia W, et al. Diabetes in Asia: epidemiology, risk factors, and
pathophysiology. Jama. May 27 2009;301(20):2129-2140.

20. International Diabetes Federation. IDF Atlas. IDF diabetes Atlas 9th edition 2015.

21. Shaw JE, Sicree RA, Zimmet PZ. Global estimates of the prevalence of diabetes
for 2010 and 2030. Diabetes Res Clin Pract. Jan 2010;87(1):4-14.

22. Mbanya JC, Motala AA, Sobngwi E, et al. Diabetes in sub-Saharan Africa.
Lancet. Jun 26 2010;375(9733):2254-2266.

23. Abubakari AR, Lauder W, Jones MC, et al. Prevalence and time trends in
diabetes and physical inactivity among adult West African populations: the
epidemic has arrived. Public Health. Sep 2009;123(9):602-614.

24. Chen L, Magliano DJ, Zimmet PZ. The worldwide epidemiology of type 2
diabetes mellitus--present and future perspectives. Nat Rev Endocrinol. Nov 8
2011;8(4):228-236.

25. Rodríguez JE, Campbell KM. Racial and Ethnic Disparities in Prevalence and
Care of Patients With Type 2 Diabetes. Clin Diabetes. Jan 2017;35(1):66-70.
26. American Diabetes Association. Understanding A1C- Diagnosis 2022.

27. National Institute of Diabetes and Digestive and Kidney Diseases. Insulin
Resistance & Prediabetes. 2018.

28. Pinhas-Hamiel O, Zeitler P. The global spread of type 2 diabetes mellitus in
children and adolescents. J Pediatr. May 2005;146(5):693-700.

14
29. Liese AD, D'Agostino RB, Jr., Hamman RF, et al. The burden of diabetes mellitus
among US youth: prevalence estimates from the SEARCH for Diabetes in Youth
Study. Pediatrics. Oct 2006;118(4):1510-1518.

30. Prevention CfDCa. Rates of New Diagnosed Cases of Type 1 and Type 2
Diabetes continue to Rise Among Children, Teens. 2020.

31. Dabelea D, Bell RA, D'Agostino RB, Jr., et al. Incidence of diabetes in youth in
the United States. Jama. Jun 27 2007;297(24):2716-2724.

32. Dabelea D, Mayer-Davis EJ, Saydah S, et al. Prevalence of type 1 and type 2
diabetes among children and adolescents from 2001 to 2009. Jama. May 7
2014;311(17):1778-1786.

33. Craig ME, Femia G, Broyda V, et al. Type 2 diabetes in Indigenous and non-
Indigenous children and adolescents in New South Wales. Med J Aust. May 21
2007;186(10):497-499.

34. Rokholm B, Baker JL, Sørensen TI. The levelling off of the obesity epidemic
since the year 1999--a review of evidence and perspectives. Obes Rev. Dec
2010;11(12):835-846.

35. Hales CM, Fryar CD, Carroll MD, et al. Trends in Obesity and Severe Obesity
Prevalence in US Youth and Adults by Sex and Age, 2007-2008 to 2015-2016.
Jama. Apr 24 2018;319(16):1723-1725.

36. National Center for Health Statistics. Prevalence of Overweight, Obesity, and
Severe Obesity Among Children and Adolescents Aged 2–19 Years: United
States, 1963–1965 Through 2017–2018. 2021.

37. Li C, Ford ES, Zhao G, Mokdad AH. Prevalence of pre-diabetes and its
association with clustering of cardiometabolic risk factors and hyperinsulinemia
among U.S. adolescents: National Health and Nutrition Examination Survey
2005-2006. Diabetes Care. Feb 2009;32(2):342-347.

38. Andes LJ, Cheng YJ, Rolka DB, et al. Prevalence of Prediabetes Among
Adolescents and Young Adults in the United States, 2005-2016. JAMA Pediatr.
Feb 1 2020;174(2):e194498.

39. Edelstein SL, Knowler WC, Bain RP, et al. Predictors of progression from
impaired glucose tolerance to NIDDM: an analysis of six prospective studies.
Diabetes. Apr 1997;46(4):701-710.

40. Qiao Q, Tuomilehto J, Balkau B, et al. Are insulin resistance, impaired fasting
glucose and impaired glucose tolerance all equally strongly related to age?
Diabet Med. Nov 2005;22(11):1476-1481.
15
41. Hu FB, Manson JE, Stampfer MJ, et al. Diet, lifestyle, and the risk of type 2
diabetes mellitus in women. N Engl J Med. Sep 13 2001;345(11):790-797.

42. Narayan KM, Boyle JP, Thompson TJ, et al. Effect of BMI on lifetime risk for
diabetes in the U.S. Diabetes Care. Jun 2007;30(6):1562-1566.

43. Lebovitz HE, Banerji MA. Point: visceral adiposity is causally related to insulin
resistance. Diabetes Care. Sep 2005;28(9):2322-2325.

44. Perseghin G. Lipids in the wrong place: visceral fat and nonalcoholic
steatohepatitis. Diabetes Care. May 2011;34 Suppl 2(Suppl 2):S367-370.

45. Panahi S, Tremblay A. Sedentariness and Health: Is Sedentary Behavior More
Than Just Physical Inactivity? Front Public Health. 2018;6:258.

46. Gluckman PD, Hanson MA, Cooper C, Thornburg KL. Effect of in utero and
early-life conditions on adult health and disease. N Engl J Med. Jul 3
2008;359(1):61-73.

47. Pinney SE, Simmons RA. Epigenetic mechanisms in the development of type 2
diabetes. Trends Endocrinol Metab. Apr 2010;21(4):223-229.

48. Whincup PH, Kaye SJ, Owen CG, et al. Birth weight and risk of type 2 diabetes:
a systematic review. Jama. Dec 24 2008;300(24):2886-2897.

49. Ravelli AC, van der Meulen JH, Michels RP, et al. Glucose tolerance in adults
after prenatal exposure to famine. Lancet. Jan 17 1998;351(9097):173-177.

50. Dabelea D, Pettitt DJ, Hanson RL, et al. Birth weight, type 2 diabetes, and insulin
resistance in Pima Indian children and young adults. Diabetes Care. Jun
1999;22(6):944-950.

51. Bavdekar A, Yajnik CS, Fall CH, et al. Insulin resistance syndrome in 8-year-old
Indian children: small at birth, big at 8 years, or both? Diabetes. Dec
1999;48(12):2422-2429.

52. Bellamy L, Casas JP, Hingorani AD, Williams D. Type 2 diabetes mellitus after
gestational diabetes: a systematic review and meta-analysis. Lancet. May 23
2009;373(9677):1773-1779.

53. Cauchi S, El Achhab Y, Choquet H, et al. TCF7L2 is reproducibly associated with
type 2 diabetes in various ethnic groups: a global meta-analysis. J Mol Med
(Berl). Jul 2007;85(7):777-782.

16
54. Shaw JE, Punjabi NM, Wilding JP, et al. Sleep-disordered breathing and type 2
diabetes: a report from the International Diabetes Federation Taskforce on
Epidemiology and Prevention. Diabetes Res Clin Pract. Jul 2008;81(1):2-12.

55. Mezuk B, Eaton WW, Albrecht S, Golden SH. Depression and type 2 diabetes
over the lifespan: a meta-analysis. Diabetes Care. Dec 2008;31(12):2383-2390.

56. Kivimäki M, Hamer M, Batty GD, et al. Antidepressant medication use, weight
gain, and risk of type 2 diabetes: a population-based study. Diabetes Care. Dec
2010;33(12):2611-2616.

57. Alonso-Magdalena P, Quesada I, Nadal A. Endocrine disruptors in the etiology of
type 2 diabetes mellitus. Nat Rev Endocrinol. Jun 2011;7(6):346-353.

58. Krämer U, Herder C, Sugiri D, et al. Traffic-related air pollution and incident type
2 diabetes: results from the SALIA cohort study. Environ Health Perspect. Sep
2010;118(9):1273-1279.

59. Kyrou I, Tsigos C. Stress hormones: physiological stress and regulation of
metabolism. Curr Opin Pharmacol. Dec 2009;9(6):787-793.

60. Peckett AJ, Wright DC, Riddell MC. The effects of glucocorticoids on adipose
tissue lipid metabolism. Metabolism. Nov 2011;60(11):1500-1510.

61. Andrews RC, Walker BR. Glucocorticoids and insulin resistance: old hormones,
new targets. Clin Sci (Lond). May 1999;96(5):513-523.

62. Antuna-Puente B, Feve B, Fellahi S, Bastard JP. Adipokines: the missing link
between insulin resistance and obesity. Diabetes Metab. Feb 2008;34(1):2-11.

63. National Center for Complementary and Integrative Health. Mind and Body
Practices. 2017.

64. Patricia M. Barnes EP-G, Kim McFann, Kim McFann, Richard L. Nahin
Complementary and Alternative Medicine Use Among Adults: United States,
2002. Advanced Data from Vital and health statistics CDC, . 2002.

65. Heckenberg RA, Eddy P, Kent S, Wright BJ. Do workplace-based mindfulness
meditation programs improve physiological indices of stress? A systematic
review and meta-analysis. J Psychosom Res. Nov 2018;114:62-71.

66. Botha E, Gwin T, Purpora C. The effectiveness of mindfulness based programs
in reducing stress experienced by nurses in adult hospital settings: a systematic
review of quantitative evidence protocol. JBI Database System Rev Implement
Rep. Oct 2015;13(10):21-29.
17
67. Pascoe MC, Thompson DR, Ski CF. Yoga, mindfulness-based stress reduction
and stress-related physiological measures: A meta-analysis.
Psychoneuroendocrinology. Dec 2017;86:152-168.

68. Bell RA, Suerken CK, Grzywacz JG, et al. Complementary and alternative
medicine use among adults with diabetes in the United States. Altern Ther Health
Med. Sep-Oct 2006;12(5):16-22.

69. Nahin RL, Byrd-Clark D, Stussman BJ, Kalyanaraman N. Disease severity is
associated with the use of complementary medicine to treat or manage type-2
diabetes: data from the 2002 and 2007 National Health Interview Survey. BMC
Complement Altern Med. Oct 22 2012;12:193.

70. Whitebird RR, Kreitzer MJ, Vazquez-Benitez G, Enstad CJ. Reducing diabetes
distress and improving self-management with mindfulness. Soc Work Health
Care. Jan 2018;57(1):48-65.

71. Kopf S, Oikonomou D, Hartmann M, et al. Effects of stress reduction on
cardiovascular risk factors in type 2 diabetes patients with early kidney disease -
results of a randomized controlled trial (HEIDIS). Exp Clin Endocrinol Diabetes.
Jun 2014;122(6):341-349.

72. Thind H, Lantini R, Balletto BL, et al. The effects of yoga among adults with type
2 diabetes: A systematic review and meta-analysis. Prev Med. Dec
2017;105:116-126.

73. Krau SD. The Multiple Uses of Guided Imagery. Nurs Clin North Am. Dec
2020;55(4):467-474.

74. David E Bresler MLR. Fundamentals of Interactice Guided Imagery 2016.

75. Astin JA, Shapiro SL, Eisenberg DM, Forys KL. Mind-body medicine: state of the
science, implications for practice. J Am Board Fam Pract. Mar-Apr
2003;16(2):131-147.

76. National Cancer Institute STM. Organization of the Nervous System.

77. Dusseldorp E, van Elderen T, Maes S, et al. A meta-analysis of
psychoeduational programs for coronary heart disease patients. Health Psychol.
Sep 1999;18(5):506-519.

78. Klein R, Bar-on E, Klein J, Benbenishty R. The impact of sexual therapy on
patients after cardiac events participating in a cardiac rehabilitation program. Eur
J Cardiovasc Prev Rehabil. Oct 2007;14(5):672-678.
18
79. Gruzelier JH. A review of the impact of hypnosis, relaxation, guided imagery and
individual differences on aspects of immunity and health. Stress. Jun
2002;5(2):147-163.

80. Chellew K, Evans P, Fornes-Vives J, et al. The effect of progressive muscle
relaxation on daily cortisol secretion. Stress. 2015;18(5):538-544.

19

CHAPTER TWO: Mind and Body-based Interventions Improve Glycemic Control in
Patients with Type 2 Diabetes: A Systematic Review and Meta-Analysis

1. Abstract

Aims/hypothesis: Only 51% of type 2 diabetes patients achieve the HbA1c <
7% target. Mind and body practices have been increasingly used to improve glycemic
control among type 2 diabetes patients, but studies show inconsistent efficacy. We
conducted a systematic review and meta-analysis to assess the association between
mind and body practices, and mean change in HbA1c and fasting blood glucose (FBG)
in type 2 diabetes patients.
Methods: We conducted a literature search of Ovid Medline, Embase, and
clinicaltrials.gov seeking through June 10, 2022 published articles on mind and body
practices and type 2 diabetes. Two reviewers independently appraised full text of
papers. Only intervention studies were included. Reviewers extracted data for meta-
analysis. Restricted maximum likelihood random effects modeling was used to calculate
mean differences and summary effect sizes. We assessed heterogeneity using
Cochran’s Q and I2 statistics. Funnel plots were generated for each outcome to gauge
publication bias. Weighted linear models were used to conduct study-level meta-
regression analyses of practice frequency.
Results: We identified 587 articles with 28 meeting inclusion criteria. A
statistically significant and clinically relevant mean reduction in HbA1c of 0.84% (95%
CI: 1.10%, 0.58%, p<0.0001) was estimated. Reduction was observed in all
intervention subgroups; mindfulness-based stress reduction: 0.48% (95% CI: 0.72%,
0.23%, p=0.03), qigong: -0.66% (95% CI: -1.18%, -0.14%, p=0.01), yoga: 1.00% (95%
20
CI: -1.38%, -0.63%, p<0.0001). Meta regression revealed that for every additional day
of yoga practice per week, the raw mean HbA1c differed by 0.22% (95% CI: -0.44%, -
0.003%; P=0.046) over the study period. FBG significantly improved following mind and
body practices, overall mean difference -22.81 mg/dl (95% CI: 33.07 mg/dl, 12.55
mg/dl; p<0.0001). However, no significant association was found between the frequency
of weekly yoga practice and change in FBG over the study period.
Conclusion/Interpretation: Mind and body practices are strongly associated
with improvement in glycemic control in type 2 diabetes patients. The overall mean
reduction in HbA1c and FBG was clinically significant, suggesting mind and body
practices may be an effective complementary, non-pharmacologic intervention for type
2 diabetes. Additional analyses revealed that mean decrease in HbA1c was greater in
studies requiring larger number of yoga practice sessions each week.

2. Introduction

Despite the availability of a range of therapies
1, 2
to address the management of
hyperglycemia, it is estimated only 51% of patients with type 2 diabetes achieve the
therapeutic target of HbA1c < 7%
3
. A variety of factors, including clinical inertia,
polypharmacy, overly complex medication regimens, socio-economic status, health
disparities, and psychiatric disorders are thought to reduce treatment compliance and
efficacy, thus contributing to inadequate glycemic control
4
. In addition, the prevalence
of diabetes distress, the emotional distress resulting from living with diabetes and the
anxiety associated with daily self-management, in people with type 2 diabetes is
reported to be 36%
5
and has been shown to be significantly linked to poor glycemic
control
6
and treatment compliance.
21
More than half of U.S. adults use some form of complementary and alternative
medicine (CAM)
7
for health reasons. Studies have shown that CAM helps a variety of
conditions including relieving stress, improving sleep, decreasing chronic pain, and
improving mental and emotional health
8
. According to the National Center for
Complementary and Integrative Health, mind and body practices are a large and
diverse subset of CAM procedures and techniques
8
. The most common mind and body
modalities used in the U.S. are mindfulness-based stress reduction (MBSR) and other
forms of meditation, yoga, guided imagery, and qigong
7
. In the United States, it is
estimated 66% of patients with type 2 diabetes use mind and body practices
9
, of whom
6-20% are using it specifically to treat their diabetes
10
.
Studies have shown inconsistent results regarding the association between mind
and body practices and improvements in glycemic control, measured by HbA1c or
fasting blood glucose (FBG), in patients with type 2 diabetes
11, 12
. In addition, the
biological mechanism(s) by which mind and body practices improve glycemic control
has largely been unexplored. A recent meta-analysis on the effects of yoga
interventions on glycemic control reported overall improvement reflected by improved
HbA1c and FBG
13
. However, the study also reported high heterogeneity in the findings
with no exploration of factors that could explain the heterogeneity. This earlier meta-
analysis did not include several experimental studies of yoga that are now available,
and information about the efficacy of other common modes of mind and body practices
on glycemic control has not, to our knowledge, been systematically explored.
The objective of this study was to conduct a systematic review and meta-analysis
to assess whether mind and body practices (including yoga, qigong, guided imagery,
22
MBSR, and other forms of meditation) improve glycemic control in patients with type 2
diabetes.

