ORIGINAL ARTICLE

E n d o c r i n e C a r e

Metabolic Syndrome and Insulin Resistance as Risk

Factors for Development of Chronic Kidney Disease
and Rapid Decline in Renal Function in Elderly

Hui-Teng Cheng,* Jenq-Wen Huang,* Chih-Kang Chiang, Chung-Jen Yen,

Kuan-Yu Hung, and Kwan-Dun Wu
Department of Internal Medicine (H.-T.C.), National Taiwan University Hospital, Hsin-Chu Branch;

Downloaded from https://academic.oup.com/jcem/article/97/4/1268/2833449 by guest on 03 April 2022

Departments of Internal Medicine (J.-W.H., C.-K.C., C.-J.Y., K.-Y.H., K.-D.W.), Integrated Diagnostics
and Therapeutics (C.-K.C.), and Geriatrics and Gerontology (C.-J.Y.), National Taiwan University Hospital,
College of Medicine, National Taiwan University, Taipei 10002, Taiwan

Context: Studies addressing the association of metabolic syndrome and insulin resistance with the
risks of incident chronic kidney disease (CKD) and the progression of renal function were either
lacking or inconclusive.

Objective: The aim of this study was to define the effect of metabolic syndrome and insulin re-
sistance on the development of new CKD and the decline in renal function.

Design and Setting: A prospective cohort study was conducted at a tertiary university-based hos-
pital in Taiwan.

Patients and Other Participants: We studied a total of 1456 Asians 65 or older who were followed
for an average of 3.15 yr. Within the cohort, we measured insulin resistance using the homeostasis
model assessment formula in 652 nondiabetic participants.

Interventions: There were no interventions.

Main Outcome Measures: We measured the prevalence and incidence of CKD and the annual
decline of the estimated glomerular filtration rate.

Results: We found that the adjusted odds ratio for prevalent CKD in association with metabolic
syndrome was 1.778 (95% confidence interval, 1.188 to 2.465), the hazard ratio for rapid decline
in renal function was 1.042 (0.802–1.355), and the hazard ratio for incident CKD was 1.931 (1.175–
3.174). With each one-unit increment of insulin resistance, the odds ratio of prevalent CKD and
proteinuria were raised 1.312-fold (1.114 to 1.545) and 1.278-fold (1.098 to 1.488), respectively.
Insulin resistance was not associated with incident CKD. Increment of insulin resistance per unit was
associated with 1.16-fold (1.06 to 1.26) elevation in the hazard ratios of the decline in renal
function.

Conclusions: Metabolic syndrome predicts the risks of prevalent and incident CKD, whereas insulin
resistance is associated with prevalent CKD and rapid decline in renal function in elderly individuals.
(J Clin Endocrinol Metab 97: 1268 –1276, 2012)

etabolic syndrome (Mets) is characterized as ab- 43.5% in subjects aged 60 – 69 yr and 42.0% in those 70
M dominal obesity, dyslipidemia, hypertension, and or older (1). Association between metabolic syndrome and
hyperglycemia. Its prevalence increases with age, reaching the development of chronic kidney disease has been dem-

ISSN Print 0021-972X ISSN Online 1945-7197 * H.-T.C. and J.-W.H. contributed equally to this work.
Printed in U.S.A. Abbreviations: CI, Confidence interval; CKD, chronic kidney disease; eGFR, estimated GFR;
Copyright © 2012 by The Endocrine Society GFR, glomerular filtration rate; HDL, high-density lipoprotein; HOMA, homeostasis model
doi: 10.1210/jc.2011-2658 Received September 26, 2011. Accepted January 12, 2012. assessment; HOMA-IR, HOMA for insulin resistance; HR, hazard ratio; Mets, metabolic
First Published Online February 15, 2012 syndrome; WBC, white blood cell.

