SEARCH

SEARCH BY CITATION

Keywords:

  • red wine;
  • alcoholic beverages;
  • prostate cancer;
  • cohort study

Abstract

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

In light of recent, strong inverse findings between lifetime red wine consumption and prostate cancer among younger men, we revisited our previous cohort analysis to more thoroughly investigate red wine consumption and prostate cancer in the Health Professionals Follow-up Study (HPFS). In 1986, HPFS participants reported their average consumption of red wine, white wine, beer and liquor during the past year, and their change in alcohol consumption over the prior 10 years. Prostate cancer diagnoses were ascertained on each biennial questionnaire and confirmed by medical record review. Between 1986 and 2002, 3,348 cases of prostate cancer were diagnosed among 45,433 eligible participants. Using men who did not consume red wine as the reference, no linear trend was observed between red wine consumption and prostate cancer in the full analytic cohort (p-trend = 0.57). Among men with unchanged alcohol consumption in the prior 10 years, and those additionally <65 years of age, slightly lower risks were observed for men who consumed ≤4 glasses of red wine/week, whereas null or slight increased risks were observed for men who consumed ≫4 glasses/week, resulting in a lack of linear trend. These findings suggest that red wine does not contribute appreciably to the etiology of prostate cancer. © 2006 Wiley-Liss, Inc.

While many epidemiologic studies have investigated associations between alcohol consumption and prostate cancer with largely null results,1, 2 fewer have investigated associations between specific alcoholic beverages, such as red wine,3, 4, 5 and prostate cancer. This focus of investigation was recently highlighted by Schoonen et al.4 in a large, population-based case–control study of prostate cancer conducted among younger men (40–64 years of age). In that study, the authors observed a statistically significant inverse trend between lifetime red wine consumption and prostate cancer, which was particularly strong for aggressive disease (odds ratio = 0.39, 95% confidence interval (CI): 0.16–0.97 for ≥8 glasses of red wine/week). When we previously investigated associations between alcohol and prostate cancer in the Health Professionals Follow-up Study (HPFS), we observed slight, nonstatistically significantly increased risks for baseline total alcohol, white wine, beer and liquor consumption, but no association for red wine.5 However, in light of Schoonen and colleagues' findings and the suggested anticancer properties of free and conjugated red wine polyphenols in in vitro and animal experiments,6, 7, 8 we revisited this question using more detailed red wine consumption characterization, a longer follow-up period, and specific investigations among younger men.

Material and methods

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Study population

The HPFS is an ongoing, prospective study of 51,529 American male health professionals 40–75 years of age at enrollment in 1986. At that time, participants completed a baseline epidemiologic questionnaire on demographics, lifestyle and medical history, and a 131 item semi-quantitative food frequency questionnaire. Since 1986, participants have completed questionnaires every 2 years to update exposure and disease status, and food frequency questionnaires every 4 years to update diet. The average response rate to follow-up questionnaires is 94% per follow-up cycle. Information on death was obtained from the National Death Index, U.S. Postal Service, or next of kin in response to mailed biennial questionnaires. We estimate that 98% of deaths have been ascertained by these methods.

In the present analysis, we excluded participants diagnosed with cancer (except nonmelanoma skin cancer) as of the date of return of the baseline questionnaire (4.3%), and those who provided invalid or incomplete baseline food frequency (3.0%) or alcohol information (4.6%). After these exclusions, 45,433 participants remained in the analytic cohort. This study was approved by the Human Subjects Committee at the Harvard School of Public Health and the Committee on Human Research at the Johns Hopkins Bloomberg School of Public Health.

Assessment of alcohol consumption

On the baseline and 1990 food frequency questionnaires, participants were asked to report their average frequencies of consumption of beer (1 glass, bottle, can), red wine (4 oz glass), white wine (4 oz glass) and liquor, e.g., whiskey, gin, etc. (1 drink or shot) during the past year: never, or less than once per month; 1–3 per month; 1 per week; 2–4 per week; 5–6 per week; 1 per day; 2–3 per day; 4–5 per day and 6+ per day. On the 1994 and 1998 food frequency questionnaires, participants were additionally asked to report their average consumption of light beer (1 glass, bottle, can) using the same frequency categorization. We have previously observed this method of alcohol consumption assessment to be both valid (r = 0.69 for wine consumption assessed on the baseline food frequency questionnaire vs. 2 one-week diet records) and reproducible (r = 0.87 for wine consumption assessed on food frequency questionnaires completed 1 year apart) in the HPFS cohort.9 On the baseline questionnaire, participants were additionally asked whether their use of alcohol had decreased, increased or stayed the same over the past 10 years.

To allow for comparisons across beverage type, we converted frequencies of consumption of specific portion sizes to grams of ethanol/day by multiplying daily beverage-specific values of consumption by their respective gram of ethanol equivalents: regular beer, 12.8 g; light beer, 11.3 g; red wine, 11.0 g; white wine, 11.0 g and liquor, 14.0 g.10 We then summed these values to obtain total daily ethanol consumption. For ease of description, we re-categorized daily consumption of specific alcoholic beverages to correspond to glasses of wine/day: 0 g, never or less than 1 glass/month; 0.01–1.31 g/day, 1–3 glasses/month; 1.32–2.41 g/day, 1 glass/week; 2.42–7.03 g/day, 2–4 glasses/week; 7.04–16.49 g/day, 5–7 glasses/week and ≥16.5 g/day, >1 glasses/day. We defined nondrinkers as men who reported no alcohol consumption in the year prior to completion of the baseline questionnaire and unchanged alcohol consumption in the prior 10 years, and former drinkers as men who reported no alcohol consumption in the year prior to completion of the baseline questionnaire and a decrease in alcohol consumption over the prior 10 years.

