Diabetologia

, Volume 54, Issue 4, pp 767–775

Serum C-reactive protein level and prediabetes in two Asian populations

  • C. Sabanayagam
  • A. Shankar
  • S. C. Lim
  • J. Lee
  • E. S. Tai
  • T. Y. Wong
Article

Abstract

Aims/Hypothesis

Prediabetes, an early stage in the hyperglycaemic continuum, increases the future risk of developing diabetes and cardiovascular disease (CVD). C-reactive protein (CRP), a marker of inflammation, is associated with diabetes and CVD. However, studies examining the association between CRP and prediabetes among participants without diabetes are limited.

Methods

We analysed data from two large population-based studies in Singapore: the Singapore Prospective Study Programme (SP2, n = 4,252 Chinese, Malay and Indians aged ≥24 years) and the Singapore Malay Eye Study (SiMES, n = 2,337 Malays aged 40–80 years), participants of which were free of diabetes mellitus. Prediabetes was defined as glycated haemoglobin of 5.7–6.4% in SiMES (n = 1,231); fasting plasma glucose of 5.6–6.9 mmol/l in SP2 (n = 386).

Results

Elevated high sensitivity CRP (hsCRP) levels were found to be associated with prediabetes after adjusting for age, sex, race–ethnicity, education, smoking, alcohol consumption, hypertension, BMI and total cholesterol. Comparing those with hsCRP <1 mg/l (referent), the OR (95% confidence interval) of prediabetes in persons with hsCRP 1–3 mg/l and >3 mg/l was 1.31 (0.99–1.74) and 2.17 (1.61–2.92), ptrend < 0.0001 in SP2; 1.23 (1.00–1.52) and 1.31 (1.06–1.64), ptrend = 0.02 in SiMES. In subgroup analysis, the association was stronger in women, Chinese and Malays, and participants with BMI <25 kg/m2.

Conclusions

Data from two population-based Asian cohorts suggest that elevated serum hsCRP levels are associated with prediabetes.

Keywords

Asia BMI C-reactive protein Cross-sectional Inflammation Multi-ethnicity Population-based Prediabetes Sex Singapore 

Abbreviations

CRP

C-reactive protein

CVD

Cardiovascular disease

hsCRP

High-sensitivity C-reactive protein

SiMES

Singapore Malay Eye Study

SP2

Singapore Prospective Study Programme

Introduction

Diabetes mellitus is a growing public health problem worldwide [1] associated with substantial cardiovascular morbidity and mortality [2, 3]. Inflammation has been suggested to play a central role in the pathogenesis of diabetes and atherosclerosis [4, 5]. Prospective and cross-sectional studies have shown that C-reactive protein (CRP), a specific marker of inflammation, is associated with increased risk of diabetes [6, 7, 8] and cardiovascular disease (CVD) [9, 10]. Recent studies have also shown that CRP is associated with fasting [11, 12, 13, 14] or post-load glucose [12, 15] or glycated haemoglobin levels [14]. According to recent ADA guidelines [16], prediabetes is an earlier stage in the hyperglycaemic continuum that is associated with increased future risk of developing diabetes and CVD [17, 18]. However, studies examining the association between CRP and prediabetes among participants without diabetes are limited [12, 19, 20]. Few studies have shown an association between CRP and prediabetes among specific population groups, including older black and white participants in the USA [12], middle-aged Japanese [19] and clinical patients in China [20]. The prevalence of diabetes is increasing in epidemic proportions among Asians [1, 16] and it has also been shown that Asians have lower levels of CRP than Western populations [21, 22]. In this context, we examined the association between CRP and prediabetes in a multi-ethnic Asian population of Chinese, Malay and Indian participants from two population-based cohorts, the Singapore Prospective Study Programme (SP2) and Singapore Malay Eye Study (SiMES) in Singapore.

Methods

Details of the study design and methods of SP2 and SiMES have been published elsewhere [23, 24]. SP2 is a population-based cross-sectional study of participants aged 24–95 years living in Singapore. Of the 7,742 eligible participants, 5,157 attended the health examination. Of the 4,834 participants with CRP and fasting glucose measurements—after excluding those with diabetes (n = 568), ethnic groups other than Chinese, Malay and Indians (n = 5) and those with missing information on variables included in the multivariable model including education, smoking, drinking status, systolic and diastolic BP, and total cholesterol (n = 9)—4,252 were included in the final analysis.

SiMES is a population-based cross-sectional study of Malay adults aged 40–80 years living in Singapore [24]. Of the 4,168 eligible participants, 3,280 participated in the study. Of the 3,136 participants with CRP and HbA1c measurements, after excluding those with diabetes (n = 717), and those with missing information on variables included in the multivariable model (n = 82), 2,337 were included in the final analysis.

Participants from both SP2 and SiMES cohorts were examined in the same study clinic (Singapore Eye Research Institute). Questionnaire and clinic examinations were similar in both studies except that blood samples were collected in non-fasting state in SiMES and fasting state in SP2. Informed consent was obtained from all study participants, and ethics approval was obtained from the institutional review boards of National University of Singapore and Singapore General Hospital for SP2 and from the institutional review board of Singapore Eye Research Institute for SiMES.

