Fat and meat intake and prostate cancer risk: The multiethnic cohort study

Authors


Abstract

Dietary fat and meat as potential risk factors for prostate cancer have been the focus of many epidemiologic investigations, and findings from recent studies in particular have been inconsistent. Therefore, we examined the association between these exposures and prostate cancer risk in the Multiethnic Cohort Study. The analyses included 82,483 men in Hawaii and Los Angeles aged ≥45, who completed a detailed quantitative food frequency questionnaire in 1993–1996. During the follow-up period of 8 years, a total of 4,404 incident cases, including 1,278 nonlocalized or high-grade cancer cases, were identified. Cox proportional hazard models were used to estimate relative risks of prostate cancer after adjustment for time on study, ethnicity, family history of prostate cancer, education, body mass index, smoking status and energy intake. Intake of different types of fat (total, saturated, monounsaturated or polyunsaturated), n-6 fatty acid, cholesterol, various meats, and fats from meat showed no association with overall prostate cancer risk or with nonlocalized or high-grade prostate cancer. Furthermore, we found little evidence of any relation of fat and meat intake with prostate cancer risk within any of the 4 racial/ethnic groups (African Americans, Japanese Americans, Latinos and Whites). There was a suggestion of a protective effect of n-3 fatty acid intake that was limited to Latinos and Whites. However, overall, our findings from a large cohort study of ethnically diverse population give no indication that intake of fat and meat substantially affects prostate cancer risk. © 2007 Wiley-Liss, Inc.

In the last 3 decades, epidemiologic studies on prostate cancer have extensively investigated dietary risk factors.1, 2 More than any other dietary components, fat and meat have been the focus of these studies. Although findings from early studies were generally supportive of a positive association, findings from recent studies have been much less consistent.3, 4 Indeed, the biologic role of fat in prostate cancer incidence and progression still remains unclear. Meat consumption may be related to prostate cancer risk through its content of several potential carcinogens in addition to its fat content.2, 3

In studies of prostate cancer, fat has been classified in several ways, including total, saturated, monounsaturated, polyunsaturated fat and animal product fat; n-3 and n-6 polyunsaturated fat; and more recently, individual fatty acids.2 Most studies showed either no association5, 6, 7, 8, 9, 10, 11 or a positive association12, 13, 14, 15, 16 of total, saturated or animal fat with risk, though some studies suggested that consumption of fatty fish containing abundant long-chain polyunsaturated fatty acids may be protective.17, 18, 19, 20 Experimental studies have suggested that specific fatty acids promote or suppress prostate tumor development, but epidemiologic investigations on individual fatty acids have yielded inconsistent findings.7, 21, 22, 23, 24 For meat intake, the findings have also been mixed.5, 6, 7, 9, 10, 25, 26, 27, 28, 29, 30, 31, 32, 33 To further explore these relationships, we examined the association between prostate cancer risk and the intake of fat (including total, saturated, monounsaturated and polyunsaturated fat, and n-3 and n-6 fatty acids), cholesterol, meat (including total, red, processed and poultry), fish and fats from meat in the Multiethnic Cohort Study.

Abbreviations:

ALA, α-linolenic acid; CI, confidence interval; DHA, docosahexaenoic acid; EPA, eicosapentaenoic acid; HPFS, the Health Professionals Follow-up Study; QFFQ, quantitative food frequency; RR, relative risk.

Material and methods

Study cohort

The Multiethnic Cohort was established in Hawaii and Los Angeles in 1993–1996 to study diet and cancer, and includes more than 215,000 men and women aged ≥45 at entry. The cohort's design targeted 5 racial/ethnic groups: African Americans, Native Hawaiians, Japanese Americans, Latinos and Whites. Participants completed a 26-page mailed questionnaire on diet, medical history and other lifestyle exposures at baseline. Details of the design and implementation of the cohort have been published previously.34 The study was approved by the respective institutional review boards (University of Hawaii, University of Southern California).

For the current analyses, men were excluded if they self-identified as other than 1 of the 5 targeted groups (n = 5,944), had a previous prostate cancer either by self report on the questionnaire or through linkage to tumor registries (n = 2,890) or had invalid dietary information based on total energy intake or its components (n = 3,653). Additionally, men who did not provide complete information on height, weight, educational level,or smoking status were excluded (n = 1,988), which left 82,483 men available for analysis.

Identification of prostate cancer cases

Incident cases of prostate cancer were identified by linkage of the cohort to 3 population-based cancer registries in Hawaii and California that are all members of the National Cancer Institute's Surveillance, Epidemiology and End Results (SEER) Program (the statewide Hawaii Tumor Registry, the Los Angeles County Cancer Surveillance Program, and the statewide California Cancer Registry). Case ascertainment was complete through December 31, 2002. The cohort was also linked to the Hawaii and California state death files and to the National Death Index file through December 31, 2002. In this analysis, we considered all prostate cancer cases as events, except in situ cases. Nonlocalized prostate cancer cases were defined as all cancers that had regional or distant spread. High-grade cancer cases were defined as cancers with Gleason score ≥ 7. During the follow-up period of 8 years, a total of 4,404 incident cases, including 1,278 advanced (nonlocalized or high-grade) cases, were identified.

