Dietary Supplement Use and Prostate Cancer Risk in the Carotene and Retinol Efficacy Trial

Dietary supplements are used by more than half of American adults (1-4). Although the motivations for supplement use vary, many individuals use these products believing that their health will be improved and risk for chronic diseases, including cancer, will be reduced (5, 6). Because dietary supplements are sold over the counter and are largely unregulated, a common opinion is that these preparations are beneficial or at the very least innocuous. However, accumulating evidence suggests that dietary supplement use may increase, instead of decrease, risk of several cancers, including prostate cancer (7, 8).

The Cancer Prevention II study showed that men with 5 or more years of multivitamin use had a 30% increased risk of fatal prostate cancer compared with nonusers (7). Similarly, the AARP study reported that among male AARP members ages 50 to 71 years, those using excessive multivitamins (more than seven times per week or the equivalent of more than one type of multivitamin per day) had an increased risk of advanced and fatal prostate cancer {relative risk (RR), 1.32 [95% confidence interval (95% CI), 1.04-1.67] and 1.98 (95% CI, 1.07-3.66), respectively; ref. 8}. More recently, Selenium and Vitamin E Clinical Trial (SELECT), a double-blind, placebo-controlled, randomized trial testing supplemental selenium, vitamin E, both or neither for the primary prevention of prostate cancer was halted early due to the interim finding that the vitamin E supplements may increase, instead of decrease, prostate cancer risk (9). In this report, we examined associations of dietary supplement use with prostate cancer risk in the Carotene and Retinol Efficacy Trial (CARET) during both the active intervention and the postintervention follow-up period.

CARET was a multicenter randomized, double-blind, placebo-controlled chemoprevention trial testing whether daily supplementation with 30 mg β-carotene + 25,000 IU retinyl palmitate would reduce lung cancer risk among 18,314 heavy smokers, former heavy smokers, and asbestos-exposed workers. The intervention ended 21 mo early in January 1996 when interim analysis found evidence that the supplements increased the risk of lung cancer and total mortality in this high-risk population by 28% and 17%, respectively (10). Although the CARET intervention ended in 1996, active follow-up of all participants continued until 2005, including the collection of end point data. CARET had excellent retention and follow-up; 94% of participants remained in active follow-up until death or within 2 y of CARET closeout.

Eligible CARET participants were men and women ages 50 to 69 y who were current or former smokers (within the previous 6 y) with a history of at least 20 pack-years of cigarette smoking (n = 14,254, 55.9% of whom were male) and men ages 45 to 69 y (n = 4,060) who were current or former smokers (within the previous 15 y) with occupational exposure to asbestos within the previous 15 y. CARET participants were recruited from health insurance rolls, managed care organizations, labor unions, workmen's compensation programs, and occupational physicians. Participants attended yearly clinic visits and completed follow-up telephone calls every 4 mo throughout the intervention period (11, 12). Other details about the design and primary results of CARET are published (10, 12, 13). The Institutional Review Board of the Fred Hutchinson Cancer Research Center and each of the five other participating institutions approved all procedures for the study, and participants provided written informed consent. This report is restricted to male participants.

At baseline and each annual clinic visit, CARET participants were queried about use of dietary supplements. Participants were permitted to use supplemental vitamins during the intervention but were asked to limit use to products containing <5,500 IU/d retinol and no β-carotene. Staff recorded the name of the supplement bottles brought to the clinic and transcribed the doses for vitamin A (retinol), vitamin E, and β-carotene. If any supplement was found to contain >5,500 IU/d retinol or any β-carotene, staff counseled participants appropriately about the CARET protocol. During the years of the CARET intervention, multivitamins were the most common supplement used by American men (2, 14). For these analyses, dietary supplement use was defined as any supplement use versus no use at baseline.

Age, sex, race/ethnicity, education, personal health history, height, and weight were collected at each participant's first CARET clinic visit and updated annually. Body mass index (BMI) was calculated as weight (kg)/height (m)². Family history of prostate cancer was defined as at least one first-degree relative with prostate cancer. Detailed data were collected on current smoking status and smoking history, including age at smoking initiation, total years smoked, and average number of cigarettes smoked per day.

