Table 2.
Association of Plasma 25(OH)D Metabolites With Incident T2D: EPIC-InterAct Study
| HR (95% CI)a |
|||||||
|---|---|---|---|---|---|---|---|
| Below the LLQb | Q1 | Q2 | Q3 | Q4 | Q5 | Per 1 SD | |
| Total 25(OH)D [nonepimeric 25(OH)D3 + 25(OH)D2], nmol/L | |||||||
| Median (range)c | 20.1 (5.68 to <25.9) | 30.9 (25.9–35.1) | 39.4 (35.1 to <43.7) | 48.7 (43.7 to <54.6) | 63.6 (54.6 to <118.7) | ||
| No. cases per subcohort | 2185/2460 | 1885/2477 | 1612/2488 | 1460/2479 | 1189/2441 | ||
| Model 1 | 1 (ref) | 0.84 (0.71, 0.98) | 0.65 (0.57, 0.73) | 0.55 (0.48, 0.64) | 0.43 (0.39, 0.48) | 0.71 (0.68, 0.73) | |
| Model 2 | 1 (ref) | 0.82 (0.70, 0.96) | 0.67 (0.59, 0.76) | 0.61 (0.53, 0.71) | 0.53 (0.46, 0.60) | 0.76 (0.73, 0.80) | |
| Model 3 | 1 (ref) | 0.91 (0.75, 1.10) | 0.79 (0.67, 0.93) | 0.79 (0.64, 0.98) | 0.72 (0.60, 0.87) | 0.86 (0.81, 0.91) | |
| Model 4 | 1 (ref) | 0.90 (0.74, 1.09) | 0.76 (0.65, 0.90) | 0.74 (0.61, 0.89) | 0.62 (0.54, 0.72) | 0.81 (0.77, 0.86) | |
| Nonepimeric 25(OH)D3, nmol/L | |||||||
| Median (range)c | 19.6 (5.30 to <25.5) | 30.5 (25.5 to <34.8) | 39.0 (34.8 to <43.3) | 48.4 (43.3 to <54.4) | 63.3 (54.4–118.7) | ||
| No. cases per subcohort | 2188/2461 | 1879/2476 | 1595/2491 | 1474/2474 | 1195/2443 | ||
| Model 1 | 1 (ref) | 0.83 (0.72, 0.95) | 0.64 (0.57, 0.72) | 0.57 (0.50, 0.65) | 0.44 (0.39, 0.49) | 0.71 (0.69, 0.74) | |
| Model 2 | 1 (ref) | 0.82 (0.71, 0.94) | 0.67 (0.58, 0.77) | 0.63 (0.55, 0.71) | 0.54 (0.47, 0.61) | 0.77 (0.74, 0.80) | |
| Model 3 | 1 (ref) | 0.89 (0.76, 1.04) | 0.76 (0.63, 0.91) | 0.80 (0.66, 0.97) | 0.72 (0.61, 0.85) | 0.86 (0.82, 0.91) | |
| Model 4 | 1 (ref) | 0.88 (0.75, 1.04) | 0.73 (0.61, 0.87) | 0.74 (0.62, 0.88) | 0.62 (0.54, 0.72) | 0.81 (0.77, 0.86) | |
| 3-epi-25(OH)D3, nmol/L | |||||||
| Median (range)c | <1 | 1.11 (1.00 to <1.25) | 1.40 (1.25 to <1.57) | 1.76 (1.57 to <1.98) | 2.29 (1.98 to <2.74) | 3.49 (2.74–15.4) | |
| No. cases per subcohort | 5143/7312 | 653/1007 | 678/1000 | 599/1006 | 619/1015 | 639/1005 | |
| Model 1 | 1.06 (0.95, 1.19) | 1 (ref) | 1.01 (0.87, 1.17) | 0.89 (0.76, 1.03) | 0.87 (0.75, 1.01) | 0.85 (0.73, 1.00) | 0.99 (0.94, 1.04) |
| Model 2 | 1.04 (0.92, 1.18) | 1 (ref) | 1.06 (0.89, 1.26) | 0.94 (0.79, 1.11) | 0.94 (0.80, 1.11) | 0.99 (0.83, 1.17) | 1.03 (0.97, 1.09) |
| Model 3 | 0.98 (0.83, 1.15) | 1 (ref) | 1.08 (0.86, 1.35) | 0.95 (0.79, 1.14) | 1.00 (0.84, 1.20) | 1.12 (0.91, 1.37) | 1.09 (1.02, 1.17) |
| Model 4 | 0.92 (0.79, 1.08) | 1 (ref) | 1.09 (0.85, 1.38) | 0.99 (0.82, 1.19) | 1.10 (0.92, 1.31) | 1.36 (1.08, 1.71) | 1.16 (1.09, 1.25) |
| 25(OH)D2, nmol/L | |||||||
