We thank Snijder et al. (1) for their interest in our article (2) and for sharing their data relating C-reactive protein (CRP) to the risk of diabetes development in the Hoorn Study. We readily agree that it is unclear why inflammation might be more important to the pathogenesis of type 2 diabetes in women compared with men or indeed vice versa. We have reanalyzed our CRP data using the same covariates as Snijder et al. and essentially our results are unchanged—the relation between CRP and diabetes remains strong in women, even after inclusion of waist-to-hip ratio (WHR), but is absent in men (Table 1). Interestingly, on reanalysis of similarly relevant data from the Insulin Resistance Atherosclerosis Study (IRAS) (3), a study of similar size to the Mexico City Diabetes Study (MCDS), the relation between baseline CRP and diabetes development was also stronger in women compared with men using the same covariate adjustments (data not shown). Moreover, in both the MCDS and IRAS, adjustment for BMI attenuated the link between CRP and diabetes more so than WHR.
One possibility of the divergent results could be the younger baseline age (∼46 years) in MCDS in comparison to that of the Hoorn study (∼61 years). In this respect it is important to consider the additional potential confounding influence of hormonal replacement therapy (HRT) in women. HRT use increases CRP concentrations in women (4), but paradoxically recent data from the Heart and Estrogen/Progestin Replacement Study suggest that HRT reduces risk of diabetes development by 35% (5). In MCDS, only 4% of women were on HRT at baseline and hormonal use did not confound the link between CRP and diabetes. However, because of higher age, a potentially much higher proportion of women in the Hoorn Study may have been on HRT at baseline, which could thus disguise any potential association between CRP and diabetes risk.
We acknowledge that our inability to find an association between CRP and diabetes risk in men may simply be due to a lack of power. A recent relevant study (6) from the West of Scotland Coronary Prevention Study (WOSCOPS) group that included 127 new cases (predominantly Caucasian) of diabetes reported a strong association between baseline CRP and diabetes risk independent of BMI, blood pressure, smoking, and fasting lipids and glucose concentrations. Alternatively and as Snijder et al. suggest, ethnicity may also play a role in explaining inconsistent results.
Clearly, further studies are needed to examine the relation between inflammatory markers and diabetes risk. It would be useful if such studies were to report associations separately in men and women in order to enable identification of any potential gender differences.
ORs (with 95% CIs) associated with high C-reactive protein (second and third tertile compared to first tertile) for developing type 2 diabetes after 6.5 years of follow-up in the Mexico City Diabetes Study
| Model . | Men (35 incident cases/n = 460) . | . | Women (54 incident cases/n = 672) . | . | ||
|---|---|---|---|---|---|---|
| . | Second tertile (0.91–1.64 ng/ml) . | Third tertile (>1.64 ng/ml) . | Second tertile (1.24–2.16 ng/ml) . | Third tertile (>2.16 ng/ml) . | ||
| 1 Age | 0.7 (0.3–1.7) | 1.0 (0.5–2.3) | 3.1 (1.2–7.9)* | 5.7 (2.3–13.9)† | ||
| 2 Model 1 + BMI | 0.6 (0.2–1.4) | 0.8 (0.4–1.9) | 2.2 (0.9–5.8)* | 3.1 (1.2–8.1)* | ||
| 3 Model 1 + WHR | 0.6 (0.2–1.5) | 0.9 (0.4–2.1) | 3.0 (1.2–7.7)* | 5.4 (2.2–13.3)† | ||
| 4 Model 1 + smoking | 0.7 (0.3–1.7) | 1.0 (0.5–2.3) | 3.1 (1.2–8.0)* | 5.7 (2.3–14.1)† | ||
| 5 Model 1 + physical activity | 0.7 (0.3–1.6) | 0.9 (0.4–2.0) | 2.7 (1.0–7.1)* | 5.3 (2.2–13.2)† | ||
| Model . | Men (35 incident cases/n = 460) . | . | Women (54 incident cases/n = 672) . | . | ||
|---|---|---|---|---|---|---|
| . | Second tertile (0.91–1.64 ng/ml) . | Third tertile (>1.64 ng/ml) . | Second tertile (1.24–2.16 ng/ml) . | Third tertile (>2.16 ng/ml) . | ||
| 1 Age | 0.7 (0.3–1.7) | 1.0 (0.5–2.3) | 3.1 (1.2–7.9)* | 5.7 (2.3–13.9)† | ||
| 2 Model 1 + BMI | 0.6 (0.2–1.4) | 0.8 (0.4–1.9) | 2.2 (0.9–5.8)* | 3.1 (1.2–8.1)* | ||
| 3 Model 1 + WHR | 0.6 (0.2–1.5) | 0.9 (0.4–2.1) | 3.0 (1.2–7.7)* | 5.4 (2.2–13.3)† | ||
| 4 Model 1 + smoking | 0.7 (0.3–1.7) | 1.0 (0.5–2.3) | 3.1 (1.2–8.0)* | 5.7 (2.3–14.1)† | ||
| 5 Model 1 + physical activity | 0.7 (0.3–1.6) | 0.9 (0.4–2.0) | 2.7 (1.0–7.1)* | 5.3 (2.2–13.2)† | ||
The main effect of gender is significant (P < 0.05) for each model. Gender-CRP tertile interaction:
P < 0.05,
P < 0.01.
References
Address correspondence to Steven M. Haffner, MD, Department of Medicine #7873, University of Texas Health Science Center at San Antonio, San Antonio, TX 78229-3900. E-mail: [email protected].
