Content uploaded by Alicia M. DeLouize
Author content
All content in this area was uploaded by Alicia M. DeLouize on Apr 03, 2020
Content may be subject to copyright.
ORIGINAL RESEARCH ARTICLE
Inflammation and central adiposity as mediators
of depression and uncontrolled diabetes in the study
on global AGEing and adult health (SAGE)
Allison C. Dona
1
| Alicia M. DeLouize
1
| Geeta Eick
1
| Elizabeth Thiele
2
|
Aarón Salinas Rodríguez
3
| Betty Soledad Manrique Espinoza
3
|
Ricardo Robledo
4
| Salvador Villalpando
4
| Nirmala Naidoo
5
|
Somnath Chatterji
5
| Paul Kowal
1,5,6
| J. Josh Snodgrass
1
1
Department of Anthropology, University
of Oregon, Eugene, Oregon
2
Department of Biology, Vassar College,
Poughkeepsie, New York
3
Centre for Evaluation Research and
Surveys, National Institute of Public
Health, Mexico
4
Nutrition and Health Investigation
Center, National Institute of Public
Health Laboratory, Cuernavaca, Morelos,
Mexico
5
Department of Health Statistics and
Information Systems, World Health
Organization SAGE, Geneva, Switzerland
6
Research Centre for Generational Health
and Ageing, University of Newcastle,
Newcastle, Australia
Correspondence
J. Josh Snodgrass, Department of
Anthropology, 1218 University of Oregon,
Eugene, OR 97403.
Email: jjosh@uoregon.edu
Funding information
Ministry of Health in Mexico; NIH NIA
Interagency Agreement YA1323-08-CN-
0020
Abstract
Objectives: Diabetes and depression are commonly present in the same
individuals, suggesting the possibility of underlying shared physiological
processes. Inflammation, as assessed with the biomarker C-reactive protein
(CRP), has not consistently explained the observed relationship between diabe-
tes and depression, although both are associated with inflammation and share
proposed inflammatory mechanisms. Central adiposity has also been associ-
ated with both conditions, potentially by causing increased inflammation. This
study uses the World Health Organization's Study on global AGEing and adult
health (SAGE) Mexico Wave 1 biomarker data (n= 1831) to evaluate if inflam-
mation and central adiposity mediate the relationship between depression and
diabetes.
Methods: Depression was estimated using a behavior-based diagnostic algo-
rithm, inflammation using venous dried blood spot (DBS) CRP, central adipos-
ity using waist-to-height ratio (WHtR), and uncontrolled diabetes using
venous DBS-glycated hemoglobin (HbA1c).
Results: The association between depression and uncontrolled diabetes was
partially mediated by CRP before but not after WHtR was considered. When
WHtR was added to the model, it partially mediated the relationship between
diabetes and depression while fully mediating the relationship between depres-
sion and CRP.
Conclusions: These findings suggest that central adiposity may be a more sig-
nificant mediator between diabetes and depression than inflammation and
account for the relationship between these disorders and inflammation.
Depression may cause an increase in central adiposity, which then may lead to
diabetes, but the increase in known systemic inflammatory pathways caused
by central adiposity may not be the key pathological mechanism.
Received: 10 April 2019 Revised: 22 February 2020 Accepted: 9 March 2020
DOI: 10.1002/ajhb.23413
Am J Hum Biol. 2020;1–12. wileyonlinelibrary.com/journal/ajhb © 2020 Wiley Periodicals, Inc. 1
1|INTRODUCTION
Mental health conditions are major contributors to long-
term morbidity and disability globally. Hundreds of mil-
lions of people suffer from mental health conditions such
as depression, yet the resources allocated to respond to
morbidity and associated impairment are disproportion-
ately low, particularly in low- and middle-income coun-
tries (Eaton et al., 2011; Saraceno et al., 2007).
Conventional public health metrics also fail to account
for the risk of acquiring a second health disorder.
Multiple studies have found that mental health conditions—
including depression, substance abuse, suicide, and post-
traumatic stress disorder (PTSD)—are related to physical
illness (Farmer, Kim, Kleinman, & Basilico, 2013) includ-
ing cardiovascular diseases, respiratory diseases, chronic
pain conditions, gastrointestinal illnesses, and cancer
(Jones et al., 2004; Sareen et al., 2007; Van der Kooy
et al., 2007).
Type 2 diabetes and depression—which impact over
400 million (WHO, 2018) and 250 million (WHO, 2019)
people worldwide, respectively—are commonly present
in the same people (Roy & Lloyd, 2012) and may share
underlying physiological mechanisms and social determi-
nants. This relationship is thought to be bidirectional
because of similarities in the underlying pathological mech-
anisms in these conditions, including inflammation, as well
as health behaviors, psychosocial factors, and brain-derived
neurotrophic factor. People with depression have also been
shown to have higher rates of diabetes and vice versa
(Berge & Riise, 2015; Stuart & Baune, 2012). The present
study was designed to investigate potential key underlying
mechanisms in the comorbidity between diabetes and
depression as well as to identify relationships related to
aging and sociodemographic parameters. We specifically
consider if depression is associated with greater inflamma-
tion and/or central adiposity and then if the inflammation
and/or central adiposity are associated with uncontrolled
diabetes.
The number of individuals with comorbid diabetes and
depression varies greatly between studies. This is in part
because of methodological differences and limitations,
including self-reported prevalence data vs biomarker or
clinical assessment, samples that do not distinguish
between type 1 and type 2 diabetes, lack of documentation
regarding relevant factors associated with the disease state,
and population differences such as race/ethnicity (Katon,
2008). In a 39-study meta-analysis, 11% of patients with
diabetes met the criteria for comorbid major depressive
disorder and 31% experienced significant depressive symp-
toms. Furthermore, depression prevalence in patients with
diabetes was significantly higher in women than men, and
the odds of having depression were twice as great in
patients with diabetes as in their nondiabetic counterparts
(Anderson, Freedland, Clouse, & Lustman, 2001).
Although the relationship between depression and diabe-
tes is bidirectional, the risk of developing diabetes after a
depression diagnosis is higher than the risk of developing
depression after a diabetes diagnosis (Mezuk, Eaton,
Albrecht, & Golden, 2008). This pattern highlights the
need to identify underlying physiological mechanisms
linking the two disorders, as the relationship is not entirely
due to the stressors of illness causing depression. In addi-
tion, individuals with type 2 diabetes and comorbid
depression have increased incidence of overall poor health
outcomes (Anderson et al., 2001; Black, Markides, & Ray,
2003) as well as worse diet and medication regimen adher-
ence, functional impairment, and higher healthcare costs
(Ciechanowski, Katon, & Russo, 2000; Katon, 2008).
Understanding how depression could lead to diabetes
physiologically could help prevent this disorder from
developing in some people, improving patient outcomes.
Epidemiological and animal studies suggest that inflam-
mation is an important mediator of the comorbidity between
diabetes and depression. Levels of both pro-inflammatory
and anti-inflammatory cytokines in the peripheral circulation
and central nervous system have been reported to rise during
depression and other brain disorders (Schwarz, Chiang,
Müller, & Ackenheil, 2001). Chronic subclinical elevation of
interleukin 6 (IL-6), C-reactive protein (CRP), orosomucoid,
and sialic acid (all of which play a role in the inflammatory
process) might be causes of diabetes in middle-aged adults
(Duncan et al., 2003).