3. Materials and Methods

We adhered to the Preferred Reporting Items for Systematic Reviews and Meta-
Analyses (PRISMA) guidelines, and a detailed protocol is available on Prospero
14
. We
applied the population, intervention, comparison, outcome (PICO) method, defining
population as patients with type 2 diabetes only and intervention as mind and body
interventions that include yoga, qigong, meditation, MBSR, or guided imagery. We
distinguish other forms of meditation from MBSR, the standardized evidence-based
mindfulness training developed by Dr. Jon Kabat Zin and used in clinical practice
15
. The
comparator depended on study design. In randomized control trials (RCTs) it was
glycemic control in individuals who did not receive the intervention during the study
period and in matched pre-post studies it was glycemic control before the intervention.
The primary outcome of interest was glycemic control as measured by mean change in
HbA1c, with mean change in FBG serving as a secondary outcome of interest.

3.1 Data Sources and Searches

We conducted literature search of Ovid Medline, Embase, and clinicaltrials.org
seeking all published articles on common mind and body practices in patients with type 2
diabetes through June 10, 2022. We conducted separate searches using both controlled
23
vocabulary (MeSH terms), keywords, clinically and commonly used terms as well as
definite and possible terms for each of type 2 diabetes, mind and body practices, and
glycemic outcome (FBG and HbA1c). Results of each search were then intersected using
the Boolean operator ‘AND’ to capture articles relevant to our systematic review.
Additional details, including specific search terms, can be found in the supplemental
materials.

3.2 Inclusion Criteria

We included only human intervention studies meeting the PICO criteria, identifying a
mind and body practice intervention, and reporting measures of glycemic control
needed to estimate raw mean difference for intervention versus comparator groups.
Studies were excluded if they provided incomplete data, reported only median values of
the outcome, included participants with type 1 diabetes or other conditions, or were not
available as English language full text reports.

3.3 Study Selection

All citations identified in the search were imported into Covidence
16
and screened
for duplicate entries, which were removed in accordance with PRISMA guidelines. The
remaining abstracts were independently reviewed by two members of the research
team to identify those with the potential to meet criteria for inclusion. These articles
were reviewed in full by independent reviewers, with critical appraisal of study quality
and to identify those that met study inclusion criteria. Disagreements between reviewers
were resolved by consensus. Publication authors were contacted to request missing
24
information for studies that met inclusion criteria but did not report data needed for the
meta-analysis.

3.4 Data Extraction and Quality Assessment

Two independent reviewers extracted the data for the meta-analysis from qualifying
reports. Redundant data from repeated publications were eliminated. Data were
extracted for study design, subject recruitment, sample size for intervention and control
groups, intervention type, outcome assessment, demographic variables, potential
confounders, and mean baseline and follow up measures of glycemic control and
corresponding standard deviation for each study group. Data were managed using
Research Electronic Data Capture (REDCap) hosted at the University of Southern
California (Los Angeles, CA)
17
.

3.5 Data Synthesis and Analysis

We performed meta-analysis of raw mean differences assuming equal variance
as there was no evidence to assume variances are unequal. The primary advantage of
the raw mean difference is that it is inherently meaningful, since the results are reported
on a known scale. Means and standard deviations for the treatment and control groups
were used for randomized control trials, and the baseline and follow up means were
used for matched pre-post studies. The standard correlation of r=0.5 was used to
account for the within subject correlation between pre- and post-scores for matched
studies
18
. Negative effect sizes indicate participants who received the intervention
improved on measures of glycemia (i.e., showed reduction in glycemic measures).
25
The restricted maximum likelihood (REML) random effect model was used to calculate
the weighted means and corresponding 95% confidence intervals. Estimates for
individual studies and summary estimates were displayed as forest plots stratified
according to specific intervention type. Heterogeneity was assessed using Cochran’s Q
and I
2
statistics
19
. Funnel plots were generated for each outcome to gauge publication
bias.
We observed high heterogeneity for the yoga intervention and subsequently
explored the influence of duration of intervention and frequency of yoga practice
(days/week) on effect size by conducting univariate meta-regression using data from all
included studies, as well as from the subset of studies that reported on both HbA1c and
FBG. We additionally conducted cumulative meta-analyses ordered on relative study
weight (large to small) to explore the robustness of available data addressing each
outcome by determining the minimum number of studies needed to achieve statistical
significance. Finally, we performed a separate meta-analysis of mean change in HbA1c
stratifying by intervention type (pre-post studies vs. RCTs) to assess the effect of study
design in the observed heterogeneity. All analyses were conducted using Stata 16
(College Station, TX)

3.6 Results

Figure 2 illustrates the flow of data through our study. We initially identified 587
citations from which 159 duplicates were detected and removed. Independent review of
titles and abstracts of the remaining 428 reports identified 164 candidates which
potentially met the inclusion criteria. Critical review of these reports revealed that 71
26
were not intervention studies, 17 did not use an intervention of interest, 16 included
participants with conditions other than type 2 diabetes (type 1 diabetes or prediabetes
as defined by the individual study), 14 did not report an outcome measure of interest, 2
were not available in English, 1 was missing full text, 1 used combined interventions, 4
lacked relevant information on participants’ recruitment process and baseline
characteristics, 4 did not provide data needed for meta-analysis, and 6 reused the same
set of data re-analyzed from studies already included in the meta-analysis
20-25
. Authors
of three of the four publications with missing data who we contacted did not respond to
our queries, and authors of the fourth were unable to provide the measurements of
interest. Thus, 28 studies were included in the final meta-analysis. Studies included
used guided imagery (n=1), MBSR (n=5), meditation (n=1), qigong (n=3) or yoga (n=18)
as the intervention. The single meditation intervention study used Buddhist Walking
meditation, which distinguishes it from MBSR, the standard evidence-based
mindfulness training used in clinical practice.
Included studies, published from 1993 to 2022, are summarized in Table 1.
Eighteen studies were RCTs, and 10 were matched pre-post studies in which values of
glycemic measures were compared for individual participants at timepoints before and
after the intervention. Seven studies reported on both the primary (HbA1c) and
secondary outcome (FBG); 8 studies reported on HbA1c alone and 13 studies reported
on FBG alone. Two studies
26, 27
reported results for 3 independent subpopulations and
data for these subpopulations were analyzed separately and distinguished in our report
with suffixes A, B and C.
27
Studies included in the meta-analysis are diverse in several ways. Population
samples represented a variety of countries, including Australia, India, the United States,
Germany, South Korea, Cuba, Thailand, China, and Japan, reflecting a range of
race/ethnicity. The duration of intervention ranged from one week to three years of
follow up, while the weekly frequency ranged from once per week to daily. The reported
mean age across studies ranged from 42 to 68 years and represented both sexes
except for the study by Sreedevi and colleagues that enrolled only females
28
.
Participants were mostly recruited from clinical settings, although a few studies
extended recruitment into the community. Ascertainment schemes varied across
studies.
The studies also shared several important features. All excluded type 2 diabetes
patients treated with insulin and those with medical complications (e.g., coronary artery
disease, renal complications), thereby controlling by restriction for diabetes complication
and duration. Participants in all studies were kept on their standard medical care,
controls in most RCTs received standard of care only. However, in one study
29

participants were randomly assigned to either intervention or a waitlist group and the
waitlist group served as controls during the study, but received the intervention at
completion of the study.
Meta-analysis results for change in HbA1c are shown in Figure 3a. The overall
mean reduction in HbA1c across all intervention types was -0.84% (95%
CI: -1.10%, -0.58%; p<0.0001). The largest mean reduction in HbA1c was observed in
studies in which the intervention was yoga, -1.00% (95% CI: -1.38%, -0.63%;
p<0.0001), although reductions in mean HbA1c were also observed in studies of MBSR
28
(-0.48% (95% CI: -0.72%, -0.23%; p=0.03), qigong (-0.66% (95% CI: -1.18%, -0.14%;
p=0.01), and meditation (-0.50% (95% CI: -2.54%, 1.54%; p=0.64). The funnel plot
(Figure 3b) of studies that reported on change in mean HbA1c appeared symmetric and
provided no indication that results reflect publication bias.
Meta-analysis results for mean change in FBG were consistent with mean
change in HbA1c (Figure 3a. -22.81 mg/dl (95% CI: -33.07 mg/dl, -12.55 mg/dl;
p<0.0001). The funnel plot (Figure 3b) again provided no indication of publication bias.

We observed significant heterogeneity of effect size for HbA1c as reflected in both
Cochran’s Q and I
2
. I
2
was estimated to be 87% for HbA1c. This heterogeneity
appeared to arise from differences among the yoga studies and was not apparent for
the other interventions. A similar heterogeneity pattern was observed for mean change
in FBG. We attempted to identify the source of this heterogeneity by exploring the
influence of duration of intervention and frequency of yoga practice (days/week) on
effect size. There was no statistically significant association between duration of
intervention and either mean change in HbA1c or FBG. A statistically significant inverse
association between number of yoga sessions per week and effect size was observed
for HbA1c, but not for FBG (Figure 4a, Figure 4b). For every additional day of yoga
practice per week, the raw mean HbA1c differed by -0.22% (95% CI: -0.44%, -0.003%;
P=0.046) over the study period. Very similar results were found by limiting meta-
regression to the 7 studies that reported on both HbA1c and FBG. This sensitivity
analysis also identified statistically significant inverse association between number of
yoga sessions per week and mean reduction in HbA1c (p=0.02), but not in FBG
(p=0.75). In this subset of studies, the raw mean difference in HbA1c was estimated to
29
decrease by -0.27% (95% CI: -0.50%, -0.04%; P=0.02) for every additional day of
practice per week (Supplementary Figure 1). Finally, we performed a separate meta-
analysis of mean change in HbA1c stratifying studies by intervention type (pre-post
studies vs. RCTs) and found moderate heterogeneity in pre-post studies (60.12%), but
high heterogeneity in RCTs (89.71%). Despite variation in heterogeneity between study
types, the mean change in HbA1c was nearly identical between study types (-0.84% in
RTCs vs. -0.85% in pre-post studies).
Finally, cumulative meta-analyses reveal summary estimates of effects of yoga to
be very robust, because results for both HbA1c and FBG require data from only 1 of 9
studies and 1 of 18 studies, respectively, to achieve statistical significance. (Figure 5a.
Figure 5b.)

3.7 Limitations

There are several limitations of our study. Participants in the studies included in
the meta-analysis were not blinded to the intervention. However, while lack of blinding
may readily increase the risk of bias in the reporting of subjective outcomes such as
behavior, outcome measures used in our analyses are objective in nature, making this a
minor concern. We had to exclude data from 4 reports that may have qualified for
inclusion despite our efforts to contact the investigators to obtain the required
information. In addition, we did not include studies identified in the grey literature,
because available information did not allow us to adequately assess quality of the
methodology employed. It is unlikely these exclusions could have spuriously created the
inverse associations reported here, in light of the robust nature of the contributing data,
30
reasonably symmetrical shape of the funnel plots, and lack of a priori rationale for
excluded studies to differ from those included.

3.8 Discussion

We identified and synthesized experimental evidence regarding the effects of
mind and body practices on glycemic control among patients with type 2 diabetes. The
28 studies included in our analyses show that taken together, a range of different mind
and body practices significantly reduced both HbA1c and FBG compared to standard of
care in patients with type 2 diabetes. The overall absolute estimated decrease of 0.84%
in mean HbA1c is statistically and clinically significant
30
. Hirst et al., estimated a 1.12%
decrease in HbA1c in a meta-analysis of 35 trials of metformin monotherapy of least 12
weeks duration
31
. By comparison, the overall effect of mind and body practices on
reduction in mean HbA1c estimated in this meta-analysis is 75% that reported for
metformin. Because participants in the studies included in our meta-analysis received
standard of care before and throughout the studies and, for the most part were actively
treated with metformin, the observed effect of mind and body practices appears to
represent an additional decrease in mean HbA1c beyond the mono-therapeutic effect of
metformin. This raises the question of whether mind and body practices could be useful
if initiated early in the course of diabetes therapy along with conventional lifestyle
treatments. It further suggests mind and body practices may also be an effective
preventive measure in people at risk for type 2 diabetes.
Our results evoke the question by what mechanism may mind and body
interventions improve glycemic control. Prior studies have shown significant
31
associations between diabetes distress and poor glycemic control in patients with type 2
diabetes
6
. Four studies included in our meta-analysis reported on various measures of
patient stress obtained using the Patient’s Health Questionnaire (PHQ;
12
, Diabetes
Distress scale
11, 32
, or Perceived Stress Scale
33
). The two studies reporting specifically
on diabetes distress showed a decrease in both diabetes distress and HbA1c in the
intervention group, which is consistent with prior literature. One possible theory is that a
decrease in psychological distress followed by increased treatment and regimen
compliance may mediate the effect of mind and body practices on glycemia. Prior
studies have also shown significant associations between elevated serum cortisol and
poor glycemic control in patients with type 2 diabetes
34
. Two studies included in the
meta-analysis
32, 35
reported a significant decrease in serum cortisol and glycemia
following intervention. Cortisol could plausibly mediate the benefit of mind and body
practices on glucose control through reduced inflammation and a cascade of
homeostatic mechanisms that improve lipid profiles, insulin sensitivity and glycemia
36,
37
. This hypothesis will require additional study.
Our primary objective was to identify and synthesize experimental data on
efficacy of common forms of mind and body practices on glycemic control in humans
with type 2 diabetes, which to our knowledge has not previously been performed. Our
results accord with the findings of a previous meta-analysis
13
that reported
improvement in HbA1c and FBG to be associated with yoga. However, these
investigators did not consider other forms of mind and body practice, and they included
data from only 4 of the experimental studies that contributed to our meta-analysis. In
addition, the previous meta-analysis was based largely on studies conducted in India, in
32
contrast to the far broader geographic distribution of source populations contributing to
our report. This is, therefore, the first report providing summary estimates of efficacy of
mind and body practices on glycemic control in type 2 diabetes to extend synthesis of
human data beyond yoga. While our summary estimates for other common types of
mind and body practices are similar to those for yoga, our systematic review established
that data for MBSR, meditation, guided imagery and qigong are very limited. These
other forms of mind and body practices warrant further investigation in people with type
2 diabetes, because they may be similarly beneficial in reducing both diabetes distress
and physiologic distress and may be more accessible than yoga to some type 2
diabetes patients.
In all studies, point estimates of effect of yoga on mean change in HbA1c and
FBG were consistently negative, although a degree of heterogeneity was apparent.
Heterogeneity could reflect differing degrees of systematic error and thus bias between
studies or true differences in effect size. Most studies controlled for a priori potential
confounders through randomization, matching, and/or restriction to patients without
complications and those treated without insulin. These measures also make greatly
varying degrees of participation bias unlikely. Nonetheless, differing degrees of
measurement error, and thus information bias, may have been present in contributing
studies. FBG has high day to day variability and is a more variable indicator of control,
compared to HbA1c. The clear inverse association between number of yoga sessions
per week and effect size observed for HbA1c, but not for FBG, may reflect true
differences in effect size that were obscured by error inherent in measuring glycemic
control by FBG. In this scenario, while heterogeneity in FBG data may be due largely to
33
bias, heterogeneity in HbA1c data may in part reflect greater efficacy of more intensive
yoga practice. This interpretation is supported by results of the sensitivity analysis of the
subset of studies that reported on both measures of glycemic control. These studies
revealed a pattern of inverse association between number of yoga sessions per week
and HbA1c, but not FBG, that is remarkably similar to results from all studies. This
agreement strongly implicates differences in accuracy of the outcome measure,
because all study elements and other data – participants, interventions, and covariate
data – would have been identical in these seven studies.
Other factors that could create true differences in effect size include the types of
yoga practiced. Although some studies reported the type of yoga used in the protocol,
others did not, so we could not address this as a study-level variable. Lastly, differences
in ascertainment schemes for recruitment in clinical settings may have produced study
populations with differing propensity to benefit from yoga. It is worth noting, for example,
that some reports did not provide a diagram illustrating the process of recruitment and
randomization of participants and therefore do not comply to current methodological
standards of reporting intervention studies. Additionally, some studies did not report
important details on patient demographics or loss to follow up.

4. Conclusions

In conclusion, we showed by systematic review followed by critical appraisal and
meta-analysis that mind and body practices reduce HbA1c and FBG in patients with
type 2 diabetes. The overall estimated effect is clinically significant and suggests these
practices may be an effective, complementary, non-pharmacologic intervention for type
34
2 diabetes. The results further suggest early initiation of mind and body practices along
with conventional lifestyle intervention could be useful in mitigating hyperglycemia or an
effective preventive measure in those at risk for type 2 diabetes.

5. References

1. American Diabetes Association. 9. Pharmacologic Approaches to Glycemic
Treatment: Standards of Medical Care in Diabetes-2019. Diabetes Care. Jan
2019;42(Suppl 1):S90-s102.

2. American Diabetes Association. 5. Lifestyle Management: Standards of Medical
Care in Diabetes-2019. Diabetes Care. Jan 2019;42(Suppl 1):S46-s60.

3. Fang M, Wang D, Coresh J, Selvin E. Trends in Diabetes Treatment and Control
in U.S. Adults, 1999-2018. N Engl J Med. Jun 10 2021;384(23):2219-2228.

4. Bailey CJ, Kodack M. Patient adherence to medication requirements for therapy
of type 2 diabetes. Int J Clin Pract. Mar 2011;65(3):314-322.

5. Perrin NE, Davies MJ, Robertson N, et al. The prevalence of diabetes-specific
emotional distress in people with Type 2 diabetes: a systematic review and meta-
analysis. Diabet Med. Nov 2017;34(11):1508-1520.