Identification of prostate cancer cases

On each biennial follow-up questionnaire, participants were asked to report medical diagnoses, including prostate cancer, during the past 2 years. Men who reported prostate cancer (or their next of kin in the case of fatal prostate cancer) were then asked for permission to review relevant portions of their medical records. Of those whose medical records and/or pathology reports could be retrieved (90%), 99% supported a diagnosis of prostatic adenocarcinoma. Additionally, most of the remaining 10% of cases whose medical records could not be reviewed provided supporting evidence for their diagnosis (e.g., evidence of treatment). Therefore, because of the observed high level of reporting accuracy in this cohort, all prostate cancer diagnoses were included in the analyses. Overall, we estimate that 96% of prostate cancer diagnoses have been ascertained in this cohort. Disease stage (TNM classification) and Gleason sum were abstracted from participants' medical records by study investigators blinded to exposure status. We did not include participants with stage T1a prostate cancers (n = 71) as cases because their tumors are detected, by definition, at transurethral resection of the prostate for benign prostatic hyperplasia, and may be especially prone to detection bias.

Statistical analysis

To describe participants and investigate potential confounding, we calculated age-standardized means and proportions of suspected correlates of red wine consumption and previously identified risk factors for prostate cancer in the HPFS by categories of red wine consumption. We investigated associations between red wine consumption and prostate cancer using Mantel-Haenszel age-adjusted (5-year age intervals) rate ratios. Participants contributed person-time from the month of return of the 1986 questionnaire to the month of diagnosis of prostate cancer, death or end of the study period on January 31, 2002. We used Cox proportional hazards regression to estimate multivariable-adjusted associations between red wine consumption and prostate cancer. The reference group in multivariable models was men who did not consume red wine (although they could consume other types of alcoholic beverages). We considered as covariates those found to be positively or inversely associated with red wine consumption (Ashkenazi Jewish heritage and use of olive oil as a main cooking oil), those previously observed to be associated with prostate cancer incidence or progression in the HPFS (race/ethnicity (Southern European, Scandinavian, other Caucasian, African-American, Asian and other), cumulative family history of prostate cancer through 1996, height (cm), body mass index at age 21 (kg/m2), updated cigarette smoking in the past 10 years (pack-years), baseline intakes of total energy (kcal/day), tomato sauce (servings/week), red meat (servings/day), fish (servings/day), calcium (mg/day) and vitamin E (<15, ≥15 mg/day), baseline energy-adjusted intakes of fructose (g/day) and α-linolenic acid (g/day), vigorous leisure-time physical activity (metabolic equivalent-hours/week) in 1986, and updated vasectomy and diabetes mellitus type 2 status), and other alcoholic beverages (white wine, beer and liquor consumption (g/day)). Age (1-year intervals) and calendar time (2-year intervals) were controlled for as stratification variables in all regression models.

We used baseline red wine consumption in the main analyses. Additional analyses using simple-updated (i.e., time-varying) and cumulative-average updated (i.e., time-varying cumulative average) values were also performed. We investigated associations between longer-term red wine consumption and prostate cancer by performing analyses restricted to men who reported unchanged alcohol consumption in the 10 years prior to baseline, and men who additionally maintained their baseline level of consumption over the course of follow-up (allowing their consumption to vary within ±1 drinking category, and assessed in a time-varying manner). These men were selected because their baseline values were hypothesized to correlate more strongly with longer-term or lifetime consumption than those from men with varying consumption over time. We also performed analyses restricted to men <65 years of age to allow for comparisons with Schoonen et al.'s findings (conducted among men exclusively <65),4 and because of previously observed heterogeneity of findings by age.11, 12, 13, 14, 15, 16 In all analyses, linear trends between red wine consumption and prostate cancer risk were investigated by including a categorical term for red wine consumption in the regression model.

To investigate potential effect modification, we performed stratified analyses by prostate cancer screening (digital rectal examination or prostate specific antigen test), diabetes mellitus type 2 status, region of residence (West coast, non-West coast), family history of prostate cancer, smoking in the 10 years prior to baseline and Southern European heritage. We investigated potential confounding by infrequent consumption of greater amounts of alcohol, which was associated with a significantly increased risk of prostate cancer in our previous cohort analysis,5 by excluding men who regularly consumed ≥77 g of alcohol (cut-point equivalent to the amount of ethanol in 7 glasses of wine) on 1–2 days/week. To distinguish the effects of red wine consumption from concomitant consumption of other alcoholic beverages, we also performed multivariable-adjusted analyses including joint effect terms for red wine consumption (≤4 glasses/week, >4 glasses/week) and consumption of other alcoholic beverages (<7, ≥7 g/day, cut-point roughly equivalent to half an alcoholic beverage/day).

The main outcome for each analysis was total prostate cancer. Additional outcomes considered were organ-confined (≤T2 and N0M0), advanced (T3b or worse), low-grade (Gleason sum < 7) and high-grade (Gleason sum ≥ 7) prostate cancer. All statistical analyses were performed using SAS version 9 (SAS Institute, Cary, NC). Reported p-values are 2-sided.