In particular, CRP is one component of the inflamma-
tory process that may be most directly related to this
comorbidity. The liver secretes CRP into the blood after
macrophages and T cells secrete IL-6 as a typical
response to cell damage, pathogens, burns, heart attack,
cancer, and other damage. The main biological function
of CRP appears to be host defense against bacterial path-
ogens and clearance of apoptotic and necrotic cells
(Volanakis, 2001). Chronically elevated CRP levels are
associated with many physical health conditions, includ-
ing depression and diabetes, as well as cardiovascular dis-
ease (Dehghan et al., 2007; Howren, Lamkin, & Suls,
2009; Pearson et al., 2003; Pradhan, Manson, Rifai,
Buring, & Ridker, 2001), although its role as a causal fac-
tor remains debated (Li & Fang, 2004; Yousuf et al., 2013;
Zimmermann et al., 2014). Depression has also been
associated with CRP in patients with type 1 and type
2 diabetes (Herder et al., 2018). However, adjustment for
markers of inflammation (including CRP) has not consis-
tently attenuated the positive association between diabe-
tes and depression, although diabetes and depression are
each associated with inflammation and share proposed
inflammatory mechanisms (Stuart & Baune, 2012).
2DONA ET AL.
Central adiposity may influence this comorbidity and
account for CRP involvement. Diabetes and depression
are both associated with obesity (Luppino et al., 2010;
Mokdad et al., 2003; Pan et al., 2012; Yudkin, Stehouwer,
Emeis, & Coppack, 1999) and specifically central adipos-
ity (Bray et al., 2008; Thakore, Richards, Reznek,
Martin, & Dinan, 1997; Vogelzangs et al., 2008), although
depression has been associated with being underweight
as well (Carey et al., 2014). Depressive symptoms have
been shown to predict greater abdominal obesity inde-
pendent of overall obesity, indicating that specific patho-
logical mechanisms of depression may lead to visceral
fat accumulation (Vogelzangs et al., 2008). Multiple
measures of central adiposity have been shown to predict
diabetes as well (Bray et al., 2008). CRP and systemic
inflammation may be a part of central adiposity pathol-
ogy; adipose tissue secretes IL-6 into circulation, which
stimulates CRP production by the liver (Trayhurn &
Wood, 2004; Voleti & Agrawal, 2005; Wisse, 2004;
Yudkin et al., 1999). However, a recent study found that
change in IL-6 was not associated with change in depres-
sive symptoms in individuals with type 1 or type 2 diabe-
tes, even when CRP was (Herder et al., 2018). Another
study showed that when controlling for central adipos-
ity, only IL-6, and not CRP, predicted the development
of type 2 diabetes (Duncan et al., 2003). These incon-
sistencies could be due to multicollinearity, which can
be detected and further understood with mediation
analyses.
Metabolic syndrome—which generally involves
increased blood pressure, high blood sugar, central adi-
posity, and abnormal cholesterol or triglyceride levels
(Huang, 2009)—combines the variables investigated in
the present article and more; therefore, related research
may provide insight into the comorbidity of diabetes and
depression. Increased oxidative stress, defined as a distur-
bance in the balance between the production of reactive
oxygen species (free radicals) and antioxidant defenses
(Betteridge, 2000), in accumulated fat has been shown to
be an early instigator of metabolic syndrome (Furukawa
et al., 2017). In addition, CRP is significantly elevated in
metabolic syndrome and is a predictor of multiple meta-
bolic disorders including type 2 diabetes (Ridker, 2003;
Sattar et al., 2003). A prospective study found that indi-
viduals with both depressive symptoms and metabolic
dysregulation at baseline were at higher risk for develop-
ing diabetes than participants with only one of the condi-
tions, and this increase was more than the sum of the
individual effects (Schmitz et al., 2016). Overall, central
adiposity may be a pathological mechanism causing both
diabetes and depression, and each condition may also
increase central adiposity and through this pathway bidi-
rectionally put individuals at greater risk for the other.
The present study investigates uncontrolled diabetes
and depression comorbidity in older Mexican adults
using biomarker data (CRP and HbA1c [glycated hemo-
globin]) from the World Health Organization's Study on
global AGEing and adult health (SAGE) (Kowal et al.,
2012). Specifically, we evaluate CRP and waist-to-height
ratio (WHtR) as mediators of this comorbidity to better
understand the pathology of both diseases. This study
focuses on adults over 50 in Mexico because most major
studies have been conducted in Europe and the United
States (Herder et al., 2018; Schmitz et al., 2016; Stuart &
Baune, 2012), diabetes and depression are highly preva-
lent in the Mexican population (IHME, 2019; WHO,
2016), and socioeconomic differences and similarities to
previous studies may contribute to meaningful cross-
study comparison. In Mexico, type 2 diabetes is the lead-
ing cause of death and disability, with more than 13 of
the 127 million people (10.4%) living with diabetes
(WHO, 2016). In addition, 12.5% of the population, over
15 million people, experienced major depressive disor-
der in 2002. Older adults are burdened disproportion-
ately; depression prevalence in individuals over 80 years
old has been estimated to range from 21.7% to 25.3%
(Alvarez-Monjaras & Gonzalez, 2016).
The aims of this article are twofold. First, we will
evaluate if CRP mediates the relationship between
uncontrolled diabetes and depression. Second, we will
evaluate if central adiposity mediates the comorbidity
between uncontrolled diabetes and depression and if it
impacts CRP as a mediator. We hypothesize that the first
model will show that depression is associated with higher
levels of CRP, which will then be associated with the
presence of uncontrolled diabetes. In the second model,
we predict that depression will be associated with greater
central adiposity, which then will be associated with the
presence of uncontrolled diabetes. We also expect that
the positive association between central adiposity and
CRP will mediate the co-occurrence of depression and
uncontrolled diabetes.
2|METHODS
2.1 |Study on global AGEing and adult
health (SAGE)
SAGE is a comprehensive longitudinal study of the
health and well-being of older adult populations and
the aging process in six middle income nations. It was
approved by the World Health Organization's Ethical
Review Committee and review bodies within each coun-
try (Kowal et al., 2012). The present study uses biomarker
data from SAGE Mexico Wave 1 to examine if CRP and
DONA ET AL.3
WHtR mediate the relationship between depression and
diabetes.
2.2 |Participants
A subset of people participating in the SAGE study in
Mexico underwent biomarker analysis (n= 1831). Exclu-
sions included 13 people for incomplete data and 173 peo-
ple for having an elevated CRP value that might indicate
injury or infection. The sample was 60% women. Partici-
pant ages for this analysis ranged from 50 to 94 years old
(M= 67.32, SD = 8.91) and education ranged from 0 to
25 years (M= 4.38, SD = 4.14). Mean education is partic-
ularly low because 20% of the sample had no formal edu-
cation. Participants were mostly married (58%), followed
by widowed (23%), never married (9%), divorced (6%),
and cohabiting (4%).