6. Fisher L, Mullan JT, Arean P, et al. Diabetes distress but not clinical depression
or depressive symptoms is associated with glycemic control in both cross-
sectional and longitudinal analyses. Diabetes Care. Jan 2010;33(1):23-28.

7. Patricia M. Barnes EP-G, Kim McFann, Kim McFann, Richard L. Nahin
Complementary and Alternative Medicine Use Among Adults: United States,
2002. Advanced Data from Vital and health statistics CDC, . 2002.

8. National Center for Complementary and Integrative Health. Mind and Body
Practices. 2017.

9. Bell RA, Suerken CK, Grzywacz JG, et al. Complementary and alternative
medicine use among adults with diabetes in the United States. Altern Ther Health
Med. Sep-Oct 2006;12(5):16-22.

10. Nahin RL, Byrd-Clark D, Stussman BJ, Kalyanaraman N. Disease severity is
associated with the use of complementary medicine to treat or manage type-2
35
diabetes: data from the 2002 and 2007 National Health Interview Survey. BMC
Complement Altern Med. Oct 22 2012;12:193.

11. Whitebird RR, Kreitzer MJ, Vazquez-Benitez G, Enstad CJ. Reducing diabetes
distress and improving self-management with mindfulness. Soc Work Health
Care. Jan 2018;57(1):48-65.

12. Kopf S, Oikonomou D, Hartmann M, et al. Effects of stress reduction on
cardiovascular risk factors in type 2 diabetes patients with early kidney disease -
results of a randomized controlled trial (HEIDIS). Exp Clin Endocrinol Diabetes.
Jun 2014;122(6):341-349.

13. Thind H, Lantini R, Balletto BL, et al. The effects of yoga among adults with type
2 diabetes: A systematic review and meta-analysis. Prev Med. Dec
2017;105:116-126.

14. Fatimata Sanogo KX, Victoria Cortessis Mind-Body modalities and the control of
type 2 diabetes (t2d): a systematicreview and meta-analysis. 2020.

15. Niazi AK, Niazi SK. Mindfulness-based stress reduction: a non-pharmacological
approach for chronic illnesses. N Am J Med Sci. Jan 2011;3(1):20-23.

16. Covidence. https://www.covidence.org

17. Harris PA, Taylor R, Thielke R, et al. Research electronic data capture
(REDCap)--a metadata-driven methodology and workflow process for providing
translational research informatics support. J Biomed Inform. Apr 2009;42(2):377-
381.

18. Michael Borenstein LVH, Julian P.T. Higgins, Hannah R. Rothstein Introduction
to meta-analysis 2009.

19. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat
Med. Jun 15 2002;21(11):1539-1558.

20. Hartmann M, Kopf S, Kircher C, et al. Sustained effects of a mindfulness-based
stress-reduction intervention in type 2 diabetic patients: design and first results of
a randomized controlled trial (the Heidelberger Diabetes and Stress-study).
Diabetes Care. May 2012;35(5):945-947.

21. Hegde SV, Adhikari P, Kotian SM, Shastry R. Effects of Yoga Versus Sham
Yoga on Oxidative Stress, Glycemic Status, and Anthropometry in Type 2
Diabetes Mellitus: A Single-Blinded Randomized Pilot Study. Int J Yoga Therap.
Jan 1 2019;30(1):33-39.

36
22. Kyizom T, Singh S, Singh KP, et al. Effect of pranayama & yoga-asana on
cognitive brain functions in type 2 diabetes-P3 event related evoked potential
(ERP). Indian J Med Res. May 2010;131:636-640.

23. Gordon L, Morrison EY, McGrowder DA, et al. Changes in clinical and metabolic
parameters after exercise therapy in patents with type 2 diabetes. Archives of
Medical Science. 2008;4(4):427-437.

24. Gordon L, Morrison EY, McGrowder D, et al. Effect of yoga and traditional
physical exercise on hormones and percentage insulin binding receptor in
patients with type 2 diabetes. American Journal of Biochemistry and
Biotechnology. 2008;4(1):35-42.

25. Vijayakumar V, Kannan S. Effect of yoga on reducing glycemic variability in
individuals with type 2 diabetes: A randomised controlled trial. Diabetes
Technology and Therapeutics. 2020;22:A-66.

26. Jain SC, Uppal A, Bhatnagar SO, Talukdar B. A study of response pattern of
non-insulin dependent diabetics to yoga therapy. Diabetes Res Clin Pract. Jan
1993;19(1):69-74.

27. Beena RK, Sreekumaran E. Yogic practice and diabetes mellitus in geriatric
patients. Int J Yoga. Jan 2013;6(1):47-54.

28. Sreedevi A, Gopalakrishnan UA, Karimassery Ramaiyer S, Kamalamma L. A
Randomized controlled trial of the effect of yoga and peer support on glycaemic
outcomes in women with type 2 diabetes mellitus: a feasibility study. BMC
Complement Altern Med. Feb 7 2017;17(1):100.

29. Hegde SV, Adhikari P, Kotian S, et al. Effect of 3-month yoga on oxidative stress
in type 2 diabetes with or without complications: a controlled clinical trial.
Diabetes Care. Oct 2011;34(10):2208-2210.

30. Standards of medical care in diabetes--2009. Diabetes Care. Jan 2009;32 Suppl
1(Suppl 1):S13-61.

31. Hirst JA, Farmer AJ, Ali R, et al. Quantifying the effect of metformin treatment
and dose on glycemic control. Diabetes Care. Feb 2012;35(2):446-454.

32. Jung HY, Lee H, Park J. Comparison of the effects of Korean mindfulness-based
stress reduction, walking, and patient education in diabetes mellitus. Nurs Health
Sci. Dec 2015;17(4):516-525.

33. Varghese MP, Balakrishnan R, Pailoor S. Association between a guided
meditation practice, sleep and psychological well-being in type 2 diabetes
mellitus patients. J Complement Integr Med. Jul 19 2018;15(4).
37
34. Ortiz R, Kluwe B, Odei JB, et al. The association of morning serum cortisol with
glucose metabolism and diabetes: The Jackson Heart Study.
Psychoneuroendocrinology. May 2019;103:25-32.

35. Gainey A, Himathongkam T, Tanaka H, Suksom D. Effects of Buddhist walking
meditation on glycemic control and vascular function in patients with type 2
diabetes. Complement Ther Med. Jun 2016;26:92-97.

36. Morais JBS, Severo JS, Beserra JB, et al. Association Between Cortisol, Insulin
Resistance and Zinc in Obesity: a Mini-Review. Biol Trace Elem Res. Oct
2019;191(2):323-330.

37. Geer EB, Islam J, Buettner C. Mechanisms of glucocorticoid-induced insulin
resistance: focus on adipose tissue function and lipid metabolism. Endocrinol
Metab Clin North Am. Mar 2014;43(1):75-102.

38

FIGURE 2. Meta-analysis flow diagram. The meta-analysis flow diagram outlines how
the final publications were selected. Literature search identified 587 studies, and 159
duplicates were detected through Covidence and removed. Two-hundred and sixty-four
studies that did not meet the study inclusion criteria were excluded. The remaining 164
studies were submitted to a full-text critical appraisal, and a final 28 studies were
included in the meta-analysis.

39

TABLE 1. Characteristics of Studies Included in the Meta-Analysis

1: C= control group, 2: tx=treatment group, 3: RCT= Randomized Control Trial, 4: pre-post= matched study, 5:
MBSR= mindfulness-based stress reduction, 6: GI= guided imagery. Each of these publications are cited in the
references. 7: intervention frequency was 1 day/week for 8 weeks, followed by booster session every 6 months. 8:
intervention was 2times/week for 3 months followed by 1 time/week for 3 months. 9: daily from 5:30 AM to 9 PM

Study Location Sample
1, 2
Study type
3, 4
Intervention details
5,6
Type Duration Frequency sessions
Jablon et al. 1997 USA 10 c, 10 tx RCT GI 4 weeks Daily
Rosenzweig et al. 2007 USA 11 Pre-post MBSR 8 weeks 1 day/week
Kopf et al. 2014 Germany 57 c, 53 tx RCT MBSR 3 years 1 day/week
7

Jung et al. 2015 South Korea 28 c, 28 tx RCT MBSR 8 weeks 2 days/week
Whitebird et al. 2017 USA 31 Pre-post MBSR 8 weeks 1 day/week
Guo et al. 2021 China 50 c, 50 tx RCT MBSR 12 weeks Daily
Gainey et al. 2016 Thailand 11c, 12 tx RCT Meditation 12 weeks 3 days/week
Tsujiuchi et al. 2002 Japan 10 c, 16 tx RCT Qigong 4 months 1 day/week
Wang et al. 2008 China 20 c, 20 tx Pre-post Qigong 4 months Daily
Lam et al. 2008 Australia 25 c, 28 tx RCT Qigong 6 months 2 times/week
8
Jain et al. 1993 (A) India 45 Pre-post Yoga 40 days 2 times/week
Jain et al. 1993 (B) India 28 Pre-post Yoga 40 days 2 times/week
Jain et al. 1993 (C) India 76 Pre-post yoga 40 days 2 times/week
Gordon et al. 2008 Cuba 77 c, 77 tx RCT Yoga 24 weeks 1 day/week
Singh et al. 2008 India 30 c, 30 tx RCT Yoga 45 days Daily
Amita et al. 2009 India 21 c, 20 tx RCT Yoga 3 months Daily
Hegde et al. 2011 India 63 c, 60 tx RCT Yoga 3 months 3 days/week
Beena et al. 2013 (A) India 37 c, 33 tx RCT Yoga 3 months 6 days/week
Beena et al. 2013 (B) India 21 c, 26 tx RCT Yoga 3 months 6 days/week
Beena et al. 2013 (C) India 12 c, 14 tx RCT Yoga 3 months 6 days/week
Vizcaino et al. 2013 USA 10 Pre-post Yoga 6 weeks 3 times/week
Popli et al. 2014 India 80 Pre-post Yoga 6 months 5 days/week
Rajani et al. 2015 India 34c, 34 tx RCT Yoga 6 months 6 days/week
Vinutha et al. 2015 India 15 Pre-post Yoga 1 week Daily
9

Mullur et al. 2016 USA 5c, 5 tx RCT Yoga 3 months Daily
Angadi et al. 2017 India 52 Pre-post Yoga 6 months Daily
Sreedevi et al. 2017 India 38 c, 35 tx RCT Yoga 3 months 2 days/week
Mondal et al. 2018 India 10 c, 10 tx RCT Yoga 12 weeks 3 days/week
Vijayakumar et al. 2018 India 189 Pre-post Yoga 10 days Daily
Vijayakumar et al. 2019 India 9 Pre-post Yoga 14 days daily
Nair et al. 2021 India 23 c, 22 tx RCT Yoga 10 weeks 4 days/week
Viswanathan et al. 2021 India 150 c, 150 tx RCT Yoga 3 months 5 days/week
40

FIGURE 3. Results of the mean change in HbA1c. (a) The forest plot of change in
HbA1c stratified by intervention type. The mean difference in HbA1c was computed as
treatment minus control. The combined result together with heterogeneity statistics was
computed for each intervention type and overall across all interventions (overall
p < 0.0001). (b) Funnel plot of studies reporting on change in HbA1c. The funnel plot
appears symmetrical and shows no significant evidence of publication bias.

41

FIGURE 4. Results of the mean change in FBG. (a) The forest plot summarizes the
meta-analysis for the change in the mean FBG stratified by intervention type. The mean
difference in FBG was computed as treatment minus control. The combined result
together with heterogeneity statistics was computed for each intervention type and
overall across all interventions (overall p < 0.0001). (b) Funnel plot appears symmetrical
and shows no significant evidence of publication bias. FBG, fasting blood glucose.

42

FIGURE 5. (a) Bubble plot of change in HbA1c. The bubble plot shows that the mean
change in HbA1c decreases as the number of weekly yoga practice increases. For
every additional day of yoga practice per week, the mean difference in HbA1c
decreases -0.22% (95% CI: -0.44% to -0.003%; p = 0.046) over the study period. (b)
The bubble plot shows no relationship between the raw mean change in FBG and the
number of weekly yoga practice. For every additional day of yoga practice per week, the
raw mean difference in FBG increases 1.9 mg/dL (95% CI: -3.14 to 6.96 mg/dL; p =
0.46) over the study period.

43

FIGURE 6. Cumulative meta-analyses by study weight (large to small) for yoga
intervention studies. (a) Results for change in HbA1c, and (b) results for change in FBG.
In both cases, only one study is required to achieve statistical significance.

44

SUPPLEMENTARY FIGURE 1. Bubble plots for studies reporting on both HbA1c and
FBG. a. Shows that the mean change in HbA1c decreases as the number of weekly
yoga practice increase. For every additional day of yoga practice per week, the mean
difference in HbA1c decreases -0.27% (95% CI: -0.50%, -0.04%; p=0.02) over the study
period. b. Shows no relationship between the mean change in FBG and the number of
weekly yoga practice. For every additional day of yoga practice per week, the raw mean
difference in FBG increases 2.44 mg/dl (95% CI: -12.30 mg/dl, 17.22mg/dl; P=0.75)
over the study period.

45

CHAPTER THREE: Cortisol Biomarkers Are Not Associated with Perceived Stress
in Predominantly Latino Adolescents

1. Abstract

Objective: We previously showed that mind and body practices improve
glycemic control in people with type 2 diabetes. A plausible mechanism is reducing
psychological stress and lowering stress biomarkers thereby reducing glycemia. But the
relationship between subjectively experienced stress and stress biomarkers, and
subsequent glycemic effects is not understood. Our primary objective was to determine
whether perceived stress is associated with cortisol levels, and our secondary objective
was to determine the relationships of both Perceived Stress Scale (PSS), and cortisol
biomarkers with fasting blood glucose (FBG) levels.
Methods: We examined 229 healthy, primarily Latino, adolescents (mean
age=15.8 years, 70 males, all BMI categories). Perceived stress was assessed using
the 14-item questionnaire of Perceived Stress Scale (PSS). Serum cortisol (sCOR) was
measured at fasting. Salivary cortisol was measured 3 times daily (awakening, 30-min
post-awakening, and evening) to assess Cortisol Awakening Response (CAR= 30-
minute post-awakening - awakening) and Diurnal Cortisol Slope (DCS= evening-
awakening). We performed multivariate linear regressions and mixed effects linear
regressions to estimate baseline associations between PSS, and outcome measures of
cortisol biomarkers (sCOR, CAR, DCS), as well as baseline associations between
fasting blood glucose (FBG), a biomarker known to be associated with stress, with each
46
measure of stress (PSS, sCOR, CAR, DCS) as predictor variables. We used
generalized least square for linear random effect regression to estimate the 12-week
longitudinal associations between PSS and outcome measures of cortisol biomarkers
(sCOR, CAR, DCS), then between FBG with each measure of stress (PSS, sCOR,
CAR, DCS) as predictor variables. Analyses were adjusted for age, sex, and BMI.
Results: At baseline, there was no association between PSS and sCOR
(p=0.40), CAR (p=0.26), DCS (p =0.38), or FBG (p = 0.47). There was a significant
association between sCOR and FBG: b=0.02 mg/dl ±0.005, p <0.001; but no significant
association between CAR and FBG (p=0.60), or between DCS and FBG (p=0.80).
Further, there was no longitudinal association between PSS and sCOR (p=0.34), CAR
(p=0.66), or DCS (p=0.12), but a significant longitudinal association between sCOR and
FBG: b=0.017 mg/dl ±0.004, p <0.001.
Conclusion: We found no evidence of association between PSS and cortisol
biomarkers of stress. We further show an association between sCOR and FBG, both
cross-sectionally and longitudinally, but no association between PSS and FBG. These
findings suggest that mind and body practices may modulate cortisol levels, and
subsequent glycemic control through a pathway independent of perceived stress, and
PSS may be related to other physiological stress responses that do not involve
activation of the HPA axis. Further investigation is needed to determine how mind and
body practices influence glycemic control and which subjective aspects of stress affect
physiological stress responses.