Dietary assessment

Information on dietary intake at baseline was obtained by a self-administered quantitative food frequency questionnaire (QFFQ) that covered the last year. The QFFQ, with over 180 items, was developed from 3-day measured food records from about 60 men and women of each ethnic group.34 The QFFQ included 8 frequency categories for foods and 9 for beverages, with 3 choices of portion sizes. Daily grams consumed from each QFFQ food item were computed as the product of the daily frequency and grams in the relevant portion size. Daily nutrient intakes from each item were calculated by applying the food composition table that has been developed and maintained at the Cancer Research Center of Hawaii (CRCH) for use in the Multiethnic Cohort Study. A calibration study was conducted and showed satisfactory correlations (0.55–0.74) between the QFFQ and three 24-hr recalls for all ethnic and sex groups being studied; the energy-adjusted correlations were 0.52–0.76 for total fat and 0.54–0.78 for saturated fat.35 Intake of meat was calculated as the sum of grams from single food items of meat and relevant portions of mixed dishes, estimated from recipes; fat intake from meat was similarly computed. Total meat consisted of red meat (beef, pork and lamb), poultry and processed meat (processed red meat and processed poultry). Individual meat groups (beef, pork, red meat, poultry and processed meat), fish and shellfish were also examined.

Statistical analysis

We used Cox proportional hazards models36 with age as the time metric to estimate relative risks (RR) of prostate cancer and 95% confidence intervals (CI) for foods and nutrients. Observation began at questionnaire completion or at age 45 for the few individuals who were less than 45 at questionnaire completion (0.5%), and ended at the earliest of these dates: date of prostate cancer diagnosis (including in situ cancer), date of death and date of closure at December 31, 2002. Dietary intakes were divided into quintiles based on the distributions of all male participants. RRs were estimated for quintiles of intake relative to the lowest quintile. The proportionality assumption was tested by Schoenfeld residuals and found to hold. All Cox models were adjusted for the following strata variables: time since cohort entry (≤2 years, 2–5 years and >5 years), ethnicity (5 racial/ethnic groups), body mass index [BMI; <22.46, 22.46–24.99, 25.00–29.99 and ≥30.00, where 22.46 is the median of the normal weight group (BMI ≤ 24.99)], educational level (≤8th grade, 9th–12th grade, some college or vocational school, and ≥graduated college), family history of prostate cancer (yes and no/do not know) and smoking status (never smoker; former smoker; current smoker of <10 cigarettes/day, 10–19/day and ≥20/day).37 RRs were also adjusted for energy intake (log transformed kcal/day). The adjustment factors were selected, because they were either related to prostate cancer risk (ethnicity and family history), to PSA screening (smoking)37 or to dietary assessment.35 We also considered other factors (e.g., age at cohort entry to account for generational effects, vitamin and mineral supplement use) as risk factors or potential confounders, but did not include them in the final models, because these variables did not alter the RRs. We tested the linear trends in dietary intake by inclusion of a trend variable, assigned the racial/ethnic specific median values within the appropriate overall quintile of intake.

To lessen the effect of measurement error,35, 38 we used the dietary variables as densities, i.e. values per 1,000 kcal or as a percentage of total energy intake.39 We also reran the analyses using calibration-corrected intake values for fat density variables (g/1,000 kcal); these are the predicted levels from a regression of the average intake from three 24-hr dietary recalls on QFFQ intake.35 The standard errors of the parameters were adjusted to account for the uncertainty in the prediction model based on the delta method.40

The models were run considering all invasive incident prostate tumors as cases, and then rerun, considering only nonlocalized or high-grade tumors as cases. The overall models were repeated separately for 4 ethnic groups, and a joint model was run to test interaction terms for ethnicity and the fat intake trend variables. Tests for interactions were based on the likelihood ratio test. p values reported were 2-sided, and statistical significance was set at p < 0.05. All analyses were conducted using SAS statistical software (version 9.1, SAS Institute, Cary, NC).

Results

Table I gives the distribution by major baseline characteristics and intake of selected fat and meat variables for all male participants in the cohort and for cases (subjects diagnosed with an incident prostate cancer during follow-up, n = 4,404). The proportion of African Americans among cases (26.9%) was more than twice as high as in noncases (12.2%). Cases were older compared with men in the cohort overall. Overweight and obesity rates were similar in the 2 groups. Somewhat fewer cases smoked cigarettes (15.4% vs. 18.0%), but in general the smoking patterns were similar in the 2 groups. A higher proportion of cases (10.1%) than the cohort as a whole (6.8%) had a family history of prostate cancer. Finally, mean dietary intakes were almost uniformly identical in the 2 groups.

Table I. Baseline Characteristics of Male Participants by Prostate Cancer Status in the Multiethnic Cohort Study, 1993–2002
CharacteristicsCohort (n = 82,483)Cases (n = 4,404)
  • 1

    Column percentage.

  • 2

    Means (SD).

  • 3

    Total meat consisted of red meat (beef, pork and lamb), poultry and processed meat (processed red meat and processed poultry).