2.3 |Sociodemographic measures
Trained interviewers performed face-to-face computer-
assisted personal interviews and undertook anthropomet-
ric measurement. The interviews took place from 2009 to
2010 in the participants' homes and lasted approximately
1.5 hours on average (Kowal et al., 2012). The self-report
depression diagnosis, answers to depression symptoms
questions, and participant demographic information were
used in the present study.
Gender,age,maritalstatus,education,perceived
health, and wealth were determined using standard
measures (Kowal et al., 2012). The wealth variable was a
standardized composite variable including household
assets and housing amenities used to assess wealth inde-
pendently of employment status. It included a 21-item
self-report measure that asked about items owned by
the household and household characteristics such as
typeoffloors,walls,watersupply,andsanitation.
2.4 |Biomarkers
Venous blood was collected in an EDTA tube using stan-
dard venipuncture. These whole blood samples were then
homogenized before being pipetted in 20 μL aliquots onto
standard Whatman 903 filter paper so that they could be
analyzed using DBS procedures (Brindle, O'Connor, &
Garrett, 2014; de Waal, Driver, & Warner, 2019; Lacher,
Berman, Chen, & Porter, 2013; Lehmann, Delaby,
Vialaret, Ducos, & Hirtz, 2013; Li & Lee, 2014; McDade
et al., 2012). The samples were analyzed after 24 hours of
drying at room temperature. A 6 mm spot was punched
out from the DBS card and eluted for 14 hours with
250 μL of PBS buffer pH = 7. DBS eluates were analyzed
using the Abbott Architect CI8200 chemistry analyzer for
CRP (Lacher et al., 2013). In addition, CRP values were
obtained from an additional serum separator tube col-
lected from 91 participants for comparison and analyzed
using the Architect CI8200. To create serum equivalents,
CRP DBS values were regressed onto CRP serum controls
for the comparison subsample using a Passing and
Bablok regression. The regression equation was:
CRP serum = 2:41 CRP DBSðÞ−1:38
The CRP DBS values and the CRP serum values were
highly correlated, indicating sufficient validity for analy-
sis, Pearson's r= 0.99. Participants were excluded for pos-
sible infection/injury if they had a serum conversion
value greater than 5 mg/L. Analyses run with a cutoff of
10 mg/L produced similar results, indicating the model is
robust to the choice of CRP cutoff value. All analyses
were run using raw DBS CRP values; however, serum
conversion values are reported in the manuscript (unless
otherwise noted) for ease of interpretation.
HbA1c was run via blood chemistry analysis using the
Architect CI8200 (de Waal et al., 2019; Lacher et al., 2013).
A 6 mm DBS punch was eluted 14 hours in 400 μLMUL-
TIAGEN Hemoglobin Denaturant. A cuvette with the elu-
ent was loaded into the Architect blood chemistry analyzer,
where HbA1c and total hemoglobin were each determined
by measuring absorbance at 700 nm and 604 nm, respec-
tively. Percent HbA1c was calculated by the analyzer's pro-
gram as [(HbA1c/TotHb) ×100] −3+(0.2 ×TotHb).
Elevated levels of HbA1c can indicate type 1 and type 2 dia-
betes; therefore, we were not able to distinguish between
them in this study. A 6.5% cutoff value was used to create
the diabetes variable for analyses. Participants who were
diagnosed with diabetes but had an HbA1c below 6.5%
were not included in the diabetes variable because treat-
ments that lower HbA1c also lower inflammation.
2.5 |Depression
Two methods were used to identify the presence of
depression. First, participants were considered to have
depression if they answered yes to, “Have you ever been
diagnosed with depression?”Second, because depression
is often underreported and underdiagnosed, symptoms
for depression were also assessed using the World Mental
Health Survey version of the Composite International
Diagnostic Interview (Kessler & Üstün, 2004). This mea-
sure was then transformed to a symptom-based diagnos-
tic variable using a previously validated algorithm to
4DONA ET AL.
include participants who had a depressive episode in the
past 12 months but no diagnosis. Participants who
responded “yes”to the first question or had a positive result
for depression as indicated by the diagnostic interview were
categorized into the depression group. This combined mea-
sure has also been used in previous studies, including others
from SAGE (eg, Hsieh, 2015; Kamenov et al., 2016).
2.6 |Anthropometrics
Height and waist circumference were recorded using
standard measures (Kowal et al., 2012). Waist circumfer-
ence was measured at the top of the iliac crest. More
detailed information about measurements can be found
online in the SAGE survey manual (https://www.who.
int/healthinfo/survey/SAGESurveyManualFinal.pdf).
WHtR has been used in previous studies and has been
shown to be an effective indicator of diabetes risk and
other health conditions (Ashwell, Gunn, & Gibson, 2012;
Browning, Hsieh, & Ashwell, 2010; Łopaty
nski,
Mardarowicz, & Szcze
sniak, 2003; Sayeed et al., 2003).
2.7 |Statistical analyses
Before analysis, depression and wealth composites were com-
puted. To evaluate the diagnostic validity of the symptom-
based depression variable, slightly different variations of an
algorithm computing the DSM-IV diagnoses for a depressive
episode were created. They were then compared to the clini-
cal diagnoses of a separate sample of participants. The algo-
rithm that was most consistent with the clinical diagnosis of
the sample population was used. For the wealth variable,
each question was treated as an independent observation of
wealth. The data were then reshaped, a pure random effects
model was fit, and data were transformed using Bayes post-
estimation to create a standardized composite wealth variable
(Schrock et al., 2017).
A hierarchical linear regression was performed in SPSS
v. 25, and mediation analyses were performed using struc-
tural equation modeling (SEM) in the lavaan package
(Rosseel, 2012) with diagonally weighted least squares esti-
mation in R (R Core Team, 2019). The alpha level was set at
0.05. Statistical assumptions were checked prior to analysis.
Variable skewness ranged from −0.47 to 1.86, and kurtosis
ranged from −1.78 to 3.61, which indicates sufficient nor-
mality (Warner, 2012). However, review of the histograms
showed that CRP was exponentially distributed; there-
fore, it was ln-transformed. In addition, all variables
were approximately linear in relation to one another
and appeared homoscedastic. There were no multivari-
ate outliers. Therefore, all assumptions were met.
A hierarchical regression was performed to, first, eval-
uate if the health variables of interest, depression and
diabetes were related to CRP in the presence of demo-
graphic controls (model 3). Second, the hierarchical
regression detected potential mediation relationships by
evaluating the extent of suppression effects between
depression and diabetes (models 1 and 2) and between
health variables, depression, diabetes, and CRP (models
3 and 4). Four models were compared to evaluate the
suppression effects between depression and diabetes and
on depression and diabetes by WHtR. Depression was
included as a predictor of CRP in model 1, diabetes was
added as a predictor in model 2, demographics were
added in model 3, and self-reported health and WHtR
were added in model 4. CRP was the outcome in all four
models.