2. Introduction

47
Previous studies
1-4
have shown that chronic stress can lead to activation of the
hypothalamic pituitary-adrenal (HPA) axis
5
through neuro-endocrine mechanisms. This
activation can cause an increase in cortisol secretion, resulting in accumulation of fat
and weight gain
6
that may lead to insulin resistance
7, 8
and hyperglycemia. We showed
by meta-analysis that mind and body practices improve glycemic control in people with
type 2 diabetes
9
, but the mechanism is not known. Prior studies suggest that mind and
body practices may also decrease perceived stress
10
and improve mental and
emotional health
11
. Others show elevated serum cortisol to be associated with poor
glycemic control in people with
12
and without
13
type 2 diabetes. We found change in
perceived stress to be positively correlated with change in cortisol awakening
response
14
in 17 adolescents without diabetes, following a six-week guided imagery
intervention. We hypothesize that mind and body practices reduce perceived stress,
which lowers cortisol biomarkers, thereby reducing glycemia.
The perceived stress scale (PSS) quantifies subjective stress; the degree to
which one’s life is perceived to be unpredictable, uncontrollable, and/or overloading.
The PSS measures integrated levels of perceived stress of the preceding month, and
therefore measures chronic or non-acute subjective stress. In general, individual self-
rating of the stressfulness of events has been shown to be a better predictor of various
health outcomes than objective measures of stressful life events
15-17
.
Cortisol is a steroid hormone involved in physiological processes including
metabolism
18, 19
, immune function
20
, and the body's response to stress. Activation of the
stress-related hypothalamic-pituitary adrenal (HPA) contributes to increased cortisol
secretion, weight gain, risk of chronic insulin resistance, and risk of type 2 diabetes
7, 8
.
48
Measures of cortisol used in clinical research include serum cortisol, the cortisol
awakening response (CAR), and diurnal cortisol slope (DCS)
21
. CAR is the increase in
cortisol levels within 30-60 min after awakening in the morning
22, 23
and DCS is the
decrease from morning to evening over the waking day
24
.
Despite widespread use of the PSS, few studies have evaluated the association
between the PSS and cortisol biomarkers, particularly in adolescent populations. A
study of adult women with fibromyalgia identified no association between PSS and
CAR
25
, but did not explore other cortisol biomarkers. A study of adolescents identified a
significant correlation between acute performance-related social stress
26
assessed by
the Groningen Social Stress Test, rather than PSS, and mean salivary cortisol. A study
of male college students in China identified higher perceived stress and higher CAR in
students 2-days prior to taking an exam compared to students not planning to take the
exam
27
. However, associations between PSS and CAR were not evaluated. Important
gaps therefore remain in understanding the relationship between subjective stress and
cortisol biomarkers.
Our primary objective was to determine whether subjective stress, as measured
by the PSS, is associated with cortisol levels measured by serum cortisol, CAR, and
DCS. Our secondary objective was to assess the associations between fasting glucose,
a measure known to be affected by stress
28
, and both PSS and cortisol biomarkers.

3. Methods

3.1 Study Design and Participants

49
We used data from the Imagine Healthy Eating Active Living Total Health
(HEALTH) study, which was a 12-week randomized controlled trial. The protocols and
details of the Imagine HEALTH study have been previously reported
29
. The trial aimed
to test the effectiveness of a lifestyle education program combined with the mind-body
modality of guided imagery to address obesity-related lifestyle behaviors and stress
biomarkers in adolescents. The study sample included 229 predominantly Latino male
and female adolescents aged 14-17 from four inner-city high schools. Participants who
agreed to attend up to 3 after-school classes per week for 12 weeks were included.
Those with serious chronic illness, taking medications known to affect the HPA axis
(e.g. oral, nasal, or inhaled glucocorticoids), cognitive behavioral disability, or prior
diagnosis of clinical eating disorder or psychiatric disorder were excluded. Informed
consent was obtained from parents and youth assent from participants.
The 12-week intervention was conducted during the spring semester for three
consecutive school years from 2015-2017. Cluster randomization was performed at the
school level, with blocking to ensure no school had the same intervention arm more
than once across the three waves. The intervention arms during the study included a
non-intervention control group, a lifestyle education group, a stress reduction guided
imagery group, and a lifestyle behavior guided imagery group. The specific details of the
guided imagery delivery and content have been previously described
29
. The study was
approved by the Internal Review Board of the University of Southern California.

3.2 Outcome Measures

50
3.2.1 Measurement Visits

Measurement visits were conducted on weekend mornings at baseline and at
follow-up after the 12-week intervention. Participants arrived following an overnight fast
and fasting blood was drawn for serum cortisol and plasma glucose and surveys were
administered by trained research staff. Participants’ perceived stress was assessed
using the standard 14-item PSS questionnaire
15
, which assessed perception of stress in
the preceding month. Age, sex, ethnicity, race, weight, height, parental education, and
parental income were self-reported. BMI was calculated as weight (kg) divided by height
(m)
2
.

3.2.2 Diurnal Salivary Cortisol Pattern

In the week following measurement visits, saliva samples were collected by
participants in their home environments using salivettes at baseline and follow-up over
three consecutive weekdays at three specific times: upon awakening, 30 minutes after
awakening, and just before bedtime. Prior to baseline saliva collection, participants
received in-person training on how to collect saliva samples and used a previously
validated mobile application called ZEMI
30
that was equipped to emit personalized
audible wake-up alarms, prompt participants to provide photos of the collected saliva
sample, and transmit data in real-time through mobile data or Wi-Fi connection. The
timing of sample collection was assessed using the timestamp of the uploaded pictures.
Samples were kept in participants’ freezers until retrieved by study personnel, at which
point they were thawed and centrifuged.

51
3.2.3 Assays

Saliva supernatant was aliquoted and stored in cryovials at -80
0
C until
subsequent assay for cortisol using a commercially available ELISA (Sal metrics, Inc;
inter-assay CV = 3.75% [high], 6.41% [low]). No individual cortisol values were more
than 3SD above the mean, and therefore no samples were excluded from analysis for
this reason. Samples were excluded (n= 25) if time of the awakening sample as
documented by the cell phone app ZEMI was before 4AM or after 12 noon. A total of
993 and 972 salivary cortisol samples were available for calculation of CAR and DCS
respectively and were included in data analysis. Timing of salivary sampling was
confirmed by ZEMI on the participant samples which had documented timing, showing
an average time of 32±12.4 min (n=700) for CAR, and 14.7 ±2.8 hours (n=596) for DCS.
Serum cortisol was assessed using commercial ELISA assay, and fasting blood glucose
using a YSI autoanalyzer using the glucose oxidase method. Assays were performed in
the USC Diabetes and Obesity Research Institute Metabolic Lab.

3.2.4 Statistical Analyses

The CAR was estimated as the difference between 30 minutes post-awakening
and awakening of salivary cortisol (i.e. mean level increase method)
23, 31
and the DCS
as the difference between evening and awakening salivary cortisol. Average CAR and
DCS were estimated as the average of 3-day measures of CAR and DCS for both
baseline and follow up. The specific interventions are not pertinent to the hypotheses
under examination in this study and therefore pre- and post-intervention participant data
from all 4 intervention arms were pooled for analysis.
52
Characteristics of study participants were summarized as means and standard
deviations for continuous variables and frequency and percentage for categorical
variables. Multivariate linear regression was used to determine baseline associations
between PSS (predictor variable) and serum cortisol (outcome variables), while mixed
effects linear regression was used to determine baseline associations between PSS
and outcome measures of CAR and DCS. Additionally, multivariate linear regressions
were also used to determine the baseline association between fasting blood glucose as
outcome variable and predictor variables of stress, including PSS, serum cortisol, CAR,
and DCS.
Generalized least square (GLS) for linear random effect regressions were used
to estimate 12-week longitudinal associations of PSS, as a predictor variable, and
cortisol biomarkers, as outcome variables. GLS for linear random effect regression was
additionally used to determine 12-week longitudinal associations between PSS, cortisol
biomarkers, as predictor variables, and fasting blood glucose (outcome variable).
Cortisol biomarkers included CAR, DCS, and serum cortisol. All analyses were adjusted
for age, sex, and BMI. In addition, longitudinal associations were adjusted for
intervention group. Residual plots of predicted vs. observed variables were used to
assess the linearity, constant variance, and normality assumptions of the models.

4. Results

Table 2 shows characteristics of the study participants. Because the
interventions were not relevant for these analyses, data from all intervention groups
were combined. Of the 229 study participants, 67% were female and 95% self-identified
53
as Hispanic or Latino. Participants’ mean (±SD) age was 15.8± 0.7. Mean BMI was
23.6± 7.4 kg/m
2
with three percent of study participants less than the 5
th
percentile
(underweight), 50% percent in the 5
th
-85
th
percentile (normal), 18% in the 85
th
-94
th

percentile (overweight), and 22% greater than the 95
th
percentile (obese). Supplemental
Table 1 provides the participant characteristics stratified by intervention type and show
no differences across intervention groups.
Table 3 shows baseline and 12-week longitudinal associations between PSS and
cortisol biomarkers. There was no significant association between PSS and any cortisol-
based biomarker at baseline (all p³ 0.26). Similarly, in longitudinal analysis, PSS was
not associated with serum cortisol (p=0.34), CAR (p=0.66), or DCS (p=0.12]).
Table 4 shows baseline and longitudinal associations between fasting blood
glucose levels and each of PSS and individual cortisol biomarkers. There was no
association between fasting blood glucose and PSS, or any salivary cortisol measures
at baseline or longitudinally (all p³ 0.47). However, we observed a statistically
significant association between serum cortisol and fasting blood glucose at both
baseline (p<0.001) and longitudinally (p<0.001). Mean fasting blood glucose increased
0.02 mg/dl ±0.005 for every nmol/l increase in serum cortisol at baseline, and 0.02
mg/dl ±0.004 for every nmol/l increase in serum cortisol at follow up. The predicted vs.
residual plots revealed no violations of linearity, constant variance, and normality
assumptions of the models.

5. Discussion

54
The objective of this study was to assess the association between subjective
stress, as measured by the PSS, and objective stress, as measured by multiple cortisol
biomarkers. We found no cross-sectional or longitudinal association between PSS and
cortisol biomarkers in this population of predominantly Latino adolescents. Our findings
have significant implications for the predictive capacity of the PSS regarding cortisol
biomarkers and stress-related physiological responses. Contrary to our hypothesis, we
found no significant association between perceived stress and cortisol biomarkers,
suggesting that the PSS may not capture HPA-related stress response. Our observation
that PSS is not associated with cortisol biomarkers accord with findings from a cross-
sectional study in a female adult population with fibromyalgia
25
. Together these findings
suggest that PSS may capture a different aspect of chronic stress than that reflected in
cortisol biomarkers. It is possible that the PSS may be associated with other
physiological stress responses that do not involve HPA activation, such as the
sympathetic nervous system
32, 33
.
The human stress response is regulated by two major physiological systems: the
autonomic nervous system and the hypothalamic-pituitary-adrenal (HPA) axis
34
. The
ability to measure salivary cortisol levels has led to the characterization of daily HPA
axis activity as represented by a typical diurnal cortisol pattern
35
, which includes CAR
and DCS
23
. Studies showed CAR to be unrelated to cortisol reactivity to experimentally
induced psychological stress
26, 36
. However, there is not enough research to determine
whether CAR or DCS are affected by chronic psychological stress, as measured by
PSS. Most studies investigating the association between CAR and psychological stress
used instruments other than PSS. Such studies have shown associations between CAR
55
and psychological factors including fatigue, burnout, exhaustion
37
, lack of social
recognition
38
, overload
39
, financial strain
40
, employment related stress
41
, and early life
adversity
42
. These findings suggest that specialized instruments for measuring
psychological stress may be more effective at capturing the physiological stress
response reflected in HPA activity than the PSS.
The American Psychological Association defines stress as the physiological and
psychological response to a condition that threatens or challenges a person and
requires some form of adaptation or adjustment
43
. The three broad classifications of
stress in clinical research are environmental (objective stressors or life events),
psychological (subjective appraisal and affective reactions) and biological
44
. The PSS is
one of the most widely used tools to assess psychological stress and has been
validated in people age 12 and above
45
across different sex, racial, ethnic
46
, and
linguistic groups
47
. Yet, few studies have evaluated the relationship between
psychological and objective stress in general, and even fewer between PSS and cortisol
biomarkers, especially in adolescents. Of the few available studies that assessed the
relationship between subjective and objective stress, most either use a stress biomarker
other than cortisol or a different instrument to measure psychological stress. This is the
case in the study by Föhr and colleagues who found a positive correlation between PSS
and heart rate variability, not cortisol, as the objective measure of stress
48
. Another
study
49
looked at the association between acutely induced stress and a single measure
of salivary cortisol and found a higher stress perception to be correlated with lower
concentration of salivary cortisol. Acutely induced stress was achieved by exposure to
socially evaluated cold pressor test, which is different from the PSS. One occupational
56
cross-sectional study in police officers, an occupation susceptible to high stress,
reported significantly higher mean values of PSS and serum cortisol in police officers
compared to their counterparts in the general population
50
. But whether PSS and serum
cortisol levels were correlated was not directly assessed. Because the PSS is widely
used, there is a need to identify its relationship to current biological measures of stress
in order to understand which activation of physiological systems, if any, involved in
stress response corresponds to the psychological dimension of stress captured by the
PSS.
Although we found no association between CAR, DCS, and fasting blood
glucose, we did find a significant association between serum cortisol and fasting blood
glucose. In these analyses, fasting blood glucose was used as a representative
outcome known to be affected by stress. Therefore, these results suggest that serum
cortisol levels may be a more sensitive indicator of HPA activity than CAR and DCS in
this population.
Our study is novel since it is the first, to our knowledge, to examine both cross
sectional and longitudinal associations between perceived stress, serum cortisol, and
diurnal cortisol patterns in adolescents. Our recent study builds upon the SOLAR
51

(Study of Latino Adolescents at Risk) Diabetes Project by demonstrating a consistent
association between serum cortisol and fasting blood glucose in adolescents. However,
it goes beyond the scope of the SOLAR Diabetes Project by examining the longitudinal
relationship between serum cortisol and fasting blood glucose. The consistency of
association between serum cortisol and fasting blood glucose at baseline and
longitudinally helps us to understand not only the association between those variables,
57
but also their patterns over time. Based on our findings, it seems that PSS does not
mediate the effect of mind and body practices on cortisol biomarkers, as was previously
hypothesized. Thus, mind and body practices may impact serum cortisol, and
subsequent blood sugar levels through a currently unknown physiological mechanism.
Alternatively, the effects of mind and body practices on serum cortisol, and glycemic
control could be behavioral rather than physiological. Further research is needed to
clarify these gaps.
In this study, we endeavored to provide valid findings by random selection of
subjects from four inner city high schools, making participation bias unlikely; minimizing
information bias by use of technical rigor, validated measurement techniques in the
assessment of biomarkers, and implementing analyses controlled for a priori potential
confounders including age, sex, and BMI.

6. Conclusion:

We found no evidence that PSS is associated with cortisol stress biomarkers or
fasting blood glucose. However, we did find an association between serum cortisol and
fasting blood glucose. These findings suggest that mind and body practices may lower
cortisol levels and improve glycemic control through a pathway independent of
perceived stress, and that PSS may be related to other physiological stress responses
that do not involve activation of the HPA axis. Further studies are needed to investigate
these gaps in knowledge.

58
7. References

1. Shively CA, Register TC, Clarkson TB. Social stress, visceral obesity, and
coronary artery atherosclerosis in female primates. Obesity (Silver Spring). Aug
2009;17(8):1513-1520.

2. Rosmond R, Dallman MF, Björntorp P. Stress-related cortisol secretion in men:
relationships with abdominal obesity and endocrine, metabolic and hemodynamic
abnormalities. J Clin Endocrinol Metab. Jun 1998;83(6):1853-1859.

3. Epel ES, McEwen B, Seeman T, et al. Stress and body shape: stress-induced
cortisol secretion is consistently greater among women with central fat.
Psychosom Med. Sep-Oct 2000;62(5):623-632.

4. Marniemi J, Kronholm E, Aunola S, et al. Visceral fat and psychosocial stress in
identical twins discordant for obesity. J Intern Med. Jan 2002;251(1):35-43.

5. Kyrou I, Tsigos C. Stress hormones: physiological stress and regulation of
metabolism. Curr Opin Pharmacol. Dec 2009;9(6):787-793.

6. Peckett AJ, Wright DC, Riddell MC. The effects of glucocorticoids on adipose
tissue lipid metabolism. Metabolism. Nov 2011;60(11):1500-1510.

7. Andrews RC, Walker BR. Glucocorticoids and insulin resistance: old hormones,
new targets. Clin Sci (Lond). May 1999;96(5):513-523.

8. Antuna-Puente B, Feve B, Fellahi S, Bastard JP. Adipokines: the missing link
between insulin resistance and obesity. Diabetes Metab. Feb 2008;34(1):2-11.

9. Sanogo F, Xu K, Cortessis VK, et al. Mind- and Body-Based Interventions
Improve Glycemic Control in Patients with Type 2 Diabetes: A Systematic
Review and Meta-Analysis. J Integr Complement Med. Sep 7 2022.

10. Weigensberg MJ, Wen CK, Sanogo F, et al. Imagine HEALTH: Changes in
diurnal salivary cortisol patterns following a 12-week guided imagery RCT
lifestyle intervention in predominantly Latino adolescents.
Psychoneuroendocrinology. May 2023;151:106053.

11. Mind and Body Practices 2017. https://www.nccih.nih.gov/health/mind-and-body-
practices

12. Ortiz R, Kluwe B, Odei JB, et al. The association of morning serum cortisol with
glucose metabolism and diabetes: The Jackson Heart Study.
Psychoneuroendocrinology. May 2019;103:25-32.
59
13. Turakitwanakan W, Mekseepralard C, Busarakumtragul P. Effects of mindfulness
meditation on serum cortisol of medical students. J Med Assoc Thai. Jan 2013;96
Suppl 1:S90-95.

14. Weigensberg MJ, Provisor J, Spruijt-Metz D, et al. Guided Imagery Council:
Feasibility, Acceptability, and Preliminary Effects of a Novel Group-Based
Lifestyle Intervention in Predominantly Latino Adolescents. Glob Adv Health Med.
2019;8:2164956119844477.

