TABLE 2.
Vitamin D metabolites at baseline according to race1
| Unadjusted median (IQR) |
Adjusted median (95% CI)2 |
||||||
| Overall (n = 208) | Black (n = 57) | White (n = 151) | P3 | Black (n = 57) | White (n = 151) | P4 | |
| Total 25(OH)D, ng/mL | 26.0 (19.4–31.3) | 22.0 (14.0–32.1) | 26.8 (22.0–31.0) | 0.011 | 20.3 (16.2, 24.5) | 26.7 (25.2, 28.1) | 0.026 |
| Total 25(OH)D <20 ng/mL, % (n) | 26.0 (54) | 45.6 (26) | 18.5 (28) | <0.001 | — | — | — |
| Directly measured free 25(OH)D, pg/mL | 5.45 (4.17–6.76) | 4.88 (3.40–6.26) | 5.73 (4.54–6.84) | 0.006 | 4.54 (3.72, 5.37) | 5.66 (5.40, 5.92) | 0.044 |
| Measured free 25(OH)D as % total 25(OH)D | 0.022 (0.019–0.026) | 0.023 (0.02–0.026) | 0.021 (0.019–0.025) | NS | 0.024 (0.021, 0.026) | 0.022 (0.021, 0.023) | NS |
| Calculated free 25(OH)D,5 pg/mL | |||||||
| Monoclonal assay | 9.17 (7.28–12.55) | 11.89 (8.57–17.09) | 8.84 (7.23–11.22) | <0.001 | 11.83 (9.05, 14.61) | 8.88 (8.36, 9.40) | 0.080 |
| Polyclonal assay | 9.14 (6.88–11.43) | 7.75 (5.99–11.68) | 9.52 (7.63–11.41) | 0.008 | 8.32 (6.74, 9.90) | 9.41 (8.85, 9.96) | NS |
| Vitamin D–binding protein, mg/L | |||||||
| Monoclonal assay | 194 (133–247) | 105 (82–154) | 212 (172–254) | <0.001 | 109 (90, 129) | 220 (207, 232) | <0.001 |
| Polyclonal assay | 199 (174–235) | 195 (176–234) | 199 (173–235) | NS | 192 (173, 211) | 201 (192, 209) | NS |
| PTH, pg/mL | 54.9 (43.6–69.5) | 60.9 (46.3–85.1) | 52.5 (42.8–65.3) | 0.065 | 58.8 (47.5, 70.0) | 51.2 (47.8, 54.6) | NS |
1
CaDDM, Calcium and Vitamin D for Diabetes Mellitus; DDM2, Vitamin D for Established Type 2 Diabetes; PTH, parathyroid hormone; 25(OH)D, 25-hydroxyvitamin D.
2
Estimated from race-stratified quantile regression models adjusted for age, sex, BMI, season, vitamin D intake, and study (CaDDM or DDM2).
3
P values of group differences (black vs. white) for continuous variables were analyzed with the use of nonparametric Kruskal-Wallis tests for medians. Proportions were analyzed with the use of chi-square tests.
4
P values of group differences (black vs. white) for continuous variables are based on 2-sample z tests.
5
Based on the method described by Bikle et al. (7).