The first mediation analysis explores the potential
mediation of depression and diabetes by CRP. The SEM
for the i
th
person, where xis the predictor, zis the media-
tor, and yis the outcome is as follows:
Diabetesi=β0y+βzylnCRPi+βxyDepi+εyi
lnCRPi=β0z+βxz Depi+εzi
The second mediation analysis explores whether the
links between CRP, depression, and diabetes are further
mediated by WHtR. The SEM for the i
th
person, where
xis a predictor, vis the mediation by WHtR, zis the
mediation by CRP, and yis the outcome is as follows:
Diabetesi=β0y+βzylnCRPi+βxyDepi+βvy WHtRi+εyi
lnCRPi=β0z+βxz Depi+βvzWHtRi+εzi
WHtRi=β0v+βxvDepi+εvi
CRP Bvalues were converted to serum equivalents
with the following equation:
serum CRP B =2:41 eDBS lnCRP B
−1:38:
3|RESULTS
Uncontrolled diabetes and depression both pose a signif-
icant health burden to people living in Mexico, with an
estimated 26% and 17% of individuals from this study
suffering from these conditions, respectively (Table 1).
DONA ET AL.5
The median HbA1c value was 5.8% (the diabetic range is
6.5% or higher; Table 2). The median CRP value in the
sample was 1.79 mg/L; there is currently no cutoff stan-
dard, but generally low risk for cardiovascular disease is
considered to be a CRP concentration of <1.0 mg/L, aver-
age risk a CRP concentration between 1.0 and 3.0 mg/L,
and high risk a CRP concentration > 3.0 mg/L (Nehring &
Patel, 2018; Pearson et al., 2003). WHtR in the population
averaged 0.63 (0.56 and above is considered high; Rodea-
Montero et al., 2014).
Using hierarchical regression, all models accounted
for a significant portion of the variance in CRP
(F's = 4.48-9.61, P's < 0.05). Model 1 showed that the
presence of depression increased CRP by an average of
1.44 mg/L (t= 2.22, P= 0.03). This value remained sig-
nificant in model 2, when controlling for diabetes
(t= 2.02, P= 0.04). The presence of diabetes increased
CRP by 1.49 mg/L (t= 2.78, P= 0.005). Diabetes
maintained its significance as a predictor in model 3, even
after controlling for demographic variables (t= 2.42,
P= 0.02). Diabetes did, however, become nonsignificant
in model 4 when WHtR was controlled for (t= 1.26,
P= 0.21). This indicates that WHtR may be more closely
related to inflammation than depression or diabetes risk.
For every 0.10 increase in WHtR, CRP increased by
3.38 mg/L (t= 8.08, P< 0.001; Table 3).
In the first mediation model, the results for direct
effects showed that the presence of depression was associ-
ated with higher CRP levels (z= 2.46, P= 0.01), higher
CRP levels were associated with the presence of diabetes
(z= 3.32, P< 0.001), and the presence of depression was
associated with the presence of diabetes (z= 2.53,
P= 0.01). The indirect effect indicated that there was par-
tial mediation of the relationship between depression and
diabetes by CRP (z= 1.98, P< 0.05; Figure 1). Because
the model was saturated, no fit statistics are reported.
WHtR was then added to the model. Adjusted direct
effects showed that the presence of depression was no
longer associated with higher CRP levels (z= 1.59,
P= 0.08). New paths showed that the presence of depres-
sion was associated with a higher WHtR (z= 4.00,
P< 0.001), and a higher WHtR was associated with
higher CRP values (z= 10.03, P< 0.001) and the pres-
ence of diabetes (z= 5.57, P< 0.001). In the mediation of
depression and inflammation by waist-to-height ratio,
the presence of WHtR accounts for an average of
1.15 mg/L of the higher CRP values seen with the pres-
ence of depression (indirect effect β= 0.02, z= 3.71,
P< 0.001). In addition, the presence of depression and
diabetes was partially mediated by WHtR (indirect effect
β= 0.02, z= 3.25, P= 0.001). However, the positive rela-
tionship between depression and diabetes was not medi-
ated by the path through WHtR and CRP (indirect effect
β= 0.002, z= 1.71, P= 0.09), and the positive relation-
ship between depression and diabetes was not mediated
by CRP (indirect effect β= 0.003, z= 1.22, P= 0.22).
Because the model was saturated, no fit statistics are
reported (Figure 2).
4|DISCUSSION
Our results suggest that central adiposity may be a more
significant mediator between diabetes and depression
TABLE 1 HbA1c and CRP statistics for the sample population
Measure Median Min Max Lower quartile Upper quartile
HbA1c 5.80% 3.97% 14.97% 5.42% 6.57%
CRP (DBS) 1.21 mg/L 0.09 mg/L 4.72 mg/L 0.76 mg/L 2.06 mg/L
CRP (serum equivalent) 1.53 mg/L 0.00 mg/L 10.00 mg/L 0.45 mg/L 3.57 mg/L
TABLE 2 Population percentage in risk categories for HbA1c,
CRP (serum equivalent), and WHtR before exclusion for infection/
injury
Measure/risk category Cut-off values
Population
percentage (%)
HbA1c
Diabetes ≥6.50% 24
Impaired glucose
tolerance
5.70-6.49% 28
Normal <5.70 48
CRP (serum equivalent)
High ≥10.00 mg/L 9
Medium/high 3.00-9.99 mg/L 27
Medium/low 1.00-2.99 mg/L 29
Low <1.00 mg/L 36
Waist to height ratio
High ≥0.56 81
Low <0.56 19
Note: Waist-to-height ratio (WHtR) reference ranges (Rodea-Montero, Evia-
Viscarra, & Apolinar-Jiménez, 2014; WHO, 2011b), C-reactive protein (CRP)
reference ranges (Nehring & Patel, 2018; Pearson et al., 2003), and HbA1c
reference ranges (WHO, 2011a).
6DONA ET AL.
than inflammation and account for the relationship
between these disorders and inflammation. The first
mediation model is consistent with the hypothesized rela-
tionship that depression is associated with higher levels
of CRP, which is then associated with uncontrolled dia-
betes. When WHtR was not considered, the relationship
between diabetes and depression was partially mediated
by CRP, although the association was modest. However,
when WHtR was included in the second mediation
model, CRP was no longer a significant mediator of the
comorbidity between depression and diabetes. Instead,
central adiposity mediated the relationship between
depression and uncontrolled diabetes.
The results from the second model were partially con-
sistent with the second hypothesis that central adiposity
would mediate depression and diabetes as well as that
TABLE 3 Regression table of serum C-reactive protein (CRP) B and standardized βvalues predicting lnCRP
Model 1 Model 2 Model 3 Model 4
Variables of interest
Depression 1.45 (0.06)* 1.41 (0.05)* 1.18 (0.02) 1.08 (0.01)
Diabetes 1.49 (0.07)** 1.42 (0.06)** 1.23 (0.03)
Demographic
Age −1.02 (−0.04) −1.02 (−0.04)
Female 1.84 (0.13)*** 1.55 (0.09)**
Education −1.01 (−0.03) −1.02 (−0.01)
Wealth 1.06 (0.01) −0.99 (−0.01)
Marital status (married vs)
Never married −0.60 (−0.05) −0.77 (−0.03)
Cohabiting 1.49 (0.03) 1.46 (0.03)
Separated/divorced −0.89 (−0.01) −0.90 (−0.01)
Widowed −0.80 (−0.04) −0.79 (−0.04)
General health
Waist/height 33.84 (0.20)***
Current health 1.17 (0.04)
R
2
0.00 0.01 0.03 0.07
ΔR
2
0.00 0.01 0.02 0.04
F4.91* 6.34** 4.49*** 9.61***
*P< 0.05.; **P< 0.01.; ***P< 0.001. All Bvalues were transformed to represent changes in serum CRP.