1268 jcem.endojournals.org J Clin Endocrinol Metab, April 2012, 97(4):1268 –1276

J Clin Endocrinol Metab, April 2012, 97(4):1268 –1276
onstrated; however, this is based primarily on middle-aged
populations with no special focus on the elderly (2–5). For
example, Kurella et al. (3) showed in a prospective study
that chronic kidney disease is more likely to develop in
nondiabetic, 45- to 64-yr-old individuals who had meta-
bolic syndrome. A study to elucidate the relationship be-
tween metabolic syndrome and incident chronic kidney
disease in an elderly population is therefore needed.
Insulin resistance is defined clinically in terms of the
failure of insulin to maintain glucose homeostasis (6) and
is believed to play a central role in the pathogenesis of the
metabolic syndrome (7). In nondiabetic individuals aged
20 or older, increasing insulin resistance was associated
with progressively higher prevalence of chronic kidney
disease (8). Another cross-sectional study also found that
the presence of insulin resistance was associated with
chronic kidney disease in nondiabetic individuals aged
70 –79 (9). There has been ample evidence indicating the
association of insulin resistance with type 2 diabetes, obe-
sity, essential hypertension, and cardiovascular disorders,
all of which are contributors to the development of chronic
kidney disease (10, 11). However, a study to directly dem-
onstrate the link between insulin resistance and the devel-
opment of chronic kidney disease is indeed lacking.
To the best our knowledge, only two reports examined
at small scale the effect of the insulin resistance on the
progression of chronic kidney disease, and the results were
inconsistent. Kobayashi et al. (12) studied 41 nondiabetic
hypertensive individuals and found insulin resistance as a
risk factor for the progression of chronic kidney disease
(CKD). In contrast, Bastürk and Unsal (13) did not find the
same effect in a study including 79 nondiabetic CKD pa-
tients. Similarly, surprisingly few studies have analyzed
the contribution of the metabolic syndrome to the deteri-
oration of renal function in CKD patients, and yet the
results were inconclusive. In a secondary analysis of the
African-American Study of Hypertension and Kidney Dis-
ease that included 842 individuals, Lea et al. (14) found no
independent association between metabolic syndrome
and the CKD progression in hypertensive African-Amer-
icans. Lee et al. (15) analyzed 762 CKD subjects, and
showed that metabolic syndrome predicted the deteriora-
tion of renal function only in nondiabetic early-stage CKD
patients, and not in nondiabetic late-stage CKD patients
or in diabetic patients.
In this report, we investigate the effect of metabolic
syndrome and insulin resistance on the development of
new CKD and the progression of renal function by con-
ducting a prospective cohort study. The subjects were age
65 or older. We hypothesized that metabolic syndrome
and/or insulin resistance is a risk factor for the develop-
ment of CKD and/or rapid decline in renal function.
jcem.endojournals.org 1269
Subjects and Methods
Study protocol
We organized an elderly cohort by collecting the data of
participants who received voluntary health check-ups at the
National Taiwan University Hospital (NTUH) between Jan-
uary 2002 and December 2005 and were aged 65 or above.
The design for studying the association between metabolic
syndrome and CKD is shown as Supplemental Fig. 1 (pub-
lished on The Endocrine Society’s Journals Online web site at
http://jcem.endojournals.org). At baseline, 1456 of 2163 indi-
viduals signed the informed consent. During the follow-up pe-
riod, all participants were invited to receive the annual elderly
Downloaded from https://academic.oup.com/jcem/article/97/4/1268/2833449 by guest on 03 April 2022
health check-up, and 916 of 1456 (62.9%) participants received
an additional health exam until August 2007.
The design for studying the insulin resistance and CKD is
shown in Supplemental Fig. 2. From the same elderly cohort, 800
individuals agreed to participate in the study. A total of 652
nondiabetic participants were included in the final analysis. Dur-
ing the follow-up period, 363 participants had received further
health exams before August 2007. The mean follow-up periods
are 1 to 5 yr (mean ⫽ 3.15 ⫾ 1.21 yr).
Serum creatinine was measured by Jaffe’s kinetic method on
a Hitachi 7170 autoanalyzer (Hitachi, Tokyo, Japan) in the uni-
versity-based laboratory. A morning spot midstream urine spec-
imen was also collected and analyzed by semiquantitative Mul-
tistix urinary strip (Bayer Clinitek Atlas; Bayer, Leverkusen,
Germany). Results of urine protein were reported as trace, 1⫹,
2⫹, 3⫹, and 4⫹. All subjects completed a questionnaire docu-
menting their sociodemographic status (age, sex, income, and
education), personal health history (hypertension, diabetes, and
hyperlipidemia), lifestyle and behaviors (e.g. smoking and alco-
hol consumption), with the assistance of general practitioners.
This entire clinical study followed the Declaration of Helsinki. It
was permitted by the Medical Ethics Committee of the NTUH
(NCT00173940), and written informed consent for each par-
ticipant was obtained.
Definition of metabolic syndrome
We used the modified National Cholesterol Education Pro-
gram Adult Treatment Panel III (NCEP ATP III) criteria with
Asian cutoff of waist circumference (16). The diagnosis of Mets
was made when three or more of the following criteria were
present: 1) central obesity with waist circumference of at least 80
cm in females and at least 90 cm in males; 2) triglyceride level of
at least 150 mg/dl or specific treatment for this lipid abnormality;
3) high-density lipoprotein (HDL) cholesterol level below 40
mg/dl in males, below 50 mg/dl in females, or specific treatment
for this lipid abnormality; 4) systolic blood pressure of at least
130 mm Hg or diastolic blood pressure of at least 85 mm Hg, or
treatment of previously diagnosed hypertension; and 5) in-
creased fasting glucose of at least 110 mg/dl or previously diag-
nosed type 2 diabetes.
Measurement of the insulin resistance
The homeostasis model assessment (HOMA) was used to
evaluate insulin resistance (17). Assuming that normal subjects
less than 35 yr of age with normal weight have an insulin resis-
tance of 1, the values for a patient can be calculated from the
fasting concentrations of insulin and glucose using the formula:
1270 Cheng et al. Metabolic Syndrome and IR Predict Renal Function Decline
fasting serum insulin (␮U/ml) ⫻ fasting plasma glucose (mmol/
liter)/22.5 (17).
Definition of CKD and rapid decline in kidney
function
The Kidney Disease Quality Outcome Initiative (K/DOQI)
group defines CKD as kidney damage or glomerular filtration
rate (GFR) below 60 ml/min/1.73 m2 for 3 months or more and
recommends that patients who test positive for albuminuria un-
dergo repeated testing within 3 months to confirm its presence.
This elderly cohort has serum creatinine and urinary protein
measured only once at the baseline. Data from the Third Na-
tional Health and Nutrition Examination Survey found persis-
tent microalbuminuria in 53.9% of patients with estimated GFR
(eGFR) of 90 ml/min/1.73 m2 or greater and 72.7% of patients
with eGFR of 60 to 89 ml/min/1.73 m2. All patients (100%) with
macroalbuminuria had persistent albuminuria on repeated mea-
surement (18). Thus, to meet the K/DOQI criteria of persistent
albuminuria, we used ⫹1 of proteinuria as the cutoff for CKD
staging in this study. eGFR was calculated using the four-vari-
able isotope dilution mass spectrometry-traceable Modification
of Diet in Renal Disease Study equation (where Scr is serum
creatinine level) (19): eGFR (ml/min/1.73 m2) ⫽ 186.3 ⫻ Scr
(mg/dl)⫺1.154 ⫻ age (yr)⫺0.203 ⫻ (0.742 if female) ⫻ (1.210 if
African-American).
Prevalent CKD was defined as the presence of either protein-
uria or the eGFR below 60 ml/min/1.73 m2 (20). In the follow-up
analysis, rapid decline in kidney function and incident CKD were
used as kidney outcome. We defined rapid decline in kidney
function as an annual decline in GFR greater than 3 ml/min/1.73
m2. This threshold was chosen based on its established use in
prior studies (21, 22). The definition of incident CKD was an
eGFR less than 60 ml/min/1.73 m2 at follow-up period with the
baseline eGFR of 60 ml/min/1.73 m2 or greater or new onset of
proteinuria (21, 22).
Statistical analysis
In the baseline analysis, we applied ␹2 analysis for categorical
variables and independent t test for continuous variables to eval-
uate the association of baseline characteristics and risk factors
with Mets and CKD. Data were presented as mean ⫾ SD for
continuous variables and as proportions for categorical vari-
ables. In the baseline analysis, the association of Mets with CKD
was examined, and logistic regression was used. In the follow-up
analysis, we examined the association of rapid kidney function
decline and incident CKD with baseline Mets status after ad-
justing by confounding factors. The baseline association analysis
of the HOMA for insulin resistance (HOMA-IR) index with
CKD and proteinuria was examined, and logistic regression was
used. Variables in univariate analysis with a P value ⬍0.1 were
considered to be potential confounders of Mets or insulin resis-
tance in multivariate models. To clarify whether isolated or sum-
marized Mets components may predict CKD events, we further
added the continuous variables, including body weight, fasting
blood glucose, systolic blood pressure, and serum creatinine in
the regression models. We adjusted those variables as indicated
in Tables 2, 3, and 4 for the multivariate analyses. In the fol-
low-up analysis, we examined the association of incident CKD
and rapid decline in kidney function with baseline HOMA-IR
using the Cox proportional hazard regression. We also arranged
linear regression between the annual eGFR decline rate and the
J Clin Endocrinol Metab, April 2012, 97(4):1268 –1276
baseline HOMA-IR. The 95% confidence intervals (CI) were
presented in parentheses as indicated. Statistical analyses were
performed using SPSS 16.0 software for Windows (SPSS Inc.,
Chicago, IL). P value ⬍0.05 is considered statistically significant.
Results
Characteristics of study populations
The general characteristics of the elderly participants
(n ⫽ 1456) were classified by Mets status (Table 1). On
average, persons with Mets were older and more likely to
be female. The mean uric acid, glutamyl pyruvic transam-
Downloaded from https://academic.oup.com/jcem/article/97/4/1268/2833449 by guest on 03 April 2022
inase, and albumin levels were higher among persons with
Mets. The percentages of proteinuria and CKD were
higher, and eGFR was lower among persons with Mets. In
Supplemental Fig. 3, the presence of Mets was stratified by
gender and age group. Elderly female participants have a
higher prevalence of Mets compared with elderly men
(45.8 vs. 33.3%). Supplemental Fig. 4 shows the distri-
bution of the study participants by the number of Mets
components present. Overall, 38.7% of the elderly par-
ticipants had metabolic syndrome.
Supplemental Table 1 presents the general charac-
teristics of nondiabetic geriatric participants in the in-
sulin resistance study by the tertiles of HOMA-IR index
(n ⫽ 652). On average, persons with the higher
HOMA-IR had higher body mass index, mean serum
albumin, uric acid, triglyceride, blood urea nitrogen, cre-
atinine, white blood cell (WBC) counts, high-sensitivity
C-reactive protein, and the chance to have a history of
hypertension. The HDL cholesterol was higher among
persons with low HOMA-IR index. The percentage of
proteinuria was higher, and eGFR was lower among the
elderly group with higher a HOMA-IR index.
Associations of metabolic syndrome with
prevalent CKD
Elderly male participants had a higher prevalence of CKD
compared with elderly women (26.4 vs. 21.8%), and both
genders increased in CKD prevalence as age increased. As
shown in Supplemental Fig. 5A, the presence of individual
Mets components, including high blood pressure, serum tri-
glyceride, fasting plasma glucose, waist circumference, and a
low HDL cholesterol level was associated with higher prev-
alence of CKD. The prevalence of CKD is 36.7% under
NCEP criteria of Mets. There was a significant graded rela-
tionship between the number of components and the corre-
sponding prevalence of CKD (␹2 ⫽ 41.213; P ⬍ 0.001) as
shown in Supplemental Fig. 5B.
As shown in Table 2, the crude and multivariate-ad-
justed odds ratios of CKD were associated with individual
and combined components of the Mets. In the multivariate
J Clin Endocrinol Metab, April 2012, 97(4):1268 –1276 jcem.endojournals.org 1271