Results

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Between 1986 and 2002, 3,348 cases of prostate cancer were diagnosed among 45,433 eligible participants. Ten percent of participants reported no current or prior alcohol consumption (nondrinkers), 13.5% reported prior alcohol consumption only (former drinkers), 35.4% reported no current red wine consumption but current consumption of other alcohol beverages and 40.7% reported current red wine consumption (Table I). Men who reported greater red wine consumption were more likely to be of Southern European or Scandinavian ethnicity, less likely to be of African-American, Asian or other non-Caucasian race/ethnicity, less likely to be Ashkenazi Jewish and more likely to report higher intakes of total energy (partially due to consumption of alcohol), white wine, beer, liquor, tomato sauce and fish, lower intakes of fructose and α-linolenic acid and more use of olive oil as a main cooking oil at home. They were also more likely to have had a vasectomy. Moderate red wine drinkers were more likely to report health-conscious behaviors (less cigarette smoking in the past 10 years and more prostate cancer screening) than light or heavier red wine drinkers.

Table I. Age-standardized characteristics by categories of red wine consumption in the Health Professionals Follow-Up Study, 1986
 NondrinkerFormer drinkerNever or <1 glass/mo1–3 glasses/mo1 glass/wk2–4 glasses/wk5–7 glasses/wk>1 glass/day
  • 1

    Assessed in 1998.

  • 2

    Assessed in 1990 through 1996.

  • 3

    Adjusted for total energy intake.

  • 4

    Assessed through 2002.

No. of participants4,6816,15616,0839,6164,4223,0841,070321
Mean age (yr)53.355.655.153.352.553.353.654.2
Race/ethnicity (%)
 Southern European22.023.622.624.125.325.826.926.2
 Scandinavian12.310.29.89.810.610.711.79.5
 Other Caucasian55.354.058.657.756.655.854.659.0
 African-American0.91.51.01.00.50.70.30.0
 Asian2.83.31.51.40.90.90.80.7
 Other6.67.36.56.06.16.05.84.6
Ashkenazi Jewish (%)19.99.612.212.612.410.07.37.6
Family history of prostate cancer2(%)13.411.811.712.111.612.911.311.5
Mean height (cm)177.9177.7178.1178.2178.2178.4178.0177.6
Mean body mass index at age 21 (kg/m2)23.023.223.023.023.023.023.122.9
Smoked cigarettes in the past 10 yr (%)9.821.025.322.221.019.224.530.4
Mean intakes of:
 Total energy (kcal/day)1,9441,9111,9362,0502,0902,0812,1162,248
 Total energy without the contribution  from alcohol (kcal/day)1,9941,9111,8481,9611,9801,9351,9071,868
 White wine (g/day)0.00.01.71.92.74.15.58.8
 Beer (g/day)0.00.04.74.24.95.26.26.4
 Liquor (g/day)0.00.06.15.86.66.98.38.8
 Tomoto sauce (servings/wk)1.41.41.41.61.71.71.71.8
 Red meat (servings/day)0.970.890.961.000.990.960.950.98
 Fish (servings/day)0.260.320.340.350.370.370.380.38
 Fructose (g/day)354.954.947.648.847.445.843.640.2
 Calcium (mg/day)925929858891894865882875
 Alpha-linolenic acid (g/day)31.141.111.071.061.061.041.000.93
Use of olive oil as a main cooking oil at home (%)1.63.02.94.05.68.610.714.2
Vitamin E intake (<15 mg/day,%)34.341.739.238.639.640.141.738.1
Regular (≥2 times/wk) use of nonsteroidal anti-inflammatory drugs (%)26.632.333.832.433.233.937.134.0
Any vigorous leisure-time physical activity (metabolic equivalent-hours/wk)10.011.913.012.714.115.314.614.3
Screening prostate specific antigen test (%)478.974.880.081.683.082.582.077.8
Screening digital rectal examination (%)483.382.386.087.088.088.187.384.4
Vasectomy (%)16.519.922.421.821.724.223.628.6
Diabetes mellitus type 2 (%)3.56.42.82.21.81.91.41.8

Full cohort analyses

No linear trends were observed between baseline red wine consumption and prostate cancer in age-adjusted analyses (p-trend = 0.27), analyses further adjusted for correlates of red wine consumption and prostate cancer risk factors (p-trend = 0.35) and analyses additionally adjusted for other alcoholic beverages (p-trend = 0.57, Table II). Null results were also observed in simple- and cumulative-average updated analyses. Exclusion of nondrinkers did not alter any of the results (data not shown).

Table II. Baseline specific alcoholic beverage consumption and prostate cancer in the Health Professionals Follow-up Study, 1986–2002
Alcohol consumption (g/day)Cases/person-timeAge-adjusted RR1 (95% CI)Multivariable-adjusted RR2 (95% CI)Multivariable-adjusted RR3 (95% CI)
  • 1

    RR, relative risk; CI, confidence interval.

  • 2

    Adjusted for age, race ethnicity (Southern European, Scandinavian, other Caucasian, African-American, Asian and other), body mass index at age 21 (kg/m2, continuous), cumulative family history of prostate cancer through 1996 (no/yes), height (cm, <80th percentile, ≥80th percentile), updated cigarette smoking in the past 10 yr (none, 1–5, 6–10, 11–20 and >20 pack-yr), baseline intakes of total energy (kcal/day, continuous), tomato sauce (servings/wk, <75th percentile, ≥75th percentile), red meat (servings/day, continuous) fish (servings/day, <75th percentile, ≥75th percentile), calcium (<500, 500–749, 750–999, 1,000–1,499, 1500–1,999 and ≥2,000 mg/day) and vitamin E (<15, ≥15 mg/day), baseline energy-adjusted intakes of fructose (g/day, <80th percentile, ≥80th percentile) and α-linolenic acid (g/day, quintiles), baseline vigorous physical activity (metabolic equivalent-hrs/wk, ordinal quintiles) and updated diabetes mellitus type 2 (no/yes) and vasectomy (no/yes) status.