Age (years)  
 45–4916.613.1
 50–5414.96.3
 55–5915.911.9
 60–6417.521.1
 65–6917.929.3
 70–7414.523.9
 ≥752.74.3
Ethnicity  
 African Americans13.026.9
 Native Hawaiians7.14.4
 Japanese Americans30.424.1
 Latinos24.023.6
 Whites25.621.0
BMI  
 <22.4615.915.5
 22.46–24.9927.227.6
 25.00–29.9942.644.7
 ≥30.0014.212.2
Education  
 ≤8 years11.010.7
 9–12 years30.234.5
 Some college or vocational school29.328.8
 College graduate or more29.626.1
Regular multivitamin supplement use36.637.9
Smoking  
 Never30.030.1
 Former52.054.5
 Current, <10 cigarettes/day5.66.0
 Current, 10–19 cigarettes/day7.06.3
 Current, ≥20 cigarettes/day5.43.0
Family history of prostate cancer6.810.1
Daily dietary intake  
 Total fat (% energy)30.2 (7.2)230.2 (7.2)
 Saturated fat (% energy)8.9 (2.7)8.9 (2.6)
 Monounsaturated fat (% energy)11.1 (2.8)11.2 (2.9)
 Polyunsaturated fat (% energy)7.5 (1.9)7.4 (1.9)
 n-3 fatty acids (g/1,000 kcal)0.86 (0.22)0.85 (0.22)
 α-Linolenic acid (g/1,000 kcal)0.81 (0.21)0.80 (0.21)
 Eicosapentaenoic acid (g/1,000 kcal)0.02 (0.02)0.02 (0.02)
 Docosahexaenoic acid (g/1,000 kcal)0.04 (0.03)0.04 (0.03)
 n-6 fatty acids (g/1,000 kcal)7.5 (2.0)7.5 (2.0)
 Cholesterol (g/1,000 kcal)109 (44.3)109 (45.1)
 Total meat3 (g/1,000 kcal)51.4 (24.9)51.1 (25.2)
 Red meat (g/1,000 kcal)20.5 (13.2)19.3 (12.7)
 Processed meat (g/1,000 kcal)10.3 (8.3)10.3 (8.6)
 Fish (g/1,000 kcal)8.6 (8.4)8.5 (9.1)

Multivariate RRs of prostate cancer for quintiles of fat intake are shown in Table II. None of the nutrients investigated exhibited a significant dose–response relation with prostate cancer risk, whether the analysis included all cases or was restricted to nonlocalized or high-grade cases. Indeed, most of the RR estimates were close to 1 and were nonsignificant. The RRs for n-3 fatty acid intake, overall and as α-linolenic acid (ALA), showed a suggested protective effect that was limited to moderate intake (the second and third quintiles) and was somewhat stronger for advanced cancer. Total fat and saturated fat from meat were also examined, and no significant associations were found (data not shown).

Table II. RRs (95% CIs)1 of Prostate Cancer According to Quintile (Q) of Fat Intake in the Multiethnic Cohort Study, 1993–2002
 Q12Q2Q3Q4Q5p for trend3
  • 1

    Adjusted for time on study, ethnicity, family history of prostate cancer, education, BMI, smoking status and energy intake.

  • 2

    Reference category.

  • 3

    The p for trend is based on the Wald statistic for a trend variable assigned the racial/ethnic specific median values within the appropriate overall quintiles of intake.

  • 4

    Median value.

  • 5

    Nonlocalized prostate cancer cases were defined as all cancers that were regional or distant, and high-grade cancer cases were based on Gleason score (≥7).

