Table 3.

Adjusted HRs and 95% CIs for an increment of 24.96 nmol/L (equivalent to 10 ng/mL) of 25(OH)D for CRC-specific and overall mortality across strata of potential effect modifiers

CRC-specific mortalityOverall mortality
Risk factorMultivariablea HR (95% CI)Pinteraction or PtrendMultivariablea HR (95% CI)Pinteraction or Ptrend
All participants0.92 (0.85–1.00)0.07b0.91 (0.84–0.99)0.03b
Sensitivity analyses
 Participants with complete CRC stage data0.93 (0.84–1.02)0.11b0.92 (0.84–1.01)0.08b
 All participants with imputed CRC stage datac0.93 (0.85–1.02)d0.15d0.92 (0.84–1.00)d0.05d
 ≥2 y0.90 (0.81–1.00)0.06b0.89 (0.81–0.98)0.02b
 ≥3 y0.92 (0.81–1.04)0.16b0.92 (0.83–1.03)0.15b
 ≥5 y0.88 (0.73–1.06)0.17b0.90 (0.76–1.06)0.21b
 Women0.87 (0.76–1.00)0.260.86 (0.76–0.99)0.29
 Men0.97 (0.86–1.09)0.94 (0.84–1.04)
Age at diagnosis, y
 <62.40.86 (0.76–0.99)0.380.89 (0.79–1.01)0.80
 ≥62.40.96 (0.85–1.08)0.90 (0.81–1.01)
Anatomical site
 Colon0.99 (0.88–1.11)0.280.95 (0.86–1.06)0.53
 Rectum0.81 (0.69–0.95)0.85 (0.73–0.98)
Colon subsitef
 Proximal0.91 (0.77–1.09)0.720.94 (0.80–1.10)0.57
 Distal1.05 (0.86–1.29)0.95 (0.79–1.15)
 I and II0.78 (0.61–1.00)0.250.80 (0.66–0.98)0.23
 III and IV0.97 (0.87–1.08)0.96 (0.87–1.06)
Year of diagnosis
 1993–19990.89 (0.77–1.03)0.530.84 (0.73–0.96)0.24
 1999–20040.95 (0.85–1.07)0.96 (0.86–1.07)
Season of blood collectionh
 Summer/autumn0.97 (0.84–1.11)0.230.96 (0.85–1.09)0.14
 Winter/spring0.89 (0.78–1.01)0.87 (0.77–0.98)
Season of diagnosish
 Summer/autumn0.92 (0.80–1.06)0.710.89 (0.78–1.01)0.44
 Winter/spring0.92 (0.82–1.04)0.92 (0.82–1.03)
Dietary calcium, mg/di
 <9281.00 (0.89–1.13)0.010.99 (0.89–1.11)0.01
 ≥9280.85 (0.74–0.97)0.84 (0.74–0.94)
Smoking status
 Never smoker0.86 (0.73–1.02)0.130.87 (0.74–1.02)0.24
 Former smoker0.99 (0.86–1.13)0.97 (0.86–1.10)
 Current smoker0.74 (0.61–0.91)0.80 (0.67–0.95)
BMI, kg/m2
 <250.87 (0.74–1.01)0.530.86 (0.75–0.99)0.31
 25–29.90.96 (0.85–1.09)0.94 (0.84–1.06)
 ≥300.96 (0.74–1.24)1.00 (0.80–1.26)

NOTE: Subgroup analyses were conducted by time between blood collection and cancer diagnosis, sex, age at diagnosis (median-dichotomized; <62, ≥62 years), location of tumor (colon vs. rectum; and within the colon, proximal vs. distal), cancer stage (I–II vs. III–IV), season of diagnosis (winter/spring vs. summer/autumn), year of diagnosis (median-dichotomized; <1999, ≥1999), prediagnostic BMI (WHO categories: <25, normal; 25–29, overweight; ≥30 kg/m2, obese), smoking status (current, former, never), and dietary calcium intake (median-dichotomized; <928, ≥928 mg/d). In interaction analyses, adjusted HRs and 95% CI for an increment of 24.96 nmol/L of 25(OH)D levels for CRC-specific and overall mortality were reported. Tests of statistical interaction between 25(OH)D and relevant factors were assessed by including in the model the cross product of 25(OH)D levels as a continuous variable and the covariate as a continuous or dichotomous variable, as appropriate.

  • aMultivariable HRs, 95% CIs, and P values are adjusted for age at diagnosis (in years as a continuous variable), sex (men or women), cancer stage (I–IV, unknown), grade of tumor differentiation (well differentiated, moderately differentiated, poorly differentiated, or unknown), location of primary tumor (colon or rectum), smoking status (current, former, never smoker, or unknown), BMI (in kg/m2 as a continuous variable), physical activity (in METs as a continuous variable), season of blood collection (winter, spring, summer, and autumn), and year of diagnosis (as a continuous variable), and stratified by country of residence. In the stratified models, the stratifying variable was not adjusted for. P value for interaction is presented unless otherwise indicated.

  • bPtrend.

  • cMissing stage data were imputed using the algorithm described in the Statistical analysis section.

  • dCombined HR estimate and 95% CI; P value from a t test for the hypothesis that the parameter is equal to its null value.

  • eMultivariable model including all CRC cases with time interval between blood collection and cancer diagnosis (follow-up) of more than 2, 3, and 5 years.

  • fOnly colon tumors with known locations were included. Unspecified (N = 63) and overlapping lesion of colon (N = 9) tumors were excluded.

  • gMissing data were not included in the analyses.

  • hSummer/autumn period included June, July, August, September, October, and November; winter/spring period included December, January, February, March, April, and May.

  • iIn the analyses by predefined categories, participants with high dietary calcium intake (≥928 mg/d) and high prediagnostic vitamin D levels (>100 nmol/L) had an adjusted HR of 0.24 (95% CI: 0.11–0.54; Ptrend = 0.012) for CRC-specific mortality and 0.26 (95% CI: = 0.13–0.53; Ptrend = 0.002) for overall mortality compared with participants with the lowest 25(OH)D levels (<25 nmol/L). Whereas among participants with low calcium intake, the corresponding HRs were 0.86 (95% CI: 0.41–1.82; Ptrend = 0.877) for CRC-specific and 0.92 (95% CI: 0.46–1.86; Ptrend = 0.733) for overall mortality.