Main outcome of interest: prediabetes

In SP2, venous blood was drawn from all participants in the morning following a 10 h overnight fast. Fasting plasma glucose was measured using enzymatic methods (ADVIA 2400, Bayer Diagnostics) at the National University Hospital Reference Laboratory, which is accredited by the College of American Pathologists. In SiMES, fasting plasma glucose was not measured. All participants had venous blood drawn in the non-fasting state. Serum HbA1c assay was carried out using HPLC cation exchange chromatography system implemented on a Biorad variant II analyser at the National University Hospital Reference Laboratory. The assay was accredited by the National Glycoprotein Standardisation Program with controls traceable to Diabetes Control and Complications Trial. Prediabetes was defined as fasting plasma glucose of 5.6–6.9 mmol/l (110–125 mg/dl) in SP2 and HbA1c value of 5.7–6.4% in SiMES following the recommendations of the ADA to include HbA1c in identifying individuals at increased risk for future diabetes [16].

Measurement of exposure variables

Exposure measurements were similar in both studies. Information on participants’ demographic characteristics, lifestyle factors and medical history were obtained using a standardised questionnaire. Age was defined as the age at the time of clinical examination. Education was categorised into: (1) primary or lower (≤6 years); (2) secondary (7–10 years); and (3) post-secondary (≥11 years, including university education). Cigarette smoking was categorised into current smokers and former or non-smokers, and alcohol consumption into ever drinkers and non-drinkers. BMI, calculated as weight in kilograms divided by the square of height in metres (kg/m2), was categorised into <25 and ≥25 kg/m2. BP was measured twice and the average of the two systolic and diastolic BP measurements were taken as the systolic and diastolic BP value. Mean arterial BP was calculated as two-thirds of the diastolic plus one-third of the systolic value. Hypertension was defined as systolic BP ≥140 mmHg or diastolic BP ≥90 mmHg, or self-reported physician-diagnosed hypertension. Diabetes mellitus was defined as fasting plasma glucose ≥7 mmol/l (≥126 mg/dl) in SP2 and non-fasting plasma glucose ≥11.1 mmol/l (≥200 mg/dl) in SiMES, or self-reported physician-diagnosed diabetes or use of glucose-lowering medication in either SP2 or SiMES. For the current analysis we excluded those with diabetes (n = 568 in SP2 and n = 717 in SiMES). CVD was defined as self-reported myocardial infarction, stroke or angina. High-sensitivity CRP (hsCRP) was assayed on a Roche/Integra 400 Analyser (Roche Diagnostics, Rotkreuz, Switzerland) by a particle-enhanced immunoturbidimetric method. The intra- and inter-assay precision were 0.6–1.3% and 2.3–3.1%, respectively. Serum lipids and hsCRP were measured from fasting blood samples in SP2 and from non-fasting blood samples in SiMES. All serum biochemistry tests were carried out at the National University Hospital Reference Laboratory.

Statistical analysis

As prediabetes was defined differently in the two studies owing to differences in data available, we performed all analyses separately for the two study cohorts. hsCRP levels were analysed both as categorical and continuous variables. For the categorical analysis, hsCRP levels were grouped into categories (<1, 1–3, and >3 mg/l) based on American Heart Association recommendations [16], and also as quartiles. For the continuous analysis hsCRP was log transformed owing to its skewed distribution. We compared selected characteristics of the study population by race–ethnicity and by prediabetes status using analysis of variance or χ2 tests as appropriate. We then examined the association between hsCRP and prediabetes in two logistic regression models. In the first model, we adjusted for age (years) and sex (female, male). In the second multivariable model, we additionally adjusted for ethnicity (Chinese, Malay, Indian), education (primary/below, secondary/higher), current smoking (absent, present), alcohol consumption (absent, present), hypertension status (absent, present), BMI (kg/m2) and total cholesterol (mmol/l). Tests for trend were performed modelling hsCRP categories as an ordinal variable in the corresponding multivariable logistic regression models. To examine the consistency of the association between hsCRP categories and prediabetes, we performed subgroup analysis stratified by sex, race–ethnicity and BMI (<25, ≥25 kg/m2). Statistical interaction between hsCRP categories and each of the stratifying variables was examined in the corresponding logistic regression model by including cross-product interaction terms.

To examine the dose–response relationship of the observed association between serum hsCRP level and prediabetes without linearity assumptions, we used flexible nonparametric logistic regression employing the generalised additive modelling approach (R system for statistical computing, available from Comprehensive R Archive Network [www.CRAN.R-project.org]) to calculate odds of prediabetes, adjusting for all covariates in the multivariable model; the predicted odds of prediabetes were then plotted against increasing hsCRP levels (on the log scale).

We performed several supplementary analyses. We repeated the multivariable model in Table 2: (1) using quartiles of CRP (<0.4, 0.4–1.0, 1.1–2.6, >2.6 mg/l); (2) after excluding participants with hsCRP levels suggestive of clinical inflammation (>10 mg/l); and (3) after excluding participants with history of CVD. As the prevalence of prediabetes in these populations was not low (9.1% in SP2 and 52.7% in SiMES), we repeated the analysis using log-binomial regression, an alternative model to logistic regression in cross-sectional studies [25]. All analyses were performed using SAS version 9.1 (SAS Institute Inc., Cary, NC, USA).