Total fat (% energy)20.5426.430.334.039.1 
 Total prostate cancer1.000.94 (0.85, 1.03)0.97 (0.88, 1.07)0.93 (0.84, 1.03)0.99 (0.89, 1.09)0.79
 Nonlocalized or high-grade cancer51.000.92 (0.76, 1.10)0.93 (0.77, 1.11)0.88 (0.73, 1.06)0.90 (0.75, 1.09)0.26
Saturated fat (% energy)5.57.48.910.312.3 
 Total prostate cancer1.000.96 (0.87, 1.06)0.94 (0.85, 1.04)0.98 (0.89, 1.09)0.94 (0.85, 1.04)0.39
 Nonlocalized or high-grade cancer1.000.93 (0.77, 1.12)0.93 (0.77, 1.12)0.92 (0.76 1.11)0.87 (0.71, 1.06)0.17
Monounsaturated fat (% energy)7.49.611.212.614.7 
 Total prostate cancer1.001.01 (0.91, 1.11)0.99 (0.89, 1.09)0.96 (0.87, 1.06)1.01 (0.91, 1.12)0.92
 Nonlocalized or high-grade cancer1.001.11 (0.92, 1.33)0.99 (0.92, 1.20)0.99 (0.82, 1.20)1.03 (0.85, 1.25)0.88
Polyunsaturated fat (% energy)5.06.47.48.49.9 
 Total prostate cancer1.001.07 (0.97, 1.18)0.98 (0.89, 1.09)1.02 (0.92, 1.12)1.01 (0.91, 1.11)0.75
 Nonlocalized or high-grade cancer1.001.09 (0.91, 1.31)1.08 (0.90, 1.31)1.07 (0.89, 1.29)1.01 (0.84, 1.23)0.99
Ratio of polyunsaturated to saturated fat0.590.730.850.981.21 
 Total prostate cancer1.000.98 (0.89, 1.08)1.08 (0.98, 1.19)0.96 (0.87, 1.06)1.01 (0.91, 1.12)0.96
 Nonlocalized or high-grade cancer1.001.08 (0.90, 1.29)1.04 (0.87, 1.25)1.01 (0.84, 1.22)1.11 (0.91, 1.34)0.50
n-3 fatty acids (g/1,000 kcal)0.590.740.850.961.13 
 Total prostate cancer1.000.85 (0.78, 0.94)0.91 (0.82, 1.00)0.94 (0.85, 1.03)0.95 (0.86, 1.05)0.85
 Nonlocalized or high-grade cancer1.000.78 (0.65, 0.93)0.82 (0.69, 0.98)0.88 (0.74, 1.06)0.90 (0.76, 1.08)0.60
α-Linolenic acid (g/1,000 kcal)0.550.700.800.901.07 
 Total prostate cancer1.000.89 (0.81, 0.98)0.94 (0.85, 1.03)0.94 (0.85, 1.04)0.92 (0.84, 1.02)0.29
 Nonlocalized or high-grade cancer1.000.87 (0.73, 1.04)0.91 (0.76, 1.09)0.85 (0.70, 1.02)0.89 (0.74, 1.06)0.20
Eicosapentaenoic acid (g/1,000 kcal)0.0040.0090.0140.0220.037 
 Total prostate cancer1.000.98 (0.89, 1.08)1.07 (0.97, 1.18)1.04 (0.94, 1.15)1.01 (0.91, 1.13)0.63
 Nonlocalized or high-grade cancer1.001.01 (0.84, 1.20)1.07 (0.89, 1.28)1.03 (0.85, 1.24)1.05 (0.86, 1.28)0.62
Docosahexaenoic acid (g/1,000 kcal)0.010.020.030.050.07 
 Total prostate cancer1.000.98 (0.89, 1.08)1.04 (0.94, 1.14)1.06 (0.96, 1.17)0.99 (0.89, 1.09)0.89
 Nonlocalized or high-grade cancer1.001.05 (0.87, 1.25)1.05 (0.88, 1.27)1.16 (0.96, 1.39)1.07 (0.88, 1.30)0.37
n-6 fatty acids (g/1,000 kcal)5.06.47.48.410.0 
 Total prostate cancer1.001.08 (0.98, 1.19)1.03 (0.93, 1.14)1.02 (0.92, 1.12)1.03 (0.93, 1.14)0.94
 Nonlocalized or high-grade cancer1.001.07 (0.88, 1.29)1.20 (1.00, 1.44)1.08 (0.89, 1.30)1.04 (0.86, 1.27)0.70
Ratio of n-6 to n-3 fatty acids7.38.18.79.210.3 
 Total prostate cancer1.000.97 (0.88, 1.08)1.02 (0.92, 1.13)1.01 (0.92, 1.12)1.04 (0.95, 1.15)0.28
 Nonlocalized or high-grade cancer1.001.10 (0.91, 1.33)1.10 (0.91, 1.32)1.13 (0.94, 1.37)1.10 (0.92, 1.33)0.32
Cholesterol (m g/1,000 kcal)59.585.3104125162 
 Total prostate cancer1.001.08 (0.98, 1.19)0.98 (0.88, 1.08)1.03 (0.94, 1.14)0.93 (0.84, 1.03)0.08
 Nonlocalized or high-grade cancer1.001.00 (0.84, 1.20)0.86 (0.72, 1.04)0.97 (0.80, 1.17)1.03 (0.86, 1.24)0.71

We also examined food groups, namely intakes of total, red, and processed meat, individual meat items and fish (Table III). These analyses did not show a statistically significant association with risk.

Table III. RRs (95% CIs)1 of Prostate Cancer According to Quintile (Q) of Meat Intake in the Multiethnic Cohort Study, 1993–2002
 Q12Q2Q3Q4Q5p for trend3
  • 1

    Adjusted for time on study, ethnicity, family history of prostate cancer, education, BMI, smoking status and energy intake.

  • 2

    Reference category.

  • 3

    The p for trend is based on the Wald statistic for a trend variable assigned the racial/ethnic specific median values within the appropriate overall quintiles of intake.

  • 4

    Total meat consisted of red meat (beef, pork, and lamb), poultry, and processed meat (processed red meat and processed poultry).

  • 5

    Median value.

  • 6

    Nonlocalized prostate cancer cases were defined as all cancers that were regional or distant, and high-grade cancer cases were based on Gleason score (≥7).