  • 3

    Additionally adjusted for all other specific alcoholic beverage types (none, 0.01–1.31, 1.32–2.41, 2.42–7.03, 7.04–16.4 and ≥16.5 g/day).

  • 4

    All regression models include a term for former drinkers; RR =1.08, 95% CI:0.96–1.22 for former drinkers in fully-adjusted models. Categories were determined by calculating the midpoint between each successive drinking frequency and multiplying this value by 11 g (the amount of ethanol in one glass of red wine).

Red wine consumption4
 No red wine consumption1,598/292,4561.001.001.00
 0.01–1.31 (1–3 glasses/mo)655/138,2170.95 (0.87–1.04)0.95 (0.87–1.04)0.91 (0.82–1.01)
 1.32–2.41 (1 glass/wk)300/63,7871.00 (0.88–1.13)0.98 (0.87–1.11)0.89 (0.77–1.03)
 2.42–7.03 (2–4 glasses/wk)232/44,3901.05 (0.92–1.21)1.04 (0.90–1.19)0.91 (0.78–1.07)
 7.04–16.4 (5–7 glasses/wk)90/15,2691.16 (0.94–1.43)1.16 (0.94–1.44)1.12 (0.89–1.40)
 ≥16.5 (>1 glass/day)27/4,5661.11 (0.76–1.62)1.12 (0.76–1.65)1.06 (0.72–1.56)
 p-trend 0.270.350.57
White wine consumption4
 No white wine consumption1,041/197,3031.001.001.00
 0.01–1.31 (1–3 glasses/mo)806/164,7001.00 (0.91–1.10)1.02 (0.93–1.12)1.05 (0.94–1.17)
 1.32–2.41 (1 glass/wk)425/86,2021.09 (0.97–1.22)1.08 (0.96–1.22)1.12 (0.98–1.28)
 2.42–7.03 (2–4 glasses/wk)430/73,4151.20 (1.07–1.34)1.19 (1.06–1.33)1.22 (1.06–1.39)
 7.04–16.4 (5–7 glasses/wk)151/29,0361.03 (0.87–1.22)1.04 (0.87–1.23)1.02 (0.84–1.22)
 ≥16.5 (>1 glass/day)49/8,0281.23 (0.92–1.64)1.18 (0.88–1.58)1.19 (0.89–1.60)
 p-trend 0.00550.0130.031
Beer consumption4
 No beer consumption1,069/194,6411.001.001.00
 0.01–1.31627/114,4361.06 (0.96–1.17)1.05 (0.95–1.16)1.04 (0.94–1.16)
 1.32–2.41385/74,7841.12 (1.00–1.26)1.11 (0.98–1.25)1.08 (0.96–1.23)
 2.42–7.03450/92,5601.09 (0.98–1.22)1.06 (0.95–1.19)1.03 (0.92–1.16)
 7.04–16.4256/54,9561.12 (0.98–1.29)1.09 (0.95–1.26)1.07 (0.92–1.23)
 ≥16.5115/25,3071.11 (0.91–1.34)1.09 (0.89–1.33)1.09 (0.89–1.34)
 p-trend 0.0370.120.27
Liquor consumption4
 No liquor consumption1,007/222,2221.001.001.00
 0.01–1.31463/104,4301.01 (0.90–1.13)1.02 (0.91–1.14)1.01 (0.90–1.13)
 1.32–2.41284/56,3451.12 (0.98–1.27)1.10 (0.96–1.26)1.08 (0.94–1.23)
 2.42–7.03458/77,3781.14 (1.02–1.27)1.10 (0.98–1.24)1.08 (0.96–1.21)
 7.04–16.4399/59,7701.10 (0.98–1.24)1.09 (0.96–1.22)1.07 (0.94–1.21)
 ≥16.5291/38,5391.15 (1.01–1.31)1.12 (0.97–1.28)1.10 (0.96–1.27)
 p-trend 0.00330.0330.093
Total alcohol consumption4
 No alcohol consumption307/67,1221.001.00 
 0.01–1.31158/37,6160.86 (0.71–1.04)0.90 (0.74–1.09) 
 1.32–2.41205/42,9300.99 (0.83–1.19)1.02 (0.85–1.22) 
 2.42–7.03594/121,5961.09 (0.95–1.25)1.12 (0.97–1.29) 
 7.04–16.4840/151,6591.16 (1.01–1.32)1.16 (1.01–1.32) 
 ≥16.5798/137,7611.15 (1.01–1.31)1.14 (0.99–1.31) 
 p-trend 0.00030.0030 

Analyses restricted to men with unchanged alcohol consumption

Among men who reported unchanged alcohol consumption in the 10 years prior to baseline, no linear trends were observed between red wine consumption and prostate cancer in age-adjusted analyses, and in analyses additionally adjusted for correlates of red wine consumption and prostate cancer risk factors (Table III). After further adjustment for other alcoholic beverages, slight reduced risks of prostate cancer were observed for men who consumed ≤4 glasses of red wine/week, whereas no association or slight nonsignificantly increased risks were observed for those who consumed >4 glasses/week (Table III). A similar pattern was observed among men who additionally maintained constant red wine consumption over the course of follow-up (data not shown).

