Table 4.
The association of NMR-measured ionized Mg with the risk of developing T2DM in the PREVEND study (sensitivity analyses).
| Continuous, per 0.1 mmol/L decrease | Tertiles of NMR-Measured Mg, mmol/L | |||
|---|---|---|---|---|
| 1 | 2 | 3 | ||
| Total (n = 5747) | ||||
| Events, n (%) | 289 (5.0) | 108 (5.6) | 99 (5.2) | 82 (4.3) |
| Crude analysis | 1.50 (1.19–1.89) | 1.36 (1.02–1.82) | 1.23 (1.02–1.82) | 1.00 (reference) |
| Age and sex adjusted | 1.54 (1.23–1.92) | 1.46 (1.09–1.94) | 1.36 (1.01–1.82) | 1.00 (reference) |
| Model 1 | 1.32 (1.04–1.67) | 1.21 (0.90–1.62) | 1.18 (0.90–1.62) | 1.00 (reference) |
| Model 2 | 1.37 (1.08–1.74) | 1.26 (0.93–1.69) | 1.20 (0.89–1.62) | 1.00 (reference) |
| Model 3 | 1.19 (0.94–1.51) | 1.12 (0.83–1.51) | 1.21 (0.90–1.63) | 1.00 (reference) |
| Men (n = 2848) | ||||
| Events, n (%) | 186 (6.5) | 66 (6.7) | 66 (7.3) | 54 (5.7) |
| Crude analysis | 1.25 (0.94–1.67) | 1.21 (0.84–1.73) | 1.31 (0.91–1.87) | 1.00 (reference) |
| Age adjusted | 1.27 (0.96–1.69) | 1.26 (0.88–1.80) | 1.41 (0.98–2.02) | 1.00 (reference) |
| Model 1 | 1.05 (0.78–1.40) | 1.01 (0.70–1.45) | 1.32 (0.92–1.90) | 1.00 (reference) |
| Model 2 | 1.04 (0.78–1.41) | 1.01 (0.70–1.45) | 1.32 (0.92–1.89) | 1.00 (reference) |
| Model 3 | 0.89 (0.67–1.20) | 0.84 (0.58–1.22) | 1.26 (0.87–1.81) | 1.00 (reference) |
| Women (n = 2899) | ||||
| Events, n (%) | 103 (3.6) | 42 (4.5) | 33 (3.3) | 28 (2.9) |
| Crude analysis | 2.02 (1.37–2.99) | 1.65 (1.02–2.66) | 1.15 (0.70–1.91) | 1.00 (reference) |
| Age adjusted | 2.33 (1.58–3.42) | 1.99 (1.23–3.22) | 1.28 (0.78–2.13) | 1.00 (reference) |
| Model 1 | 1.70 (1.14–2.52) | 1.52 (0.92–2.49) | 1.12 (0.67–1.87) | 1.00 (reference) |
| Model 2 | 1.89 (1.27–2.81) | 1.71 (1.04–2.83) | 1.18 (0.70–1.99) | 1.00 (reference) |
| Model 3 | 1.81 (1.22–2.69) | 1.75 (1.06–2.89) | 1.29 (1.06–2.17) | 1.00 (reference) |
Hazard ratios and 95% confidence intervals were derived from Cox proportional hazards regression models. T2DM, type 2 diabetes mellitus; GFR indicates glomerular filtration rate; HDL, high-density lipoprotein. Model 1: Adjusted for age, sex, body mass index, alcohol consumption, smoking status, triglyceride:HDL cholesterol ratio, presence of cardiovascular disease, and parental history of T2DM. Model 2: Model 1 and additionally adjusted for plasma levels of albumin, potassium, and calcium and urinary albumin excretion. Model 3: Model 2 and additionally adjusted for CRP, fasting glucose, and presence of chronic kidney disease.