High-sensitivity C-reactive protein, but not white blood cell count, independently predicted incident diabetes in a Japanese health screening population.

AIMS: To compare high-sensitivity C-reactive protein (hs-CRP) and white blood cell count (WBC) as a predictor of incident diabetes in a population where obesity is not prevalent.
METHODS: This is a retrospective 6-year follow-up study in a Japanese health screening population including 1874 men and 1094 women. Using Cox regression methods, hazard ratios (HRs) of incident diabetes for hs-CRP and WBC adjusting for fasting plasma glucose (FPG), hemoglobin A1c (HbA1c) and other confounders were calculated, and using areas under receiver operating characteristic curve (AUCs), diabetes-predicting abilities of hs-CRP and WBC were compared. Diabetes was defined as FPG ≥ 126 mg/dL and HbA1c ≥ 6.5 % or use of antidiabetic medication.
RESULTS: During the 6-year follow-up period (mean ± SD, 4.8 ± 1.7 years), 71 men (3.8 %) and 19 women (1.7 %) developed incident diabetes. The fully adjusted HRs [95 % confidence intervals (CIs)] of incident diabetes for each 1 SD increase in log hs-CRP and WBC were 1.20 (0.92-1.56) (p = 0.174) and 1.01 (0.78-1.30) (p = 0.946), respectively. The fully adjusted HRs (95 % CIs) of incident diabetes for the highest tertile of hs-CRP and WBC compared with the lowest tertile were 2.57 (1.05-6.27) (p = 0.039) and 1.20 (0.53-2.70) (p = 0.665), respectively. The AUCs (95 % CIs) of hs-CRP and WBC for the discrimination of incident diabetes were 0.73 (0.68-0.77) and 0.67 (0.62-0.72), respectively.
CONCLUSIONS: Hs-CRP, but not WBC, was independently associated with incident diabetes in a Japanese health screening population where obesity is not prevalent.