TABLE 1. Clinical characteristics of study subjects with and without Mets

Without
Characteristics All participants Mets With Mets P value
n 1456 893 563
Age (yr) 73.1 (5.4) 72.7 (5.3) 73.5 (5.4) ⬍0.001
Men, n (%) 831 (57.1%) 554 (62.0%) 278 (42.9%) ⬍0.001
SBP (mm Hg) 138.4 (18.1) 134.8 (17.4) 144.1 (17.9) ⬍0.001
DBP (mm Hg) 78.6 (11.5) 77.2 (11.3) 80.7 (11.5) ⬍0.001
Antihypertensive agents, n (%) 421 (28.9%) 217 (24.2%) 204 (36.2%) ⬍0.001
BMI (kg/m2) 24.0 (3.2) 22.9 (2.9) 25.6 (2.9) ⬍0.001
Waist circumference (cm) 85.9 (9.6) 82.6 (8.9) 91.1 (8.3) ⬍0.001
Fasting blood glucose (mg/dl) 103.7 (24.8) 97.2 (15.0) 113.9 (32.7) ⬍0.001
HDL cholesterol (mg/dl) 48.2 (10.7) 52.0 (10.7) 42.2 (7.6) ⬍0.001
Triglyceride (mg/dl) 123.5 (65.7) 97.2 (40.4) 165.3 (75.6) ⬍0.001

Downloaded from https://academic.oup.com/jcem/article/97/4/1268/2833449 by guest on 03 April 2022