Table III. Baseline specific alcoholic beverage consumption and prostate cancer among men with unchanged alcohol consumption from 1976 to 1986 in the Health Professionals Follow-up Study, 1986–2002
Alcohol consumption (g/day)Cases/person-timeAge-adjusted RR1 (95% CI)Multivariable-adjusted RR2 (95 % CI)Multivariable-adjusted RR3 (95% CI)
  • 1

    RR, relative risk; CI, confidence interval.

  • 2

    Adjusted for age, race/ethnicity, body mass index at age 21, cumulative family history of prostate cancer through 1996, height, updated cigarette smoking in the past 10 years, baseline intakes of total energy, tomato sauce, red meat, fish, calcium and vitamin E, baseline energy-adjusted intakes of fructose and α-linolenic acid, baseline vigorous physical activity and updated diabetes mellitus type 2 and vasectomy status.

  • 3

    Additionally adjusted for all other specific alcoholic beverage types.

  • 4

    Categories were determined by calculating the midpoint between each successive drinking frequency and multiplying this value by 11 g (the amount of ethanol in one glass of red wine).

Red wine consumption4
 No red wine consumption964/175,2711.001.001.00
 0.01–1.31 (1–3 glasses/mo)341/74,7640.88 (0.78–1.000.87 (0.76–0.98)0.80 (0.69–0.92)
 1.32–2.41 (1 glass/wk)179/36,4671.0 (0.86–1.180.98 (0.83–1.15)0.84 (0.69–1.02)
 2.42–7.03 (2–4 glasses/wk)139/26,9150.99 (0.83–1.19)0.96 (0.80–1.16)0.79 (0.64–0.97)
 7.04–16.4 (5–7 glasses/wk)57/9,5841.09 (0.84–1.43)1.08 (0.82–1.43)1.01 (0.76–1.35)
 >16.5 (1 glass/day)19/2,8351.20 (0.76–1.89)1.26 (0.79–2.01)1.17 (0.73–1.87)
P-trend 0.680.890.17
White wine consumption4
 No white wine consumption637/126,8661.001.001.00
 0.01–1.31 (1–3 glasses/mo)421/84,5401.03 (0.91–1.16)1.02 (0.90–1.16)1.12 (0.97–1.30)
 1.32–2.41 (1 glass/wk)249/48,1111.14 (0.98–1.32)1.13 (0.97–1.32)1.24 (1.04–1.48)
 2.42–7.03 (2–4 glasses/wk)265/43,6311.25 (1.08–1.44)1.27 (1.09–1.48)1.42 (1.19–1.69)
 7.04–16.4 (5–7 glasses/wk)95/18,1071.05 (0.84–1.30)1.03 (0.82–1.29)1.07 (0.84–1.36)
 ≥16.5 (> 1 glass/day)32/4,5811.32 (0.93–1.89)1.26 (0.87–1.82)1.32 (0.90–1.91)
P-trend 0.00480.00960.0040
Beer consumption4
 No beer consumption542/125,9841.001.001.00
 0.01–1.31211/50,9330.97 (0.82–1.13)0.96 (0.82–1.13)0.96 (0.81–1.14)
 1.32–2.41155/32,6561.07 (0.90–1.28)1.07 (0.89–1.28)1.04 (0.86–1.27)
 2.42–7.03304/50,1321.17 (1.01–1.35)1.12 (0.97–1.30)1.11 (0.95–1.29)
 7.04–16.4290/41,3001.15 (0.99–1.33)1.10 (0.94–1.28)1.09 (0.93–1.28)
 > 16.5197/24,8301.19 (1.01–1.41)1.14 (0.95–1.36)1.13 (0.94–1.36)
P-trend 0.00530.0450.062
Total alcohol consumption4
 No alcohol consumption239/56,4661.001.00 
 0.01–1.3141/14,0540.64 (0.46–0.90)0.68 (0.48–0.95) 
 1.32–2.4178/16,7841.03 (0.80–1.33)1.06 (0.82–1.38) 
 2.42–7.03295/57,3021.15 (0.97–1.36)1.16 (0.97–1.38) 
 7.04–16.4502/91,1641.13 (0.97–1.32)1.09 (0.93–1.29) 
 P-trend 0.00340.030 

When only men <65 years of age with unchanged alcohol consumption in the 10 years prior to baseline were considered, no linear trends were observed between red wine consumption and prostate cancer in age-adjusted analyses, and in analyses adjusted for correlates of red wine consumption and prostate cancer risk factors (Table IV). When the analyses were further adjusted for other alcoholic beverages, slightly more pronounced reduced risks were observed for men who consumed ≤4 glasses of red wine/week, whereas no association or slight increased risks were observed for men who consumed greater than this amount (Table IV). Similar results were observed when only men who reported both unchanged alcohol consumption in the 10 years prior to baseline and routine prostate cancer screening were considered (relative risk (RR) adjusted for correlates of red wine consumption, prostate cancer risk factors and other alcoholic beverages = 0.74, 95% CI: 0.64–0.87 for ≤4 glasses of red wine/day; RR = 1.11, 95% CI: 0.82–1.51 for >4 glasses/week). No differences were observed when the analyses were stratified by history of type 2 diabetes mellitus, region of residence, family history of prostate cancer, smoking and Southern European heritage (all p-interaction >0.10), and when men who consumed larger quantities of alcohol over 1–2 days/week were excluded from the analyses (data not shown).

