Table 2. Odds ratios for diabetes in participants grouped by serum 25-hydroxyvitamin D3 quartiles.
| Serum 25-hydroxyvitamin D3, ng/mL | ||||
|
| ||||
| 1st quartile group | 2nd quartile group | 3rd quartile group | 4th quartile group | |
| ≤18.10 (2.08–18.10) | 18.11–22.90 | 22.91–28.17 | ≥28.18 (28.18–78.83) | |
| (n = 120) | (n = 120) | (n = 120) | (n = 120) | |
| Cases of diabetes, n | 12 | 9 | 8 | 6 |
| Prevalence, % | 10.0 | 7.5 | 6.7 | 5.0 |
| Adjusted odds ratio (95% CI), model 1 | 3.14 (1.08–9.11) | 2.06 (0.68–6.24) | 1.63 (0.53–4.98) | 1.00 (Reference) |
| Adjusted odds ratio (95% CI), model 2 | 4.02 (1.25–12.89) | 2.48 (0.76–8.05) | 1.89 (0.59–6.02) | 1.00 (Reference) |
| Adjusted odds ratio (95% CI), model 3 | 4.02 (1.25–12.92) | 2.50 (0.77–8.10) | 1.91 (0.60–6.09) | 1.00 (Reference) |
CI, confidence interval.
Diabetes was defined as a fasting plasma glucose level ≥126 mg/dL and/or a HbA1c level ≥6.5% [reference 24].
Three different logistic regression models were used to calculate odds ratio (95% CI) with the 4th quartile group serving as the reference group: model 1 adjusted for age, sex, and months of blood sample collection; model 2 adjusted for the same covariates used in model 1, in addition to smoking habits, estimated glomerular filtration rate, work status, exercise habits, alcohol intake, protein intake, fat intake, and total dietary fiber intake; model 3 adjusted for the same covariates used in model 2, in addition to body mass index.