Content uploaded by Christian Heinz Anderwald
Author content
All content in this area was uploaded by Christian Heinz Anderwald on May 07, 2017
Content may be subject to copyright.
The Relationship between Insulin Resistance and the
Cardiovascular Biomarker Growth Differentiation
Factor-15 in Obese Patients
Greisa Vila,
1
Michaela Riedl,
1
Christian Anderwald,
1
Michael Resl,
1
Ammon Handisurya,
2
Martin Clodi,
1
Gerhard Prager,
3
Bernhard Ludvik,
1
Michael Krebs,
1
and Anton Luger
1*
BACKGROUND:Growth differentiation factor-15 (GDF-
15) is a stress-responsive cytokine linked to obesity co-
morbidities such as cardiovascular disease, inflamma-
tion, and cancer. GDF-15 also has adipokine properties
and recently emerged as a prognostic biomarker for
cardiovascular events.
METHODS:We evaluated the relationship of plasma
GDF-15 concentrations with parameters of obesity, in-
flammation, and glucose and lipid metabolism in a co-
hort of 118 morbidly obese patients [mean (SD) age
37.2 (12) years, 89 females, 29 males] and 30 age- and
sex-matched healthy lean individuals. All study partici-
pants underwent a 75-g oral glucose tolerance test; 28
patients were studied before and 1 year after Roux-
en-Y gastric bypass surgery.
RESULTS:Obese individuals displayed increased plasma
GDF-15 concentrations (P⬍0.001), with highest con-
centrations observed in patients with type 2 diabetes.
GDF-15 was positively correlated with age, waist-to-
height ratio, mean arterial blood pressure, triglycer-
ides, creatinine, glucose, insulin, C-peptide, hemoglo-
bin A
1c
, and homeostatic model assessment insulin
resistance index and negatively correlated with oral
glucose insulin sensitivity. Age, homeostatic model as-
sessment index, oral glucose insulin sensitivity, and
creatinine were independent predictors of GDF-15
concentrations. Roux-en-Y gastric bypass led to a sig-
nificant reduction in weight, leptin, insulin, and insu-
lin resistance, but further increased GDF-15 concen-
trations (P⬍0.001).
CONCLUSIONS:The associations between circulating
GDF-15 concentrations and age, insulin resistance,
and creatinine might account for the additional cardio-
vascular predictive information of GDF-15 compared
to traditional risk factors. Nevertheless, GDF-15
changes following bariatric surgery suggest an indirect
relationship between GDF-15 and insulin resistance.
The clinical utility of GDF-15 as a biomarker might be
limited until the pathways directly controlling GDF-15
concentrations are better understood.
© 2010 American Association for Clinical Chemistry
The global expansion of obesity counts among the par-
amount healthcare concerns of this century (1 ). Excess
weight is associated with increased health risks and es-
pecially with significantly increased cardiovascular
mortality (2 ). Therefore, individual cardiovascular risk
stratification and respective therapy are important
tasks in the management of obese patients. Growth dif-
ferentiation factor-15 (GDF-15),
4
also known as mac-
rophage inhibitory cytokine-1, is a promising new car-
diovascular biomarker (3, 4 ). GDF-15, a product of
macrophages, cardiomyocytes, and endothelial cells, is
released in response to tissue injury, anoxia, and proin-
flammatory cytokines, and which exerts antiapoptotic
effects (5, 6, 7 ). Several studies have revealed a strong
prognostic value of GDF-15 in patients with coronary
heart disease and heart failure, and also in apparently
healthy women (8–11 ). GDF-15 is directly associated
with measurements of endothelial and cardiovascular
dysfunction and is proposed to carry predictive infor-
mation that outranks that of traditional cardiovascular
risk factors (3 ). Recently Ding et al. found that GDF-15
is also expressed in and released from adipocytes, and
contributes to increasing adiponectin production (12 ).
1
Divisions of Endocrinology and Metabolism, Department of Internal Medicine III,
and
2
Nephrology, Department of Medicine III, and
3
Department of Surgery,
Medical University of Vienna, Vienna, Austria.
* Address correspondence to this author at: Division of Endocrinology and Metabo-
lism, Department of Internal Medicine III, Medical University of Vienna, Vienna,
Austria. Fax ⫹43-140400-7790; e-mail anton.luger@meduniwien.ac.at.
