Table 2.
Risk of diabetes according to quartiles of hs-CRP: The Singapore Chinese Health Study
| Quartiles of hs-CRPa | P for trendb | Per 1 log mg/L increase | ||||
|---|---|---|---|---|---|---|
| Q1 | Q2 | Q3 | Q4 | |||
| Median (range) | 0.4 (0.1–0.6) | 0.9 (0.7–1.2) | 1.6 (1.3–2.3) | 3.6 (2.4–11.0) | ||
| Total diabetes | ||||||
| Cases/controls | 78/154 | 107/139 | 150/138 | 236/140 | ||
| Model 1 | 1.00 | 1.51 (1.03–2.20) | 2.14 (1.47–3.11) | 3.25 (2.28–4.64) | <0.001 | 1.59 (1.40–1.81) |
| Model 2 | 1.00 | 1.42 (0.96–2.11) | 1.89 (1.27–2.80) | 2.70 (1.85–3.95) | <0.001 | 1.50 (1.31–1.71) |
| Model 3 | 1.00 | 1.23 (0.81–1.86) | 1.50 (0.99–2.27) | 1.97 (1.32–2.94) | <0.001 | 1.34 (1.16–1.54) |
| Model 4 | 1.00 | 1.24 (0.79–1.96) | 1.22 (0.77–1.92) | 1.74 (1.12–2.70) | 0.016 | 1.27 (1.09–1.48) |
| Undiagnosed diabetes | ||||||
| Cases/controls | 32/81 | 43/67 | 73/54 | 131/77 | ||
| Model 1 | 1.00 | 1.48 (0.84–2.62) | 3.63 (2.02–6.53) | 4.39 (2.59–7.45) | <0.001 | 1.84 (1.52–2.24) |
| Model 2 | 1.00 | 1.29 (0.71–2.37) | 3.17 (1.72–5.82) | 3.63 (2.08–6.36) | <0.001 | 1.75 (1.43–2.14) |
| Model 3 | 1.00 | 1.21 (0.64–2.28) | 2.41 (1.26–4.59) | 2.55 (1.42–4.58) | <0.001 | 1.53 (1.24–1.89) |
| Model 4 | 1.00 | 1.20 (0.59–2.44) | 2.17 (1.07–4.40) | 2.43 (1.25–4.71) | 0.003 | 1.53 (1.20–1.94) |
| Incident diabetes | ||||||
| Cases/controls | 46/73 | 64/72 | 77/84 | 105/63 | ||
| Model 1 | 1.00 | 1.45 (0.87–2.43) | 1.45 (0.88–2.39) | 2.54 (1.54–4.20) | <0.001 | 1.38 (1.16–1.64) |
| Model 2 | 1.00 | 1.50 (0.85–2.65) | 1.24 (0.70–2.18) | 2.20 (1.26–3.84) | 0.008 | 1.31 (1.08–1.59) |
| Model 3 | 1.00 | 1.31 (0.73–2.36) | 1.07 (0.60–1.93) | 1.70 (0.95–3.07) | 0.14 | 1.20 (0.98–1.47) |
| Model 4 | 1.00 | 1.24 (0.65–2.39) | 0.75 (0.39–1.45) | 1.24 (0.64–2.39) | 0.93 | 1.06 (0.85–1.33) |
Abbreviation: hs-CRP high-sensitive C-reactive protein, OR odds ratio, CI confidence interval
aQuartiles of hs-CRP were created based on controls in the whole population
bLinear trend was tested by using the median level of each quartile of hs-CRP as continuous variables
Model 1: adjusted for age (continuous variable), matched on age (±3 years), sex, dialect, and date of blood collection (±6 months)
Model 2: model 1 plus education level, smoking status, alcohol intake, physical activity, hypertension, and fasting status
Model 3: model 2 plus adjusted for body mass index (continuous)
Model 4: model 3 plus plasma triglycerides and HDL-cholesterol levels (both in quartiles)