Results

Table 1 shows characteristics of the study population by study cohort and ethnicity. Among the three ethnic groups, Malays in both SP2 and SiMES were more likely to be primary/below educated, current smokers, have higher systolic and diastolic BP, BMI and total cholesterol, and less likely to consume alcohol; Chinese were more likely to be females, and Indians had higher levels of CRP and lower levels of HDL cholesterol. In general, participants from the SiMES cohort were older and had a higher prevalence of all risk factors examined, compared with those from the SP2 cohort. Table 2 shows characteristics of the study cohorts by prediabetes status. The prevalence of prediabetes was 9.1% in SP2 cohort and 52.7% in SiMES cohort. In both the cohorts, compared with those without prediabetes, those with prediabetes were more likely to be older, have higher systolic BP and BMI, and were less likely to be female. Further, in the SP2 cohort, those with prediabetes had higher diastolic BP and CRP, lower levels of HDL cholesterol and were less likely to consume alcohol; in the SiMES cohort, those with prediabetes had higher levels of total cholesterol.
Table 1

Characteristics of the study participants by race/ethnicity and by study cohort

Characteristic

SP2

SiMES

p valuea

Chinese (n = 2,900)

Malay (n = 737)

Indian (n = 615)

Malay (n = 2,337)

Age (years)

48.7 (11.7)

48.3 (11.2)

50.3 (11.1)

57.3 (11.2)

<0.0001

Sex, female (%)

56.5

50.8

52.9

51.2

0.0005

Education, primary/below (%)

23.9

29.4

23.7

71.4

<0.0001

Current smoking (%)

10.4

18.9

10.7

21.9

<0.0001

Alcohol consumption (%)

46.0

6.8

33.3

2.1

<0.0001

Systolic BP (mmHg)

129.4 (20.1)

134.7 (19.3)

131.3 (20.4)

144.7 (23.5)

<0.0001

Diastolic BP (mmHg)

76.8 (10.9)

78.9 (10.3)

78.5 (10.5)

80.0 (11.3)

<0.0001

BMI (kg/m2)

22.7 (3.7)

26.0 (4.8)

25.7 (4.5)

26.1 (5.2)

<0.0001

Total cholesterol (mmol/l)

5.2 (0.9)

5.5 (0.9)

5.2 (0.9)

5.7 (1.1)

<0.0001

HDL-cholesterol (mmol/l)

1.4 (0.4)

1.3 (0.3)

1.2 (0.3)

1.4 (0.3)

<0.0001

Fasting plasma glucose (mmol/l)

4.8 (0.5)

4.8 (0.6)

4.9 (0.6)

<0.0001

HbA1c (%)

5.9 (0.6)

CRP (mg/l)

2.0 (4.9)

3.1 (4.5)

4.4 (8.3)

5.9 (0.6)

<0.0001

Data are mean (SD) unless stated otherwise

aThe p value represents difference in characteristics by ethnicity based on analysis of variance or χ2 test as appropriate

Table 2

Characteristics of the study participants by prediabetes status

Characteristic

SP2 cohort

SiMES cohort

Prediabetes absent (n = 3,866)

Prediabetes present (n = 386)

p valuea

Prediabetes absent (n = 1,106)

Prediabetes present (n = 1,231)

p valuea

Age (years)

48.2 (11.4)

55.9 (10.5)

<0.0001

56.5 (11.5)

58.1 (10.8)

0.001

Sex, female (%)

56.1

43.8

<0.0001

54.7

48.1

0.001

Education, primary/below (%)

23.6

36.5

<0.0001

70.1

72.6

0.2

Current smoking (%)

12.1

9.8

0.2

21.5

22.3

0.7

Alcohol consumption (%)

38.4

27.2

<0.0001

2.6

1.5

0.07

Systolic BP (mmHg)

129.2 (19.4)

144.1 (21.5)

<0.0001

143.5 (24.3)

145.8 (22.6)

0.02

Diastolic BP (mmHg)

76.9 (10.7)

82.4 (10.9)

<0.0001

79.4 (11.5)

80.3 (11.1)

0.05

BMI (kg/m2)

23.5 (4.1)

26.0 (4.8)

<0.0001

25.7 (5.2)

26.5 (5.1)

0.0004

Total cholesterol (mmol/l)

5.2 (0.9)

5.3 (0.9)

0.3

5.6 (1.1)

5.8 (1.1)

<0.0001

HDL-cholesterol (mmol/l)

1.4 (0.4)

1.3 (0.3)

<0.0001

1.4 (0.3)

1.4 (0.3)

0.2

Fasting plasma glucose (mmol/l)

4.7 (0.4)

6.0 (0.4)

<0.0001

HbA1c (%)

<0.0001

5.7 (0.8)

6.0 (0.2)

<0.0001

CRP (mg/l)

2.3 (5.1)

4.2 (9.0)

<0.0001

3.7 (7.6)

4.2 (8.5)

0.1

Data are mean (SD) unless stated otherwise

aThe p value represents difference in characteristics by ethnicity based on analysis of variance or χ2 test as appropriate

Table 3 presents the ORs of prediabetes by categories of hsCRP in both the study cohorts. The prevalence of prediabetes increased with increasing categories of hsCRP in a dose-dependent manner (ptrend < 0.0001 in SP2 and ptrend = 0.02 in SiMES). hsCRP was significantly associated with prediabetes in both the age, sex-adjusted and the multivariable model. This positive association between hsCRP and prediabetes persisted when CRP was analysed as a continuous variable in both the cohorts.
Table 3

