Table 2. Multivariable-adjusted HRs and 95% CIs of incidence of pancreatic cancer by levels of total magnesium intake, VITAL Cohort Study, 2000–2008a.
| Linear trend | |||||
|---|---|---|---|---|---|
| ⩾100% RDA | 75–99% RDA | <75% RDA | ↓100 mg per day | P-valueb | |
|
Total magnesium intake | |||||
|
Magnesium (mg per day) | |||||
| Mean (s.d.) | 518.51 (148.94) | 326.74 (50.79) | 219.43 (53.53) | ||
| Range | 320–1805 | 240–420 | <315 | ||
| No. of participant at risk | 35 348 | 17 063 | 14 395 | ||
| No. of events | 64 | 43 | 44 | ||
| Model 1c | 1 (Reference) | 1.39 (0.94, 2.05) | 1.71 (1.16, 2.52) | 1.23 (1.06, 1.41) | <0.01 |
| Model 2d | 1 (Reference) | 1.42 (0.91, 2.21) | 1.76 (1.04, 2.96) | 1.24 (1.02, 1.50) | 0.03 |
|
Magnesium supplement users | |||||
|
Magnesium (mg per day) | |||||
| Mean (s.d.) | 529.99 (156.02) | 325.85 (50.94) | 242.62 (43.27) | ||
| Range | 320–1805 | 240–420 | <315 | ||
| No. of participant at risk | 27 016 | 8838 | 3201 | ||
| No. of events | 45 | 21 | 11 | ||
| Model 1c | 1 (Reference) | 1.44 (0.85, 2.43) | 2.13 (1.09, 4.16) | 1.30 (1.04, 1.63) | 0.02 |
| Model 2d | 1 (Reference) | 1.51 (0.80, 2.84) | 2.21 (0.93, 5.26) | 1.33 (0.98, 1.80) | 0.07 |
|
Magnesium supplement non-users | |||||
|
Magnesium (mg per day) | |||||
| Mean (s.d.) | 481.30 (115.39) | 328.04 (50.55) | 213.12 (54.37) | ||
| Range | 320–1309 | 240–420 | <315 | ||
| No. of participant at risk | 8166 | 8112 | 11 010 | ||
| No. of events | 18 | 20 | 30 | ||
| Model 1c | 1 (Reference) | 1.07 (0.56, 2.02) | 1.12 (0.62, 2.03) | 1.05 (0.83, 1.31) | 0.70 |
| Model 2d | 1 (Reference) | 0.99 (0.47, 2.12) | 0.92 (0.34, 2.43) | 0.93 (0.65, 1.33) | 0.88 |
Abbreviations: CI=confidence interval; HR=hazard ratio; MET=metabolic equivalent; NSAIDs=non-steroidal anti-inflammatory drugs; RDA=recommended dietary allowance; VITAL=Vitamins and Lifestyle.
The magnesium intake RDA cutoff points used are in females 240 mg (75% RDA) and 320 mg (100%RDA) and in males 315 mg (75% RDA) and 420 mg (100% RDA).
All models were constructed using Cox proportional hazards regression analysis.
P for trend was calculated using the continuous values (after deleting the extreme values above the 99th percentile) of the exposure.
Model 1: adjusted for age (time variable), gender, ethnicity (White and non-White), and education (high school graduate or less, some college, college or advanced degree).
Model 2: additionally adjusted for body mass index (<25, 25–29, or ⩾30 kg m−2), physical activity (0, tertiles of MET hours per week), smoking (0, tertiles of pack-years), alcohol consumption (tertiles), diabetes mellitus (yes or no), family history of pancreatic cancer (yes or no), use of NSAIDs (yes or no), and total intakes (tertiles) of calcium, selenium, long-chain omega-3 fatty acids, and total calories.