Previous presentations: This work was presented at the 92nd Annual Meeting of
the Endocrine Society, June 19–22, 2010, San Diego, CA.
Received July 18, 2010; accepted November 30, 2010.
Previously published online at DOI: 10.1373/clinchem.2010.153726
4
Nonstandard abbreviations: GDF-15, growth differentiation factor-15; BMI,
body mass index; CRP, C-reactive protein; RYGB, Roux-en-Y gastric bypass;
MAP, mean arterial pressure; Hb A
1c
, hemoglobin A
1c
; OGTT, oral glucose
tolerance test; HOMA, homeostasis model assessment; OGIS, oral glucose
insulin sensitivity; IQR, interquartile range; CLIX, clamp-like insulin resistance
index; NGT, normal glucose tolerance; IGT, impaired glucose tolerance; DM,
type 2 diabetes mellitus.
Clinical Chemistry 57:2
309–316 (2011)
Lipids, Lipoproteins, and Cardiovascular Risk Factors
309
In women, circulating GDF-15 concentrations are in-
creased with type 2 diabetes and correlate with body
mass index (BMI), body fat, glucose, and C-reactive
protein (CRP) (13 ).
The relation of GDF-15 to BMI, and to obesity
comorbidities such as diabetes, inflammation, endo-
thelial dysfunction, and cardiovascular disease, high-
light the importance of characterizing GDF-15 in obese
patients. Here we studied the relationship of GDF-15
with anthropometrical measurements of obesity,
blood pressure, parameters of glucose and lipid metab-
olism, inflammation, and renal function in a cohort of
118 morbidly obese patients vs 30 age- and sex-
matched healthy individuals, and in 28 patients who
underwent laparoscopic Roux-en-Y gastric bypass sur-
gery (RYGB).
Study Participants and Methods
STUDY PARTICIPANTS AND DESIGN
The study protocol was approved by the institutional
review board of the Medical University of Vienna.
Thirty healthy individuals and 118 obese individuals
were evaluated in a cross-sectional study. Inclusion cri-
teria for the healthy individuals were BMI ⬍25 kg/m
2
and no previous medical history. The obese patients
were recruited from the obesity outpatient clinic of the
Division of Endocrinology and Metabolism, and inclu-
sion criteria were BMI ⬎35 kg/m
2
and no previously
diagnosed diabetes mellitus. Exclusion criteria were
positive medical history for coronary heart disease,
heart failure, peripheral artery disease, stroke, malig-
nancy, and chronic liver, renal, or endocrine disease.
During the study day, participants underwent a thor-
ough medical examination. Weight was measured to
the nearest 100 g. Height, waist, and hip circumference
were measured to the nearest centimeter. BMI was cal-
culated as weight in kilograms divided by the square of
height in meters. Blood pressure was measured on the
left arm by use of a sphygmomanometer and a cuff
appropriate for the arm circumference, after the study
participant had been sitting for 10 min. Mean arterial
pressure (MAP) was calculated as (2 ⫻diastolic blood
pressure ⫹systolic blood pressure)/3. Blood samples
were withdrawn for the measurement of triglycerides,
total cholesterol, LDL cholesterol, HDL cholesterol,
CRP, creatinine, albumin, and hemoglobin A
1c
(Hb
A
1c
) at baseline. Blood samples for the measurement of
GDF-15 were collected in tubes containing EDTA, cen-
trifuged at 1500gfor 10 min, and immediately frozen at
⫺20 °C. Then, an oral glucose tolerance test (OGTT)
was performed using 75 g glucose. The homeostasis
model assessment (HOMA) insulin resistance index
was calculated as the product of fasting glucose (in mg/
dL) and insulin (in mU/L) divided by the constant 405.
The oral glucose insulin sensitivity (OGIS) was calcu-
lated as explained in http://webmet.pd.cnr.it/ogis (14 ).
The clamp-like insulin resistance index (CLIX) was
calculated as previously reported (15 ). We calculated
the glomerular filtration rate (GFR) using the Modifi-
cation of Diet in Renal Disease formula (16 ).
In an interventional study, 28 obese patients
scheduled to undergo RYGB surgery were studied at 2
time points: before and 1 year after the intervention. At
both study days, a clinical examination was performed;
weight, height, and waist circumference were mea-
sured; and blood samples were withdrawn according to
the same protocol used in the cross-sectional study and
explained above.