Depression Diabetes
lnCRP
.08*
.11***
.06*
E1
E2
FIGURE 1 Mediation model, including βvalues, showing
diabetes and depression were mediated by lnCRP. (Indirect effect
β= 0.01 P< 0.05.) *P< 0.05, **P< 0.01, and ***P< 0.001
.04
.19***
WHtR
Depression
lnCRP
Diabetes
.09***
.25***
.07
.07*
E1
E3
E2
FIGURE 2 Waist-to-height ratio (WHtR) was added to the
mediation model with βvalues for direct effects noted. This model
shows that WHtR mediated the positive relationship between
depression and lnCRP (Indirect effect β= 0.02, P< 0.001), WHtR
partially mediated the positive relationship between depression and
diabetes (Indirect effect β= 0.02, P= 0.001), the path through
WHtR and lnCRP did not mediate the positive relationship
between depression and diabetes (Indirect effect β= 0.002,
P= 0.09), and lnCRP did not mediate the positive relationship
between depression and diabetes (Indirect effect β= 0.003,
P= 0.22) *P< 0.05 **P< 0.01 ***P< 0.001
DONA ET AL.7
the positive association between central adiposity and
CRP would mediate depression and diabetes. As
predicted, the presence of depression was associated with
a greater WHtR, and a greater WHtR was associated with
the presence of diabetes. However, CRP no longer medi-
ated the relationship between depression and diabetes
and the positive association between WHtR and CRP did
not mediate the relationship between depression and
diabetes. Therefore, the relationship between depression
and uncontrolled diabetes appears to have a central adi-
posity component but not a systemic inflammatory com-
ponent as indicated by CRP. In addition, the relationship
between depression and CRP was fully mediated by
WHtR. This suggests that central adiposity may be a
more significant mediator of the relationship between
uncontrolled diabetes and depression as well as account
for the observed mediation of depression and diabetes
by CRP.
These results suggest that central adiposity is one of
the causal factors in the relationship between depression
and diabetes: the presence of depression may cause an
increase in visceral adipose tissue, which then may cause
an increase in diabetes risk. It is also possible that this
relationship could be bidirectional. Although inflamma-
tion did appear to be an intermediate pathway step
between these two diseases when assessed alone, central
adiposity fully accounted for the previously observed rela-
tionship between depression and CRP, supporting that
this inflammation results from central adiposity and is
not involved in the association between depression and
diabetes. However, CRP no longer mediating central adi-
posity and diabetes suggests that other pathways caused
by central adiposity should also be further evaluated as
potential mediators of the two disorders.
Although some of these findings are comparable to
previously published studies, others are inconsistent.
First, the association between WHtR and diabetes found
in this study using HbA1c to identify participants with
diabetes suggests that the results of an earlier analysis of
SAGE Wave 1 data—which did not observe an associa-
tion between WHtR and diabetes in Mexico (but did
observe this relationship in other countries) (Tyrovolas
et al., 2015)—may have been due to self-report bias or
lack of reliable reporting of clinical diagnosis in the sam-
ple (biomarker data were unavailable at the time of that
publication). Second, although previous studies indicate
that inflammation could be a causal factor in the comor-
bidity of diabetes and depression, the present study does
not and suggests that central adiposity is a more signifi-
cant factor in mediating the relationship between inflam-
mation and depression as well as depression and
diabetes. Although many studies have shown the rela-
tionship between chronic illness and inflammation, not
all studies share these results. For example, several stud-
ies from the Tsimane Health and Life History Project
have found that higher levels of inflammation, and spe-
cifically CRP, co-occur with better health outcomes such
as lower cholesterol levels and coronary artery calcium,
potentially because of parasitic burden driving high CRP
and protective lifestyle factors keeping chronic disease
risk low (Gurven et al., 2017; Kaplan et al., 2017). These
studies as well as the present study indicate that the rela-
tionship between inflammation and chronic disease is
complex and that inflammation is not always the key
pathophysiological mechanism.
Furthermore, this study adds to the field of mental
health and aging research from middle-income countries.
Infact,itmaybethefirsttousealargesampleofolder
adults in a middle-income nation with biomarker informa-
tion to investigate physiological processes that might be
involved in diabetes and depression. This is significant
because studies of populations from low- and middle-
income countries are underrepresented in high-impact
psychiatry journals (Farmer et al., 2013). Indeed, a 2001 sur-
vey found that, in six leading psychiatry journals, only 6%
of the papers originated from countries outside of Western
Europe, North America, Australia, and New Zealand.
Within this 6%, 4% of published studies were from Latin
America (0.0024% of the total papers) (Patel & Sumathipala,
2001). Although representation of research from low- and
middle-income countries is improving, these limitations
were also identified in papers investigating comorbid diabe-
tes and depression, namely the need for research in
populations that are not of European descent and not from
high-income nations (Herder et al., 2018; Schmitz et al.,
2016). This study begins to clarify conflicting relationships
between these conditions observed in smaller and/or less
diverse samples (Herder et al., 2018; Hood et al., 2012). In
addition, the worldwide number of people aged 60 years
and older is growing faster than all younger age groups and
is expected to more than double by 2050 and to more than
triple by 2100, rising from 962 million globally in 2017 to
2.1 billion in 2050 and 3.1 billion in 2100 (United Nations,
2017). With the number of older people increasing world-
wide, making additional years healthy is critical.
4.1 |Limitations
The main limitations of this study are that directionality
cannot be assessed without longitudinal data, and these
results cannot demonstrate causation based on our study
design and data. In other words, although we were able
to demonstrate that diabetes was associated with an
increase in WHtR and diabetes, we were not able to con-
clude that depression causes an increase in WHtR, which
8DONA ET AL.
in turn causes diabetes or vice versa (indeed, this path-
way might be bidirectional). In addition, this study may
only be representative of Mexico, and the results cannot
necessarily be generalized to all countries and communi-
ties. Overall, though, the diversity and size of the sample
are strength of this study.
Another limitation is the identification of a partici-
pant's sex. Because the interviewer made a determination
of sex based on appearance and recorded either “Male”or
“Female,”this variable is a measure of the participant's
gender presentation as perceived by the interviewer, not a
measure of biological sex. This measure would be consis-
tent with biological sex for most individuals; however, it
fails to take into account transgender, non-binary, and
intersex individuals. Therefore, this variable has been
renamed gender for the purposes of this article, and the
limitation of true gender identification due to the method-
ology is noted.
Finally, this study was unable to differentiate between
type 1 and type 2 diabetes, as HbA1c levels increase in both
conditions. The underlying pathological mechanisms differ
significantly between type 1 and type 2 diabetes (Cnop
et al., 2005), and it is possible that the inability to distin-
guish between the two could have increased our analyses'
variance and made our results less significant. However, we
would expect type 2 diabetes to be vastly more common,
especially in older Mexican individuals. Other health condi-
tions have also been shown to impact HbA1c levels, includ-
ing iron deficiency among adults without diabetes (Kim,
Bullard, Herman, & Beckles, 2010). In addition, we did not
differentiate between people being treated for uncontrolled
diabetes (HbA1c > 6.5%) and people who were not.
