Prospective Relation of C-Reactive Protein With Type 2 Diabetes: Response to Han et al.

We read with great interest the article by Han et al., which was recently published in Diabetes Care (1). The authors studied the association between baseline levels of C-reactive protein (CRP) and the 6-year incidence of the metabolic syndrome and type 2 diabetes in Mexican subjects. They found an odds ratio (OR) of 4.1 (95% CI 2.1–8.0) for developing the metabolic syndrome and an OR of 5.4 (2.2–13.4) for incident type 2 diabetes among women in the highest compared with the lowest tertile of CRP. ORs were adjusted for age, smoking, physical activity, and alcohol intake. In men, no such association was found. Additional adjustment for BMI slightly lowered the ORs; adjustment for waist-to-hip ratio (WHR) did not.

We have studied baseline CRP levels and their relation to incident type 2 diabetes in an age-, sex-, and glucose-stratified sample of the Hoorn Study, a population-based cohort study of glucose tolerance among Caucasian people (2). The study methods and follow-up duration are closely similar to the Mexico City Diabetes Study (MCDS) described by Han et al., except that our subjects were ∼15 years older at baseline. Glucose tolerance status was assessed by a 75-g oral glucose tolerance test at baseline and after 6.4 years of follow-up. CRP was measured in plasma by high-sensitive enzyme-linked immunosorbent assay, while information was available on BMI, WHR, smoking, and physical activity. Of the 140 men and 139 women who had follow-up measurements and were free of diabetes at baseline, 17.8% of the men and 20.9% of the women developed diabetes. In contrast to the findings of Han et al., we did not observe an association between baseline CRP levels and incident diabetes in women, while in men we found an OR of 3.0 (95% CI 1.0–9.3) in the highest compared with the lowest tertile of CRP, after adjustment for age (Table 1). Further adjustment for BMI, smoking, or physical activity did not materially change these results, while adding WHR to the model substantially lowered the OR in men. Thus, in the Hoorn Study, CRP is not a very strong determinant of the development of type 2 diabetes, in contrast to WHR and impaired glucose metabolism (3). We realize that our study sample of 279 subjects is much smaller than the study presented by Han et al. (n = 1,244). The relationship in men, however, was strong, whereas in the MCDS it was strong only in women. In the previous prospective study by Barzilay et al. (4) an OR of 2.0 (1.4–2.9) was found when extreme quartiles of CRP were compared. They did not analyze men and women or black and white subjects separately. Pradhan et al. (5) found an OR of 4.2 (1.5–12.0) in women. Both studies did not take WHR into account. Ethnicity may also play an important role in explaining the inconsistent results.

In conclusion, we suggest caution in the interpretation of the results because of inconsistent findings, in particular between sexes. It is unclear why inflammation would be important in the pathogenesis of type 2 diabetes only in women or only in men.