ASSAYS
We measured human GDF-15 using a quantitative sand-
wich ELISA kit (# DGD150, R&D Systems) with intra-
and interassay CVs of ⬍2.8% and ⬍6%, respectively.
Insulin and C peptide were determined by using com-
mercially available RIAs (LINCO Research). Leptin was
measured by using the Human Fluorokine MAP Base Kit
(Obesity Panel) and the Leptin Fluorokine MAP (R&D
Systems). Fasting glucose, triglycerides, total cholesterol,
LDL-cholesterol, HDL-cholesterol, albumin, CRP, creat-
inine, and Hb A
1c
were quantified by using routine
tests in a certified clinical laboratory.
STATISTICAL ANALYSIS
Data distributions were tested for normality by using
histograms. Normally distributed data are expressed as
mean (SE), nonnormally distributed data are pre-
sented as median and interquartile range (IQR). Dif-
ferences between the groups were tested by using the
Bonferroni-Holm– corrected 2-sided independent-
samples t-test for parametric data and the Mann–
Whitney U-test for nonnormally distributed data (such
as GDF-15). Spearman rank correlations were com-
puted to assess the relationship between variables.
Multiple regression analyses were performed for iden-
tifying independent relationships and adjusting the ef-
fects of covariates. Nonnormally distributed parame-
ters (GDF-15, creatinine, insulin, C-peptide, HOMA
insulin resistance index, and triglycerides) were loga-
rithmically transformed before regression analyses.
Differences in GDF-15 between the 4 subgroups
[healthy, obese with normal glucose tolerance (NGT),
obese with impaired glucose tolerance (IGT), and
obese with type 2 diabetes mellitus (DM)] were tested
by use of one-way ANOVA followed by post hoc t-tests
with Bonferroni correction for multiple testing. In the
interventional study, differences between baseline and
post-RYGB values were tested using a Bonferroni-Holm–
310 Clinical Chemistry 57:2 (2011)
corrected paired Student t-test. The statistical software
package SPSS release 15.0.1 (SPSS) was used. Pvalues ⬍
0.05 were considered statistically significant.
Results
Clinical, biochemical, and metabolic characteristics of
participants of the cross-sectional study are given in
Table 1. Median (IQR) plasma GDF-15 concentra-
tions were 309 (275– 411) ng/L in healthy individu-
als and 427 (344 – 626) ng/L in obese patients (P⬍
0.001) (Fig. 1).
In the obese cohort, GDF-15 concentrations were
significantly correlated with age, waist circumference
(and waist-to-height ratio), MAP, fasting glucose, fast-
ing insulin, fasting C-peptide, Hb A
1c
, HOMA insulin
resistance index, and fasting triglycerides and creati-
nine and negatively correlated with OGIS (Table 2, Fig.
2 A-B). GDF-15 was not associated with renal function
(GFR) or CRP (Table 2). Multiple regression analysis
revealed that age, HOMA insulin resistance index,
OGIS, and creatinine were independent predictors of
circulating GDF-15 concentrations (Table 3). The cor-
relations between GDF-15 and MAP and fasting trig-
lycerides and fasting glucose disappeared when
GDF-15 was adjusted for age. The correlations between
GDF-15 and waist circumference and fasting insulin
and fasting C-peptide remained significant after we ad-
justed GDF-15 for age and creatinine, but disappeared
after an additional adjustment for HOMA insulin re-
sistance index and OGIS. Arterial hypertension was
present in 49 patients (41%). There were no significant
differences in plasma GDF-15 between patients with
and without hypertension.
When data from all participants (healthy and
obese individuals) were taken together, all the above
relationships between GDF-15 and anthropometric or
metabolic parameters remained significant. In addi-
tion, GDF-15 was weakly but significantly related to
BMI, CRP, and CLIX (Table 2), but not to GFR.
Obese patients were subdivided according to
OGTT results: 69 patients with NGT, 35 patients with
IGT, and 14 patients with newly diagnosed DM (Fig.