Association between CRP and prediabetes

Serum CRP categories (mg/l)

Number at risk (cases)

Prediabetes (%)

Age–sex adjusted OR (95% CI)

Multivariable OR (95% CI)a

SP2 population (n = 4,252)

 <1

1,943 (105)

5.4

1.00 (referent)

1.00 (referent)

 1–3

1,379 (130)

9.4

1.70 (1.30–2.23)

1.31 (0.99–1.74)

 >3

930 (151)

16.2

3.35 (2.56–4.38)

2.17 (1.61–2.92)

 ptrend

  

<0.0001

<0.0001

Log-transformed CRP, mg/l

  

1.48 (1.35–1.62)

1.28 (1.16–1.42)

SiMES population (n = 2,337)

 <1

645 (301)

46.7

1.00 (Referent)

1.00 (Referent)

 1–3

889 (479)

53.9

1.35 (1.10–1.65)

1.23 (1.00–1.52)

 >3

803 (451)

56.2

1.51 (1.22–1.86)

1.31 (1.06–1.64)

 ptrend

  

0.0002

0.02

Log transformed CRP, mg/l

  

1.15 (1.08–1.24)

1.11 (1.03–1.19)

aAdjusted for age (years), sex (female, male), race–ethnicity (Chinese, Malay, Indian) in SP2, education (primary/below, secondary/higher), BMI (kg/m2), current smoker (absent, present), ever drinker (absent, present), hypertension (absent, present), total cholesterol (mmol/l)

Table 4 shows the association between hsCRP and prediabetes in subgroups of sex, race–ethnicity and BMI categories in both the study cohorts. In the SP2 cohort, similar to the main results in Table 3, a positive association was observed between CRP and prediabetes within subgroups of sex, race–ethnicity and BMI categories; however, the association was stronger in females (p for interaction = 0.03); although the association was stronger in Chinese and Malays and among those with BMI <25 kg/m2, there was no significant interaction by race–ethnicity (p for interaction = 0.2) or by BMI (p for interaction = 0.2). In the SiMES cohort, the association was significant in females (p for interaction = 0.02) and those with BMI <25 kg/m2 (p for interaction = 0.02).
Table 4

Association between serum CRP and prediabetes within selected subgroups

Subgroups

CRP categories (mg/l)

<1

1–3

>3

ptrend

n (cases)

Multivariable OR (95% CI) a,b

n (cases)

Multivariable OR (95% CI) a

n (cases)

Multivariable OR (95% CI) a

 

SP2 population

 Sexb

  Male (n = 1,914)

881 (73)

1.00

668 (80)

1.22 (0.86–1.74)

365 (64)

1.72 (1.16–2.55)

0.007

  Female (n = 2,338)

1,062 (32)

1.00

711 (50)

1.51 (0.94–2.45)

565 (87)

2.84 (1.75–4.59)

<0.0001

 Race–ethnicityb

  Chinese (n = 2,900)

1,556 (72)

1.00

898 (70)

1.23 (0.86–1.76)

446 (71)

2.51 (1.72–3.67)

<0.0001

  Malay (n = 737)

229 (18)

1.00

286 (33)

1.40 (0.74–2.65)

222 (37)

2.11 (1.09–4.09)

0.02

  Indian (n = 615)

158 (15)

1.00

195 (27)

1.19 (0.59–2.40)

262 (43)

1.33 (0.66–2.68)

0.4

 BMIb

  <25 kg/m2 (n = 2,875)

1,645 (70)

1.00

855 (63)

1.30 (0.90–1.88)

375 (49)

2.17 (1.43–3.30)

0.0005

  ≥25 kg/m2 (n = 1,377)

298 (35)

1.00

524 (67)

1.01 (0.65–1.59)

555 (102)

1.49 (0.94–2.34)

0.05

SiMES population

 Sexb

       

  Male (n = 1,140)

336 (176)

1.00

452 (265)

1.25 (0.93–1.68)

352 (198)

1.11 (0.82–1.52)

0.5

  Female (n = 1,197)

309 (125)

1.00

437 (214)

1.20 (0.88–1.64)

451 (253)

1.48 (1.07–2.03)

0.02

 BMIb

  <25 kg/m2 (n = 1,041)

424 (174)

1.00

360 (189)

1.45 (1.08–1.96)

257 (145)

1.58 (1.14–2.20)

0.004

  ≥25 kg/m2 (n = 1,296)

221 (127)

1.00

529 (290)

0.89 (0.65–1.23)

546 (306)

0.91 (0.66–1.27)

0.7

aAdjusted for age (years), sex (female, male), race–ethnicity (Chinese, Malay, Indian) in SP2, education (primary/below, secondary/higher), BMI (kg/m2), current smoker (absent, present), ever drinker (absent, present), hypertension (absent, present), total cholesterol (mmol/l)

bp interaction for serum CRP categories and female is equal to 0.03 in SP2 and 0.02 in SiMES; CRP categories and race–ethnicity = 0.2 in SP2; CRP categories and BMI = 0.2 in SP2 and 0.02 in SiMES