Health status, including any cancer diagnosis, was updated up to four times/year during the active intervention and annually thereafter. Outcomes were reviewed by the CARET Endpoints Committee; prostate cancer was a secondary end-point of the trial. For this study, we conducted an augmented review of medical records and Surveillance Epidemiology and End Results cancer registry files to obtain Gleason score and stage of disease. Approximately 92% of the prostate cancer cases were adjudicated and confirmed. “Aggressive prostate cancer” was defined as Gleason score ≥7 and/or stage III/IV (distant) at diagnosis (15). All other prostate cancer cases with stage and grade available were considered “nonaggressive.” Prostate-specific antigen (PSA) data were not available to distinguish screen-detected versus clinically detected cases because widespread PSA screening was not in use when CARET began. Noncases included all other men in the cohort, except those excluded due to a history of prostate cancer at baseline (n = 25). CARET had only n = 46 prostate cancer–specific deaths in CARET, so fatal prostate cancer is not used as an outcome in this report.

We used Cox proportional hazards regression to estimate RR (and 95% CIs) of prostate cancer for dietary supplement users compared with nonusers. Models included time-dependent covariates and cross-product interaction terms to allow for separate estimates by follow-up phase (intervention or postintervention phase) and intervention arm assignment (CARET study vitamins or placebo). Analyses restricted to aggressive and nonaggressive disease cases were done separately to assess the extent to which the associations of supplemental vitamin intake on prostate cancer risk varied by disease stage and grade classification. All models included adjustment for age, race, BMI, family history of first-degree relative with prostate cancer, and smoking.

Table 1 gives the baseline demographic and health characteristics of male CARET participants by intervention arm assignment. Age, race, BMI, use of supplemental vitamins, and family history of prostate cancer were similar across the CARET intervention and placebo groups. As a proportion of the total prostate cancer cases, more men who were randomized to the active arm developed high-grade prostate cancer (Gleason 7-10) than in the placebo arm (44.6% versus 40.1%, respectively).

Table 2 gives results for total, aggressive cases, and nonaggressive prostate cancer cases in the intervention and postintervention phases. The referent group in these models is the placebo group taking no supplemental vitamins. There were no remarkable associations for total prostate cancer in either the intervention or postintervention phases. For aggressive prostate cancer, men using a dietary supplement at baseline who were randomized to the CARET intervention arm had a RR for aggressive prostate cancer of 1.36 (95% CI, 0.87-2.13) relative to the placebo arm/no supplement use. Although this finding was not statistically significant, the decline to a postintervention RR = 0.68 (95% CI, 0.45-1.04) provides a modest suggestion that withdrawal of the combination of CARET vitamins plus other vitamins may have biological relevance. For nonaggressive prostate cancer, men in the active CARET arm not using other supplements had a statistically significant 35% reduced risk of cancer compared with the referent, but there were no other significant associations of the CARET vitamins (with or without other supplements) with nonaggressive prostate cancer.

Our a priori hypothesis was that excessive supplements would increase prostate cancer risk; in particular, those who took dietary supplements at baseline who also took the high-dose β-carotene + retinyl palmitate CARET supplements might be at the highest risk because they received the largest combined doses. Thus, we combined all men in the placebo arm plus men in the active arm who did not use supplements at baseline (referent) and compared the prostate cancer risk of men in the active arm who also used dietary supplements (highest risk group) with the combined referent group (Table 2). Compared with the referent, men taking the CARET study vitamins plus another dietary supplement had a prostate cancer RR of 1.52 (95% CI, 1.03-2.24; P = 0.03), which declined to 0.75 (95% CI, 0.51-1.09) in the postintervention period. In these models, the findings were restricted to aggressive cancers as there were no associations with either total or nonaggressive prostate cancers.

Additional analyses tested whether smoking pack-years was an effect modifier of the relationship between supplement use and prostate cancer risk. In analyses testing interaction of pack-years with supplement use (with and without the CARET vitamins), results did not support effect modification of the main effects by smoking duration or intensity (data not shown). Other analyses examined whether time on the CARET intervention influenced results. We fit models that included time-dependent variables representing 2-year increments during and after the CARET intervention, but evidence did not support an increased prostate cancer risk according to time on intervention (data not shown).

In this randomized controlled trial of β-carotene + retinyl palmitate, men who used the CARET study vitamins plus another dietary supplement had a 52% increased risk of aggressive prostate cancer compared with all men in the placebo arm or those who used only the CARET vitamins. This risk diminished on discontinuation of study vitamins. Any suggestion of increased risk due to supplement use was for aggressive disease only and was only observed during the active CARET intervention. Conversely, a modest protective association for nonaggressive disease was observed among men using only the CARET vitamins during the active intervention. Because aggressive disease has a poor prognosis, whereas the clinical relevance of nonaggressive disease is less certain, the results for increased risk of aggressive disease have potential clinical relevance.