4.2 |Summary and conclusions
In summary, this study found that, although CRP alone
partially mediated the co-occurrence of depression and
diabetes, controlling for WHtR showed that CRP is not
involved in the mediation of diabetes and depression and
only appears associated at first because of its strong rela-
tionship with central adiposity. Instead, central adiposity
alone mediated the relationship between depression and
diabetes. In addition, central adiposity fully mediated the
relationship between depression and CRP. These results
suggest that depression may cause an increase in central
adiposity which then leads to diabetes and is incidentally
associated with systemic inflammation. Because CRP did
not mediate the relationship between central adiposity
and diabetes, it is possible that there are other pathways
that are also causally involved in the presence of both
depression and diabetes. Knowledge of how central
adiposity contributes to this comorbidity may help in the
design of preventative strategies or improve treatment.
ACKNOWLEDGMENTS
We thank all the participants in this study. SAGE was
supported by NIH NIA Interagency Agreement
YA1323-08-CN-0020 and the Ministry of Health in
Mexico.
AUTHOR CONTRIBUTIONS
Allison C. Dona: Writing of abstract, introduction, parts of
methods, and discussion; data interpretation; and main
editor/composer of manuscript. Alicia M. DeLouize: Statis-
tical analysis, writing of most of the methods and results
sections; data interpretation; and critical feedback to
manuscript. Geeta Eick: Important biomarker insight
and critical feedback to manuscript. Elizabeth Thiele:
SAGE consultant and biomarker insight. Aarón Salinas
Rodríguez, Betty Soledad Manrique Espinoza, Ricardo
Robledo, Salvador Villalpando: Data collection and
biomarker analysis. Paul Kowal: SAGE co-director and
critical feedback to manuscript. Nirmala Naidoo: SAGE
co-director. Somnath Chatterji: SAGE co-director. Josh
Snodgrass: Project management and critical feedback to
manuscript.
ORCID
Allison C. Dona https://orcid.org/0000-0002-5830-626X
Geeta Eick https://orcid.org/0000-0001-7512-3265
REFERENCES
Alvarez-Monjaras, M., & Gonzalez, D. (2016). Depression in Mex-
ico: Stigma and its policy implications. The Yale Global Health
Review,4,13–16.
Anderson,R.J.,Freedland,K.E.,Clouse,R.E.,&Lustman,P.J.
(2001). The prevalence of comorbid depression in adults
with diabetes: A meta-analysis. Diabetes care,24(6), 1069–1078.40.
Ashwell, M., Gunn, P., & Gibson, S. (2012). Waist-to-height ratio is
a better screening tool than waist circumference and BMI for
adult cardiometabolic risk factors: Systematic review and meta-
analysis. Obesity Reviews,13(3), 275–286.
Berge, L. I., & Riise, T. (2015). Comorbidity between type 2 diabetes
and depression in the adult population: directions of the associ-
ation and its possible pathophysiological mechanisms. Interna-
tional Journal of Endocrinology,2015,1–7.
Betteridge, D. J. (2000). What is oxidative stress? Metabolism,49
(2), 3–8.
Black, S. A., Markides, K. S., & Ray, L. A. (2003). Depression pre-
dicts increased incidence of adverse health outcomes in older
Mexican Americans with type 2 diabetes. Diabetes Care,26(10),
2822–2828.
Bray, G. A., Jablonski, K. A., Fujimoto, W. Y., Barrett-Connor, E.,
Haffner, S., Hanson, R. L., …Pi-Sunyer, F. X. (2008). Relation
of central adiposity and body mass index to the development of
DONA ET AL.9
diabetes in the Diabetes Prevention Program. The American
Journal of Clinical Nutrition,87(5), 1212–1218.
Brindle, E., O'Connor, K. A., & Garrett, D. A. (2014). Applications
of dried blood spots in general human health studies. Dried
Blood Spots: Applications and Techniques,1, 114–129.
Browning, L. M., Hsieh, S. D., & Ashwell, M. (2010). A systematic
review of waist-to- height ratio as a screening tool for the pre-
diction of cardiovascular disease and diabetes: 05 could be a
suitable global boundary value. Nutrition Research Reviews,
23(2), 247–269.
Carey, M., Small, H., Yoong, S. L., Boyes, A., Bisquera, A., &
Sanson-Fisher, R. (2014). Prevalence of comorbid depression
and obesity in general practice: A cross- sectional survey. The
British Journal of General Practice,64(620), e122-e127.
Ciechanowski, P. S., Katon, W. J., & Russo, J. E. (2000). Depression
and diabetes: Impact of depressive symptoms on adherence, func-
tion, and costs. Archives of Internal Medicine,160(21), 3278–3285.
Cnop, M., Welsh, N., Jonas, J. C., Jörns, A., Lenzen, S., &
Eizirik, D. L. (2005). Mechanisms of pancreatic β-cell death in
type 1 and type 2 diabetes: Many differences, few similarities.
Diabetes,54(suppl 2), S97–S107.
de Waal, E., Driver, K., & Warner, J. (2019). Measurement of
HbA1c from dried blood spot samples with the Abbott architect
enzymatic HbA1c method. Pathology,51, S109.
Dehghan, A., Kardys, I., de Maat, M. P., Uitterlinden, A. G.,
Sijbrands, E. J., Bootsma, A. H., …Witteman, J. C. (2007).
Genetic variation, C-reactive protein levels, and incidence of
diabetes. Diabetes,56(3), 872–878.
Duncan, B. B., Schmidt, M. I., Pankow, J. S., Ballantyne, C. M.,
Couper,D.,Vigo,A.,…Heiss, G. (2003). Low-grade systemic
inflammation and the development of type 2 diabetes: The athero-
sclerosis risk in communities study. Diabetes,52(7), 1799–1805.
Eaton, J., McCay, L., Semrau, M., Chatterjee, S., Baingana, F.,
Araya, R., …& Saxena, S. (2011). Scale up of services for mental
health in low-income and middle-income countries. The Lan-
cet,378(9802), 1592–1603.
Farmer, P., Kim, J. Y., Kleinman, A., & Basilico, M. (2013).
Reimagining global health: An introduction (Vol. 26). Univ of
California Press, California and London.
Furukawa, S., Fujita, T., Shimabukuro, M., Iwaki, M., Yamada, Y.,
Nakajima, Y., …Shimomura, I. (2017). Increased oxidative
stress in obesity and its impact on metabolic syndrome. The
Journal of Clinical Investigation,114(12), 1752–1761.
Gurven, M., Stieglitz, J., Trumble, B., Blackwell, A. D., Beheim, B.,
Davis, H., …Kaplan, H. (2017). The Tsimane health and life
history project: integrating anthropology and biomedicine. Evo-
lutionary Anthropology: Issues, News, and Reviews,26(2), 54–73.
Herder, C., Schmitt, A., Budden, F., Reimer, A., Kulzer, B.,
Roden, M., …Hermanns, N. (2018). Association between pro-
and anti-inflammatory cytokines and depressive symptoms in
patients with diabetes—potential differences by diabetes type
and depression scores. Translational Psychiatry,7(11), 1.