Figure 1 shows the dose–response relationship between serum hsCRP levels and prediabetes employing nonparametric models without linearity assumptions. A continuous positive association was observed between serum hsCRP levels and prediabetes in both the study cohorts; however, the dose–response curve was steeper in the SP2 cohort.
Fig. 1

Multivariable-adjusted odds of prediabetes according to serum CRP level (mg/l) in SP2 (a) and SiMES (b) populations. The black line represents the predicted odds of prediabetes from non-parametric logistic regression; the dashed lines represent 95% confidence limits for the nonparametric logistic regression estimates. The non-parametric logistic regression was adjusted for age (years), sex (male, female), race–ethnicity (Chinese, Malay, Indians) in SP2, education (primary/below, secondary/higher), BMI (kg/m2), current smoker (absent, present), ever drinker (absent, present), hypertension (absent, present), total cholesterol (mmol/l)

In a supplementary analysis, when we repeated the main analysis in Table 3 using quartiles of CRP, the results were largely similar; in the SP2 cohort, compared with quartile 1 (referent) of hsCRP, the multivariable OR (95% CI) of prediabetes was 1.15 (0.81–1.62) in quartile 2, 1.21 (0.84–1.74) in quartile 3 and 2.25 (1.59–3.20) in quartile 4; ptrend < 0.0001. In the SiMES cohort, compared with quartile 1 (referent) of hsCRP, the multivariable OR (95% CI) of prediabetes was 1.13 (0.86–1.48) in quartile 2, 1.15 (0.88–1.51) in quartile 3 and 1.39 (1.06–1.82) in quartile 4; ptrend = 0.01. Additional supplementary analyses are shown in Table 5. Excluding those with evidence of clinical inflammation (>10 mg/l) and a history of CVD did not materially alter the results. When we repeated the main analysis in Table 3 using log-binomial regression, the effect estimates (prevalence ratios, PR) were slightly attenuated, but the association still remained positive and significant.
Table 5

Supplementary analyses for association between CRP and prediabetes

Model

Multivariable OR (95% confidence interval) a

ptrend

CRP categories (mg/l)

<1b

1–3

>3

SP2 population

 Using log binomial model instead of logistic regression

1.00

1.28 (0.98–1.67)

1.94 (1.47–2.55)

<0.0001

 Excluding those with clinical inflammation, CRP >10 mg/l (n = 4,084)

1.00

1.27 (0.96–1.68)

1.86 (1.35–2.57)

0.0002

 Excluding those with history of CVD (n = 4,127)

1.00

1.27 (0.95–1.69)

1.86 (1.36–2.55)

0.0001

SiMES population

 Using log binomial model instead of logistic regression

1.00

1.12 (0.97–1.29)

1.21 (1.05–1.40)

0.01

 Excluding those with clinical inflammation, CRP >10 mg/l (n = 2,174)

1.00

1.24 (1.00–1.53)

1.31 (1.04–1.65)

0.03

 Excluding those with history of CVD (n = 2,126)

1.00

1.28 (1.03–1.60)

1.34 (1.07–1.69)

0.02

aAdjusted for age (years), sex (female, male), race–ethnicity (Chinese, Malay, Indian) in SP2, education (primary/below, secondary/higher), BMI (kg/m2), current smoker (absent, present), ever drinker (absent, present), hypertension (absent, present), total cholesterol (mmol/l)

bReferent

Discussion

The current study, analysing data from two large population-based Asian cohorts, showed that higher hsCRP levels were associated with prediabetes among participants without diabetes. This association was found to be independent of age, sex, race–ethnicity, education, current smoking, alcohol consumption, hypertension, BMI, and total cholesterol, and a positive dose–response trend was observed between increasing categories of hsCRP and prediabetes. The positive association between hsCRP and prediabetes was found to be stronger in women, Chinese and Malays, and participants with BMI <25 kg/m2. Finally, despite using two different populations with different ways of measuring glycaemia, the results for the association between hsCRP and prediabetes were similar.

A higher level of hsCRP is a sensitive marker of subclinical inflammation. Our finding of an association between CRP and prediabetes is largely consistent with previous studies that report an association between CRP and diabetes [6, 7, 12, 26] or elevated blood glucose levels [11, 13, 14] or HbA1c [14]. It is also consistent with previous studies that report an association between CRP and prediabetes among specific populations [12, 19, 20, 27] and extends the association to a multi-ethnic Asian population. A positive association between elevated CRP and prediabetes was reported among the Chinese Han population (n = 1,730) in a clinic-based study [20], 3,075 well-functioning black and white participants aged 70–79 years in the Health, Aging and Body Composition Study in the USA [12], 2,127 middle-aged Japanese men and women in the Hisayama Study [19] and among 232 non-smoking elderly Korean women [27].

The mechanism linking elevated CRP levels and prediabetes could be partially explained by its association with insulin resistance [28]. Treatment with insulin-sensitising agents has been shown to lower CRP levels in type 2 diabetic patients [29]. Subclinical inflammation leads to endothelial dysfunction and increased peripheral resistance that promotes insulin resistance further [30, 31].