Total cholesterol (mg/dl) 202.6 (38.2) 204.4 (37.3) 199.7 (39.4) 0.022
Serum albumin (g/dl) 4.42 (0.24) 4.40 (0.24) 4.45 (0.25) 0.002
Serum GPT (U/liter) 25.3 (18.0) 24.1 (14.2) 27.3 (22.5) 0.001
Serum uric acid (mg/dl) 6.2 (1.5) 6.1 (1.5) 6.4 (1.5) ⬍0.001
Hemoglobin (g/dl) 13.6 (1.3) 13.6 (1.3) 13.7 (1.3) 0.718
WBC count (⫻103/␮l) 5.5 (1.4) 5.3 (1.4) 5.8 (1.4) 0.067
Platelet count (⫻103/␮l) 197 (54) 195 (53) 200 (55) 0.277
Alcohol intake, n (%) 289 (19.8%) 179 (20.0%) 110 (19.5%) 0.813
Current smoking, n (%) 82 (5.6%) 49 (5.5%) 33 (5.9%) 0.763
eGFR (ml/min/1.73 m2) 72.5 (15.3) 73.2 (14.8) 71.4 (16.1) 0.033
Proteinuria, n (%)a 147 (10.1%) 62 (6.9%) 85 (15.1%) ⬍0.001
eGFR ⬍60 ml/min/1.73 m2, n (%) 267 (18.3%) 141 (15.8%) 126 (22.4%) 0.001
CKD, n (%)b 355 (24.4%) 176 (19.7%) 179 (35.8%) ⬍0.001
Data are expressed as number (percentage) or mean (SD). SBP, Systolic blood pressure; DBP, diastolic blood pressure; BMI, body mass index; GPT,
glutamyl pyruvic transaminase.
a
Proteinuria denotes ⫹1 (⬎300 mg/liter) or greater by dipstick.
b
CKD is defined as GFR less than 60 ml/min/1.73 m2 or the presence of proteinuria.

models, high blood pressure, low HDL cholesterol level, subjects with zero components of Mets, participants with
high triglyceride level, and high fasting plasma glucose three, four, and five components had increased odds ratios
level were all significantly associated with an increased of 2.684, 2.355, and 4.115 for CKD, respectively. Persons
odds ratio of prevalent CKD (P ⬍ 0.05). Compared with with the NCEP ATP III criteria-defined Mets had a 2.306-

TABLE 2. Crude and multivariate-adjusted odds ratios of chronic kidney disease associated with individual or several
components of the metabolic syndrome
Odds ratio of chronic kidney disease (95% CI)
Variable Univariate P value Multivariate-adjusteda P value
Blood pressure ⭌130/85 mm Hg 2.038 (1.406 –2.953) ⬍0.001 2.040 (1.408 –2.956) ⬍0.001
Serum HDL cholesterol level ⬍40 mg/dl in 1.632 (1.256 –2.121) ⬍0.001 1.576 (1.062–2.340) 0.024
men or ⬍50 mg/dl in women
Serum triglyceride level ⭌150 mg/dl 1.922 (1.460 –2.530) ⬍0.001 1.702 (1.126 –2.572) 0.012
Plasma glucose level ⭌110 mg/dl 1.675 (1.258 –2.229) ⬍0.001 1.464 (1.079 –1.988) 0.014
Waist circumference ⭌90 cm in men and 1.318 (1.015–1.713) 0.038 1.390 (0.931–2.076) 0.107
⭌80 cm in women
One componentb 1.465 (0.769 –2.789) 0.246 1.293 (0.656 –2.547) 0.458
Two componentsb 1.653 (0.878 –3.110) 0.119 1.658 (0.846 –3.249) 0.140
Three componentsb 2.750 (1.464 –5.163) 0.002 2.684 (1.382–5.012) 0.004
Four componentsb 3.558 (1.848 – 6.849) ⬍0.001 2.355 (1.176 – 4.716) 0.016
Five componentsb 4.473 (2.066 –9.680) ⬍0.001 4.115 (1.878 –9.013) ⬍0.001
Four and five componentsb 3.779 (2.004 –7.124) ⬍0.001 3.888 (1.925–7.853) ⬍0.001
Mets, NCEP ATP III 2.151 (1.656 –2.796) ⬍0.001 2.306 (1.588 –3.347) ⬍0.001
Mets, NCEP ATP IIIc 1.880 (1.271–2.779) 0.002
Mets, NCEP ATP IIId 1.778 (1.188 –2.465) 0.004
a
Adjusted for age, gender, hemoglobin, serum albumin, globulin, and uric acid.
b
Compared with those with zero components of the metabolic syndrome.
c
Metabolic syndrome adjusted for other covariates and proteinuria.
d
Metabolic syndrome further adjusted for body weight, systolic blood pressure, fasting blood glucose, and serum creatinine.
1272 Cheng et al. Metabolic Syndrome and IR Predict Renal Function Decline J Clin Endocrinol Metab, April 2012, 97(4):1268 –1276

TABLE 3. Univariate and multivariate-adjusted HR of rapid kidney function declinea associated with metabolic
syndrome and individual components
HR of rapid kidney function decline (95% CI)
Variable Exp (B) univariate P value Exp (B) multivariateb P value
Blood pressure ⭌130/85 mm Hg 0.863 (0.616 –1.209) 0.392 0.942 (0.666 –1.332) 0.733
Serum HDL cholesterol level ⬍40 mg/dl in 1.075 (0.805–1.436) 0.625 1.014 (0.747–1.375) 0.931
men or ⬍50 mg/dl in women
Serum triglyceride level ⭌150 mg/dl 1.031 (0.709 –1.420) 0.849 1.195 (0.855–1.671) 0.296
Plasma glucose level ⭌110 mg/dl 1.701 (1.224 –2.363) 0.002 1.583 (1.174 –2.135) 0.003
Waist circumference ⭌90 cm in men and 1.204 (0.907–1.597) 0.199 1.238 (0.912–1.680) 0.171
⭌80 cm in women
Mets, NCEP ATP III 1.263 (1.000 –1.615) 0.049 1.406 (1.036 –1.907) 0.029
Mets, NCEP ATP IIIc 1.385 (1.018 –1.885) 0.038

Downloaded from https://academic.oup.com/jcem/article/97/4/1268/2833449 by guest on 03 April 2022

Mets, NCEP ATP IIId 1.042 (0.802–1.355) 0.757
a
A significant decline in renal function was defined as annual decline in GFR larger than 3 ml/min/1.73 m2.
b
Adjusted for age, gender, hemoglobin, serum albumin, globulin, and uric acid.
c
Metabolic syndrome adjusted for other covariates and proteinuria.
d
Metabolic syndrome further adjusted for body weight, systolic blood pressure, fasting blood glucose, and serum creatinine.