Table IV. Baseline specific alcoholic beverage consumption and prostate cancer among men <65 years of age with unchanged alcohol consumption from 1976 to 1986 in the Health Professionals Follow-up Study, 1986–2002
Alcohol consumption (g/day)Cases/person-timeAge-adjusted RR1 (95% CI)Multivariable-adjusted RR2(95% CI)Multivariable-adjusted RR3(95% CI)
  • 1

    RR, relative risk; CI, confidence interval.

  • 2

    Adjusted for age, race/ethnicity, body mass index at age 21, cumulative family history of prostate cancer through 1996, height, updated cigarette smoking in the past 10 yr, baseline intakes of total energy, tomato sauce, red meat, fish, calcium and vitamin E, baseline energy-adjusted intakes of fructose and α-linolenic acid, baseline vigorous physical activity and updated diabetes mellitus type 2 and vasectomy status.

  • 3

    Additionally adjusted for all other specific alcoholic beverage types.

  • 4

    Categories were determined by calculating the midpoint between each successive drinking frequency and multiplying this value by 11 g (the amount of ethanol in one glass of red wine).

Red wine consumption4
 No red wine consumption317/116,0891.001.001.00
 0.01–1.31 (1–3 glasses/mo)121/52,1980.86 (0.70–1.07)0.87 (0.70–1.08)0.79 (0.62–1.01)
 1.32–2.41 (1 glass/wk)70/26,5520.99 (0.77–1.29)0.94 (0.72–1.23)0.70 (0.52–0.96)
 2.42–7.03 (2–4 glasses/wk)43/18,7550.84 (0.61–1.15)0.82 (0.59–1.13)0.64 (0.45–0.93)
 7.04–16.4 (5–7 glasses/wk)20/6,4891.08 (0.69–1.70)1.12 (0.70–1.78)1.01 (0.62–1.64)
 ≥16.5 (>1 glass/day)7/1,8511.38 (0.65–2.91)1.41 (0.65–3.04)1.24 (0.57–2.74)
 p-trend 0.790.680.10
White wine consumption4
 No red wine consumption197/84,2431.001.001.00
 0.01–1.31 (1–3 glasses/mo)136/57,7731.01 (0.81–1.26)1.01 (0.81–1.26)1.16 (0.90–1.51)
 1.32–2.41 (1 glass/wk)111/34,7701.39 (1.10–1.76)1.34 (1.05–1.71)1.64 (1.23–2.20)
 2.42–7.03 (2–4 glasses/wk)92/29,9841.28 (1.00–1.64)1.26 (0.97–1.64)1.62 (1.20–2.21)
 7.04–16.4 (5–7 glasses/wk)31/12,1951.04 (0.71–1.52)1.02 (0.69–1.51)1.14 (0.75–1.74)
 ≥16.5 (>1 glass/day)11/2,9691.45 (0.79–2.66)1.33 (0.71–2.49)1.43 (0.75–2.74)
 p-trend 0.0220.0650.013
Beer consumption4
 No beer consumption202/79,4321.001.001.00
 0.01–1.31117/37,9321.20 (0.95–1.50)1.16 (0.92–1.47)1.17 (0.91–1.50)
 1.32–2.4183/27,4091.22 (0.94–1.57)1.15 (0.88–1.50)1.13 (0.85–1.50)
 2.42–7.0391/38,5050.97 (0.76–1.24)0.93 (0.72–1.21)0.92 (0.70–1.22)
 7.04–16.454/26,2490.87 (0.64–1.17)0.88 (0.64–1.20)0.88 (0.63–1.21)
 ≥16.531/12,4061.05 (0.72–1.53)1.10 (0.73–1.64)1.13 (0.76–1.70)
 p-trend 0.520.530.50
Liquor consumption4
 No liquor consumption229/93,6561.001.001.00
 0.01–1.3179/37,9170.85 (0.66–1.10)0.84 (0.65–1.10)0.80 (0.60–1.05)
 1.32–2.4169/23,7761.18 (0.90–1.54)1.12 (0.85–1.48)1.05 (0.78–1.41)
 2.42–7.0391/32,3861.04 (0.81–1.32)0.97 (0.75–1.24)0.92 (0.70–1.21)
 7.04–16.467/22,4100.99 (0.75–1.30)0.91 (0.69–1.22)0.88 (0.65–1.18)
 ≥16.543/11,7881.13 (0.82–1.58)1.18 (0.83–1.69)1.15 (0.80–1.66)
 p-trend 0.460.730.87
Total alcohol consumption4
 No alcohol consumption106/41,6921.001.00 
 0.01–1.3111/9,8410.43 (0.23–0.80)0.43 (0.23–0.81) 
 1.32–2.4126/11,8020.85 (0.55–1.30)0.83 (0.54–1.29) 
 2.42–7.03121/40,4871.15 (0.89–1.49)1.08 (0.83–1.42) 
 7.04–16.4162/60,8700.96 (0.75–1.23)0.89 (0.69–1.15) 
 ≥16.5152/57,2410.94 (0.74–1.21)0.90 (0.68–1.18) 
 p-trend 0.820.81 

Too few men drank exclusively red wine to examine associations between red wine consumption and prostate cancer without the potential influence of other alcoholic beverages. However, sufficient numbers of men drank red wine and only smaller quantities of other alcoholic beverages (<half a glass/day) to investigate the joint effects of red wine and nonred wine alcohol consumption. Using men who did not consume red wine as the reference, similar associations were observed for men who consumed ≤4 glasses of red wine/week and <half a glass of alcohol (<7 g) from sources other than red wine/day (RR adjusted for correlates of red wine consumption, prostate cancer risk factors and other alcoholic beverages = 0.78, 95% CI: 0.65–0.95) as for those who consumed ≤4 glasses of red wine/week and ≥half a glass of alcohol from other sources/day (RR = 0.81, 95% CI: 0.65–1.02). The same was true for men who consumed >4 glasses of red wine/week and <half a glass of alcohol from other sources /day (RR = 1.13, 95% CI: 0.72–1.78), and men who consumed >4 glasses of red wine/week and ≥half a glass of alcohol from other sources/day (RR = 1.01, 95% CI: 0.70–1.45).

