Re: Prospective Studies of Dairy Product and Calcium Intakes and Prostate Cancer Risk: A Meta-Analysis

In their meta-analysis, Gao et al. ( 1 ) concluded that “High intake of dairy products and calcium may be associated with an increased risk of prostate cancer, although the increase appears to be small.” We report findings from the Melbourne Collaborative Cohort Study (MCCS), a prospective study of 41 528 people (17 049 men) who were 27–75 years old at baseline (99.3% were 40–69 years old). As previously described ( 2 , 3 ) , volunteers were recruited from November 26, 1990, to October 25, 1994, from among the residents of Melbourne, Australia. At baseline, subjects completed a questionnaire about potential risk factors for cancer as well as a food frequency questionnaire (FFQ) that was developed specifically for the MCCS ( 4 ) . In the FFQ, the dairy products category included milk, cheese, yogurt, cream, and custard. Butter and margarine were categorized separately. Calcium and energy intakes were calculated by using data from the FFQ, the mean sex-specific portion sizes from weighed food records, and Australian food composition tables ( 5 ) . Calcium density was calculated by dividing calcium intake by the total energy intake. After excluding 2407 men who did not complete the FFQ ( n = 35), who had a confirmed diagnosis of prostate cancer prior to baseline ( n = 105), who reported having a previous medical condition (i.e., self-reported angina, myocardial infarction, diabetes, hypertension, cancer, or stroke; n = 1928), or whose reported energy intake was in the lowest or highest 1% of the sex-specific distribution ( n = 339), 14 642 men were available for analysis. Incident cases of prostate cancer were ascertained through the state cancer registries in Australia; deaths were identified from state death records for Victoria and the Australian National Death Index. By the end of 2004, 23 men (fewer than 0.2%) were known to have left Australia and were considered lost to follow-up. Follow-up began at baseline and continued until the date of diagnosis of prostate cancer, of death, or of emigration from Australia or December 31, 2004, whichever came first.

During an average of 10.9 years of follow-up, we identified 674 prostate cancers among the eligible subjects, including 107 cases of aggressive prostate cancer [defined as a Gleason score >7 or stage T4, N+, or M+ ( 6 ) ]. Table 1 presents the hazard ratios for total, nonaggressive, and aggressive prostate cancer from the Cox proportional hazards models ( 7 ) . Tests based on Schoenfeld residuals showed no evidence that proportional hazard assumptions were violated. All statistical tests were two-sided. The hazard ratios for total prostate cancer ranged from 0.87 to 1.28, and all but one 95% confidence interval included 1. There was no evidence of a dose–response relationship between intakes and prostate cancer risk (all P_trend ≥.1).

To examine whether prevalent cancers affected associations between intakes and prostate cancer risk, we included a time-dependent covariate in the Cox regression model. All hazard ratios for the first 2 years of follow-up were higher than the corresponding hazard ratios for the remainder of the follow-up period, although only the difference for butter intake was statistically significant ( P = .03). The hazard ratios for aggressive prostate cancer were generally lower than those for nonaggressive disease, but the differences were not statistically significant (all P ≥.1). We conclude that our findings are not consistent with any association between dairy product, butter, margarine, and calcium intakes and the risk of prostate cancer. The conclusions of the meta-analysis by Gao et al. were based on a very small effect size and relatively weak statistical evidence, and we question whether they would change if our data were included.

References