2C). GDF-15 was significantly increased in all of these
subgroups compared to the healthy control group (P⫽
0.001 for comparison between healthy and NGT; P⬍0.001
Table 1. Clinical, biochemical, and metabolic characteristics of the study participants.
a
Healthy (n ⴝ30) Obese (n ⴝ120) P
Sex, male/female 9/21 30/90
Age, years 38.2 (1.6) 37.3 (1.1) NS
b
Weight, kg 67.7 (1.9) 134.9 (2) ⬍0.001
BMI, kg/m
2
22.6 (0.4) 47.1 (0.6) ⬍0.001
Waist-to-height ratio 0.46 (0.01) 0.74 (0.01) ⬍0.001
MAP, mmHg 97.1 (2.1) 120 (1.4) ⬍0.001
Fasting glucose, mg/dL
c
86.9 (1.1) 110 (1.6) ⬍0.001
Fasting insulin, mU/L 7.4 (5.7–8.6) 26 (20–36) ⬍0.001
Fasting C-peptide,
g/L 1.6 (1.4–2) 4.1 (3.1–5.7) ⬍0.001
HOMA insulin resistance index 1.5 (1.2–1.9) 6 (4.7–9.7) ⬍0.001
OGIS 471 (8) 318 (6) ⬍0.001
Hb A
1c
,% 5.3 (0.06) 5.6 (0.05) 0.02
Triglycerides, mg/dL 81 (69–96) 137 (103–185) ⬍0.001
Total cholesterol, mg/dL 188 (6) 201 (4) NS
LDL cholesterol, mg/dL 109 (5) 123 (3) 0.04
HDL cholesterol, mg/dL 62.5 (2) 47.6 (1) ⬍0.001
CRP, mg/L 1.4 (0.2) 11.6 (0.8) ⬍0.001
Creatinine, mg/dL 0.9 (0.83–0.95) 0.85 (0.79–0.95) NS
a
Normally distributed data are expressed as mean (SE), nonnormally distributed data are presented as median (IQR). The
P
values correspond to the differences
between healthy and obese individuals.
b
NS not significant.
c
To convert concentrations to millimoles per liter, multiply by 0.0555 for glucose; by 0.0113 for triglycerides; by 0.0259 for cholesterol, LDL cholesterol, and HDL
cholesterol; and by 88.4 for creatinine.
GDF-15 and Insulin Resistance
Clinical Chemistry 57:2 (2011) 311
for comparison between healthy and IGT; P⬍0.001
for comparison between healthy and DM; Fig. 2D).
There were no significant differences in age between
healthy study participants and those in the NGT, IGT,
and DM groups. Within the obese cohort, patients with
DM had significantly higher GDF-15 concentrations
(P⫽0.016) and were significantly older (P⫽0.028)
compared to patients with NGT (Fig. 2D). Differences
in age and GDF-15 between other obese subgroups
were not found to be significant.
In the interventional study, we measured GDF-15
concentrations in 28 individuals undergoing laparoscopic
RYGB surgery, at baseline and 1 year after the interven-
tion. RYGB-induced changes in clinical, biochemical, and
metabolic parameters are presented in Table 4. GDF-15
significantly increased from 474 (31) to 637 (52) ng/L af-
ter bariatric surgery, P⫽0.001 both before and after ex-
clusion of the outlier value (173% increase in GDF-15
after bariatric surgery) (Fig. 2E). One year after RYGB, the
correlation between GDF-15 and age remained signifi-
cant (R⫽0.495, P⫽0.009), whereas all other associations
did not. The RYGB-induced increase in GDF-15 was pos-
itively associated with the decreases in BMI (R⫽0.541,
P⫽0.004) and in the HOMA insulin resistance index
(R⫽0.622, P⫽0.003) (Fig. 2F).
Discussion
GDF-15 is known as a stress-induced cytokine that in-
creases in response to cardiovascular dysfunction and
carries prognostic information on cardiovascular mor-
tality in healthy people and in patients with known car-
diovascular disease (3, 11 ). The main finding of this
study was that GDF-15 is related to all parameters char-
acterizing glucose metabolism and is positively corre-
lated to glucose, insulin, C-peptide, Hb A
1c
, and
HOMA insulin resistance index, and negatively corre-
lated to the oral glucose insulin sensitivity (measured as
OGIS). HOMA insulin resistance index and OGIS were
both independent predictors of GDF-15 in obese pa-
tients. We included both HOMA and OGIS in the mul-
tiple regression analysis because they are used to esti-
mate different processes. The HOMA insulin resistance
index is a parameter that is calculated by using fasting
glucose and insulin concentrations and reflects mainly
hepatic, but not peripheral, insulin resistance (17 ).