15. Cohen S, Kamarck T, Mermelstein R. A global measure of perceived stress. J
Health Soc Behav. Dec 1983;24(4):385-396.

16. Sarason IG, Johnson JH, Siegel JM. Assessing the impact of life changes:
development of the Life Experiences Survey. J Consult Clin Psychol. Oct
1978;46(5):932-946.

17. Vinokur A, Selzer ML. Desirable versus undesirable life events: their relationship
to stress and mental distress. J Pers Soc Psychol. Aug 1975;32(2):329-337.

18. Lightman SL, Birnie MT, Conway-Campbell BL. Dynamics of ACTH and Cortisol
Secretion and Implications for Disease. Endocr Rev. Jun 1 2020;41(3).

19. Hewagalamulage SD, Lee TK, Clarke IJ, Henry BA. Stress, cortisol, and obesity:
a role for cortisol responsiveness in identifying individuals prone to obesity.
Domest Anim Endocrinol. Jul 2016;56 Suppl:S112-120.

20. Rook GA. Glucocorticoids and immune function. Baillieres Best Pract Res Clin
Endocrinol Metab. Dec 1999;13(4):567-581.

21. Adam EK, Quinn ME, Tavernier R, et al. Diurnal cortisol slopes and mental and
physical health outcomes: A systematic review and meta-analysis.
Psychoneuroendocrinology. Sep 2017;83:25-41.

22. Fries E, Dettenborn L, Kirschbaum C. The cortisol awakening response (CAR):
facts and future directions. Int J Psychophysiol. Apr 2009;72(1):67-73.

23. Stalder T, Kirschbaum C, Kudielka BM, et al. Assessment of the cortisol
awakening response: Expert consensus guidelines. Psychoneuroendocrinology.
Jan 2016;63:414-432.

24. Adam EK, Kumari M. Assessing salivary cortisol in large-scale, epidemiological
research. Psychoneuroendocrinology. Nov 2009;34(10):1423-1436.

25. Weissbecker I, Floyd A, Dedert E, et al. Childhood trauma and diurnal cortisol
disruption in fibromyalgia syndrome. Psychoneuroendocrinology. Apr
2006;31(3):312-324.
60
26. Bouma EM, Riese H, Ormel J, et al. Adolescents' cortisol responses to
awakening and social stress; effects of gender, menstrual phase and oral
contraceptives. The TRAILS study. Psychoneuroendocrinology. Jul
2009;34(6):884-893.

27. Duan H, Yuan Y, Zhang L, et al. Chronic stress exposure decreases the cortisol
awakening response in healthy young men. Stress. Nov 2013;16(6):630-637.

28. Surwit RS, Schneider MS, Feinglos MN. Stress and diabetes mellitus. Diabetes
Care. Oct 1992;15(10):1413-1422.

29. Weigensberg MJ, Spruijt-Metz D, Wen CKF, et al. Protocol for the Imagine
HEALTH Study: Guided imagery lifestyle intervention to improve obesity-related
behaviors and salivary cortisol patterns in predominantly Latino adolescents.
Contemp Clin Trials. Sep 2018;72:103-116.

30. Cheng K. Fred Wen SS, Marc J. Weigensberg, Bas Weerman, Donna Spruijt-
Metz. Accuracy of a Photo-based Smartphone Application to Assess Salivary
Cortisol Sampling Time in Adolescents. SAGE Journals 2022.

31. Marc J. Weigensberg CKFW, Fatimata Sanogo, Claudia M Toledo-Corral, Li
Ding, . Imagine HEALTH: Changes in diurnal salivary cortisol patterns following a
12-week guided imagery RCT lifestyle intervention in predominantly Latino
adolescents. Psychoneuroendocrinology. 2023.

32. Janczura M, Dropinski J, Gielicz A, et al. Potential roles of psychological and
oxidative stress in insulin resistance: a cohort-based study. Diabetol Metab
Syndr. 2020;12:58.

33. Yang Y, Bi M, Xiao L, et al. Perceived stress status and sympathetic nervous
system activation in young male patients with coronary artery disease in China.
Eur J Intern Med. Nov 2015;26(9):726-730.

34. Chrousos GP. The role of stress and the hypothalamic-pituitary-adrenal axis in
the pathogenesis of the metabolic syndrome: neuro-endocrine and target tissue-
related causes. Int J Obes Relat Metab Disord. Jun 2000;24 Suppl 2:S50-55.

35. Miller R, Stalder T, Jarczok M, et al. The CIRCORT database: Reference ranges
and seasonal changes in diurnal salivary cortisol derived from a meta-dataset
comprised of 15 field studies. Psychoneuroendocrinology. Nov 2016;73:16-23.

36. Schmidt-Reinwald A, Pruessner JC, Hellhammer DH, et al. The cortisol response
to awakening in relation to different challenge tests and a 12-hour cortisol
rhythm. Life Sci. 1999;64(18):1653-1660.

61
37. Chida Y, Steptoe A. Cortisol awakening response and psychosocial factors: a
systematic review and meta-analysis. Biol Psychol. Mar 2009;80(3):265-278.

38. Wüst S, Federenko I, Hellhammer DH, Kirschbaum C. Genetic factors, perceived
chronic stress, and the free cortisol response to awakening.
Psychoneuroendocrinology. Oct 2000;25(7):707-720.

39. Schlotz W, Hellhammer J, Schulz P, Stone AA. Perceived work overload and
chronic worrying predict weekend-weekday differences in the cortisol awakening
response. Psychosom Med. Mar-Apr 2004;66(2):207-214.

40. Steptoe A, Brydon L, Kunz-Ebrecht S. Changes in financial strain over three
years, ambulatory blood pressure, and cortisol responses to awakening.
Psychosom Med. Mar-Apr 2005;67(2):281-287.

41. Gustafsson PE, Janlert U, Virtanen P, Hammarström A. The association between
long-term accumulation of temporary employment, the cortisol awakening
response and circadian cortisol levels. Psychoneuroendocrinology. Jun
2012;37(6):789-800.

42. Meinlschmidt G, Heim C. Decreased cortisol awakening response after early loss
experience. Psychoneuroendocrinology. Jul 2005;30(6):568-576.

43. Stephanie Franks SMS, Ann Vanichkachorn, . Stress and Health 2023.

44. Lee EH. Review of the psychometric evidence of the perceived stress scale.
Asian Nurs Res (Korean Soc Nurs Sci). Dec 2012;6(4):121-127.

45. Child Outcomes Research Consortium. Outcome and Experience Measures,
Perceived Stress Scale 2023.

46. Baik SH, Fox RS, Mills SD, et al. Reliability and validity of the Perceived Stress
Scale-10 in Hispanic Americans with English or Spanish language preference. J
Health Psychol. Apr 2019;24(5):628-639.

47. Makhubela M. Assessing psychological stress in South African university
students: Measurement validity of the Perceived Stress Scale (PSS-10) in
diverse populations. APA PsychNet. 2020.

48. Föhr T, Tolvanen A, Myllymäki T, et al. Subjective stress, objective heart rate
variability-based stress, and recovery on workdays among overweight and
psychologically distressed individuals: a cross-sectional study. J Occup Med
Toxicol. 2015;10:39.

62
49. Buzgoova K, Balagova L, Marko M, et al. Higher perceived stress is associated
with lower cortisol concentrations but higher salivary interleukin-1beta in socially
evaluated cold pressor test. Stress. May 2020;23(3):248-255.

50. Walvekar SS, Ambekar JG, Devaranavadagi BB. Study on serum cortisol and
perceived stress scale in the police constables. J Clin Diagn Res. Feb
2015;9(2):Bc10-14.

51. Weigensberg MJ, Toledo-Corral CM, Goran MI. Association between the
metabolic syndrome and serum cortisol in overweight Latino youth. J Clin
Endocrinol Metab. Apr 2008;93(4):1372-1378.

63

TABLE 2: Participant Demographics and Clinical Characteristics

N = 229

Grade 10th 84 (36.7%)
11th 145 (63.3%)

Sex Female 153 (66.8%)
Male 76 (33.2%)

Ethnicity Hispanic or Latino 217 (94.7%)
Not Hispanic or Latino 12 (5.3%)

Race White 205 (89.9%)
Other 24 (10.1%)

Parent SES ($*) < 10,000 32 (15.5%)
> 50,000 15 (7.3%)
10,000-20,000 68 (33.0%)
20,000-30,000 52 (25.2%)
30,000-40,000 21 (10.2%)
40,000-50,000 18 (8.7%)

Mean ± SD
Age (years)

15.75 ± 0.66

BMI (Body Mass Index)

23.64 ± 7.42
64

Salivary Cortisol (nmol/L**)

Baseline Awakening 9.2 ± 0.46
+30 minutes 14.9 ± 1.2
Evening 2.4 ± 0.46

3 Month Awakening 9.2 ± 1.0
+30 minutes 14.4 ± 1.2
Evening 2.2 ± 0.91

Perceived Stress Score
Baseline 24.5 ± 1.5
3 months 24.5 ± 1.4

Serum Cortisol (nmol/L)
Baseline 409.69 ± 158.08
3 months 397.12 ±148.89

Fasting Blood Glucose (mg/dl)
Baseline 87.30 ± 11.60
3 months 88.53 ± 11.86
Continuous variables reported as mean(±SD) and categorical variables as count (%)

65
TABLE 3: Association between PSS and biomarkers of stress
Baseline Association* Longitudinal Association
Adjusted for age, sex, and BMI Adjusted for age, sex, BMI, and study group
Beta±SE Test p-value Beta±SE Test p-value
sCOR -1.42±1.68 -0.85 0.40 -1.24±1.31 -0.95 0.34
CAR 0.07±0.06 -1.12 0.26** -0.02±0.05 -0.44 0.66
DCS -0.04±0.05 -0.88 0.38** -0.06±0.04 -1.54 0.12
sCOR= serum cortisol. Predictor Variable: PSS, outcome Variable: sCOR, CAR, DCS
*Analyses performed using multivariate linear regression
**Analyses performed using mixed effect linear regression.
Longitudinal analyses performed using GLS for linear random effect regression.

66
TABLE 4: Association between PSS, cortisol biomarkers, and FBG
Baseline Association 12-Week Longitudinal Association
Adjusted for age, sex, and BMI Adjusted for age, sex, BMI, and study group
Beta±SE Test p-value Beta±SE Test p-value
sCOR 0.02±0.005 4.52 <0.001 0.017±0.004 4.59 <0.001
CAR 0.07±0.14 0.53 0.60 0.07±0.11 0.66 0.51
DCS 0.04±0.15 0.25 0.80 -0.01±0.11 -0.11 0.91
PSS 0.08±0.11 0.72 0.47 -0.003±0.09 -0.03 0.98
sCOR= serum cortisol. Predictor Variable: PSS, sCOR, CAR, DCS. Outcome Variable
Baseline analyses performed using multivariate linear regressions. Average CAR and DCS used as predictor
variables.
Longitudinal analyses performed using GLS for linear random effect regression.

67
SUPPLEMENTAL TABLE 1: Participant Demographics and Clinical Characteristics by
Treatment Group

Control
(N= 51)
N(%)
LS
(N= 60)
N(%)
LBGI
(N= 54)
N(%)
SRGI
(N= 64)
N(%)

p-value

Grade 10th 11 (21.6%) 31 (51.7%) 23 (42.6%) 19 (29.7%) 0.01
11th 40 (78.4%) 29 (48.3%) 31 (57.4%) 45 (70.3%)

Sex Female 38 (74.5%) 35 (58.3%) 40 (74.1%) 40 (62.5%) 0.16
Male 13 (25.5%) 25 (41.7%) 14 (25.9%) 24 (37.5%)

Ethnicity Hispanic or
Latino
50 (98.0%) 56 (93.2%) 51 (94.3%) 60 (93.8%) 0.67
Not Hispanic or
Latino
1 (2.0%) 4 (6.8%) 3 (5.7%) 4 (6.2%)

Race White 50 (98.0%) 54 (90.0%) 45 (83.3%) 56 (88.9%) 0.04
Other 1 (2.0%) 6 (10.0%) 9 (16.7%) 8 (11.1%)

Parent SES*
< 10,000
6
(12.5%) 12 (21.1%) 6 (13.0%) 8 (14.5%)
0.77
> 50,000 5 (10.4%) 3 (5.3%) 3 (6.5%) 4 (7.3%)
10,000-20,000 16 (33.3%) 16 (28.1%) 19 (41.3%) 17 (30.9%)

20,000-30,000
10
(20.8%) 17 (29.8%) 11 (23.9%) 14 (25.5%)

30,000-40,000
7
(14.6%) 2 (3.5%) 4 (8.7%) 8 (14.5%)

40,000-50,000
4
(8.3%) 7 (12.3%) 3 (6.5%) 4 (7.3%)

Mean ± SD Mean ± SD Mean ± SD Mean ± SD p-value
b

Age (years)

15.82 ±
0.48
15.63 ±
0.66
15.65 ±
0.70
15.89 ±
0.72
0.08
68

BMI (Body Mass
Index)

24.97 ±
6.41
24.34 ±
6.02
23.02 ±
5.96
25.87 ±
5.43
0.09

Salivary Cortisol
(nmol/L**)

Baseline Awakening 9.3 ± 6.3 9.3 ± 5.5 9.5 ± 5.6 8.8 ± 4.9 0.72
+30 minutes 15.8 ± 9.7 14.5 ± 6.1 14.8 ± 7.0 13.6 ± 6.5 0.08
Evening 2.7 ± 8.4 2.3 ± 5.0 2.8 ± 8.0 1.9 ± 2.4 0.55

3 Month Awakening 8.4 ± 4.6 9.4 ± 5.0 10.4 ± 6.5 9.4 ± 4.6 0.06
+30 minutes 14.0 ± 7.9 14.2 ± 7.2 16.0 ± 8.1 13.5 ± 6.2 0.11
Evening 2.4 ± 4.5 1.4 ± 1.5 2.4 ± 3.8 2.2 ± 4.1 0.14

Perceived Stress
Score

Baseline 24.3 ± 8.1 24.7 ± 5.8 23.5 ± 6.5 25.3 ± 6.9 0.59
3 months 25.2 ± 7.8 24.1 ± 7.9 23.2 ± 7.9 24.6 ± 5.6 0.71

Serum Cortisol
(nmol/L)

Baseline
407.32 ±
171.83
435.04 ±
170.75
397.04 ±
140.96
403.58 ±
155.05 0.82

3 months
382.69 ±
182.31
396.94 ±
138.20
409.93 ±
142.19
403.13 ±
136.87 0.96

Fasting Blood
Glucose

Baseline 89.24 ± 9.14
90.50 ±
10.16
84.31 ±
13.36
85.51 ±
12.60 0.06

3 months
88.85 ±
12.18
88.97 ±
12.02
84.30 ±
12.38
90.64 ±
10.71 0.20
a
Categorical variable: chi-squared test where >20% of expected cell counts are larger than 5. Fisher's exact test otherwise
b
Continuous variables: ANOVA F test
* Annual Income
** Conversion factor nmol/L /27.6 = µg/dL cortisol
p < 0.05 statistically significant

69

CHAPTER FOUR: Insulin Resistance is Associated with Serum Cortisol, but not
Perceived Stress Score in Latino Adolescents

1. Abstract

Background: Mind and body practices have been shown to improve glycemic
control, perceived stress, and cortisol biomarkers in diabetic and non-diabetic
populations. The mechanism behind these benefits is not fully understood, but it is
hypothesized that mind and body practices may reduce insulin resistance through a
stress pathway, which could have implications for reducing the risk of type 2 diabetes in
high-risk populations. However, research on the relationship between perceived stress
and insulin resistance in adults is inconsistent, and there is limited research on
associations of insulin resistance with perceived stress and with cortisol biomarkers in
adolescent populations.
Objective: To determine the associations between insulin resistance, perceived
stress, and cortisol biomarker measures in predominantly Latino adolescents, both
cross-sectionally and longitudinally.
Methods: We studied 229 healthy, primarily Latino adolescents (mean age=15.8
years, 70 males, 95% Latino) using data from the Imagine HEALTH study. We used
baseline and 12-week follow up outcome measures of Homeostatic Model Assessment
of Insulin Resistance (HOMA-IR) and fasting insulin as indices of insulin resistance. We
additionally used body mass index (BMI) and percent fat as distant proxies for insulin
resistance in Latino adolescents. We measured stress, the predictor variables, using the
14-item PSS questionnaire and cortisol levels using serum cortisol, salivary Cortisol
Awakening Response (CAR= 30-minute post-awakening - awakening), and Diurnal
70
Cortisol Slope (DCS= evening-awakening). We used multivariate linear regressions to
examine the baseline cross-sectional associations and generalized least square linear
random effect regressions to determine 12-week longitudinal associations. We
examined baseline associations between HOMA-IR (outcome variable), perceived
stress, and cortisol biomarkers. We additionally evaluated the baseline and longitudinal
associations between outcome variables of fasting insulin, BMI, and BIA, with perceived
stress as well as cortisol biomarkers. Analyses were adjusted for age, sex, and BMI.
Results: At baseline, there were no significant associations between HOMA-IR
and PSS (p=0.11), CAR (p=0.64), DCS (p=0.44), or serum cortisol (p=0.70). Fasting
insulin also showed no significant baseline associations with PSS (p=0.22) or any
cortisol biomarkers, including CAR (p=0.55), DCS (p=0.37), and serum cortisol
(p=0.29). Similarly, there were no significant associations between BMI or BIA and PSS,
CAR, DCS, or serum cortisol at baseline (all p-values > 0.10).
Furthermore, we observed no significant longitudinal associations between fasting
insulin and PSS (p = 0.67), CAR (p = 0.99), or DCS (p = 0.66), nor between BMI or BIA
and PSS, CAR, DCS, or serum cortisol over time (all p-values> 0.11)."
Over time, we did find a statistically significant association between fasting insulin and
serum cortisol (p = 0.004). Log mean fasting insulin increased by 4.9 x10
-4
pg/ml ± 1.7
x10
-4
for every nmol/l increase in serum cortisol over 12-week follow up time.
Conclusion: In predominantly Latino adolescents, change in serum cortisol over
time is associated with change in insulin resistance. This temporal relationship suggests
that lifestyle interventions, such as mind and body practices, that lower serum cortisol
may improve metabolic health in Latino adolescents.
71