The observed differences in risk between the CARET intervention phase (when risk of aggressive disease was increased) and the postintervention phase (when risk of aggressive disease was decreased) are consistent with the growing evidence that the effects of antioxidants on cellular processes may vary depending on the oxidative milieu of the tissues (16). Similar effects were noted in a randomized trial of antioxidant vitamins to prevent second primary cancers in head and neck cancer, wherein there was an increased risk of recurrence or second primary cancer during the supplementation period [hazard ratio (HR), 2.88; 95% CI, 1.56-5.31], but a lower risk after supplementation was discontinued (HR, 0.41; 95% CI, 0.16-1.03; ref. 17). Under this proposed mechanism, we are unclear, though, why the use of CARET vitamins resulted in a 35% reduced risk for nonaggressive prostate cancer. Clearly, much remains to be learned about the effects of antioxidant supplements on normal and neoplastic cells. Taken together, our results add to the growing body of evidence that dietary supplements, particularly multiple supplements or those used in large pharmacologic doses, should be used with caution.

Our results are consistent with some, but not all, previously published reports. Two cohort studies have reported a finding of increased risk of fatal or aggressive prostate cancer among dietary supplement users compared with nonusers (7, 8). One of these studies, the AARP cohort, reported the greatest risk among men using excessive multivitamins or more than one preparation per day. From reports of randomized, controlled trials, the Alpha-Tocopherol Beta-Carotene (ATBC) Cancer Prevention Study reported a lower risk of prostate cancer in men randomized to the α-tocopherol arm of the trial, no effect for the α-tocopherol plus β-carotene arm, and a slight nonsignificant increase in prostate cancer risk for the β-carotene only arm (RR, 1.23; 95% CI, 0.89-1.70; refs. 18, 19). Consistent with the findings we present from CARET, any β-carotene–related excess risk disappeared in the post–ATBC follow-up period (19).

There are strengths to this study. CARET was a randomized controlled trial where participants came to annual clinic visits with additional contacts by phone. At these visits, health status and personal lifestyle habits were updated using a uniform protocol across all CARET clinical centers to ensure standardized data collection. Another strength is that there was excellent follow-up of all CARET participants during and after the trial with careful adjudication of 92% of all prostate cancer cases. There are also limitations. Because our data only captured information on whether participants used dietary supplements, we are unable to investigate which particular nutrients or combinations of nutrients (especially in high doses) may place persons at risk. Additionally, we have no data on duration of supplement use, which at least one study has suggested a decrease in prostate cancer risk when supplement use is 10 years or more in duration (7). Another limitation is that we were underpowered to examine prostate cancer deaths, an end point that was associated with dietary supplement use in other cohorts (7, 8). Although our classification of “aggressive disease” included Gleason ≥7 or stage III/IV at diagnosis (both highly predictive of poor outcome), most future cohorts will have few cases diagnosed at later stages due to prostate-specific antigen screening and early detection programs (20). For comparisons across studies, use of Gleason alone might be beneficial for newer studies. Further, despite CARET's design (randomized controlled trial), we are unable to completely rule out residual confounding, particularly for variables measured with imprecision or those not assessed at all in CARET. For example, diabetes is inversely associated with prostate cancer risk, but diabetes data were not collected in CARET (21). Other lifestyle variables, such as energy intake and physical activity, may increase and decrease prostate cancer risk, respectively, but these self-report data are well known to be subject to measurement error (22). Finally, the CARET cohort was composed of heavy smokers and former heavy smokers. Although the results from this study may not be generalizable to all U.S. males, our findings are applicable to the millions of men who are current or former smokers.

In summary, our results provide modest evidence that a high dose of β-carotene (30 mg/d) and retinyl palmitate (25,000 IU/d) plus at least one other dietary supplement may increase risk of aggressive prostate cancer. These results are consistent with other recently published findings, including those using various high-dose supplements in relation to risk of other cancers or precancerous lesions (23). Men may want to carefully consider whether they should use a supplement, especially if they have existing risk factors for prostate cancer.

No potential conflicts of interest were disclosed.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.