Total meat4 (g/1,000 kcal)22.6537.348.861.683.0 
 Total prostate cancer1.001.09 (0.99, 1.20)1.03 (0.93, 1.13)1.01 (0.91, 1.11)1.01 (0.91, 1.12)0.63
 Nonlocalized or high-grade cancer61.000.96 (0.81, 1.15)0.89 (0.74, 1.07)0.86 (0.71, 1.03)1.07 (0.89, 1.28)0.67
Red meat (g/1,000 kcal)5.512.618.525.237.0 
 Total prostate cancer1.001.04 (0.94, 1.14)1.06 (0.96, 1.17)1.02 (0.92, 1.12)0.97 (0.87, 1.07)0.40
 Nonlocalized or high-grade cancer1.000.99 (0.83, 1.17)0.93 (0.78, 1.12)0.97 (0.81, 1.16)0.95 (0.79, 1.14)0.56
Processed meat (g/1,000 kcal)2.25.48.412.220.0 
 Total prostate cancer1.001.08 (0.98, 1.19)1.06 (0.96, 1.17)0.97 (0.88, 1.07)1.01 (0.91, 1.12)0.49
 Nonlocalized or high-grade cancer1.000.98 (0.82, 1.17)0.99 (0.83, 1.19)0.82 (0.68, 1.00)0.92 (0.77, 1.11)0.21
Poultry (g/1,000 kcal)5.911.716.923.839.9 
 Total prostate cancer1.001.01 (0.92, 1.11)1.08 (0.98, 1.19)1.02 (0.92, 1.13)1.01 (0.92, 1.12)0.99
 Nonlocalized or high-grade cancer1.000.95 (0.79, 1.15)0.99 (0.83, 1.19)1.02 (0.84, 1.22)1.06 (0.88, 1.28)0.33
Beef (g/1,000 kcal)3.78.713.018.127.7 
 Total prostate cancer1.000.97 (0.88, 1.07)0.99 (0.90, 1.09)0.96 (0.87, 1.06)0.98 (0.88, 1.08)0.59
 Nonlocalized or high-grade cancer1.000.90 (0.75, 1.06)0.81 (0.68, 0.97)0.86 (0.71, 1.03)0.97 (0.81, 1.16)0.76
Pork (g/1,000 kcal)0.52.13.96.210.2 
 Total prostate cancer1.000.99 (0.90, 1.09)1.08 (0.98, 1.19)0.95 (0.86, 1.05)0.97 (0.88, 1.08)0.39
 Nonlocalized or high-grade cancer1.000.98 (0.83, 1.17)1.03 (0.86, 1.23)0.87 (0.72, 1.05)0.92 (0.76, 1.11)0.22
Fish (g/1,000 kcal)1.23.96.610.117.8 
 Total prostate cancer1.001.09 (0.99, 1.20)1.05 (0.95, 1.16)1.11 (1.00, 1.22)1.04 (0.93, 1.15)0.75
 Nonlocalized or high-grade cancer1.001.11 (0.93, 1.33)1.04 (0.87, 1.26)1.15 (0.95, 1.39)1.01 (0.82, 1.23)0.89
Shellfish (g/1,000 kcal)0.050.461.382.605.10 
 Total prostate cancer1.000.93 (0.84, 1.02)0.96 (0.87, 1.05)0.99 (0.90, 1.09)1.00 (0.91, 1.10)0.44
 Nonlocalized or high-grade cancer1.000.89 (0.75, 1.06)0.92 (0.77, 1.10)1.02 (0.86, 1.22)0.97 (0.80, 1.17)0.65

In racial/ethnic-specific analyses (Tables IV and V), the lack of association between fat and meat intake and prostate cancer risk was generally consistent across the groups. However, there was evidence of heterogeneity of the effect for n-3 fatty acids, overall and as ALA (p's for interaction ≤ 0.05). While not monotonic, there was a statistically significant inverse association for n-3 fatty acids (p for trend = 0.04) and ALA (p for trend = 0.03) in Latinos, and the suggestion of one for ALA in Whites (p = 0.06). Because of the small number of cases in Native Hawaiian men, and therefore, limited statistical power, they were not included in this analysis.

Table IV. Racial/Ethnic Specific RRs (95% CIs)1 of Total Prostate Cancer According to Quintile (Q) of Fat Intake in the Multiethnic Cohort Study, 1993–2002
 African Americans (n = 10,706)Japanese Americans (n = 25,052)Latinos (n = 19,779)Whites (n = 21,090)p for interaction2
CasesRRCasesRRCasesRRCasesRR
  • 1

    Adjusted for time on study, family history of prostate cancer, education, BMI, smoking status and energy intake.

  • 2

    Interaction test between fat and meat intake and ethnicity is based on the likelihood ratio test.

  • 3

    Reference category.

  • 4

    Median value.

  • 5

    Values in parentheses indicate 95% CIs.

  • 6

    The p for trend is based on the Wald statistic for a trend variable assigned the median values within the appropriate quintiles of intake.