OGIS is an indicator of insulin sensitivity in response to
OGTT and therefore reflects mainly glucose clearance
and muscle sensitivity to insulin (14 ).
GDF-15 concentrations were not related to renal
function (measured as GFR), but were predicted by
creatinine, a parameter known to reflect muscle mass
in individuals with normal renal function (18 ).
GDF-15 was higher in obese patients with newly diag-
nosed DM compared to obese patients with NGT. Nev-
Table 2. Spearman correlations of GDF-15.
Obese
Healthy ⴙ
obese
Age 0.512
a
0.369
a
BMI 0.041 0.282
a
Waist circumference 0.349
a
0.449
a
MAP 0.196
c
0.343
a
Fasting glucose 0.336
a
0.370
a
Fasting insulin 0.270
b
0.387
a
Fasting C-peptide 0.363
a
0.455
a
Hb A
1c
0.386
a
0.394
a
HOMA insulin resistance index 0.324
a
0.421
a
OGIS ⫺0.204
c
⫺0.327
a
CLIX ⫺0.128 ⫺0.304
a
CRP ⫺0.024 0.216
c
Fasting triglycerides 0.187
c
0.328
a
Creatinine 0.401
a
0.312
a
a
P
⬍0.001.
b
P
⬍0.01.
c
P
⬍0.05.
Fig. 1. GDF-15 plasma concentrations in lean (n ⴝ
30) and obese (n ⴝ120) individuals matched for age
and sex.
Bars represent IQRs and lines mark medians. Whiskers extend
from the box up to the smallest/highest observations that lie
within 1.5 IQR from the quartiles. Observations that lie further
from the quartiles are marked by circles (1.5–3 IQR) or an
asterisk (more than 3 IQR from the quartiles).
312 Clinical Chemistry 57:2 (2011)
Fig. 2. Scatterplots representing the relationship between (A) GDF-15 and age and (B) GDF-15 and HOMA insulin
resistance index in obese individuals.
(C), Glucose concentrations in response to 75g–2h-OGTT in healthy individuals (white triangles), obese-NGT group (black
triangles), obese-IGT group (white circles) and obese-DM patients (black circles). Data are presented as mean ⫾SE. To convert
glucose concentrations to mmol/L, multiply by 0.0555. (D), GDF-15 plasma concentrations in healthy individuals, and obese
patients with NGT, IGT and DM. Data are presented as mean (SE) *
P
⬍0.05 versus healthy individuals, $
P
⬍0.05 versus the
obese-NGT group, and §
P
⬍0.05 versus the obese-DM group. (E), Individual GDF-15 plasma concentrations before and after
RYGB. (F), Scatterplot displaying RYGB-induced changes in GDF-15 and HOMA insulin resistance index.
GDF-15 and Insulin Resistance
Clinical Chemistry 57:2 (2011) 313
ertheless, obese patients with diabetes were a small and
older subgroup of our cohort. Whether GDF-15 con-
centrations are increased in patients with DM com-
pared to age- and sex-matched healthy individuals re-
mains to be evaluated in further studies.
The strongest predictor of GDF-15 in obese indi-
viduals was age, a parameter that outranks all modifi-
able cardiovascular risk factors in the cardiovascular
risk stratification (19 ). In addition, GDF-15 was
strongly associated with the waist-to-height ratio, but
not to BMI in obese individuals (despite the wide BMI
range: 37– 62 kg/m
2
). Recently, the measurements of
abdominal obesity, and especially the waist-to-height
ratio, have been identified to have a better cardiovas-
cular predictive value compared to BMI (20 ). In sum-
mary, the strong relationships between GDF-15 and
age, insulin resistance, creatinine, and waist-to-height
ratio taken together might contribute to the increased
prognostic information of GDF-15 compared with
other clinical and biochemical markers of cardiovascu-
lar risk (3 ).
In addition to cardiovascular disease, GDF-15 has
been linked to inflammation and cancer (21 ). Macro-
phages, endothelial cells, and cardiomyocytes com-
prise the main sources of GDF-15 (5, 6, 7 ). In vitro
studies have found increased GDF-15 release after tis-
sue injury, anoxia, and stimulation with proinflamma-
tory cytokines such as tumor necrosis factor-
␣
, but not
with lipopolysaccharide (5 ). Inflammation has been
implicated in the pathophysiology of atherosclerotic
plaques and therefore in cardiovascular events (22 ).