Our finding of a stronger association between CRP and prediabetes in women is consistent with previous reports [13, 14, 21, 32]. Nakanishi et al. [13] reported a stronger association between CRP and fasting plasma glucose in Japanese women compared with men [13]. Few previous studies found an association between CRP and fasting plasma glucose in women only [14, 21]. In the KORA study in Germany, the positive association between CRP and diabetes disappeared in men after adjustment for BMI, whereas it persisted in women [32]. It could be speculated that the stronger association between CRP and prediabetes in women could be as a result of higher insulin resistance resulting from higher body fat composition [32]. At comparable BMI levels, women tend to have higher body fat than men [33, 34]. Also, endogenous sex hormones modulate glucose metabolism differently in men and women [35]. Lower levels of oestrogens and higher levels of androgens increase the risk of diabetes in women [36, 37].

In the current study, the association between CRP and prediabetes was significantly present among Chinese and Malays. Although the association was in the same direction as other ethnic groups, the association was weaker and not statistically significant in Indians. This is in contrast to previous reports from the UK [38] and India [39, 40] that have reported an association between CRP and diabetes among Indians. The exact reason for this lack of association among Indians is not clear. It may be because of the relatively smaller sample size of the Indian cohort. Further studies with larger sample sizes are warranted to explore the association between CRP and prediabetes among Indians.

We found that the association between CRP and prediabetes was stronger among those with BMI <25 kg/m2. Similar to our report, few previous studies [41, 42] including a Japanese study [32], have shown that the association between CRP and diabetes is stronger among lean individuals, whereas some studies have shown that this association is consistently present among those with normal BMI and overweight individuals [20]. Adipose cells produce proinflammatory cytokines that play a role in the development of insulin resistance and glucose intolerance [43, 44]. Our finding of an association between CRP and prediabetes could be owing to the association between BMI and prediabetes in those with BMI <25 kg/m2. Consistent with our findings, previous studies have reported that Asians have higher visceral adipose tissue and higher per cent of body fat despite a lean BMI resulting in a higher risk for insulin resistance [45, 46].

Our study has some limitations. The cross-sectional nature of the study limits making causal inferences. Strengths of the study include its large sample size, population-based nature and the use of two separate Asian cohorts.

In conclusion, data from two large-population based Asian cohorts show that elevated CRP levels are associated with prediabetes. If confirmed by future prospective studies, CRP evaluation may be a useful tool in clinical practice for assessing future risk of diabetes in Asian populations.

Notes

Acknowledgements

The authors thank the staff and participants in the Singapore Malay Eye Study (SiMES), Singapore Prospective Study Program (SP2) and the Singapore Cardiovascular Cohort Study (SCCS2) for their important contributions. The authors thank S. Kalidindi, Department of Statistics, West Virginia University, for her help with the graphs. SP2/SCCS2 was supported by the Biomedical Research Council Grants No 03/1/27/18/216 and 08/1/35/19/550, National Medical Research Council Grants No 0838/2004 and the Singapore Bio Imaging Consortium (C-011/2006). SiMES was funded by the National Medical Research Council (NMRC) 0796/2003 & the Biomedical Research Council (BMRC) 501/1/25-5, with support from the Singapore Prospective Study Program and the Singapore Tissue Network, A*STAR.

Duality of Interest

The authors declare that there is no duality of interest associated with this manuscript.