2. Introduction

Mind and body practices improve glycemic control in patients with type 2
diabetes
1, 2
, but the underlying mechanism is still unknown. Several interventions
studies have demonstrated a statistically significant decrease in mean glycemia and
mean serum cortisol levels following mind and body interventions in patients with type 2
diabetes
3, 4
as well as non-diabetic populations
5
. We have previously shown that
guided imagery as a mind and body intervention reduced perceived stress in
predominantly Latino adolescents
6
, but found no evidence of associations between
perceived stress and cortisol biomarkers (chapter three).
A proposed mechanism is that mind and body practices reduce insulin resistance
and subsequently improve glycemic control through a stress pathway, either
physiological or psychological. Insulin resistance is a physiological state characterized
by impaired insulin action
7
which leads to chronic hyperglycemia
8
. Insulin resistance is
associated with a range of stress-related factors, including hypothalamic-pituitary
adrenal (HPA) axis activation
9, 10
, sympathetic nervous system overactivity
11
, and
dysfunctional adipose tissue
12
. If mind and body practices can improve insulin
resistance through a stress pathway, they may be an effective nonpharmacological
intervention to reduce the risk of type 2 diabetes in high-risk populations, such as Latino
adolescents
13
.
Although the link between stress, insulin resistance, and subsequent glycemia
has been hypothesized in the literature, available data are inconclusive
14
. Prior studies
reveal inconsistent results in the association between perceived stress and insulin
72
resistance and the risk of hyperglycemia in midlife adults
15-22
, while few studies have
investigated the effects of psychological stress on the risk of insulin resistance
development in adolescents, both cross-sectionally and longitudinally. Similarly, while
cortisol biomarkers have been shown to be associated with insulin resistance in
adults
23, 24
, few studies have investigated these associations in adolescent
populations
25, 26
. Additionally, the limited existing research on cortisol levels and insulin
resistance in adolescents focuses on those who are overweight/obese. Moreover, these
studies mostly investigated morning serum cortisol levels; no other crucial cortisol
biomarkers, such as cortisol awakening response and diurnal cortisol slope have been
evaluated.
Our primary objective is to determine the baseline and longitudinal associations
between insulin resistance, perceived stress, and different measures of cortisol
biomarkers in predominantly Latino adolescents. Based on previous research, we
hypothesize that cortisol biomarkers will be positively associated with insulin resistance.
Additionally, we predict no association between perceived stress levels and insulin
resistance in this population.
3. Methods

3.1 Study Design and Participants

We used data from the Imagine Healthy Eating Active Living Total Health
(HEALTH) study, a 12-week randomized controlled trial aimed at evaluating the
effectiveness of a lifestyle education program combined with guided imagery for stress
management and obesity prevention in predominantly Latino adolescents aged 14-17.
73
The study sample included 229 adolescents from four inner-city high schools who
attended up to three after-school classes per week for 12 weeks. Participants with
chronic illness, taking medications known to affect the HPA axis (e.g. oral, nasal, or
inhaled glucocorticoids), cognitive behavioral disability, clinical eating disorders, or
psychiatric disorders were excluded. Informed consent was obtained from parents with
youth assent from participants.
The intervention took place during the spring semester for three consecutive
school years from 2015-2017, with a cluster randomization performed at the school
level to ensure no school had the same intervention arm more than once. The
intervention arms included a non-intervention control group, a lifestyle education group,
a stress reduction guided imagery group, and a lifestyle behavior guided imagery group.
The study was approved by the Internal Review Board of the University of Southern
California. The specifics of guided imagery delivery and content have been previously
reported
27
. It is important to note that the specific interventions are not relevant to the
hypotheses being examined in this study. Therefore, participant data collected before
and after the interventions across all four intervention groups were merged for the
analyses reported here.

3.2 Measurement Visits

A measurement visit refers to various assessments, or tests conducted on study
participants during a scheduled visit or appointment. Measurement visits were carried
out during the mornings on weekends at baseline and follow-up, with participants
74
arriving after an overnight fast. Fasting blood samples were taken by certified
phlebotomists to measure serum cortisol and fasting insulin levels.
Following the blood draw, electronic tablets and the REDCap database platform
were used by trained research staff to administer surveys and capture and manage
responses. Participants' perceived stress levels were evaluated using the 14-item PSS
questionnaire, which assessed their perception of stress over the preceding month.
Self-reported demographic data, including age, sex, ethnicity, race, weight, height,
parental education, and parental income, were collected. BMI was calculated as weight
(kg) divided by height (m)
2
.
To assess body composition, we used bioimpedance analysis (BIA) to measure
the percentage of body fat and lean tissue. BIA measurements were conducted using
the Tanita TBF-310GS Body Composition Analyzer/Scale, which is a commonly used
device for non-invasive measurement of body composition
28
.

3.3 Diurnal Salivary Cortisol Pattern Measurement Visits

Saliva samples were collected from participants at two different time points
(baseline and follow-up) over three consecutive weekdays at specific times: upon
awakening, 30 minutes after awakening, and just before bedtime. Prior to the baseline
saliva collection, participants received in-person training on how to collect saliva
samples and use a mobile application called ZEMI. This application was installed on a
study cell phone loaned to the participants and emitted personalized audible alarms,
reminding them to provide photos of the collected saliva sample and transmit data in
real-time through a mobile data or Wi-Fi connection. Samples were collected using the
75
Salivette system and stored in participants' freezers until retrieved by study personnel,
at which point they were thawed and centrifuged. The validity of this method has been
reported elsewhere
29
.

3.4 Assays

Saliva supernatant was collected and stored in cryovials at -80°C until analysis
via a commercially available ELISA (Sal metrics, Inc; inter-assay CV = 3.75% [high],
6.41% [low]). None of the individual cortisol values exceeded three standard deviations
from the mean. Thus, no samples were excluded from the analysis for this reason.
Twenty-five samples where the time of the awakening collection, as documented by the
ZEMI cell phone app, was before 4AM or after 12 noon for the awakening salivary
cortisol were excluded. In total, 993 and 972 salivary cortisol samples available for
calculation of the cortisol awakening response (CAR) and diurnal cortisol slope (DCS),
respectively, and included these samples in our data analysis. The timing of sample
collection was assessed using the timestamp of the uploaded pictures. ZEMI confirmed
the timing of salivary sampling, showing an average time of 32±12.4 minutes (n=700)
for cortisol awakening response (CAR) and 14.7±2.8 hours (n=596) for diurnal cortisol
slope (DCS). To measure serum cortisol levels, a fasting serum sample was collected
between 7:30AM and 9:30AM on the day of the measurement visit. A commercial
ELISA assay was used to assess serum cortisol, and fasting insulin was measured
using the EMD Millipore Luminex xMAP multi-analyte platform. These assays were
conducted in the USC Diabetes and Obesity Research Institute Metabolic Lab.

76
3.5 Statistical Analyses

We used the Homeostatic Model Assessment of Insulin Resistance (HOMA-IR)
to assess whole-body insulin resistance computed as: HOMA-IR = (fasting glucose
(mg/dL) ´ fasting insulin (pg/mL))/405
30
. The CAR was calculated as the difference
between salivary cortisol levels 30 minutes after awakening and levels upon
awakening
31
. The DCS was determined as the difference between evening and
awakening cortisol levels. The 3 consecutive days of measurements were averaged
both at baseline and follow-up to estimate the average CAR and DCS.
Characteristics of the study participants were summarized using means and
standard deviations for continuous variables, and frequency and percentages for
categorical variables. Outcome measures of HOMA-IR, fasting insulin were used as
indices of insulin resistance. Additionally, BMI and BIA which have been associated with
insulin sensitivity in Latino adolescents
32
were also used as distant proxies for insulin
resistance. For these analyses, HOMA-IR and fasting insulin were natural log-
transformed due to their skewed distributions.
Multivariate linear regressions were used to examine the baseline associations of
each of the outcome variables mentioned above with perceived stress, serum cortisol,
CAR, and DCS. Additionally, generalized least square linear random effect regressions
were used to determine the 12-week longitudinal associations between the outcome
measures of fasting insulin, BMI, and BIA, each with the different measures of stress
including PSS, serum cortisol, CAR and DCS. All analyses were adjusted for age, sex,
and BMI. However, for the outcome analyses of BMI and BIA, adjustment was made
only for age and sex. Longitudinal associations were additionally adjusted for
77
intervention group. The linearity, constant variance, and normality assumptions of the
models was assessed by examining residual plots of the predicted versus observed
variables.

4. Results

The characteristics of the study participants are presented in Table 5. Of the total
participants, 67% were female and 94% self-identified as Hispanic or Latino with mean
(±SD) age of the participants was 16.8 years (± 0.7). The average BMI of study
participants was 23.6(± 7.4) kg/m
2
. Baseline HOMA-IR ranged from 3.11 to 6.41 units,
with a mean of 4.6 (± 0.6) units. When stratifying the analyses by intervention type, we
found that, apart from grade level, there were no significant differences between the
intervention groups with respect to participant characteristics, measures of stress, or
metabolic measures at baseline or follow-up (supplemental table 2).
Table 6 shows associations estimated at baseline between log (HOMA-IR) and
each of the measures of stress. No significant association was identified at this
timepoint between HOMA-IR and PSS (p =0.11) or any cortisol biomarker (CAR
[p=0.64], DCS [p=0.44], serum cortisol [p=0.70]).
Table 7 shows associations estimated at baseline and longitudinally between
fasting insulin and each measure of stress. There was no significant baseline
association between fasting insulin and PSS (p =0.22) or any cortisol biomarker (CAR [p
=0.55], DCS [p =0.37], serum cortisol [p=0.29]). Additionally, there was no longitudinal
association between change in fasting insulin and change in: PSS (p=0.67), CAR
(p=0.99), or DCS (p=0.66). However, we observed a statistically significant association
78
between change in fasting insulin and change in serum cortisol over time, in which Log
mean fasting insulin increased by 4.9 x10
-4
pg/ml ± 1.7 x10
-4
for every nmol/l increase in
serum cortisol over 12-week follow up time.
Table 8 shows the baseline and longitudinal associations between BMI and the
various measures of stress. There was no significant association between BMI and PSS
(p=0.23) or any cortisol biomarker (CAR [p =0.10], DCS [p =0.88], serum cortisol
[p=0.54]) at baseline. These results were consistent over time [PSS (p =0.22), CAR
(p=0.15), DCS (p =0.93), serum cortisol (p=0.77).]
Table 9 shows the baseline and longitudinal associations between BIA and the
various measure of stress. No significant association were found between BIA and PSS
(p=0.26) or any of the cortisol biomarkers (CAR [p =0.77], DCS [p =0.55], serum cortisol
[p=0.86]) at baseline. These results were consistent over time, with no significant
associations observed between BIA and PSS (p =0.94), CAR (p=0.16), DCS (p =0.11),
or serum cortisol (p=0.11). The predicted vs. residual plots revealed no violations of
linearity, constant variance, and normality assumptions of the models.

5. Discussion

In a predominantly Latino adolescent population, we evaluated the associations
between insulin resistance using outcome measures of fasting insulin levels, HOMA-IR,
BMI, and BIA, and predictor variables of both perceived stress and biomarkers of stress.
Assessing these associations both cross-sectionally at baseline and over time, we
found a statistically significant association between fasting insulin and serum cortisol
levels over 12-week follow up time. However, we did not observe any significant
79
association between fasting insulin and diurnal cortisol patterns or perceived stress at
baseline or over time. Furthermore, we found no evidence of cross-sectional
associations between HOMA-IR and the different measures of stress. Our results
additionally reveal no significant association between BMI, BIA and the different
measures of stress, both cross sectionally and overtime.
Our findings carry significant implications for understanding the potential
mechanisms underlying the impact of mind-body practices on insulin resistance, the
sensitivity of different cortisol biomarkers in reflecting HPA stress-related changes, and
the predictive value of the PSS in relation to metabolic outcomes associated with type 2
diabetes. Based on our current findings of significant associations over time between
serum cortisol and fasting insulin, it is possible that the previously observed reduction in
cortisol levels resulting from mind and body interventions
3-5, 33, 34
could lead to a
subsequent decrease in insulin resistance. This could reduce the risk of developing type
2 diabetes in a high-risk population or improve glycemic control in those already
affected by the disease. Our findings suggest that the previously observed association
between mind-body practices and serum cortisol levels is irrespective of one's
perceived stress levels. This raises the question of how these practices affect
physiological outcomes like cortisol, fasting insulin, and glycemia. It is plausible that the
pathway through which mind-body practices influence serum cortisol levels involves
either an unidentified physiological mechanism or a behavioral pathway, such as the
physical activity component of some of those modalities. The mechanism behind this
unknown phenomenon requires further investigation.
80
The observation that change in fasting insulin is associated with change in serum
cortisol but not with diurnal cortisol patterns such as CAR and DCS over time could
indicate that in this population, serum cortisol levels may be a more sensitive indicator
of HPA stress-related changes than either diurnal cortisol patterns or self-reported
perceived stress. Further research is warranted to explore the potential mechanisms
underlying these associations and their implications for adolescent health.
Previous research has produced mixed results regarding the relationship
between perceived stress and insulin resistance, with consistent patterns observed
within certain race/ethnicity groups. In particular, perceived stress has been positively
associated with insulin resistance in white adults (including Swedish and Finn men
15, 16

and Australian women
17
), independent of adiposity, whereas in American women, this
association appears to be mediated through adiposity
35
. By contrast, studies conducted
in Hispanic
18
adults in the US have found no significant association between perceived
stress and insulin resistance, after adjusting for adiposity. To our knowledge, our study
is the first to investigate this relationship over time in predominantly Latino adolescents,
and our findings indicate that perceived stress is not associated with insulin resistance,
as measured by HOMA-IR, fasting insulin, BMI, and BIA in this population. These
results align with previous research in adult Hispanic populations and suggest ethnic
differences on the effect of perceived stress on insulin resistance, as was suggested in
earlier studies
35, 36
. The present findings also suggest that the previously observed
benefits of mind and body interventions, reducing perceived stress in predominantly
Latino adolescents
6
, do not necessarily lead to subsequent improvements in cortisol
levels or insulin resistance.
81
In this study, we used HOMA-IR, fasting insulin, BMI and BIA as indices of insulin
resistance. HOMA-IR is a widely used index of insulin resistance and its estimates have
been shown to correlate with glucose clamp techniques
30, 37
. However, it may not be as
accurate in certain groups, such as the elderly or individuals with impaired glucose
tolerance or diabetes. Nonetheless, HOMA-IR is appropriate for measuring insulin
resistance in younger individuals with normal glucose tolerance, as is the case in our
population. While it may not be reliable for measuring insulin resistance longitudinally
38
,
we used fasting glucose as an index for this purpose. It is important to note that fasting
insulin may not always accurately reflect insulin sensitivity, as compensatory
mechanisms can maintain normal fasting insulin levels in some individuals with insulin
resistance. Nevertheless, in our study, it can still help us capture changes in fasting
insulin over time in relation to changes in PSS and cortisol biomarkers. BMI is a
commonly used and useful measure of obesity that is strongly correlated with adiposity
and related metabolic disorders, including insulin resistance. In this study, we found a
strong correlation of 0.80 between BMI and BIA. However, BMI is not a direct measure
of insulin resistance and may not be a reliable indicator of insulin resistance in certain
populations, such as individuals of Asian or African descent. It is important to note,
however, that in our study, 95% of participants were White Latinos. BIA is a useful and
relatively non-invasive method for estimating body composition and assessing insulin
resistance. Unlike BMI, BIA takes into account differences in lean body mass and fat
mass, providing a more accurate estimate of body fat percentage. Additionally, BIA has
been shown to be a useful predictor of insulin resistance in many populations. Although
it may not be a reliable indicator of insulin resistance in those with high skeletal muscle
82
mass, extreme obesity, or medical conditions that affect body composition, it is
important to note that in our study, participants had normal body fat and skeletal muscle
mass on average.
To minimize the potential influence of confounding factors in this study, we
conducted multivariate analyses that were adjusted for potential confounders. However,
residual confounding may be possible, as factors such as family history of diabetes or
physical activity levels were not accounted for in our analyses. To ensure the validity of
our findings, we randomly selected participants from four inner city high schools,
thereby reducing the likelihood of participation bias. Additionally, we employed technical
rigor and time point validation to minimize the possibility of information bias in our
assessment of metabolic outcomes and biomarkers.