Total fat (% energy)
 Q1320.541561.003481.001611.001821.000.33
 Q226.41660.88 (0.70, 1.11)52700.98 (0.83, 1.16)1940.89 (0.72, 1.10)1740.89 (0.72, 1.10) 
 Q330.31920.83 (0.66, 1.04)2261.11 (0.94, 1.32)2140.83 (0.67, 1.03)1890.98 (0.79, 1.21) 
 Q434.02550.82 (0.67, 1.02)1320.90 (0.73, 1.11)2420.86 (0.70, 1.06)1920.99 (0.80, 1.22) 
 Q539.14170.96 (0.79, 1.17)861.08 (0.84, 1.38)2290.83 (0.67, 1.03)1870.92 (0.74, 1.14) 
  p for trend6  0.98 0.79 0.11 0.75 
Saturated fat (% energy)
 Q15.51431.004111.001221.001751.000.36
 Q27.41830.94 (0.74, 1.18)3001.03 (0.88, 1.20)1640.89 (0.70, 1.14)1610.85 (0.68, 1.06) 
 Q38.92100.84 (0.67, 1.06)1880.97 (0.81, 1.16)2140.88 (0.70, 1.11)1760.87 (0.70, 1.08) 
 Q410.32850.92 (0.74, 1.14)1081.03 (0.83, 1.28)2700.86 (0.69, 1.08)2100.96 (0.77, 1.18) 
 Q512.33650.90 (0.73, 1.11)551.18 (0.88, 1.59)2700.80 (0.64, 1.00)2020.89 (0.72, 1.10) 
  p for trend  0.48 0.55 0.06 0.59 
Monounsaturated fat (% energy)
 Q17.41491.003291.001781.001761.000.29
 Q29.61710.98 (0.77, 1.23)2771.06 (0.90, 1.24)1940.88 (0.71, 1.09)1861.02 (0.82, 1.26) 
 Q311.21900.86 (0.68, 1.08)2191.05 (0.88, 1.26)2190.88 (0.72, 1.08)1851.05 (0.85, 1.30) 
 Q412.62580.89 (0.72, 1.10)1400.92 (0.75, 1.13)2230.82 (0.66, 1.00)2001.12 (0.91, 1.39) 
 Q514.74181.00 (0.82, 1.22)971.04 (0.82, 1.31)2260.88 (0.71, 1.07)1770.95 (0.76, 1.19) 
  p for trend  0.80 0.88 0.17 0.99 
Polyunsaturated fat (% energy)
 Q15.01621.002241.002081.002131.000.35
 Q26.42000.95 (0.76, 1.19)2501.27 (1.06, 1.53)2291.02 (0.84, 1.24)2171.03 (0.85, 1.25) 
 Q37.42070.94 (0.75, 1.16)2181.18 (0.98, 1.43)2310.93 (0.77, 1.13)1600.82 (0.67, 1.02) 
 Q48.42810.98 (0.80, 1.20)2051.19 (0.98, 1.45)2020.90 (0.74, 1.10)1700.94 (0.76, 1.16) 
 Q59.83361.02 (0.83, 1.24)1651.03 (0.84, 1.27)1700.93 (0.75, 1.16)1640.92 (0.74, 1.14) 
  p for trend  0.67 0.81 0.28 0.28 
n-3 fatty acids (g/1,000 kcal)
 Q10.592071.002271.002481.002511.000.04
 Q20.742150.87 (0.71, 1.07)1880.87 (0.72, 1.07)2100.80 (0.66, 0.96)2050.86 (0.71, 1.05) 
 Q30.852180.87 (0.71, 1.07)2031.01 (0.83, 1.22)2250.84 (0.70, 1.02)1710.83 (0.68, 1.02) 
 Q40.962580.97 (0.80, 1.18)2301.15 (0.95, 1.39)1810.71 (0.58, 0.87)1600.91 (0.74, 1.12) 
 Q51.132881.05 (0.87, 1.28)2140.95 (0.79, 1.16)1760.85 (0.69, 1.04)1370.87 (0.70, 1.08) 
  p for trend  0.25 0.65 0.04 0.24 
α-Linolenic acid (g/1,000 kcal)
 Q10.552051.002401.002111.002431.000.05
 Q20.702110.84 (0.68, 1.03)2140.98 (0.81, 1.18)2070.84 (0.69, 1.02)2090.88 (0.73, 1.07) 
 Q30.802370.91 (0.75, 1.12)2001.04 (0.86, 1.26)2230.87 (0.72, 1.06)1730.83 (0.68, 1.02) 
 Q40.902600.73 (0.76, 1.13)2201.18 (0.98, 1.43)1960.73 (0.60, 0.90)1640.91 (0.74, 1.12) 
 Q51.072731.00 (0.82, 1.21)1880.97 (0.80, 1.18)2030.83 (0.67, 1.01)1350.79 (0.64, 0.99) 
  p for trend  0.59 0.67 0.03 0.06 
Eicosapentaenoic acid (g/1,000 kcal)
 Q10.0042101.00811.004091.002231.000.75
 Q20.0092160.92 (0.75, 1.13)1440.88 (0.67, 1.17)2690.97 (0.83, 1.14)2311.12 (0.92, 1.35) 
 Q30.0142961.17 (0.96, 1.41)2230.94 (0.72, 1.22)1861.03 (0.86, 1.24)1941.10 (0.90, 1.34) 
 Q40.0222551.06 (0.87, 1.29)3030.95 (0.75, 1.23)1110.93 (0.75, 1.16)1581.10 (0.89, 1.36) 
 Q50.0372091.02 (0.83, 1.25)3110.88 (0.68, 1.13)650.92 (0.70, 1.20)1181.02 (0.81, 1.29) 
  p for trend  0.62 0.41 0.52 0.95 
Docohahexaenoic acid (g/1,000 kcal)
 Q10.011731.001091.003851.002391.000.56
 Q20.022161.11 (0.90, 1.38)1550.88 (0.68, 1.12)2861.03 (0.88, 1.21)1990.91 (0.75, 1.11) 
 Q30.032691.17 (0.96, 1.44)2200.92 (0.73, 1.16)1740.92 (0.76, 1.10)2031.11 (0.91, 1.35) 
 Q40.052811.15 (0.94, 1.41)2890.94 (0.75, 1.18)1281.04 (0.84, 1.28)1641.07 (0.87, 1.32) 
 Q50.072471.08 (0.87, 1.32)2890.85 (0.68, 1.07)670.89 (0.68, 1.17)1190.98 (0.78, 1.23) 
  p for trend  0.75 0.28 0.46 0.69 
n-6 fatty acids (g/1,000 kcal)
 Q15.01561.002221.002031.002071.000.38
 Q26.41951.05 (0.84, 1.31)2481.25 (1.04, 1.51)2321.05 (0.86, 1.28)2060.97 (0.80, 1.19) 
 Q37.42211.03 (0.83, 1.28)2111.17 (0.96, 1.42)2361.00 (0.82, 1.22)1610.84 (0.68, 1.04) 
 Q48.42750.99 (0.80, 1.22)2141.23 (1.01, 1.49)2010.92 (0.75, 1.13)1750.92 (0.75, 1.14) 
 Q510.03391.09 (0.89, 1.33)1671.05 (0.85, 1.29)1680.98 (0.79, 1.21)1750.91 (0.74, 1.13) 
  p for trend  0.53 0.61 0.50 0.32 
Ratio of n-6 to n-3 fatty acids
 Q17.31451.002671.001821.001531.000.76
 Q28.11650.91 (0.71, 1.15)2381.13 (0.94, 1.35)2030.95 (0.77, 1.17)1390.87 (0.68, 1.10) 
 Q38.72521.04 (0.84, 1.30)2121.18 (0.98, 1.41)2050.97 (0.79, 1.19)1670.83 (0.74, 1.16) 
 Q49.22820.95 (0.76, 1.17)1871.15 (0.95, 1.39)2221.00 (0.81, 1.22)1950.94 (0.75, 1.17) 
 Q510.33421.01 (0.82, 1.24)1581.01 (0.83, 1.24)2281.13 (0.92, 1.39)2700.98 (0.79, 1.20) 
  p for trend  0.75 0.67 0.18 0.77 
Table V. Racial/Ethnic Specific RRs (95% CIs)1 of Total Prostate Cancer According to Quintile (Q) of Meat Intake in the Multiethnic Cohort Study, 1993–2002
 African Americans (n = 10,706)Japanese Americans (n = 25,052)Latinos (n = 19,779)Whites (n = 21,090)p for interaction2
CasesRRCasesRRCasesRRCasesRR
  • 1