Analyses characterized by prostate cancer stage and grade

Similar results were observed for organ-confined, low-grade and high-grade prostate cancer as for total prostate cancer in the full analytic cohort (data not shown) and among men with unchanged alcohol consumption in the 10 years prior to baseline (Table V). Slightly lower risks were observed for advanced stage prostate cancer, particularly among men with unchanged alcohol consumption in the 10 years prior to baseline (Table V).

Table V. Baseline red wine consumption and prostate cancer characterized by stage and grade among men with unchanged alcohol consumption from 1976 to 1986 in the Health Professionals Follow-up Study, 1986–2002.
 Cases/person-timeAge-adjusted RR1 (95 % CI)Multivariable-adjusted RR2 (95 % CI)Multivariable-adjusted RR3 (95 % CI)
  • 1

    RR, relative risk; CI, confidence interval.

  • 2

    Adjusted for age, race/ethnicity, body mass index at age 21, cumulative family history of prostate cancer through 1996, height, updated cigarette smoking in the past 10 yr, baseline intakes of total energy, tomato sauce, red meat, fish, calciumand vitamin E, baseline energy-adjusted intakes of fructose and α-linolenic acid, baseline vigorous physical activity and updated diabetes mellitus type 2 and vasectomy status.

  • 3

    Additionally adjusted for all other specific alcoholic beverage types.

  • 4

    Categories were determined by calculating the midpoint between each successive drinking frequency and multiplying this value by 11 g (the amount of ethanol in one glass of red wine).

Organ-confined (≤T2 and NOMO) prostate cancers4
No red wine consumption576/175,6191.001.001.00
0.01–1.31 (1–3 glassed/mo)226/74,8730.98 (0.84–1.14)0.93 (0.80–1.09)0.80 (0.67–0.96)
1.32–2.41 (1 glasses/wk)119/36,5231.12 (0.92–1.36)1.03 (0.84–1.26)0.85 (0.67–1.08)
2.42–7.03 (2–4 glasses/wk)80/26,9690.96 (0.76–1.21)0.92 (0.72–1.17)0.71 (0.54–0.93)
7.04–16.4 (5–7 glasses/wk)37/9,5991.19 (0.85–1.65)1.14 (0.81–1.60)1.03 (0.72–1.47)
> 16.5 (> 1 glass/day)11/2,8421.16 (0.64–2.11)1.07 (0.58–1.98)0.96 (0.52–1.79)
p-trend 0.430.920.11
Advanced (T3b or worse) prostate cancers4
No red wine consumption438/175,7331.001.001.00
0.01–1.31 (1–3 glasses/mo)161/74,9330.92 (0.76–1.10)0.87 (0.72–1.04)0.82 (0.66–1.02)
1.32–2.41 (1 glass/wk)73/36,5620.89 (0.70–1.14)0.83 (0.64–1.07)0.78 (0.58–1.04)
2.42–7.03 (2–4 glasses/wk)58/26,9890.91 (0.69–1.19)0.85 (0.64–1.13)0.73 (0.53–1.00)
7.04–16.4 (5–7 glasses/wk)28/9,6091.18 (0.80–1.72)1.06 (0.72–1.57)1.06 (0.70–1.60)
≥16.5 (>1 glass/day)7/2,8460.97 (0.46–2.06)1.04 (0.48–2.23)1.01 (0.46–2.19)
Low-grade (Gleason sum < 7) prostate cancers4
No red wine consumption438/175,7331.001.001.00
0.01–1.31 (1–3 glasses/mo)161/74,9630.92 (0.76–1.10)0.87 (0.72–1.04)0.82 (0.66–1.02)
1.32–2.41 (1 glass/wk)73/36,5620.89 (0.70–1.14)0.83 (0.64–1.07)0.78 (0.58–1.04)
2.42–7.03 (2–4 glasses/wk)58/26,9890.91 (0.69–1.19)0.85 (0.64–1.13)0.73 (0.53–1.00)
7.04–16.4 (5–7 glasses/wk)28/9,6091.18 (0.80–1.72)1.06 (0.72–1.57)1.06 (0.70–1.60)
>16.5 (>1 glass/day)7/2,8460.97 (0.46–2.06)1.04 (0.48–2.23)1.01 (0.46–2.19)
p-trend 0.700.310.19
High-grade (Gleason sum ≥ 7) prostate cancers4
No red wine consumption305/175,8601.001.001.00
0.01–1.31 (1–3 glasses/mo)113/74,9630.92 (0.74–1.15)0.88 (0.71–1.11)0.76 (0.59–0.97)
1.32–2.41 (1 glass/wk)70/36,5651.24 (0.96–1.61)1.13 (0.86–1.48)0.80 (0.58–1.10)
2.42–7.03 (2–4 glasses/wk)45/27,0001.01 (0.74–1.38)0.96 (0.70–1.33)0.75 (0.52–1.08)
7.04–16.4 (5–7 glasses/wk)17/9,6241.03 (0.63–1.68)1.00 (0.60–1.68)0.86 (0.50–1.48)
>16.5 (>1 glass/day)10/2,8432.00 (1.06–3.75)2.06 (1.07–3.97)1.72 (0.88–3.36)
p-trend 0.180.390.42