6. Conclusion

In predominantly Latino adolescents, change in serum cortisol over time is
associated with change in insulin resistance. This temporal relationship suggests that
lifestyle interventions such as mind and body practices aimed to lower serum cortisol
may improve metabolic health in Latino adolescents. The previously observed lack of
association between perceived stress and serum cortisol, and present lack of
association between perceived stress and insulin resistance suggests that the potential
impact of mind and body practices on serum cortisol and subsequent insulin resistance
is achieved through a process that is not influenced by perceived stress. These findings
raise the question by which mechanism mind and body practices effect serum cortisol
83
levels. To fully understand how mind and body practices lower serum cortisol levels,
whether through physiological or behavioral means, further research is necessary.

7. References

1. Thind H, Lantini R, Balletto BL, et al. The effects of yoga among adults with type
2 diabetes: A systematic review and meta-analysis. Prev Med. Dec
2017;105:116-126.

2. Sanogo F, Xu K, Cortessis VK, et al. Mind- and Body-Based Interventions
Improve Glycemic Control in Patients with Type 2 Diabetes: A Systematic
Review and Meta-Analysis. J Integr Complement Med. Sep 7 2022.

3. Jung HY, Lee H, Park J. Comparison of the effects of Korean mindfulness-based
stress reduction, walking, and patient education in diabetes mellitus. Nurs Health
Sci. Dec 2015;17(4):516-525.

4. Gainey A, Himathongkam T, Tanaka H, Suksom D. Effects of Buddhist walking
meditation on glycemic control and vascular function in patients with type 2
diabetes. Complement Ther Med. Jun 2016;26:92-97.

5. Turakitwanakan W, Mekseepralard C, Busarakumtragul P. Effects of mindfulness
meditation on serum cortisol of medical students. J Med Assoc Thai. Jan 2013;96
Suppl 1:S90-95.

6. Marc J. Weigensberg CKFW, Fatimata Sanogo, Claudia M Toledo-Corral, Li
Ding, . Imagine HEALTH: Changes in diurnal salivary cortisol patterns following a
12-week guided imagery RCT lifestyle intervention in predominantly Latino
adolescents. Psychoneuroendocrinology. 2023.

7. National Institute of Diabetes and Digestive and Kidney Diseases. Insulin
Resistance & Prediabetes. 2018.

8. Castro AV, Kolka CM, Kim SP, Bergman RN. Obesity, insulin resistance and
comorbidities? Mechanisms of association. Arq Bras Endocrinol Metabol. Aug
2014;58(6):600-609.

9. Andrews RC, Walker BR. Glucocorticoids and insulin resistance: old hormones,
new targets. Clin Sci (Lond). May 1999;96(5):513-523.

10. Antuna-Puente B, Feve B, Fellahi S, Bastard JP. Adipokines: the missing link
between insulin resistance and obesity. Diabetes Metab. Feb 2008;34(1):2-11.
84
11. Lambert GW, Straznicky NE, Lambert EA, et al. Sympathetic nervous activation
in obesity and the metabolic syndrome--causes, consequences and therapeutic
implications. Pharmacol Ther. May 2010;126(2):159-172.

12. Item F, Konrad D. Visceral fat and metabolic inflammation: the portal theory
revisited. Obes Rev. Dec 2012;13 Suppl 2:30-39.

13. Goran MI, Lane C, Toledo-Corral C, Weigensberg MJ. Persistence of pre-
diabetes in overweight and obese Hispanic children: association with progressive
insulin resistance, poor beta-cell function, and increasing visceral fat. Diabetes.
Nov 2008;57(11):3007-3012.

14. Spruijt-Metz D, O'Reilly GA, Cook L, et al. Behavioral contributions to the
pathogenesis of type 2 diabetes. Curr Diab Rep. Apr 2014;14(4):475.

15. Räikkönen K, Keltikangas-Järvinen L, Adlercreutz H, Hautanen A. Psychosocial
stress and the insulin resistance syndrome. Metabolism. Dec 1996;45(12):1533-
1538.

16. Novak M, Björck L, Giang KW, et al. Perceived stress and incidence of Type 2
diabetes: a 35-year follow-up study of middle-aged Swedish men. Diabet Med.
Jan 2013;30(1):e8-16.

17. Williams ED, Magliano DJ, Tapp RJ, et al. Psychosocial stress predicts abnormal
glucose metabolism: the Australian Diabetes, Obesity and Lifestyle (AusDiab)
study. Ann Behav Med. Aug 2013;46(1):62-72.

18. McCurley JL, Mills PJ, Roesch SC, et al. Chronic stress, inflammation, and
glucose regulation in U.S. Hispanics from the HCHS/SOL Sociocultural Ancillary
Study. Psychophysiology. Aug 2015;52(8):1071-1079.

19. Cutrona CE, Abraham WT, Russell DW, et al. Financial strain, inflammatory
factors, and haemoglobin A1c levels in African American women. Br J Health
Psychol. Sep 2015;20(3):662-679.

20. Elizabeth O. Hedgeman. Perceived Stress, Insulin Resistance, and the Effects of
Physical Activity. 2017.

21. Philip P, Facchini FS. Self-perceived stress, hostility, and insulin resistance: lack
of association. Diabetes Care. Aug 1995;18(8):1198-1199.

22. Janczura M, Dropinski J, Gielicz A, et al. Potential roles of psychological and
oxidative stress in insulin resistance: a cohort-based study. Diabetol Metab
Syndr. 2020;12:58.
85
23. Fernandez-Real JM, Pugeat M, Grasa M, et al. Serum corticosteroid-binding
globulin concentration and insulin resistance syndrome: a population study. J
Clin Endocrinol Metab. Oct 2002;87(10):4686-4690.

24. Rafacho A, Ortsäter H, Nadal A, Quesada I. Glucocorticoid treatment and
endocrine pancreas function: implications for glucose homeostasis, insulin
resistance and diabetes. J Endocrinol. Dec 2014;223(3):R49-62.

25. Adam TC, Hasson RE, Ventura EE, et al. Cortisol is negatively associated with
insulin sensitivity in overweight Latino youth. J Clin Endocrinol Metab. Oct
2010;95(10):4729-4735.

26. Weigensberg MJ, Toledo-Corral CM, Goran MI. Association between the
metabolic syndrome and serum cortisol in overweight Latino youth. J Clin
Endocrinol Metab. Apr 2008;93(4):1372-1378.

27. Weigensberg MJ, Spruijt-Metz D, Wen CKF, et al. Protocol for the Imagine
HEALTH Study: Guided imagery lifestyle intervention to improve obesity-related
behaviors and salivary cortisol patterns in predominantly Latino adolescents.
Contemp Clin Trials. Sep 2018;72:103-116.

28. Bioelectrical impedance analysis in body composition measurement: National
Institutes of Health Technology Assessment Conference Statement. Am J Clin
Nutr. Sep 1996;64(3 Suppl):524s-532s.

29. Cheng K. Fred Wen SS, Marc J. Weigensberg, Bas Weerman, Donna Spruijt-
Metz. Accuracy of a Photo-based Smartphone Application to Assess Salivary
Cortisol Sampling Time in Adolescents. SAGE Journals 2022.

30. Matthews DR, Hosker JP, Rudenski AS, et al. Homeostasis model assessment:
insulin resistance and beta-cell function from fasting plasma glucose and insulin
concentrations in man. Diabetologia. Jul 1985;28(7):412-419.

31. Stalder T, Kirschbaum C, Kudielka BM, et al. Assessment of the cortisol
awakening response: Expert consensus guidelines. Psychoneuroendocrinology.
Jan 2016;63:414-432.

32. Kobaissi HA, Weigensberg MJ, Ball GD, et al. Relation between acanthosis
nigricans and insulin sensitivity in overweight Hispanic children at risk for type 2
diabetes. Diabetes Care. Jun 2004;27(6):1412-1416.

33. Bansal A, Mittal A, Seth V. Osho Dynamic Meditation's Effect on Serum Cortisol
Level. J Clin Diagn Res. Nov 2016;10(11):Cc05-cc08.

86
34. Sudsuang R, Chentanez V, Veluvan K. Effect of Buddhist meditation on serum
cortisol and total protein levels, blood pressure, pulse rate, lung volume and
reaction time. Physiol Behav. Sep 1991;50(3):543-548.

35. Elizabeth O. Hedgeman. Perceived Stress, Insulin Resistance and the Effects of
Physical Activity. 2017.

36. Ajibewa TA, Toledo-Corral C, Miller AL, et al. Racial differences in psychological
stress and insulin sensitivity in non-Hispanic Black and White adolescents with
overweight/obesity. Physiol Behav. Mar 1 2022;245:113672.

37. Hanson RL, Pratley RE, Bogardus C, et al. Evaluation of simple indices of insulin
sensitivity and insulin secretion for use in epidemiologic studies. Am J Epidemiol.
Jan 15 2000;151(2):190-198.

38. Shaibi GQ, Cruz ML, Ball GD, et al. Effects of resistance training on insulin
sensitivity in overweight Latino adolescent males. Med Sci Sports Exerc. Jul
2006;38(7):1208-1215.

87
TABLE 5: Participant Demographics and Clinical Characteristics

N = 229

Grade 10th 84 (36.7%)
11th 145 (63.3%)

Sex Female 153 (66.8%)
Male 76 (33.2%)

Ethnicity Hispanic or Latino 217 (94.7%)
Not Hispanic or Latino 12 (5.3%)

Race White 205 (89.9%)
Other 24 (10.1%)

Parent SES ($*) < 10,000 32 (15.5%)
> 50,000 15 (7.3%)
10,000-20,000 68 (33.0%)
20,000-30,000 52 (25.2%)
30,000-40,000 21 (10.2%)
40,000-50,000 18 (8.7%)

Mean ± SD
Age (years)

15.75 ± 0.66

HOMA-IR (baseline) 4.6 ± 0.60
Fasting Blood Glucose (mg/dl) Baseline 87.30 ± 11.52
88
3 months 88.38 ± 11.88
Log Fasting Insulin Baseline 6.11± 0.58
3 months 6.14 ± 0.63

BMI (Body Mass Index)
Baseline
24.54 ± 5.96
3 months 24.32 ± 5.24
BIA (Bioelectrical Impedance Analysis) Baseline 28.46 ± 10.81
3 months 26.64 ±10.74
Salivary Cortisol (nmol/L**)

Baseline Awakening 9.2 ± 0.46
+30 minutes 14.9 ± 1.2
Evening 2.4 ± 0.46

3 Month Awakening 9.2 ± 1.0
+30 minutes 14.4 ± 1.2
Evening 2.2 ± 0.91

Perceived Stress Score
Baseline 24.5 ± 1.5
3 months 24.5 ± 1.4

Serum Cortisol (nmol/L)
Baseline 409.69 ± 158.08
3 months 397.12 ±148.89

Continuous variables reported as means (±SD) and categorical variables reported as count (%).

89
TABLE 6: Baseline Associations of log HOMA-IR with PSS and Cortisol Biomarkers

Beta SE P-value
PSS 0.0086 0.0055 0.11
CAR -0.0033 0.0070 0.64
DCS -0.0058 0.0075 0.44
sCOR -0.0022 0.0057 0.70
Outcome Variable: HOMA-IR. Predictor variables: PSS, CAR, DCS, sCOR. sCOR= serum cortisol
Analyses performed using multivariate linear regressions. Average CAR and DCS used as predictor
variables.
All models adjusted for age, sex, and BMI. Beta=log (HOMA-IR).

90
TABLE 7: Associations Between Fasting Insulin, PSS and biomarkers of stress
Baseline Association

Adjusted for age, sex, and BMI
12-Week Longitudinal Association

Adjusted for age, sex, BMI, and study
group
Beta SE p-value Beta SE p-value
PSS 0.0067 0.0054 0.22 -0.0017 0.0040 0.40
CAR -0.0042 0.0069 0.55 -0.000044 0.0049 0.99
DCS -0.0066 0.0074 0.37 -0.0023 0.0052 0.66
sCOR 0.00025 0.00023 0.29 0.00048 0.00017 0.004
Outcome Variable: Fasting insulin. Predictor variables: PSS, CAR, DCS, sCOR. sCOR= serum cortisol
Baseline analyses performed using multivariate linear regressions. Average CAR and DCS used as
predictor variables.
Longitudinal associations performed using generalized least square for linear random effect regression
All models adjusted for age, sex, and BMI. Beta=log (fasting insulin).

91
TABLE 8: Associations Between BMI, PSS and biomarkers of stress
Baseline Association

Adjusted for age and sex
Longitudinal Association

Adjusted for age, sex, and study
group
Beta SE p-value Beta SE p-value
PSS 0.073 0.061 0.23 0.031 0.025 0.22
CAR -0.11 0.069 0.10 -0.037 0.025 0.15
DCS 0.012 0.080 0.88 -0.0025 0.028 0.93
sCOR -0.037 0.061 0.54 0.0059 0.020 0.77
Outcome Variable: BMI. Predictor variables: PSS, CAR, DCS, sCOR. sCOR= serum cortisol
Baseline analyses performed using multivariate linear regressions. Average CAR and DCS used as
predictor variables.
Longitudinal associations performed using generalized least square for linear random effect regression
All models adjusted for age and sex.

92
TABLE 9: Associations Between BIA, PSS and Biomarkers of Stress
Baseline Association

Adjusted for age and sex
Longitudinal Association

Adjusted for age, sex, and study
group
Beta SE p-value Beta SE p-value
PSS 0.12 0.11 0.26 0.0033 0.041 0.94
CAR -0.035 0.12 0.77 0.058 0.041 0.16
DCS 0.082 0.14 0.55 0.072 0.045 0.11
sCOR -0.018 0.10 0.86 0.051 0.032 0.11
Outcome Variable: BIA. Predictor variables: PSS, CAR, DCS, sCOR. sCOR= serum cortisol
Baseline analyses performed using multivariate linear regressions. Average CAR and DCS used as
predictor variables.
Longitudinal associations performed using generalized least square for linear random effect regression
All models adjusted for age and sex.

93
SUPPLEMENTAL TABLE 2: Participant Demographics and Clinical Characteristics by
Treatment Group

Control
(N= 51)
N(%)
LS
(N= 60)
N(%)
LBGI
(N= 54)
N(%)
SRGI
(N= 64)
N(%)

p-value
a
Grade 10th 11 (21.6%) 31 (51.7%) 23 (42.6%) 19 (29.7%) 0.01
11th 40 (78.4%) 29 (48.3%) 31 (57.4%) 45 (70.3%)

Sex Female 38 (74.5%) 35 (58.3%) 40 (74.1%) 40 (62.5%) 0.16
Male 13 (25.5%) 25 (41.7%) 14 (25.9%) 24 (37.5%)

Ethnicity Hispanic or Latino 50 (98.0%) 56 (93.2%) 51 (94.3%) 60 (93.8%) 0.67
Not Hispanic or
Latino
1 (2.0%) 4 (6.8%) 3 (5.7%) 4 (6.2%)

Race White 50 (98.0%) 54 (90.0%) 45 (83.3%) 56 (88.9%) 0.04
Other 1 (2.0%) 6 (10.0%) 9 (16.7%) 8 (11.1%)

Parent SES*
< 10,000
6
(12.5%) 12 (21.1%) 6 (13.0%) 8 (14.5%)
0.77
> 50,000 5 (10.4%) 3 (5.3%) 3 (6.5%) 4 (7.3%)
10,000-20,000 16 (33.3%) 16 (28.1%) 19 (41.3%) 17 (30.9%)

20,000-30,000
10
(20.8%) 17 (29.8%) 11 (23.9%) 14 (25.5%)

30,000-40,000
7
(14.6%) 2 (3.5%) 4 (8.7%) 8 (14.5%)

40,000-50,000
4
(8.3%) 7 (12.3%) 3 (6.5%) 4 (7.3%)

Mean ± SD Mean ± SD Mean ± SD Mean ± SD p-value
b

Age (years)

15.82 ±
0.48
15.63 ±
0.66
15.65 ±
0.70
15.89 ±
0.72
0.08
94

BMI (Body Mass
Index)

24.97 ±
6.41
24.34 ±
6.02
23.02 ±
5.96
25.87 ±
5.43
0.09

BIA (Bio-electrical
Impedance
Analysis)

29.21 ±
11.36
26.58 ±
10.58
27.50 ±
9.98
30.63 ±
11.05

0.20

Salivary Cortisol
(nmol/L**)

Baseline Awakening 9.3 ± 6.3 9.3 ± 5.5 9.5 ± 5.6 8.8 ± 4.9 0.72
+30 minutes 15.8 ± 9.7 14.5 ± 6.1 14.8 ± 7.0 13.6 ± 6.5 0.08
Evening 2.7 ± 8.4 2.3 ± 5.0 2.8 ± 8.0 1.9 ± 2.4 0.55

3 Month Awakening 8.4 ± 4.6 9.4 ± 5.0 10.4 ± 6.5 9.4 ± 4.6 0.06
+30 minutes 14.0 ± 7.9 14.2 ± 7.2 16.0 ± 8.1 13.5 ± 6.2 0.11
Evening 2.4 ± 4.5 1.4 ± 1.5 2.4 ± 3.8 2.2 ± 4.1 0.14