    Adjusted for time on study, family history of prostate cancer, education, BMI, smoking status and energy intake.

  • 2

    Interaction test between fat and meat intake and ethnicity is based on the likelihood ratio test.

  • 3

    Total meat consisted of red meat (beef, pork and lamb), poultry and processed meat (processed red meat and processed poultry).

  • 4

    Reference category.

  • 5

    Median value.

  • 6

    Values in parentheses indicate 95% CIs.

  • 7

    The p for trend is based on the Wald statistic for a trend variable assigned the median values within the appropriate quintiles of intake.

Total meat (g/1,000 kcal)3
 Q1422.551531.002951.001691.002561.000.53
 Q237.32051.17 (0.93, 1.46)62741.11 (0.94, 1.32)1971.07 (0.86, 1.33)2191.02 (0.84, 1.23) 
 Q348.82021.00 (0.80, 1.26)2271.04 (0.87, 1.25)2121.02 (0.83, 1.27)1911.00 (0.82, 1.22) 
 Q461.62561.13 (0.91, 1.39)1701.02 (0.83, 1.24)2110.93 (0.75, 1.15)1470.92 (0.74, 1.14) 
 Q583.03701.02 (0.84, 1.25)960.98 (0.77, 1.24)2510.99 (0.81, 1.22)1111.00 (0.79, 1.27) 
  p for trend7  0.72 0.79 0.54 0.70 
Red meat (g/1,000 kcal)
 Q15.52521.002471.001841.002451.000.45
 Q212.62631.13 (0.94, 1.36)2731.14 (0.96, 1.36)1700.86 (0.70, 1.07)2080.96 (0.79, 1.17) 
 Q318.52080.95 (0.78, 1.16)2421.16 (0.97, 1.39)2211.07 (0.87, 1.31)1861.00 (0.82, 1.23) 
 Q425.22371.10 (0.91, 1.33)1911.09 (0.90, 1.33)1950.82 (0.67, 1.01)1701.05 (0.85, 1.28) 
 Q537.02261.05 (0.86, 1.27)1091.04 (0.82, 1.31)2700.87 (0.72, 1.06)1150.83 (0.65, 1.05) 
  p for trend  0.76 0.71 0.16 0.29 
Processed meat (g/1,000 kcal)
 Q12.21961.001901.002471.002161.000.35
 Q25.41840.91 (0.74, 1.13)2311.10 (0.91, 1.34)2741.16 (0.97, 1.39)2141.15 (0.95, 1.40) 
 Q38.42001.00 (0.81, 1.24)2371.12 (0.92, 1.36)2161.05 (0.87, 1.27)1881.08 (0.88, 1.32) 
 Q412.22330.89 (0.72, 1.09)2231.04 (0.85, 1.27)1690.99 (0.81, 1.22)1490.93 (0.74, 1.15) 
 Q520.03731.00 (0.83, 1.20)1811.09 (0.88, 1.34)1340.86 (0.69, 1.08)1571.02 (0.82, 1.27) 
  p for trend  0.78 0.71 0.06 0.57 
Fish (g/1,000 kcal)
 Q11.22451.00731.004071.001741.000.56
 Q23.92911.12 (0.93, 1.34)1470.97 (0.73, 1.29)2901.14 (0.97, 1.33)2001.05 (0.85, 1.30) 
 Q36.62491.03 (0.86, 1.25)2301.02 (0.78, 1.34)1640.97 (0.80, 1.17)2051.09 (0.88, 1.34) 
 Q410.12251.11 (0.91, 1.35)2880.95 (0.73, 1.23)1161.12 (0.91, 1.39)1951.21 (0.98, 1.50) 
 Q517.81761.06 (0.87, 1.31)3240.90 (0.69, 1.16)630.91 (0.69, 1.21)1501.07 (0.85, 1.34) 
  p for trend  0.72 0.17 0.76 0.40 

Since meat cooked at high temperature may be a source of carcinogenic compounds, we also investigated usual meat preparation, but found no association with prostate cancer risk: RR of well done vs. rare or medium = 1.10, 95% CI = 0.91–1.33.