Obesity is associated with a mild systemic inflamma-
tion and, as expected, we found a mild but significant
relationship between GDF-15 and CRP in the whole
cohort comprising healthy and obese individuals. Nev-
ertheless, this relationship disappeared within the
obese cohort, revealing the independence of GDF-15
concentrations from the degree of systemic inflamma-
tion in obesity. Given the fact that GDF-15 is secreted
by adipocytes and therefore considered to be an adipo-
kine, we assumed that GDF-15 concentrations are al-
tered in obese individuals (12 ). Nevertheless, results of
a recent study demonstrated increased circulating
Table 3. Determinants of log GDF-15 (standardized

-coefficient and
P
value) in multiple linear
regression analysis.
Variable

Coefficient P
Age 0.437 ⬍0.001
Log creatinine 0.319 ⬍0.001
Log HOMA-insulin resistance 0.343 ⬍0.001
OGIS 0.177 0.019
Waist circumference 0.113 0.143
MAP ⫺0.072 0.332
Fasting glucose ⫺0.044 0.664
Log insulin ⫺0.220 0.544
Log C-peptide 0.190 0.078
Log triglycerides ⫺0.108 0.175
Table 4. Clinical and biochemical parameters of morbidly obese individuals before and 1 year after RYGB.
a
Baseline 1 year after surgery P
Age, male/female 42.9 (1.9) (3/25)
GDF-15, ng/L 474 (31) 637 (52) ⬍0.001
Weight, kg 128 (3) 95 (3) ⬍0.001
Fasting insulin, mU/L 32.6 (4) 12.5 (0.7) ⬍0.001
HOMA insulin resistance index 6.9 (0.9) 2.7 (0.2) ⬍0.001
Fasting triglycerides, mg/dL
b
166 (18) 123 (16) NS
c
Total cholesterol, mg/dL 190 (5) 160 (6) ⬍0.001
LDL cholesterol, mg/dL 121 (5) 87 (5) ⬍0.001
CRP, mg/L 11.6 (1) 4.5 (1) ⬍0.001
Creatinine, mg/dL 0.8 (0.02) 0.79 (0.02) NS
Albumin, g/L 42.4 (0.4) 40.8 (0.4) ⬍0.001
Leptin,
g/L 110 (7) 36 (4) ⬍0.001
a
Data are presented as mean (SE)
P
for comparison between preoperative and postoperative values (Bonferroni-Holm corrected paired
t
-tests).
b
To convert concentrations to millimoles per liter, multiply by 0.0113 for triglycerides; by 0.0259 for cholesterol, LDL cholesterol, and HDL cholesterol; and by 88.4
for creatinine.
c
NS, not significant.
314 Clinical Chemistry 57:2 (2011)
GDF-15 concentrations in obese individuals, but no
differences at the level of gene expression within the
adipose tissue (13 ). The pathophysiological mecha-
nism underlying increased GDF-15 concentrations in
obesity remains unknown and may not be linked only
to adipose tissue. Endothelial dysfunction, cardiac
stress,

-cell function, and insulin resistance may all
contribute to the changes in GDF-15. In the light of the
strong association between GDF-15 and parameters of
glucose metabolism, it is important to identify the in-
fluence of GDF-15 on

-cell function and glucose up-
take and vice versa, an eventual effect of glucose and
insulin on GDF-15 release.
Bariatric surgery is to date the only efficient ther-
apeutic means for achieving weight loss in individuals
with severe obesity. Our observation that RYGB sur-
gery significantly decreased body weight, leptin, CRP,
insulin, and HOMA insulin resistance index confirmed
the results of previous studies (23, 24 ). Nevertheless,
GDF-15 concentrations increased further. The RYGB-
induced increase in GDF-15 was significantly corre-
lated with age. The strong association with insulin re-
sistance was noticeable even during the changes
following bariatric surgery, because obese patients with
larger reductions in weight and insulin resistance had
smaller increases in GDF-15 (Fig. 2F). Nevertheless,
these results suggest an indirect association between
GDF-15 and insulin resistance, and the pathophysio-
logical mechanisms that control postoperative GDF-15
concentrations remain unknown. It is interesting to
note that GDF-15 concentrations also increase after
diet-induced weight loss and in patients with anorexia
nervosa (13, 25 ). To date, it is not known whether cir-
culating GDF-15 concentrations depend on albumin
or any carrier proteins. It is important to emphasize
that the increase in GDF-15 is not in line with the sig-
nificant improvement in cardiovascular function that
occurs following bariatric surgery (26 ). Therefore,
GDF-15 is highly likely to be an unreliable cardiovas-
cular biomarker in patients who have undergone gas-
tric bypass surgery.