Hood, K. K., Lawrence, J. M., Anderson, A., Bell, R., Dabelea, D.,
Daniels, S., …SEARCH for Diabetes in Youth Study Group.
(2012). Metabolic and inflammatory links to depression in
youth with diabetes. Diabetes Care, 35(12), 2443–2446.
Howren, M. B., Lamkin, D. M., & Suls, J. (2009). Associations of
depression with C- reactive protein, IL-1, and IL-6: A meta-
analysis. Psychosomatic Medicine,71(2), 171–186.
Hsieh, N. (2015). Economic security, social cohesion, and depres-
sion disparities in post- transition societies: A comparison of
older adults in China and Russia. Journal of Health and Social
Behavior,56(4), 534–551.
Huang, P. L. (2009). A comprehensive definition for metabolic syn-
drome. Disease models & mechanisms,2(5-6), 231–237.
Institute of Health Metrics and Evaluation (IHME). (2019). Mexico.
University of Washington. http://www.healthdata.org/mexico
Jones, D. R., Macias, C., Barreira, P. J., Fisher, W. H.,
Hargreaves, W. A., & Harding, C. M. (2004). Prevalence, sever-
ity, and co-occurrence of chronic physical health problems of
persons with serious mental illness. Psychiatric Services,55(11),
1250–1257.
Kamenov, K., Caballero, F. F., Miret, M., Leonardi, M., Sainio, P.,
Tobiasz-Adamczyk, B., …Cabello, M. (2016). Which are the
most burdensome functioning areas in depression? A cross-
national study. Frontiers in Psychology,7, 1342.
Kaplan, H., Thompson, R. C., Trumble, B. C., Wann, L. S.,
Allam, A. H., Beheim, B., …Rodriguez, D. E. (2017). Coronary
atherosclerosis in indigenous South American Tsimane: A cross-
sectional cohort study. The Lancet,389(10080), 1730–1739.
Katon, W. J. (2008). The comorbidity of diabetes mellitus and
depression. The American Journal of Medicine,121(11), S8–S15.
Kessler, R. C., & Üstün, T. B. (2004). The world mental health (WMH)
survey initiative version of the world health organization (WHO)
composite international diagnostic interview (CIDI). International
Journal of Methods in Psychiatric Research,13(2), 93–121.
Kim, C., Bullard, K. M., Herman, W. H., & Beckles, G. L. (2010).
Association between iron deficiency and A1C levels among
adults without diabetes in the National Health and Nutrition
Examination Survey, 1999–2006. Diabetes Care,33(4), 780–785.
Kowal, P., Chatterji, S., Naidoo, N., Biritwum, R., Fan, W., Lopez
Ridaura, R., …Boerma, J. T. (2012). Data resource profile: The
World Health Organization study on global AGEing and adult
health (SAGE). International Journal of Epidemiology,41(6),
1639–1649. https://doi.org/10.1093/ije/dys210
Lacher, D. A., Berman, L. E., Chen, T. C., & Porter, K. S. (2013).
Comparison of dried blood spot to venous methods for hemo-
globin A1c, glucose, total cholesterol, high-density lipoprotein
cholesterol, and C-reactive protein. Clinica Chimica Acta,422,
54–58.
Lehmann, S., Delaby, C., Vialaret, J., Ducos, J., & Hirtz, C. (2013).
Current and future use of “dried blood spot”analyses in clinical
chemistry. Clinical Chemistry and Laboratory Medicine,51(10),
1897–1909.
Li, J. J., & Fang, C. H. (2004). C-reactive protein is not only an
inflammatory marker but also a direct cause of cardiovascular
diseases. Medical Hypotheses,62(4), 499–506.
Li, W., & Lee, M. (2014). In W. Li & M. S. V. I. I. I.-I. X. Lee (Eds.),
Dried blood spots. Applications and techniques. Wiley Hoboken.
Łopaty
nski, J., Mardarowicz, G., & Szcze
sniak, G. (2003). A com-
parative evaluation of waist circumference, waist-to-hip ratio,
waist-to-height ratio and body mass index as indicators of
impaired glucose tolerance and as risk factors for type-2 diabe-
tes mellitus. Annales Universitatis Mariae Curie-Sklodowska.
Sectio D: Medicina, Poland. Uniwersytet Marii Curie-
Sklodowskiej, 58(1), 413–419.
Luppino, F. S., de Wit, L. M., Bouvy, P. F., Stijnen, T., Cuijpers, P.,
Penninx, B. W., & Zitman, F. G. (2010). Overweight, obesity,
10 DONA ET AL.
and depression: A systematic review and meta-analysis of longi-
tudinal studies. Archives of General Psychiatry,67(3), 220–229.
McDade, T. W., Tallman, P. S., Madimenos, F. C., Liebert, M. A.,
Cepon, T. J., Sugiyama, L. S., & Snodgrass, J. J. (2012). Analysis
of variability of high sensitivity C-reactive protein in lowland
Ecuador reveals no evidence of chronic low-grade inflamma-
tion. American Journal of Human Biology,24(5), 675–681.
Mezuk, B., Eaton, W. W., Albrecht, S., & Golden, S. H. (2008).
Depression and type 2 diabetes over the lifespan: A meta-analy-
sis. Diabetes Care,31(12), 2383–2390.
Mokdad, A. H., Ford, E. S., Bowman, B. A., Dietz, W. H.,
Vinicor, F., Bales, V. S., & Marks, J. S. (2003). Prevalence of
obesity, diabetes, and obesity-related health risk factors, 2001.
JAMA,289(1), 76–79.
Nehring, S. M., & Patel, B. C. (2018). C reactive protein (CRP). In
StatPearls [Internet]. StatPearls Publishing, Treasure Island
Florida.
Pan, A., Sun, Q., Czernichow, S., Kivimaki, M., Okereke, O. I.,
Lucas, M., …Hu, F. B. (2012). Bidirectional association between
depression and obesity in middle-aged and older women. Inter-
national Journal of Obesity,36, 595–602.
Patel, V., & Sumathipala, A. (2001). International representation in
psychiatric literature: Survey of six leading journals. The British
Journal of Psychiatry,178(5), 406–409.
Pearson, T. A., Mensah, G. A., Alexander, R. W., Anderson, J. L.,
Cannon, R. O., III, Criqui, M., …Rifai, N. (2003). Markers of
inflammation and cardiovascular disease: application to clinical
and public health practice: A statement for healthcare profes-
sionals from the Centers for Disease Control and Prevention
and the American Heart Association. Circulation,107(3),
499–511.
Pradhan, A. D., Manson, J. E., Rifai, N., Buring, J. E., &
Ridker, P. M. (2001). C- reactive protein, interleukin 6, and risk
of developing type 2 diabetes mellitus. JAMA,286(3), 327–334.
R Core Team. (2019). R: A language and environment for statistical
computing. Vienna, Austria: R Foundation for Statistical Com-
puting. https://www.R-project.org/
Ridker, P. M. (2003). Clinical application of C-reactive protein for
cardiovascular disease detection and prevention. Circulation,
107(3), 363–369.