References

  1. 1.
    King H, Aubert RE, Herman WH (1998) Global burden of diabetes, 1995–2025: prevalence, numerical estimates, and projections. Diab Care 21:1414–1431Google Scholar
  2. 2.
    Fox CS, Coady S, Sorlie PD et al (2007) Increasing cardiovascular disease burden due to diabetes mellitus: the Framingham Heart Study. Circulation 115:1544–1550PubMedGoogle Scholar
  3. 3.
    Milicevic Z, Raz I, Beattie SD et al (2008) Natural history of cardiovascular disease in patients with diabetes: role of hyperglycemia. Diab Care 31(Suppl 2):S155–S160Google Scholar
  4. 4.
    Willerson JT, Ridker PM (2004) Inflammation as a cardiovascular risk factor. Circulation 109: II2–10Google Scholar
  5. 5.
    Ross R (1999) Atherosclerosis—an inflammatory disease. N Engl J Med 340:115–126PubMedGoogle Scholar
  6. 6.
    Chase HP, Cooper S, Osberg I et al (2004) Elevated C-reactive protein levels in the development of type 1 diabetes. Diabetes 53:2569–2573PubMedGoogle Scholar
  7. 7.
    Doi Y, Kiyohara Y, Kubo M et al (2005) Elevated C-reactive protein is a predictor of the development of diabetes in a general Japanese population: the Hisayama Study. Diabetes Care 28:2497–2500PubMedGoogle Scholar
  8. 8.
    Hu G, Jousilahti P, Tuomilehto J, Antikainen R, Sundvall J, Salomaa V (2009) Association of serum C-reactive protein level with sex-specific type 2 diabetes risk: a prospective Finnish study. J Clin Endocrinol Metab 94:2099–2105PubMedGoogle Scholar
  9. 9.
    Buckley DI, Fu R, Freeman M, Rogers K, Helfand M (2009) C-reactive protein as a risk factor for coronary heart disease: a systematic review and meta-analyses for the US Preventive Services Task Force. Ann Intern Med 151:483–495PubMedGoogle Scholar
  10. 10.
    Wilson PW, Pencina M, Jacques P, Selhub J, D'Agostino R Sr, O'Donnell CJ (2008) C-reactive protein and reclassification of cardiovascular risk in the Framingham Heart Study. Circ Cardiovasc Qual Outcomes 1:92–97PubMedGoogle Scholar
  11. 11.
    Aronson D, Bartha P, Zinder O et al (2004) Association between fasting glucose and C-reactive protein in middle-aged subjects. Diabet Med 21:39–44PubMedGoogle Scholar
  12. 12.
    de RN P, Peila R, Ding J et al (2006) Diabetes, hyperglycemia, and inflammation in older individuals: the health, aging and body composition study. Diabetes Care 29:1902–1908Google Scholar
  13. 13.
    Nakanishi N, Shiraishi T, Wada M (2005) Association between fasting glucose and C-reactive protein in a Japanese population: the Minoh study. Diabetes Res Clin Pract 69:88–98PubMedGoogle Scholar
  14. 14.
    Wu T, Dorn JP, Donahue RP, Sempos CT, Trevisan M (2002) Associations of serum C-reactive protein with fasting insulin, glucose, and glycosylated hemoglobin: the Third National Health and Nutrition Examination Survey, 1988–1994. Am J Epidemiol 155:65–71PubMedGoogle Scholar
  15. 15.
    Festa A, D'Agostino R Jr, Tracy RP, Haffner SM (2002) C-reactive protein is more strongly related to post-glucose load glucose than to fasting glucose in non-diabetic subjects; the Insulin Resistance Atherosclerosis Study. Diabet Med 19:939–943PubMedGoogle Scholar
  16. 16.
    Anonymous (2010) Standards of medical care in diabetes–2010. Diabetes Care 33(Suppl 1):S11–S61Google Scholar
  17. 17.
    Ford ES, Zhao G, Li C (2010) Pre-diabetes and the risk for cardiovascular disease: a systematic review of the evidence. J Am Coll Cardiol 55:1310–1317PubMedGoogle Scholar
  18. 18.
    Haffner SM (2003) Pre-diabetes, insulin resistance, inflammation and CVD risk. Diabetes Res Clin Pract 61(Suppl 1):S9–S18PubMedGoogle Scholar
  19. 19.
    Doi Y, Kiyohara Y, Kubo M et al (2005) Relationship between C-reactive protein and glucose levels in community-dwelling subjects without diabetes: the Hisayama Study. Diabetes Care 28:1211–1213PubMedGoogle Scholar
  20. 20.
    Lin J, Zhang M, Song F et al (2009) Association between C-reactive protein and pre-diabetic status in a Chinese Han clinical population. Diabetes Metab Res Rev 25:219–223PubMedGoogle Scholar
  21. 21.
    Yamada S, Gotoh T, Nakashima Y et al (2001) Distribution of serum C-reactive protein and its association with atherosclerotic risk factors in a Japanese population: Jichi Medical School cohort study. Am J Epidemiol 153:1183–1190PubMedGoogle Scholar
  22. 22.
    Kao PC, Shiesh SC, Wu TJ (2006) Serum C-reactive protein as a marker for wellness assessment. Ann Clin Lab Sci 36:163–169PubMedGoogle Scholar
  23. 23.
    Sabanayagam C, Tai ES, Shankar A, Lee J, Sun C, Wong TY (2009) Retinal arteriolar narrowing increases the likelihood of chronic kidney disease in hypertension. J Hypertens 27:2209–2217PubMedGoogle Scholar
  24. 24.
    Foong AW, Saw SM, Loo JL et al (2007) Rationale and methodology for a population-based study of eye diseases in Malay people: the Singapore Malay eye study (SiMES). Ophthalmic Epidemiol 14:25–35PubMedGoogle Scholar
  25. 25.
    Barros AJ, Hirakata VN (2003) Alternatives for logistic regression in cross-sectional studies: an empirical comparison of models that directly estimate the prevalence ratio. BMC Med Res Methodol 3:21PubMedGoogle Scholar
  26. 26.
    Barzilay JI, Abraham L, Heckbert SR et al (2001) The relation of markers of inflammation to the development of glucose disorders in the elderly: the Cardiovascular Health Study. Diabetes 50:2384–2389PubMedGoogle Scholar
  27. 27.
