Table 5.

Associations of retinol with risk of total, low-, and high-grade prostate cancer, stratified by supplemental retinol use, in the placebo arm of the PCPTa,b

Total cancerGleason 2–6Gleason 7–10
Case/controlOR (95% CI)Case/controlOR (95% CI)Case/controlOR (95% CI)
User
 Q1c86/110Ref.69/110Ref.15/110Ref.
 Q2110/1101.21 (0.81–1.79)82/1101.09 (0.71–1.68)23/1101.59 (0.78–3.23)
 Q3101/1101.06 (0.71–1.58)78/1100.98 (0.63–1.51)18/1101.24 (0.59–2.63)
 Q4147/1101.52 (1.03–2.25)107/1101.32 (0.87–2.01)36/1102.45 (1.24–4.85)
P trend0.060.260.02
Nonuser
 Q191/116Ref.63/116Ref.24/116Ref.
 Q2101/1161.06 (0.72–1.57)77/1161.15 (0.75–1.77)20/1160.86 (0.45–1.66)
 Q3108/1171.11 (0.75–1.63)88/1171.26 (0.83–1.92)16/1170.70 (0.35–1.40)
 Q4108/1161.14 (0.78–1.68)80/1161.18 (0.77–1.80)24/1161.09 (0.58–2.05)
P trend0.490.420.94
P interactiond0.890.980.65
  • aIndividuals were defined as supplement users if they reported using vitamin A supplements more than once per week. Supplement use data were missing for 215 cases and 182 controls.

  • bAssociations adjusted for age, race, BMI, family history of prostate cancer, and serum cholesterol.

  • cQuartiles of serum retinol were defined based on values in all controls (finasteride and placebo arms combined). Quartiles 1, 2, 3, and 4, respectively, were defined using the following stratum-specific cutoffs: <0.60 μg/dL, 0.60 to <0.69 μg/dL, 0.69 to <0.78 μg/dL, >0.78 μg/dL.

  • dP value tests for interaction between serum retinol quartile (as a trend ordinal value) and supplemental retinol use.