| Median (range)c | <3 | 3.36 (3.03 to <3.74) | 4.30 (3.74 to <5.06) | 6.12 (5.06 to <7.42) | 8.74 (7.42 to <11.1) | 15.0 (11.1–46.4) | |
| No. cases per subcohort | 7988/11,773 | 50/111 | 72/115 | 70/116 | 88/114 | 63/116 | |
| Model 1 | 1.45 (1.02, 2.05) | 1 (ref) | 1.16 (0.73, 1.83) | 1.07 (0.68, 1.70) | 1.32 (0.85, 2.07) | 1.01 (0.63, 1.61) | 1.00 (0.85, 1.18) |
| Model 2 | 1.36 (0.93, 1.98) | 1 (ref) | 1.04 (0.63, 1.71) | 1.12 (0.68, 1.85) | 1.44 (0.88, 2.38) | 1.02 (0.62, 1.69) | 1.02 (0.84, 1.23) |
| Model 3 | 1.30 (0.86, 1.96) | 1 (ref) | 0.90 (0.50, 1.64) | 1.18 (0.69, 2.02) | 1.18 (0.67, 2.09) | 1.08 (0.62, 1.87) | 1.03 (0.84, 1.25) |
| Model 4 | 1.29 (0.86, 1.95) | 1 (ref) | 0.93 (0.52, 1.66) | 1.15 (0.68, 1.96) | 1.10 (0.62, 1.96) | 1.00 (0.58, 1.73) | 0.94 (0.76, 1.18) |
HRs of T2D comparing quintiles (Q2 to Q5) of 25(OH)D metabolites with Q1 or per 1 SD increase of 25(OH)D metabolites, estimated from country-specific Prentice-weighted Cox regression models; estimates were combined across countries using random-effects meta-analysis. Effect estimates for 25(OH)D2 were derived from analysis of the overall EPIC-InterAct data (i.e., not country-specific) owing to limited sample size. 1 SD (calculated from the subcohort) was 17.3 nmol/L for nonepimeric 25(OH)D3, 1.31 nmol/L for 3-epi-25(OH)D3, and 6.52 nmol/L for 25(OH)D2. The present analyses were based on complete case analyses excluding participants with missing covariates based on model 4. The sample size of total cases per subcohort was 8331/12,345 for nonepimeric 25(OH)D3, 3188/5033 for 3-epi-25(OH)D3 and 343/572 for 25(OH)D2. Models were as follows: model 1, adjusted for age (as underlying timescale), sex, center, and seasonality (continuous: sine and cosine function of the day of blood draw); model 2, model 1 plus smoking status (current, former, never), physical activity (inactive, moderately inactive, moderately active, active), education (none, primary, technical or professional, secondary, higher education), alcohol drinking (never, >0 to <6 g/d, 6 to <12 g/d, 12 to <24 g/d, ≥24 g/d), total energy intake (continuous), Mediterranean diet score (low, moderate, high) and plasma lipid biomarkers (continuous: HDL-cholesterol, LDL-cholesterol); model 3, model 2 plus BMI (continuous); model 4, model 3 plus mutual adjustment for the other 25(OH)D metabolites [nonepimeric 25(OH)D3 (continuous), 3-epi-25(OH)D3 (categorical: below the LLQ, Q1, Q2, Q3, Q4, and Q5), or 25(OH)D2 (categorical: below and above the LLQ)].
The LLQs for 3-epi-25(OH)D3 and 25(OH)D2 are 1 nmol/L and 3 nmol/L, respectively.
Median and range of the 25(OH)D metabolites in each quintile in the InterAct subcohort.