    Choi KM, Lee J, Lee KW et al (2004) Comparison of serum concentrations of C-reactive protein, TNF-alpha, and interleukin 6 between elderly Korean women with normal and impaired glucose tolerance. Diabetes Res Clin Pract 64:99–106PubMedGoogle Scholar
  28. 28.
    Festa A, Hanley AJ, Tracy RP, D'Agostino R Jr, Haffner SM (2003) Inflammation in the prediabetic state is related to increased insulin resistance rather than decreased insulin secretion. Circulation 108:1822–1830PubMedGoogle Scholar
  29. 29.
    Yatagai T, Nakamura T, Nagasaka S et al (2004) Decrease in serum C-reactive protein levels by troglitazone is associated with pretreatment insulin resistance, but independent of its effect on glycemia, in type 2 diabetic subjects. Diabetes Res Clin Pract 63:19–26PubMedGoogle Scholar
  30. 30.
    Goldberg RB (2009) Cytokine and cytokine-like inflammation markers, endothelial dysfunction, and imbalanced coagulation in development of diabetes and its complications. J Clin Endocrinol Metab 94:3171–3182PubMedGoogle Scholar
  31. 31.
    Caballero AE (2004) Endothelial dysfunction, inflammation, and insulin resistance: a focus on subjects at risk for type 2 diabetes. Curr Diab Rep 4:237–246PubMedGoogle Scholar
  32. 32.
    Thorand B, Baumert J, Kolb H et al (2007) Sex differences in the prediction of type 2 diabetes by inflammatory markers: results from the MONICA/KORA Augsburg case-cohort study, 1984–2002. Diab Care 30:854–860Google Scholar
  33. 33.
    Gallagher D, Visser M, Sepulveda D, Pierson RN, Harris T, Heymsfield SB (1996) How useful is body mass index for comparison of body fatness across age, sex, and ethnic groups? Am J Epidemiol 143:228–239PubMedGoogle Scholar
  34. 34.
    Durenberg-Yap M, Chew SK, Deurenberg P (2002) Elevated body fat percentage and cardiovascular risks at low body mass index levels among Singaporean Chinese, Malays and Indians. Obes Rev 3:209–215Google Scholar
  35. 35.
    Cowie CC, Rust KF, Ford ES et al (2009) Full accounting of diabetes and pre-diabetes in the US population in 1988–1994 and 2005–2006. Diab Care 32:287–294Google Scholar
  36. 36.
    Evans DJ, Barth JH, Burke CW (1988) Body fat topography in women with androgen excess. Int J Obes 12:157–162PubMedGoogle Scholar
  37. 37.
    Samy N, Hashim M, Sayed M, Said M (2009) Clinical significance of inflammatory markers in polycystic ovary syndrome: their relationship to insulin resistance and body mass index. Dis Markers 26:163–170PubMedGoogle Scholar
  38. 38.
    Chambers JC, Eda S, Bassett P et al (2001) C-reactive protein, insulin resistance, central obesity, and coronary heart disease risk in Indian Asians from the United Kingdom compared with European Whites. Circulation 104:145–150PubMedGoogle Scholar
  39. 39.
    Mahajan A, Tabassum R, Chavali S et al (2009) High-sensitivity C-reactive protein levels and type 2 diabetes in urban North Indians. J Clin Endocrinol Metab 94:2123–2127PubMedGoogle Scholar
  40. 40.
    Mohan V, Deepa R, Velmurugan K, Premalatha G (2005) Association of C-reactive protein with body fat, diabetes and coronary artery disease in Asian Indians: the Chennai Urban Rural Epidemiology Study (CURES-6). Diabet Med 22:863–870PubMedGoogle Scholar
  41. 41.
    Pradhan AD, Cook NR, Buring JE, Manson JE, Ridker PM (2003) C-reactive protein is independently associated with fasting insulin in nondiabetic women. Arterioscler Thromb Vasc Biol 23:650–655PubMedGoogle Scholar
  42. 42.
    Ridker PM, Hennekens CH, Buring JE, Rifai N (2000) C-reactive protein and other markers of inflammation in the prediction of cardiovascular disease in women. N Engl J Med 342:836–843PubMedGoogle Scholar
  43. 43.
    Yudkin JS, Stehouwer CDA, Emeis JJ, Coppack SW (1999) C-reactive protein in healthy subjects: associations with obesity, insulin resistance, and endothelial dysfunction: a potential role for cytokines originating from adipose tissue? Arterioscler Thromb Vasc Biol 19:972–978PubMedGoogle Scholar
  44. 44.
    Hotamisligil GS, Arner P, Caro JF, Atkinson RL, Spiegelman BM (1995) Increased adipose tissue expression of tumor necrosis factor-alpha in human obesity and insulin resistance. J Clin Invest 95:2409–2415PubMedGoogle Scholar
  45. 45.
    Lear SA, Kohli S, Bondy GP, Tchernof A, Sniderman AD (2009) Ethnic variation in fat and lean body mass and the association with insulin resistance. J Clin Endocrinol Metab 94:4696–4702PubMedGoogle Scholar
  46. 46.
    Banerji MA, Faridi N, Atluri R, Chaiken RL, Lebovitz HE (1999) Body composition, visceral fat, leptin, and insulin resistance in Asian Indian men. J Clin Endocrinol Metab 84:137–144PubMedGoogle Scholar

Copyright information

© Springer-Verlag 2011

Authors and Affiliations

  • C. Sabanayagam
    • 1
  • A. Shankar
    • 1
  • S. C. Lim
    • 2
  • J. Lee
    • 3
  • E. S. Tai
    • 4
  • T. Y. Wong
    • 5
    • 6
  1. 1.Department of Community MedicineWest Virginia University School of MedicineMorgantownUSA
  2. 2.Department of MedicineAlexandra HospitalSingaporeRepublic of Singapore
  3. 3.Department of Epidemiology and Public Health, Yong Loo Lin School of MedicineNational University of SingaporeSingaporeRepublic of Singapore
  4. 4.Department of Medicine, Yong Loo Lin School of MedicineNational University of SingaporeSingaporeRepublic of Singapore
  5. 5.Singapore Eye Research InstituteSingaporeRepublic of Singapore
  6. 6.Centre for Eye Research AustraliaUniversity of MelbourneMelbourneAustralia

Personalised recommendations