Analyses of other types of alcoholic beverages

Parallel analyses were performed for white wine, beer, liquor and total alcohol consumption (Tables II–IV, and data not shown). In general, null or slight nonsignificant increased risks were observed for consumption of beer and liquor, and a significant positive trend was observed for consumption of white wine, with the greatest risks observed for moderate consumption. With respect to total alcohol consumption, significant positive trends were observed in the full analytic cohort and among men with unchanged alcohol consumption in the 10 years prior to baseline, whereas no clear pattern was observed among men additionally <65 years of age. Of note, the upper category for liquor and total alcohol consumption included a much wider range of drinking frequencies than for other alcoholic beverages.

Discussion

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

In this prospective cohort of American male health professionals, no clear associations were observed between red wine consumption and risk of prostate cancer. Among the sub-groups of men who reported unchanged alcohol consumption in the 10 years prior to baseline, and those additionally <65 years of age, reduced risks of prostate cancer were observed for men who consumed ≤4 glasses of red wine/week, whereas null or slight increased risks were observed for men who consumed greater than this amount, resulting in a lack of linear trend. Similar patterns of association were observed for prostate cancer characterized by stage and grade, with slightly more pronounced associations for advanced stage disease. These patterns of association were distinct from the generally null or slightly increased risks observed for white wine, beer, liquor and total alcohol consumption.

Our results for baseline red wine consumption and prostate cancer are consistent with the results from one previous cohort study3 and previous analyses in the HPFS cohort,5 both of which observed null or very weak inverse associations between baseline red wine consumption and prostate cancer. Our observed lack of trend for red wine consumption among men with unchanged alcohol consumption in the 10 years prior to baseline, and among those additionally <65 years of age, differs from the significant inverse trend observed by Schoonen et al.4 in their large population-based case–control study of lifetime alcohol consumption and prostate cancer. However, this difference may potentially be explained by selection bias, whereby health-conscious controls aware of the purported benefits of red wine consumption for heart disease prevention might have been more likely to participate in Schoonen and colleagues' study than less health-conscious controls. Although not the focus of this investigation, our results for white wine, beer, liquor and total alcohol consumption are consistent with previous HPFS findings,5 taking into consideration slight differences in drinking amount categorizations.

Despite our observed lack of trend for red wine consumption and prostate cancer risk, we did observe statistically significant inverse associations between some of the lower categories of red wine consumption and prostate cancer. However, the likelihood that these were true causal protective associations is limited by several factors. First, inverse associations were only observed for consumption of ≤4 glasses of red wine/week, whereas null or slight positive associations were observed for consumption of >4 glasses of red wine/week. This pattern of association is difficult to explain biologically because protective levels of free and conjugated red wine polyphenols, such as trans-resveratrol, (+)-catechin, (−)-epicatechin and quercitin, would be much more likely to accumulate at higher levels of consumption than at levels as low as 1–3 glasses/month. We also considered nonbiologic explanations for our observed pattern of association, including infrequent consumption of greater amounts of alcohol (i.e., binge-drinking), higher concomitant consumption of other alcoholic beverages and a shift in the balance between antioxidative properties of red wine polyphenols and oxidative properties of ethanol at higher than at lower levels of red wine consumption.5 However, these did not explain study findings because similar patterns of association were observed when men who consumed large amounts of alcohol infrequently were excluded, and when the joint effects of red wine and nonred wine alcohol consumption were examined.

A second factor limiting the likelihood of a causal protective association is the observation that inverse associations were only apparent after adjustment for consumption of other alcoholic beverages, particularly white wine. These findings imply that white wine confounded associations between red wine consumption and prostate cancer, and that white wine was therefore associated with both red wine consumption and an increased risk of prostate cancer. However, why white wine would be associated with a stronger increased risk of prostate cancer than consumption of other alcoholic beverages, such as beer or liquor, is difficult to explain. Other alternative explanations for our findings include residual confounding by distinct covariate patterns associated with red or white wine consumption and prostate cancer risk that could not be captured by the extensive adjustment performed in this analysis, or chance.

In the present study, we revisited our previous cohort analysis to address the question of red wine consumption and prostate cancer incidence, using more extensive red wine exposure characterization and an even greater sample size. No clear positive or inverse trends were observed between red wine consumption and prostate cancer in the full analytic cohort, among men with long-term unchanged alcohol consumption and among those additionally <65 years of age. In summary, red wine consumption is unlikely to contribute appreciably to the etiology of prostate cancer.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

We thank Jill Arnold, Elizabeth Frost-Hawes, Mira Kaufman, Yan Lydia Liu, Laura Sampson and Mildred Wolff for their continued help in conducting the Health Professionals Follow-up Study. Dr. Siobhan Sutcliffe was supported by the Fund for Research and Progress in Urology, John Hopkins Medical Institutions. Dr. Eric Rimm received an honorarium for speaking twice in the past several years at academic meetings where funding was from a nonprofit arm of an institute funded by industry.

References

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References