fold increase in odds ratio of CKD as compared with those
without after adjusting for confounding factors. We also
included proteinuria as a confounding factor to exclude its
possible exclusive influence. Mets still had 1.880-fold in-
creased odds ratio of CKD after adjusting proteinuria.
After adjusting individual Mets components as continu-
ous variables, we found Mets still had 1.778-fold (95%
CI, 1.188 to 2.465; P ⫽ 0.004) increased odds ratio of
prevalent CKD.
Association of metabolic syndrome with incident
CKD but not with rapid decline in kidney function
in the follow-up study
The participants who did or did not enter the follow-up
study had similar characteristics, laboratory findings, and
CKD prevalence (Supplemental Table 2). A total of 916
participants had at least two serum creatinine measure-
ments during the mean follow-up period of 3.15 ⫾ 1.21 yr.
The hazard ratio (HR) and 95% CI of Cox proportional
hazards models for rapid kidney function decline are
shown in Table 3. In the adjusted model for rapid decline
in kidney function, persons with the Mets showed 1.406
(95% CI, 1.036 to 1.907) times the risk of those without
Mets. Among the five traits of Mets, increased fasting glu-
cose level showed the strongest effect on rapid decline in
kidney function (HR, 1.583; 95% CI, 1.174 to 2.135).
Other components of Mets were not associated with rapid
decline in kidney function. Furthermore, we added pro-
teinuria as a confounding variable; the odds ratio of rapid
decline in kidney function slightly attenuated from 1.406
to 1.385, yet were still statistically significant. After we
adjusted individual Mets components as continuous vari-
ables, we found Mets no longer predicted rapid decline in
kidney function (HR, 1.042; 95% CI, 0.802 to 1.355),
whereas fasting blood glucose was still an independent
predictor of rapid decline in kidney function (HR, 1.007;
95% CI, 1.003 to 1.012).
During the follow-up period, there were 67 cases
(7.3%) of incident CKD (new onset of proteinuria and
eGFR less than 60 ml/min/1.73 m2). We further analyzed
the association between Mets and incident CKD of the
elderly cohort by Cox proportional hazards models using
binominal regression analysis. As shown in Table 4, per-
sons with Mets showed 2.518 (95% CI, 1.524 to 4.161;
P ⬍ 0.001) times greater risk of incident CKD than those
without Mets after adjusting by confounding variables.
When considering the individual component of Mets, in-
creased fasting glucose level showed the strongest effect on
incident CKD (HR, 2.238; 95% CI, 1.351 to 3.708; P ⫽
0.002). A strong association with incident CKD was also
observed for increased triglyceride level (HR, 1.945; 95%
CI, 1.162 to 3.254; P ⫽ 0.011). Low HDL cholesterol level
was positively associated with risks of CKD, although not
reaching statistical significance (HR, 1.612; 95% CI,
0.973 to 2.669; P ⫽ 0.064). We also adjusted proteinuria
as a confounding factor and found that Mets still had a
2.255-fold (95% CI, 1.344 to 3.783; P ⫽ 0.002) increased
HR of incident CKD. After we further adjusted individual
Mets components as continuous variables, we found that
Mets still had 1.931-fold (95% CI, 1.175 to 3.174; P ⫽
0.009) increased HR of incident CKD.
Associations of insulin resistance with prevalent
CKD and proteinuria but not with incident CKD in
nondiabetic elderly
Figure 1 shows the presence of chronic kidney disease
stratified by tertiles of HOMA-IR. Overall, the prevalence
of CKD in the nondiabetic elderly participants is 23.6%.
J Clin Endocrinol Metab, April 2012, 97(4):1268 –1276 jcem.endojournals.org 1273

TABLE 4. Univariate and multivariate-adjusted HR of new onset of chronic kidney disease associated with metabolic
syndrome and individual components
HR of new onset of chronic kidney disease (95% CI)
Variable Exp (B) univariate P value Exp (B) multivariatea P value
Blood pressure ⭌130/85 mm Hg 0.847 (0.488 –1.469) 0.554 0.882 (0.507–1.536) 0.657
Serum HDL cholesterol level ⬍40 mg/dl) 1.379 (0.845–2.252) 0.199 1.612 (0.973–2.669) 0.064
in men or ⬍50 mg/dl) in women
Serum triglyceride level ⭌150 mg/dl 1.686 (1.026 –2.766) 0.039 1.945 (1.162–3.254) 0.011
Plasma glucose level ⭌110 mg/dl 2.289 (1.397–3.752) 0.001 2.238 (1.351–3.708) 0.002
Waist circumference ⭌90 cm in men and 1.204 (0.744 –1.950) 0.4500 1.573 (0.943–2.626) 0.083
⭌80 cm in women
Mets, NCEP ATP III 1.974 (1.221–3.191) 0.006 2.518 (1.524 – 4.161) ⬍0.001
Mets, NCEP ATP IIIb 2.255 (1.344 –3.783) 0.002

Downloaded from https://academic.oup.com/jcem/article/97/4/1268/2833449 by guest on 03 April 2022

Mets, NCEP ATP IIIc 1.931 (1.175–3.174) 0.009
a
Adjusted for age, gender, hemoglobin, serum albumin, globulin, and uric acid.
b
Metabolic syndrome adjusted for other covariates and proteinuria.
c
Metabolic syndrome further adjusted for body weight, systolic blood pressure, fasting blood glucose, and serum creatinine.