Rodea-Montero, E. R., Evia-Viscarra, M. L., & Apolinar-Jiménez, E.
(2014). Waist-to-height ratio is a better anthropometric index
than waist circumference and BMI in predicting metabolic syn-
drome among obese Mexican adolescents. International Journal
of Endocrinology,2014,1–9.
Rosseel, Y. (2012). lavaan: An R package for structural equation
modeling. Journal of Statistical Software,48(2), 1–36.
Roy, T., & Lloyd, C. E. (2012). Epidemiology of depression and dia-
betes: a systematic review. Journal of affective disorders,142,
S8–S21.
Saraceno, B., van Ommeren, M., Batniji, R., Cohen, A., Gureje, O.,
Mahoney, J., …Underhill, C. (2007). Barriers to improvement
of mental health services in low- income and middle-income
countries. The Lancet,370(9593), 1164–1174.
Sareen, J., Cox, B. J., Stein, M. B., Afifi, T. O., Fleet, C., &
Asmundson, G. J. (2007). Physical and mental comorbidity, dis-
ability, and suicidal behavior associated with posttraumatic
stress disorder in a large community sample. Psychosomatic
Medicine,69(3), 242–248.
Sattar, N., Gaw, A., Scherbakova, O., Ford, I., O'Reilly, D. S. J.,
Haffner, S. M., …Shepherd, J. (2003). Metabolic syndrome with
and without C-reactive protein as a predictor of coronary heart
disease and diabetes in the West of Scotland Coronary Preven-
tion Study. Circulation,108(4), 414–419.
Sayeed, M. A., Mahtab, H., Latif, Z. A., Khanam, P. A.,
Ahsan, K. A., Banu, A., & Azad, A. K. (2003). Waist-to-height
ratio is a better obesity index than body mass index and waist-
to-hip ratio for predicting diabetes, hypertension and lipidemia.
Bangladesh Medical Research Council Bulletin,29(1), 1–10.
Schmitz, N., Deschenes, S. S., Burns, R. J., Smith, K. J., Lesage, A.,
Strychar, I., …Wang, J. L. (2016). Depression and risk of type
2 diabetes: The potential role of metabolic factors. Molecular
Psychiatry,21(12), 1726–1732.
Schrock, J. M., McClure, H. H., Snodgrass, J. J., Liebert, M. A.,
Charlton, K. E., Arokiasamy, P., …Kowal, P. (2017). Food inse-
curity partially mediates associations between social disadvan-
tage and body composition among older adults in India:
Results from the study on global AGEing and adult health
(SAGE). American Journal of Human Biology,29(6), e23033.
Schwarz, M. J., Chiang, S., Müller, N., & Ackenheil, M. (2001). T-
helper-1 and T- helper-2 responses in psychiatric disorders.
Brain, Behavior, and Immunity,15(4), 340–370.
Stuart, M. J., & Baune, B. T. (2012). Depression and type 2 diabetes:
Inflammatory mechanisms of a psychoneuroendocrine co-morbidity.
Neuroscience & Biobehavioral Reviews,36(1), 658–676.
Thakore, J. H., Richards, P. J., Reznek, R. H., Martin, A., &
Dinan, T. G. (1997). Increased intra-abdominal fat deposition
in patients with major depressive illness as measured by com-
puted tomography. Biological Psychiatry,41(11), 1140–1142.
Trayhurn, P., & Wood, I. S. (2004). Adipokines: Inflammation and
the pleiotropic role of white adipose tissue. British Journal of
Nutrition,92(3), 347–355.
Tyrovolas, S., Koyanagi, A., Garin, N., Olaya, B., Ayuso-Mateos, J. L.,
Miret, M., …Haro, J. M. (2015). Diabetes mellitus and its associa-
tion with central obesity and disability among older adults: A
global perspective. Experimental Gerontology,64,70–77.
United Nations, Department of Economic and Social Affairs, Popu-
lation Division (2017). World population prospects: The 2017
revision, key findings and advance tables. Working Paper
No. ESA/P/WP/248.
Van der Kooy, K., Van Hout, H., Marwijk, H., Marten, H.,
Stehouwer, C., & Beekman, A. (2007). Depression and the risk
for cardiovascular diseases: systematic review and meta analy-
sis. International Journal of Geriatric Psychiatry,22(7), 613–626.
Vogelzangs, N., Kritchevsky, S. B., Beekman, A. T.,
Newman,A.B.,Satterfield,S.,Simonsick,E.M.,…
Penninx, B. W. (2008). Depressive symptoms and change in
abdominal obesity in older persons. Archives of General Psy-
chiatry,65(12), 1386–1393.
Volanakis, J. E. (2001). Human C-reactive protein: Expression,
structure, and function. Molecular Immunology,38(2–3),
189–197.
Voleti, B., & Agrawal, A. (2005). Regulation of basal and induced
expression of C- reactive protein through an overlapping ele-
ment for OCT-1 and NF-κB on the proximal promoter. The
Journal of Immunology,175(5), 3386–3390.
Warner, R. M. (2012). Applied statistics: From bivariate through mul-
tivariate techniques. Sage, Thousand Oaks California.
DONA ET AL.11
Wisse, B. E. (2004). The inflammatory syndrome: The role of adipose
tissue cytokines in metabolic disorders linked to obesity. Journal
of the American Society of Nephrology,15(11), 2792–2800.
World Health Organization. (2011a). Use of glycated haemoglobin
(HbA1c) in the diagnosis of diabetes mellitus. Diabetes Research
and Clinical Practice,93(3), 299–309.
World Health Organization. (2011b). Waist circumference and
waist-hip ratio: Report of a WHO expert consultation, Geneva,
8–11 December 2008. Geneva: World Health Organization.
World Health Organization. (2016). Mexico diabetes country profile.
World Health Organization.
World Health Organization. (2019). Depression fact sheet. World
Health Organization.
World Health Organization. (2018). Diabetes fact sheet. World
Health Organization.
Yousuf,O.,Mohanty,B.D.,Martin,S.S.,Joshi,P.H.,Blaha,M.J.,
Nasir, K., …Budoff, M. J. (2013). High-sensitivity C-reactive pro-
tein and cardiovascular disease: A resolute belief or an elusive
link? Journal of the American College of Cardiology,62(5), 397–408.
Yudkin, J. S., Stehouwer, C., Emeis, J., & Coppack, S. (1999). C-
reactive protein in healthy subjects: Associations with obesity,
insulin resistance, and endothelial dysfunction: a potential role
for cytokines originating from adipose tissue? Arteriosclerosis,
Thrombosis, and Vascular Biology,19(4), 972–978.
Zimmermann, O., Li, K., Zaczkiewicz, M., Graf, M., Liu, Z., &
Torzewski, J. (2014). C-reactive protein in human atherogene-
sis: Facts and fiction. Mediators of Inflammation,2014,1–6.
How to cite this article: Dona AC,
DeLouize AM, Eick G, et al. Inflammation and
central adiposity as mediators of depression and
uncontrolled diabetes in the study on global
AGEing and adult health (SAGE). Am J Hum Biol.
2020;1–12. https://doi.org/10.1002/ajhb.23413
12 DONA ET AL.