Purpose: High calcium intake from dairy products and supplements has been associated with an increased risk of incident and lethal prostate cancer (PCa) in some, but not all, studies. We evaluated the associations of dietary (from foods only) and total (from foods and supplements) intake of calcium with PCa risk in a pooled analysis of 14 prospective cohort studies.

Methods: There were 51,896 incident cases of PCa diagnosed among the 806,969 participants including 5,061 advanced stage (equivalent to T4, or N1, or M1, or fatal) and 9,414 high grade (equivalent to poorly differentiated, undifferentiated, or Gleason score ≥8) cancers. Using the primary data, we estimated study-specific relative risks (RR) using multivariable Cox proportional hazards models, and pooled the RRs using a random effects model. We tested for between-studies heterogeneity using the Q statistic. We examined PCa risk overall and by stage (advanced and localized [equivalent to T1 or T2 and N0M0]), and grade (high and low grade [equivalent to well or moderately differentiated or Gleason score <8]) at diagnosis; we also examined fatal PCa (n = 3,692). Calcium intake was modeled as quintiles and as specified categories (<500, 500-<700, 700-<900, 900-<1,100, 1, 100-<1,300, 1,300-<1,500, ≥1,500 mg/day).

Results: We saw an increased risk of PCa overall with high dietary (pooled RRQ5 vs. Q1: 1.08, 95% confidence interval [CI]: 1.04–1.12) and total (pooled RRQ5 vs. Q1: 1.08, 95% CI: 1.03–1.13) calcium intake. The results for localized PCa were similar to those observed for overall prostate cancer; pooled RRsQ5 vs. Q1 (95% CI) were 1.07 (1.03–1.10) for dietary calcium and 1.06 (1.02–1.09) for total calcium intake.

Nonsignificant associations were observed between calcium intake and risk of advanced stage PCa; the pooled RRsQ5 vs. Q1 (95% CI) were 1.05 (0.94–1.17) for dietary calcium intake and 1.05 (0.90–1.22) for total calcium intake. Neither of these results was significantly different from those observed for localized PCa (p-value, test for common effects >0.50). There was also no significant association between calcium intake and risk of fatal PCa; the pooled RRsQ5 vs. Q1 (95% CI) were 1.03 (0.94–1.13) for dietary calcium, and 1.00 (0.80–1.25) for total calcium intake. In categorical analyses using common cutpoints comparing ≥1500 vs. <500 mg/day, no significant associations were observed between total calcium intake and advanced stage (pooled RR = 0.98, 95% CI 0.81–1.18) or fatal (RR 1.04, 95% CI 0.83–1.31) PCa.

Results for low grade PCa for both dietary and total calcium intakes were similar to those observed for localized PCa. Nonsignificant associations were observed for both dietary (pooled RRQ5 vs. Q1 1.06, 95% CI: 0.99–1.13) and total (pooled RRQ5 vs. Q1: 1.08, 95% CI: 0.99–1.16) calcium intake and risk of high grade PCa. Associations for dietary and total calcium intake did not differ significantly by grade (p-value, test for common effects for quintile 5 >0.50).

There was no evidence of between-studies heterogeneity in any of the analyses conducted for dietary or total calcium intake (p-value, test for between-studies heterogeneity for highest category ≥0.15).

Conclusion: These preliminary findings from our pooled analyses of 14 cohort studies suggest that high calcium intake may modestly increase the risk of prostate cancer, regardless of stage or grade. We did not find an increased risk of high grade or advanced stage prostate cancer with high calcium intake.

Citation Information: Cancer Prev Res 2011;4(10 Suppl):B104.