The highest tertile of HOMA-IR has a higher prevalence
of CKD compared with the lowest tertile (35.9 vs. 12.9%;
P ⬍ 0.05). Supplemental Tables 3 and 4 show the asso-
ciation of HOMA-IR with the crude and multivariate-
adjusted odds ratios of CKD and proteinuria. In the mul-
tivariate models, HOMA-IR, age, WBC counts, and lower
eGFR were all significantly associated with an increased
odds ratio of prevalent CKD (P ⬍ 0.05); HOMA-IR, hy-
pertension, high-sensitivity C-reactive protein, lower
eGFR, and lower total cholesterol level were significantly
associated with an increased odds ratio of proteinuria (P ⬍
0.05). Elderly people with each one-unit increase of
HOMA-IR had a 1.312-fold (95% CI, 1.114 to 1.545)
and 1.278-fold (95% CI, 1.098 to1.488) increased odds
ratio of prevalence CKD and proteinuria, respectively
(Supplemental Tables 3 and 4). Supplemental Table 5
shows that insulin resistance was not associated with in-
cident CKD.
FIG. 1. Prevalence of chronic kidney disease in the nondiabetic
geriatric Chinese by tertiles of HOMA-IR. In the highest tertile of
HOMA-IR, the prevalence of chronic kidney disease was 35.9%, which
is much higher than the first tertile (12.9%) in nondiabetic elderly
patients.
Association of insulin resistance with rapid decline
in kidney function in nondiabetic elderly
A total of 363 nondiabetic participants had at least two
serum creatinine measurements during the follow-up pe-
riod. In the adjusted model shown in Table 5, increment of
insulin resistance per unit was associated with 1.16-fold
increase in the HR (95% CI, 1.06 to 1.26; P ⬍ 0.01) of the
decline in renal function; and the correlation coefficient
between the annual eGRF decline and the HOMA-IR was
0.528 (P ⬍ 0.05).
Discussion
In this study, we found that Mets was significantly and
independently associated with not only risk of prevalent
CKD, as many have shown (2, 23, 24), but also risk of
incident CKD among subjects age 65 or older. High blood
pressure, low HDL cholesterol level, high triglyceride
level, and high fasting plasma glucose level were the com-
ponents that increased the risk of prevalent CKD. How-
ever, only increased fasting plasma glucose level and se-
rum triglyceride level were associated with incident CKD;
and only high fasting plasma glucose level was associated
with rapid decline in renal function. We further demon-
strated that insulin resistance was associated with in-
creased risk of prevalent CKD and faster decline in renal
function in the elderly population.
Our study showed that the metabolic syndrome is a risk
factor for prevalent CKD in the elderly as shown in a younger
population. A previous report by Tanaka et al. (5) from Ja-
pan stated that Mets is a strong and independent risk factor
of CKD only in younger men (less than 60 yr of age). The
exact reason for the inconsistency observed between these
two studies is not clear. Compared with our general popu-
1274 Cheng et al. Metabolic Syndrome and IR Predict Renal Function Decline J Clin Endocrinol Metab, April 2012, 97(4):1268 –1276

TABLE 5. HOMA-IR index acts as an independent predictor of rapid kidney function decline in the nondiabetic
elderly
Model 1 Model 2 Model 3 Model 4
HR of rapid kidney function 1.14 (1.03 to 1.26)a 1.13 (1.03 to 1.25)a 1.16 (1.06 to 1.27)b 1.16 (1.06 to 1.26)b
decline (95% CI)
Correlation coefficient (B) 0.533 (0.136 to 0.930)a 0.529 (0.130 to 0.928)a 0.529 (0.106 to 0.951)a 0.528 (0.104 to 0.952)a
of annual GFR decline
rate (95% CI)
Model 1, Univariate analysis; model 2, model 1 and age; model 3, model 2 and serum albumin, body mass index, high-sensitivity C-reactive
protein, alanine transaminase, uric acid, triglyceride, WBC count; model 4, model 3 and hypertension.
a
P ⬍ 0.05.
b
P ⬍ 0.01.

Downloaded from https://academic.oup.com/jcem/article/97/4/1268/2833449 by guest on 03 April 2022