In summary, age, insulin resistance, and creatinine
were independent predictors of GDF-15 in obese pa-
tients, and these associations might contribute to the
recently found increased cardiovascular prediction
value of GDF-15 compared with classical predictors.
Nevertheless, the increase in GDF-15 concentrations
following weight loss is not in line with a direct rela-
tionship between GDF-15 and insulin resistance
and/or clinical measurements of obesity. The utility of
GDF-15 as a biomarker might be limited until the path-
ways that directly control GDF-15 concentrations in
humans are better understood.
Author Contributions: All authors confirmed they have contributed to
the intellectual content of this paper and have met the following 3 re-
quirements: (a) significant contributions to the conception and design,
acquisition of data, or analysis and interpretation of data; (b) drafting
or revising the article for intellectual content; and (c) final approval of
the published article.
Authors’ Disclosures or Potential Conflicts of Interest: Upon
manuscript submission, all authors completed the Disclosures of Poten-
tial Conflict of Interest form. Potential conflicts of interest:
Employment or Leadership: None declared.
Consultant or Advisory Role: None declared.
Stock Ownership: None declared.
Honoraria: None declared.
Research Funding: M. Riedl, Medical Scientific Fund of the Mayor
of the City of Vienna.
Expert Testimony: None declared.
Role of Sponsor: The funding organizations played no role in the
design of study, choice of enrolled patients, review and interpretation
of data, or preparation or approval of manuscript.
References
1. Franks PW, Hanson RL, Knowler WC, Sievers ML,
Bennett PH, Looker HC. Childhood obesity, other
cardiovascular risk factors, and premature death.
N Engl J Med 2010;362:485–93.
2. Flegal KM, Graubard BI, Williamson DF, Gail MH.
Cause-specific excess deaths associated with un-
derweight, overweight, and obesity. JAMA 2007;
298:2028–37.
3. Lind L, Wallentin L, Kempf T, Tapken H, Quint A,
Lindahl B, et al. Growth-differentiation factor-15
is an independent marker of cardiovascular dys-
function and disease in the elderly: results from
the Prospective Investigation of the Vasculature
in Uppsala Seniors (PIVUS) Study. Eur Heart J
2009;30:2346–53.
4. Kempf T, Horn-Wichmann R, Brabant G, Peter T,
Allhoff T, Klein G, et al. Circulating concentra-
tions of growth-differentiation factor 15 in appar-
ently healthy elderly individuals and patients with
chronic heart failure as assessed by a new immu-
noradiometric sandwich assay. Clin Chem 2007;
53:284–91.
5. Bootcov MR, Bauskin AR, Valenzuela SM, Moore
AG, Bansal M, He XY, et al. MIC-1, a novel
macrophage inhibitory cytokine, is a divergent
member of the TGF-

superfamily. Proc Natl Acad
SciUSA1997;94:11514–9.
6. Kempf T, Eden M, Strelau J, Naguib M, Willen-
bockel C, Tongers J, et al. The transforming
growth factor-beta superfamily member growth-
differentiation factor-15 protects the heart from
ischemia/reperfusion injury. Circ Res 2006;98:
351–60.
7. Jurczyluk J, Brown D, Stanley KK. Polarised secre-
tion of cytokines in primary human microvascular
endothelial cells is not dependent on N-linked
glycosylation. Cell Biol Int 2003;27:997–1003.
8. Wollert KC, Kempf T, Peter T, Olofsson S, James S,
Johnston N, et al. Prognostic value of growth-
differentiation factor-15 in patients with non-ST-
segment elevation acute coronary syndrome. Cir-
culation 2007;115:962–71.
9. Khan SQ, Ng K, Dhillon O, Kelly D, Quinn P,
Squire IB, et al. Growth differentiation
factor-15 as a prognostic marker in patients
with acute myocardial infarction. Eur Heart J
2009;30:1057–65.
10. Kempf T, von Haehling S, Peter T, Allhoff T,
Cicoira M, Doehner W, et al. Prognostic utility of
growth differentiation factor-15 in patients with
chronic heart failure. J Am Coll Cardiol 2007;50:
1054–60.