Perceived Stress
Score

Baseline 24.3 ± 8.1 24.7 ± 5.8 23.5 ± 6.5 25.3 ± 6.9 0.59
3 month 25.2 ± 7.8 24.1 ± 7.9 23.2 ± 7.9 24.6 ± 5.6 0.71

Serum Cortisol
(nmol/L)

Baseline
407.32 ±
171.83
435.04 ±
170.75
397.04 ±
140.96
403.58 ±
155.05 0.82

3 month
382.69 ±
182.31
396.94 ±
138.20
409.93 ±
142.19
403.13 ±
136.87 0.96

Fasting Insulin 6.15 ± 0.61 6.11 ± 0.56 6.03 ± 0.50 6.16 ± 0.65 0.65

HOMA-IR

126.76 ±
102.32
115.59 ±
70.16
96.03 ±
49.20
122.89 ±
77.60 0.23

a
Categorical variable: chi-squared test where >20% of expected cell counts are larger than 5. Fisher's exact test otherwise
b
Continuous variables: ANOVA F test
95
* Annual Income
** Conversion factor nmol/L /27.6 = µg/dL cortisol
p < 0.05 statistically significant

96
CHAPTER FIVE: Public Health Implications, and Future Directions

1. Clinical and Public Health Implications

The results of our meta-analysis highlight the importance of ensuring equitable
access to integrative healthcare. Type 2 diabetes is a significant health issue that is
often associated with socio-demographic disparities
1
. Integrative medicine and health,
as defined by the Academic Consortium for Integrative Medicine and Health, is a
patient-centered approach that focuses on whole person health, utilizes evidence-based
practices, and incorporates appropriate therapeutic and lifestyle interventions to
promote optimal health and healing
2
. By combining the most effective mainstream and
complementary/lifestyle approaches, an integrative medicine and health approach can
improve overall health and wellbeing for individuals.
Increasing research suggests that integrative health therapies can be both safe
and effective in treating a variety of health conditions. The meta-analysis we conducted
as part of this dissertation work found that mind and body practices can significantly
improve glycemic control in patients with type 2 diabetes
3
. However, despite the
potential benefits, there are still significant disparities in access to integrative medicine
services among non-white, lower-income, and less educated populations.
Nationally representative data from the 2012 National Health and Interview
Survey show that 33.2% of US adults used at least one complementary or integrative
medicine therapy in the previous year
4
. However, there are significant disparities in
usage rates among different racial/ethnic and socio-economic groups. For example, use
among Hispanics (22%) and non-Hispanic blacks (19.3%) is lower compared to non-
97
Hispanic whites (37.9%). Similarly, poor individuals (household income <100% of the
US Census Bureau poverty threshold) have a lower usage rate (20.6%) compared to
non-poor adults (income >200% of the poverty threshold)
5
.
Research conducted by Robert Saper and his colleagues identified four primary
barriers to the use of integrative medicine among underserved populations: awareness,
availability, accessibility, and affordability. In particular, individuals with lower levels of
education were more likely to cite a lack of knowledge as a reason for not using
integrative therapies
6
. Additionally, studies have shown that facilities offering integrative
therapies, such as yoga studios, are often concentrated in neighborhoods with higher
median incomes and are absent in lower-income areas in the same city. For those who
are aware of the benefits of integrative therapies, factors such as lack of transportation
or a rigid work schedule can be obstacles. Although the availability of online resources
for modalities like yoga and meditation has increased since the COVID-19 pandemic,
language barriers can still impede access for non-native English speakers. Finally, the
high cost of integrative therapies remains a major barrier. Americans spent $33.8
billion
7
“out of pocket” on complementary medicine visits, classes, and products in 2007
alone. The annual amount spent per person out of pocket in 2015 varied from $568-
$895
8
. For low-income individuals with limited discretionary income, these expenses
can be unaffordable.
Integrative health equity offers a ripe opportunity to reduce health disparities by
targeting underserved populations where disparities exist and there is either preliminary
evidence or plausibility of effectiveness of therapies. For instance, using mind and body
approaches to manage type 2 diabetes shows promise in this regard. Since social
98
determinants of health are responsible for approximately 80% of health outcomes, with
clinical care accounting for the remaining 20%
9
, adopting a systems-based approach is
critical. This approach should incorporate not only clinical services but also public health
and policy reform. Increasing awareness across all socio-demographic groups can be
achieved through television, social media, school-based health education, and
community-based events. Making integrative services affordable will also require
continuous lobbying efforts to get insurance to reimburse for safe, effective integrative
therapies, group visit models, and lifestyle modification programs.
As integrative health becomes increasingly mainstream, it is important to address
two major concerns beyond the risk of it disproportionately benefiting those who are
already overserved. Firstly, there is a pressing need for research on integrative
medicine to include a more diverse range of minority groups in the United States.
Currently, studies on integrative health therapies have primarily focused on white
populations
5
. Secondly, it is crucial to ensure that the indigenous roots of these
practices are respectfully honored and not diluted in the process of mainstreaming.
Research indicates that chronic stress can have a lasting impact on one's life
course. Prenatal and childhood stress experiences have been found to increase the risk
of developing diseases in the future, as shown in previous studies
10, 11
. Lower
socioeconomic status can further exacerbate this risk, as financial strain, exposure to
neighborhood crime, and environmental pollution can all contribute to higher levels of
stress and strain on the body's stress response system (known as allostatic load), even
at a young age
12, 13
. Interestingly, insulin resistance has been linked to stress pathways
in overweight and minority youths in the United States
14, 15
. Our own findings suggest an
99
association between serum cortisol, a biomarker of stress, and insulin resistance in
predominantly Latino adolescents, who are a high-risk population for type 2 diabetes.
This highlights the potential effectiveness of lifestyle interventions that target stress
reduction, such as mind and body practices, in reducing the risk of type 2 diabetes in
this population.
Our results also showed no association between perceived stress and cortisol
biomarkers, insulin resistance, or fasting blood glucose in this population, which is
thought-provoking. This lack of association may suggest racial differences in the effects
of perceived stress on insulin resistance, as previous studies have shown positive
associations between perceived stress and insulin resistance in white
16-19
adults but no
association in Latino adults
20
. Alternatively, it may indicate the need for new tools to
better capture psychological stress in minority groups that is reflected in physiological
stress response systems.
In summary, this research highlights four key public health recommendations:

1. Increase awareness, availability, accessibility, and affordability of integrative
medicine therapies such as mind and body practices in underserved and low
socio-economic communities. This can have a positive impact on the health of
individuals living with type 2 diabetes or those at high risk for the disease.

2. Ensure that integrative medicine research conducted in the United States
includes equal participation of both minority groups and white communities.

3. Respect and safeguard the indigenous roots of these practices.

100
4. Develop effective tools to capture psychological stress in minority groups, if
needed.

2. Future Research

Our systematic review established that data for MBSR, meditation, guided
imagery, and qigong are very limited. However, given their potential to alleviate both
diabetes distress and physiological distress, and their potential to be more accessible
than yoga to some patients, further research into these other forms of mind and body
practices is warranted. The high effectiveness of mind and body practices for people
with type 2 diabetes raise the question whether such practices could also be beneficial
to those with other chronic conditions such as cancer or multiple chronic conditions self-
management. Future intervention studies could focus on the effectiveness of such
practices for chronic pain management, improved quality of life, and reduced risk of
disability among individuals with cancer or multiple chronic conditions.
While our current findings reveal no association between perceived stress and
cortisol biomarkers, future research should aim to identify which physiological stress
response system, if any, is linked to the perceived stress scale. Additionally, our
findings suggest that there may be racial differences in the associations between
perceived stress and insulin resistance, as we observed no such association in
predominantly Latino adolescents. However, larger prospective studies, and studies
using more indices of insulin resistance such as Intravenous glucose tolerance test
(IVGTT) are needed to clarify this hypothesis. It is important to note that the absence of
a relationship between perceived stress and these physiological markers does not
101
exclude the need for developing new tools to effectively measure psychological stress in
minority groups, or to better understand how psychological stress may predict
physiological stress response. Further studies can help to shed light on this important
issue.
According to our research findings, mind and body practices may improve insulin
resistance by reducing serum cortisol levels through a pathway independent of
perceived stress. However, the specific pathway through which mind and body practices
affect cortisol levels remains unclear. Further investigation through larger prospective
observational studies, or mediation analyses is necessary to shed light on this matter.
Our present work was conducted in a population of predominantly healthy,
hispanic Latino adolescents. This limits the generalizability of findings and future work
can focus on investigating the associations between PSS and cortisol biomarkers, as
well as perceived stress and insulin resistance in adult populations at higher risk for
type 2 diabetes (such as adults who are insulin resistant) or in a population with type 2
diabetes. In addition, future research can also expand measures of cortisol biomarkers
to include to daily exposure of salivary cortisol, a measure of salivary cortisol that was
not assessed in this work.
Finally, as research continues to support the safety and efficacy of integrative
health therapies for treating various health conditions, it is important to prioritize
effectiveness studies over efficacy trials. By doing so, we can focus on identifying
successful frameworks and methods for delivering these therapies to individuals who
can benefit from them, rather than simply establishing whether or not they are effective
in controlled settings.
102

3. Conclusion

This dissertation enhances our understanding of the effects of mind and body
practices in on glycemic control in patients with type 2 diabetes, and their potential
benefits for reducing type 2 diabetes risk in vulnerable populations. Evidence suggests
that incorporating mind and body practices with standard of care can be a valuable non-
pharmacological approach for individuals with type 2 diabetes who are looking to
improve their glycemic control. The effectiveness of these practices is comparable to
that of metformin monotherapy. This means that mind and body practices have the
potential to help reduce risks associated with type 2 diabetes and can be particularly
useful for high-risk populations.

This dissertation sheds light on the underlying physiological mechanism by which
mind and body interventions influence insulin resistance. Our investigation shows no
evidence of association between perceived stress and cortisol biomarkers or insulin
resistance. However, we did find a significant association between serum cortisol and
fasting insulin, as well as fasting blood glucose. This suggests that mind-body practices
may decrease serum cortisol levels and, consequently, insulin resistance through a
mechanism independent of perceived stress. Furthermore, our results suggest that
perceived stress may be related to other physiological stress responses not involving
activation of the hypothalamic pituitary-adrenal axis. Future research is necessary to
clarify these findings and gain a complete understanding of how mind-body practices
improve serum cortisol levels and insulin resistance.

103
4. References

1. Institute of Medicine Committee on U, Eliminating R, Ethnic Disparities in Health
C. In: Smedley BD, Stith AY, Nelson AR, eds. Unequal Treatment: Confronting
Racial and Ethnic Disparities in Health Care. Washington (DC): National
Academies Press (US) Copyright 2002 by the National Academy of Sciences. All
rights reserved.; 2003.

2. Academic Consortium for Integrative Medicine and Health. Integrative Medicine
and Health 2015.

3. Sanogo F, Xu K, Cortessis VK, et al. Mind- and Body-Based Interventions
Improve Glycemic Control in Patients with Type 2 Diabetes: A Systematic
Review and Meta-Analysis. J Integr Complement Med. Sep 7 2022.

4. Clarke TC, Black LI, Stussman BJ, et al. Trends in the use of complementary
health approaches among adults: United States, 2002-2012. Natl Health Stat
Report. Feb 10 2015(79):1-16.

5. Saper R. Integrative Medicine and Health Disparities. Glob Adv Health Med. Jan
2016;5(1):5-8.

6. Burke A, Nahin RL, Stussman BJ. Limited Health Knowledge as a Reason for
Non-Use of Four Common Complementary Health Practices. PLoS One.
2015;10(6):e0129336.

7. Nahin RL, Barnes PM, Stussman BJ, Bloom B. Costs of complementary and
alternative medicine (CAM) and frequency of visits to CAM practitioners: United
States, 2007. Natl Health Stat Report. Jul 30 2009(18):1-14.

8. Nahin RL, Stussman BJ, Herman PM. Out-Of-Pocket Expenditures on
Complementary Health Approaches Associated With Painful Health Conditions in
a Nationally Representative Adult Sample. J Pain. Nov 2015;16(11):1147-1162.

9. Weeks J. The End of Tinkering: International Academic Group Explores
Transformational Needs in Health Professional Education. Glob Adv Health Med.
Jul 2015;4(4):5-7.

10. Pearlin LI, Schieman S, Fazio EM, Meersman SC. Stress, health, and the life
course: some conceptual perspectives. J Health Soc Behav. Jun 2005;46(2):205-
219.

11. Cohen BE, Panguluri P, Na B, Whooley MA. Psychological risk factors and the
metabolic syndrome in patients with coronary heart disease: findings from the
Heart and Soul Study. Psychiatry Res. Jan 30 2010;175(1-2):133-137.
104
12. McEwen BS, Gianaros PJ. Central role of the brain in stress and adaptation: links
to socioeconomic status, health, and disease. Ann N Y Acad Sci. Feb
2010;1186:190-222.

13. Diez Roux AV, Mair C. Neighborhoods and health. Ann N Y Acad Sci. Feb
2010;1186:125-145.

14. Hasson RE, Adam TC, Pearson J, et al. Sociocultural and socioeconomic
influences on type 2 diabetes risk in overweight/obese African-American and
Latino-American children and adolescents. J Obes. 2013;2013:512914.

15. Adam TC, Hasson RE, Ventura EE, et al. Cortisol is negatively associated with
insulin sensitivity in overweight Latino youth. J Clin Endocrinol Metab. Oct
2010;95(10):4729-4735.

16. Räikkönen K, Keltikangas-Järvinen L, Adlercreutz H, Hautanen A. Psychosocial
stress and the insulin resistance syndrome. Metabolism. Dec 1996;45(12):1533-
1538.

17. Novak M, Björck L, Giang KW, et al. Perceived stress and incidence of Type 2
diabetes: a 35-year follow-up study of middle-aged Swedish men. Diabet Med.
Jan 2013;30(1):e8-16.

18. Williams ED, Magliano DJ, Tapp RJ, et al. Psychosocial stress predicts abnormal
glucose metabolism: the Australian Diabetes, Obesity and Lifestyle (AusDiab)
study. Ann Behav Med. Aug 2013;46(1):62-72.

19. Elizabeth O. Hedgeman. Perceived Stress, Insulin Resistance and the Effects of
Physical Activity. 2017.

20. McCurley JL, Mills PJ, Roesch SC, et al. Chronic stress, inflammation, and
glucose regulation in U.S. Hispanics from the HCHS/SOL Sociocultural Ancillary
Study. Psychophysiology. Aug 2015;52(8):1071-1079.
105

Abstract (if available)

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Mechanisms relating obesity to insulin resistance and the metabolic syndrome

PDF

No evidence for a direct effect of osteocalcin on pancreatic beta-cells in Mexican Americans

PDF

Omentectomy in the canine: a unique model to investigate the physiologic relationship between visceral fat and insulin action

PDF

The effects of mindfulness meditation on stress levels among student registered nurse anesthetists: a pilot study

PDF

Variation in CRY2 and MTNR1B have independent effects on insulin secretion in Mexican Americans

PDF

Targeted metabolic signatures and diet in associations with obesity and insulin resistance in young adults

PDF

The effects of mindfulness meditation on stress levels among student registered nurse anesthetists: a pilot study

PDF

The associations between ultra-processed food consumption and type 2 diabetes and obesity among young adults

PDF

The effects of mindfulness meditation on stress levels among student registered nurse anesthetists: a pilot study

PDF

Quantity versus quality: how adipose tissue accumulation and immune cell profile associate with risk for type 2 diabetes in minority children and adults

PDF

The influence of dietary fructose on genetic effects of GCK and GCKR in Mexican Americans

PDF

Effects of occupation on physiologic biomarkers: Doppler sonographic evaluation of intraneural blood flow

PDF

Renin-angiotensin system modulation for the prevention and treatment of metabolic dysfunction

Asset Metadata

Creator Sanogo, Fatimata (author)

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

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Epidemiology

Degree Conferral Date 2023-08

Publication Date 08/08/2023

Defense Date 05/24/2023

Publisher University of Southern California. Libraries (digital)

Tag cortisol,diabetes,fasting blood glucose,fasting insulin,guided imagery,hemoglobin A1C,HOMA-IR,imagine HEALTH,insulin resistance,integrative medicine,Meditation,meta-analysis,mind and body practices,mindfulness based stress reduction,OAI-PMH Harvest,Perceived Stress Scale,Qigong,stress,whole person health,Yoga

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Watanabe, Richard (committee chair), Cortessis, Victoria (committee member), Mack, Wendy (committee member), Mckean-Cowdin, Roberta (committee member), Weigensberg, Marc (committee member)

Creator Email fatimatasanogo@yahoo.fr,fsanogo@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC113296679

Unique identifier UC113296679

Identifier etd-SanogoFati-12206.pdf (filename)

Legacy Identifier etd-SanogoFati-12206

Document Type Dissertation

Rights Sanogo, Fatimata

Internet Media Type application/pdf

Type texts

Source 20230808-usctheses-batch-1080 (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. 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, Los Angeles, California 90089, USA

Repository Email cisadmin@lib.usc.edu