lation, the Japanese enjoy a longer survival. The phenome-
non itself or its underlying causes, such as their special life-
style, may modify the influence of Mets to the development
of CKD.
Insulin resistance may be the central hub that links met-
abolic syndrome and the deterioration of renal function.
The severity of insulin resistance was found to be higher in
older individuals (25), partly through alterations in fat-
ness (increased obesity and a detrimental pattern of fat
distribution) and fitness (physical inactivity) (26). People
with insulin resistance display impaired glucose tolerance
and hyperinsulinemia (27). It has been hypothesized that
clustering of these risk factors may result in oxidative
stress and endothelial dysfunction (28), which may con-
sequently cause glomerulosclerosis or atherosclerosis-re-
lated kidney damage in an aging population (29). On the
other hand, the observation that elderly patients with
Mets show a higher chance of having CKD may also sug-
gest that Mets and renal dysfunction may result from a
common cause linked to the aging phenomenon.
Insulin resistance may drive the overproduction of very
low-density lipoprotein cholesterol and contribute to hy-
pertriglyceridemia (30). Triglyceride-rich apolipoprotein
B-containing lipoproteins clearly promote the progression
of renal insufficiency (31). High triglyceride levels are a
risk factor for developing proteinuria (32). Thus, early
detection of metabolic syndrome or high triglyceride levels
might be beneficial if accompanied by early intervention
such as fibrate to lower triglyceride levels and suppress the
pathways for renal injury. However, studies showed a
controversial role of hypertriglyceridemia in CKD.
Hadjadj et al. (33) found that patients with hypertriglyc-
eridemia had a relative risk of 2.01 for nephropathy. Lee
et al. (34) found a higher risk of CKD in patients with
triglyceride levels greater than 200 mg/dl. However, the
Helsinki Heart Study, a randomized clinical trial of gem-
fibrozil in preventing coronary heart disease, found no
association between triglyceride levels and CKD (35).
It is intriguing that the incident CKD in our study pop-
ulation is not associated with insulin resistance. One ex-
planation is that insulin resistance alone may cause a de-
cline in renal function, but it is apparently not strong
enough to produce full-blown kidney disease. Combina-
tion with other factors such as diabetes, high blood pres-
sure, urinary tract obstruction, and/or infection (20) may
be necessary to lead to the development of new CKD.
Elevated blood pressure and visceral obesity were found
in 78.7 and 52.7% of the elder adults in this cohort, respec-
tively (Supplemental Fig. 5), but none was associated with
incident CKD or rapid decline in renal function. This finding
could suggest that the validity of the cutoff is questionable in
these two components, at least in our elderly population.
There were no significant changes between baseline and fol-
low-up in waist circumference and body weight. The possible
influence of body weight changes on the eGFR and renal
outcome might be excluded. However, there is a significantly
negative correlation between eGFR, numbers of metabolic
syndrome criteria (r ⫽ ⫺0.083; P ⫽ 0.002), and waist cir-
cumference (r ⫽ ⫺0.83; P ⫽ 0.001).
There are several limitations in this study that deserve
mention. First, differences between participants and non-
participants may have resulted in selection bias, especially
if selection factors were associated with the exposure sta-
tus. Although only 916 of 1456 (62.9%) participants re-
ceived an additional health exam during the follow-up
period, we believe that the selection bias was minimal be-
cause the baseline characteristics of the nonparticipants
were equal to followed-up participants (Supplemental Ta-
ble 2). Another selection bias may have occurred because
those who undergo regular health exams are relatively
healthy individuals. Second, measurement of proteinuria
and eGFR showed a high degree of intraindividual vari-
ability. Repeated measurements to accurately assess kid-
ney function would likely decrease the prevalence of pro-
teinuria. Similarly, a single serum creatinine value to
estimate kidney function might lead to the misclassifica-
J Clin Endocrinol Metab, April 2012, 97(4):1268 –1276
tion of CKD. Due to day-to-day variation in serum cre-
atinine levels, it is likely to bias the association toward
zero. Third, GFR was not directly measured but was es-
timated by a serum creatinine-based equation, which
might have overestimated or underestimated the actual
GFR in the Chinese elderly population (36). Also, the ATP
III definition of metabolic syndrome might not be suitable
for a Chinese elderly population. Using the International
Diabetes Federation definition, however, we also docu-
mented a positive and significant association between the
metabolic syndrome and risk of CKD in the present study.
Fourth, use of the Cox proportional hazard model re-
quires the knowledge of the exact time of events. The mod-
eling in the follow-up study actually used the time when
the event had already occurred (when a second creatinine
data were available). The result may underestimate the
association between insulin resistance and the risk of in-
cident CKD events.
In summary, our study shows that elderly individuals
with Mets have increased risk of not only prevalent CKD
but also incident CKD. High insulin resistance is associ-
ated with prevalent CKD and rapid decline in renal func-
tion. Further studies are needed to assess the impact of
treating metabolic syndrome and insulin resistance on re-
nal outcomes in elderly.
Acknowledgments
The authors thank Ms. Chen-Chih Chang for data collection. We
also thank the elderly health exam team and the staff in the
Second Core Lab, Department of Medical Research of National
Taiwan University Hospital, for technical support.
Address all correspondence and requests for reprints to:
Chung-Jen Yen, M.D., and Chih-Kang Chiang, M.D., Ph.D.,
Department of Internal Medicine, National Taiwan University
Hospital and National Taiwan University College of Medicine,
7, Chung Shan South Road, Taipei 100, Taiwan. E-mail:
ycjycj@ntuh.gov.tw and ckchiang@ntu.edu.tw.
This study was partially supported by National Taiwan Uni-
versity Hospital Grants NTUH-92-S027 and NTUH-95N02, the
Ta-Tung Kidney Foundation, and the Mrs. Hsiu-Chin Lee Kid-
ney Research Fund, Taipei.
Disclosure Summary: The authors have no commercial asso-
ciations or sources of support that might pose a conflict of in-
terest. All authors have made substantive contributions to the
study, and all authors endorse the data and conclusions.
References
1. Ford ES, Giles WH, Dietz WH 2002 Prevalence of the metabolic
syndrome among US adults: findings from the third National Health
and Nutrition Examination Survey. JAMA 287:356 –359
2. Kitiyakara C, Yamwong S, Cheepudomwit S, Domrongkitchaiporn
jcem.endojournals.org 1275
S, Unkurapinun N, Pakpeankitvatana V, Sritara P 2007 The meta-
bolic syndrome and chronic kidney disease in a Southeast Asian