11. Brown DA, Breit SN, Buring J, Fairlie WD, Bauskin
AR, Liu T, Ridker PM. Concentration in plasma of
macrophage inhibitory cytokine-1 and risk of car-
diovascular events in women: a nested case-
control study. Lancet 2002;359:2159–63.
12. Ding Q, Mracek T, Gonzalez-Muniesa P, Kos K,
Wilding J, Trayhurn P, Bing C. Identification of
macrophage inhibitory cytokine-1 in adipose tis-
GDF-15 and Insulin Resistance
Clinical Chemistry 57:2 (2011) 315
sue and its secretion as an adipokine by human
adipocytes. Endocrinology 2009;150:1688–96.
13. Dostalova I, Roubicek T, Bartlova M, Mraz M,
Lacinova Z, Haluzikova D, et al. Increased serum
concentrations of macrophage inhibitory
cytokine-1 in patients with obesity and type 2
diabetes mellitus: the influence of very low calo-
rie diet. Eur J Endocrinol 2009;161:397–404.
14. Mari A, Pacini G, Murphy E, Ludvik B, Nolan JJ. A
model-based method for assessing insulin sensi-
tivity from the oral glucose tolerance test. Diabe-
tes Care 2001;24:539–48.
15. Anderwald C, Anderwald-Stadler M, Promintzer
M, Prager G, Mandl M, Nowotny P, et al. The
clamp-like index: A novel and highly sensitive
insulin sensitivity index to calculate hyperinsu-
linemic clamp glucose infusion rates from oral
glucose tolerance tests in nondiabetic subjects.
Diabetes Care 2007;30:2374–80.
16. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N,
Roth D. A more accurate method to estimate
glomerular filtration rate from serum creatinine: a
new prediction equation. The Modification of Diet
in Renal Disease Study Group. Ann Intern Med
1999;130:461–70.
17. Hoffman RP. Indices of insulin action calculated
from fasting glucose and insulin reflect hepatic,
not peripheral, insulin sensitivity in African-
American and Caucasian adolescents. Pediatr Di-
abetes 2008;9:57–61.
18. Schuttle JE, Longhurst JC, Gaffney FA, Bastian
BC, Bloomqvist CG. Total plasma creatinine: an
accurate measure of total striated muscle mass.
J Appl Physiol 1981;51:762–6.
19. Sniderman AD, Furberg CD. Age as a modifiable
risk factor for cardiovascular disease. Lancet
2008;371:1547–9.
20. Schneider HJ, Friedrich N, Klotsche J, Pieper L,
Nauck M, John U, et al. The predictive value of
different measures of obesity for incident cardio-
vascular events and mortality. J Clin Endocrinol
Metab 2010;95:1777–85.
21. Karan D, Holzbeierlein J, Thrasher JB. Macro-
phage inhibitory cytokine-1: possible bridge mol-
ecule of inflammation and prostate cancer. Can-
cer Res 2009;69:2–5.
22. Hansson GK. Inflammation, atherosclerosis, and
coronary artery disease. N Engl J Med 2005;352:
1685–95.
23. Riedl M, Vila G, Maier C, Handisurya A, Shakeri-
Manesch S, Prager G, et al. Plasma osteopontin
increases after bariatric surgery and correlates
with markers of bone turnover but not with in-
sulin resistance. J Clin Endocrinol Metab 2008;
93:2307–12.
24. Vila G, Riedl M, Maier C, Struck J, Morgenthaler
NG, Handisurya A, et al. Plasma MR-proADM
correlates to BMI and decreases in relation to
leptin after gastric bypass surgery. Obesity 2009;
17:1184–8.
25. Dosta´ lova´ I, Kava´ lkova´ P, Papezova´ H, Domlu-
vilova´ D, Zika´ n V, Haluzı´k M. Association of
macrophage inhibitory cytokine-1 with nutritional
status, body composition and bone mineral den-
sity in patients with anorexia nervosa: the influ-
ence of partial realimentation. Nutr Metab (Lond)
2010;23:7:34.
26. Adams TD, Gress RE, Smith SC, Halverson RC,
Simper SC, Rosamond WD, et al. Long-term mor-
tality after gastric bypass surgery. N Engl J Med
2007;357:753–61.
316 Clinical Chemistry 57:2 (2011)