cohort. Kidney Int 71:693–700
3. Kurella M, Lo JC, Chertow GM 2005 Metabolic syndrome and the
risk for chronic kidney disease among nondiabetic adults. J Am Soc
Nephrol 16:2134 –2140
4. Ninomiya T, Kiyohara Y, Kubo M, Yonemoto K, Tanizaki Y, Doi
Y, Hirakata H, Iida M 2006 Metabolic syndrome and CKD in a
general Japanese population: the Hisayama study. Am J Kidney Dis
48:383–391
5. Tanaka H, Shiohira Y, Uezu Y, Higa A, Iseki K 2006 Metabolic
syndrome and chronic kidney disease in Okinawa, Japan. Kidney Int
69:369 –374
6. Haas JT, Biddinger SB 2009 Dissecting the role of insulin resistance
in the metabolic syndrome. Curr Opin Lipidol 20:206 –210
Downloaded from https://academic.oup.com/jcem/article/97/4/1268/2833449 by guest on 03 April 2022
7. Huang PL 2009 A comprehensive definition for metabolic syn-
drome. Dis Model Mech 2:231–237
8. Chen J, Muntner P, Hamm LL, Fonseca V, Batuman V, Whelton PK,
He J 2003 Insulin resistance and risk of chronic kidney disease in
nondiabetic US adults. J Am Soc Nephrol 14:469 – 477
9. Landau M, Kurella-Tamura M, Shlipak MG, Kanaya A, Strotmeyer
E, Koster A, Satterfield S, Simsonick EM, Goodpaster B, Newman
AB, Fried LF 2011 Correlates of insulin resistance in older individ-
uals with and without kidney disease. Nephrol Dial Transplant 26:
2814 –2819
10. Reaven G 2004 The metabolic syndrome or the insulin resistance
syndrome? Different names, different concepts, and different goals.
Endocrinol Metab Clin North Am 33:283–303
11. El-Atat FA, Stas SN, McFarlane SI, Sowers JR 2004 The relationship
between hyperinsulinemia, hypertension and progressive renal dis-
ease. J Am Soc Nephrol 15:2816 –2827
12. Kobayashi H, Tokudome G, Hara Y, Sugano N, Endo S, Suetsugu
Y, Kuriyama S, Hosoya T 2009 Insulin resistance is a risk factor for
the progression of chronic kidney disease. Clin Nephrol 71:643–
651
13. Bastürk T, Unsal A 2011 Is insulin resistance a risk factor for the
progression of chronic kidney disease? Kidney Blood Press Res 34:
111–115
14. Lea J, Cheek D, Thornley-Brown D, Appel L, Agodoa L, Contreras
G, Gassman J, Lash J, Miller 3rd ER, Randall O, Wang X, McClellan
W 2008 Metabolic syndrome, proteinuria, and the risk of progres-
sive CKD in hypertensive African Americans. Am J Kidney Dis 51:
732–740
15. Lee CC, Sun CY, Wu IW, Wang SY, Wu MS 2011 Metabolic syn-
drome loses its predictive power in late-stage chronic kidney disease
progression—a paradoxical phenomenon. Clin Nephrol 75:141–
149
16. 2001 Executive Summary of The Third Report of The National
Cholesterol Education Program (NCEP) Expert Panel on Detection,
Evaluation, and Treatment of High Blood Cholesterol in Adults
(Adult Treatment Panel III). JAMA 285:2486 –2497
17. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF,
Turner RC 1985 Homeostasis model assessment: insulin resistance
and ␤-cell function from fasting plasma glucose and insulin concen-
trations in man. Diabetologia 28:412– 419
18. Coresh J, Selvin E, Stevens LA, Manzi J, Kusek JW, Eggers P, Van
Lente F, Levey AS 2007 Prevalence of chronic kidney disease in the
United States. JAMA 298:2038 –2047
19. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D 1999 A
more accurate method to estimate glomerular filtration rate from
serum creatinine: a new prediction equation. Modification of Diet in
Renal Disease Study Group. Ann Intern Med 130:461– 470
20. Levey AS, Coresh J, Balk E, Kausz AT, Levin A, Steffes MW, Hogg
RJ, Perrone RD, Lau J, Eknoyan G 2003 National Kidney Founda-
tion practice guidelines for chronic kidney disease: evaluation, clas-
sification, and stratification. Ann Intern Med 139:137–147
21. Klein R, Klein BE, Moss SE, Cruickshanks KJ, Brazy PC 1999 The
1276 Cheng et al. Metabolic Syndrome and IR Predict Renal Function Decline
10-year incidence of renal insufficiency in people with type 1 dia-
betes. Diabetes Care 22:743–751
22. Fried L, Solomon C, Shlipak M, Seliger S, Stehman-Breen C, Bleyer
AJ, Chaves P, Furberg C, Kuller L, Newman A 2004 Inflammatory
and prothrombotic markers and the progression of renal disease in
elderly individuals. J Am Soc Nephrol 15:3184 –3191
23. Whaley-Connell A, Pavey BS, McCullough PA, Saab G, Li S,
McFarlane SI, Chen SC, Vassalotti JA, Collins AJ, Bakris G, Sowers
JR 2010 Dysglycemia predicts cardiovascular and kidney disease in
the Kidney Early Evaluation Program. J Clin Hypertens (Greenwich)
12:51–58
24. Chen J, Muntner P, Hamm LL, Jones DW, Batuman V, Fonseca V,
Whelton PK, He J 2004 The metabolic syndrome and chronic kidney
disease in U.S. adults. Ann Intern Med 140:167–174
25. Rowe JW, Minaker KL, Pallotta JA, Flier JS 1983 Characterization
of the insulin resistance of aging. J Clin Invest 71:1581–1587
26. Muller DC, Elahi D, Tobin JD, Andres R 1996 The effect of age on
insulin resistance and secretion: a review. Semin Nephrol 16:289 –
298
27. Gerich JE 1997 Metabolic abnormalities in impaired glucose toler-
ance. Metabolism 46:40 – 43
28. Egan BM, Greene EL, Goodfriend TL 2001 Insulin resistance and
cardiovascular disease. Am J Hypertens 14:116S–125S
29. Hostetter TH, Rennke HG, Brenner BM 1982 The case for intrare-
nal hypertension in the initiation and progression of diabetic and
other glomerulopathies. Am J Med 72:375–380
30. Tsimihodimos V, Dounousi E, Siamopoulos KC 2008 Dyslipidemia
Members receive an email digest
of endocrinology-related news selected from thousands of sources.
J Clin Endocrinol Metab, April 2012, 97(4):1268 –1276
in chronic kidney disease: an approach to pathogenesis and treat-
ment. Am J Nephrol 28:958 –973
31. Samuelsson O, Attman PO, Knight-Gibson C, Larsson R, Mulec H,
Weiss L, Alaupovic P 1998 Complex apolipoprotein B-containing
lipoprotein particles are associated with a higher rate of progression
of human chronic renal insufficiency. J Am Soc Nephrol 9:1482–
1488
32. Tozawa M, Iseki K, Iseki C, Oshiro S, Ikemiya Y, Takishita S 2002
Triglyceride, but not total cholesterol or low-density lipoprotein
cholesterol levels, predict development of proteinuria. Kidney Int
62:1743–1749
33. Hadjadj S, Duly-Bouhanick B, Bekherraz A, BrIdoux F, Gallois Y,
Mauco G, Ebran J, Marre M 2004 Serum triglycerides are a pre-
dictive factor for the development and the progression of renal and
retinal complications in patients with type 1 diabetes. Diabetes
Metab 30:43–51
Downloaded from https://academic.oup.com/jcem/article/97/4/1268/2833449 by guest on 03 April 2022
34. Lee PH, Chang HY, Tung CW, Hsu YC, Lei CC, Chang HH, Yang
HF, Lu LC, Jong MC, Chen CY, Fang KY, Chao YS, Shih YH, Lin
CL 2009 Hypertriglyceridemia: an independent risk factor of
chronic kidney disease in Taiwanese adults. Am J Med Sci 338:185–
189
35. Manninen V, Tenkanen L, Koskinen P, Huttunen JK, Mänttäri M,
Heinonen OP, Frick MH 1992 Joint effects of serum triglyceride and
LDL cholesterol and HDL cholesterol concentrations on coronary
heart disease risk in the Helsinki Heart Study. Implications for treat-
ment. Circulation 85:37– 45
36. Zuo L, Ma YC, Zhou YH, Wang M, Xu GB, Wang HY 2005 Ap-
plication of GFR-estimating equations in Chinese patients with
chronic kidney disease. Am J Kidney Dis 45:463– 472