In separate analyses for cases diagnosed within 2 years of follow-up and after 2 years, the findings on total fat and prostate cancer were similar: RR of the highest vs. lowest quintile = 1.02, 95% CI = 0.82–1.26, p for trend = 0.96 for cases diagnosed within 2 years (n = 955); RR of the highest vs. lowest quintile = 0.98, 95% CI = 0.87–1.09, p for trend = 0.77 for cases (n = 3,449) after 2 years of follow-up.

Discussion

In this large multiethnic cohort, intake of different types of fat and meat showed no strong association with overall prostate cancer risk. We also found no evidence of such an association of fat and meat with risk of nonlocalized or high-grade prostate cancer. Furthermore, we generally found little evidence of any relation of fat and meat intake with prostate cancer risk within any of the 4 racial/ethnic groups. An inverse association was found for n-3 fatty acid intake, overall and as ALA, that was limited to selective subgroups, particularly Latinos.

Findings on fat, meat and prostate cancer have been recently reviewed,4, 41, 42 and therefore the discussion here focuses on more recent additions to the literature, particularly results from prospective studies and reviews.

The association between total dietary fat and prostate cancer was examined in a recent metaanalysis.43 In that report, an analysis based on 4 cohort studies found no association between total and saturated fat and prostate cancer risk, whereas an analysis of 11 case–control studies showed a 10–21% increased risk for an increase of 45 g/day in total fat. For these 15 studies overall, a 11% increased risk (95% CI = 1.01–1.22) was found, but the heterogeneity between studies and study designs was substantial. The difference in findings between the cohort and case–control studies may reflect recall bias in the latter group. Another difference in 2 types of study is latency. The case–control studies ask about very recent diet while the timing of the diet to cancer in a cohort varies. When we ran separate models for cases diagnosed within 2 years of follow-up and after 2 years in our study, the findings were similar.

Recently, epidemiologic studies have focused more on specific types of fatty acids rather than total fat intake.20 Long-chain polyunsaturated n-3 fatty acids have been of particular interest, because experimental studies have shown that these fatty acids suppress tumor cell growth. However, data on these constituents from studies of prostate cancer in humans are neither sufficient nor consistent.19, 44 As these fatty acids are contained primarily in fatty fish, some studies investigated fish consumption related to prostate cancer incidence. The Health Professionals Follow-up Study (HPFS) found a decreased risk for metastatic prostate cancer from consumption of fish (RR for >3 times/week vs. <2/month = 0.56, 95% CI = 0.37–0.86).19 A prospective cohort study in Sweden reported a 2.3 times higher risk of prostate cancer among men who ate no fish compared with those who ate moderate amounts.18 In contrast, a prospective study in Japan found an increased risk (RR of almost daily fish consumption vs. < 2 times/week = 1.54, 95% CI = 1.03–2.31).30 Indeed, a review on long-chain n-3 fatty acids and prostate cancer risk concluded that the evidence was still too inconclusive to make a recommendation regarding fish consumption.45

Many individual fatty acids have been investigated in relation to prostate cancer risk, although long-chain n-3 fatty acids such as eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) predominate. The HPFS reported that EPA and DHA intakes were related to lower risk of total and advanced prostate cancer, while ALA intake was associated with a higher risk of advanced prostate cancer.22 Contrary to the HPFS, one case–control study nested in a Finnish cohort of male smokers found no association for EPA, DHA and ALA intake, or for these fatty acids measured in serum, while the study found an increased risk for serum myristic acid (RR of highest quartiles vs. lowest = 1.93, 95% CI = 1.02–3.64).21 One review on ALA and prostate cancer concluded that ALA is not associated with an increased risk,46 whereas a metaanalysis of 5 cohort and case–control studies showed a positive association (RR of 1.5 g/day increase in ALA = 1.26, 95% CI = 1.10–1.45).43 Another nested case–control study found an increased risk of prostate cancer associated with selected serum trans-fatty acids.23 Finally, one review (based on 16 case–control, 7 cohort and 3 nested case–control studies) concluded that long-chain n-3 fatty acids in fish are possible promising nutrients for the dietary prevention of prostate cancer, but to date with little epidemiological support.44 Unlike the HPFS, we observed no overall association for ALA, EPA or DHA in our cohort. Mean daily intake of EPA and DHA in our study was substantially lower than in the HPFS: 0.02% vs. 0.05% of energy for EPA and 0.04% vs. 0.08% of energy for DHA (in our study vs. in the HPFS).22 We found a suggested inverse association, which was largely limited to Latino men, of n-3 fatty acids and ALA (contributing to 94% of n-3 fatty acid intake in our cohort) with prostate cancer risk. The reasons for this ethnic-specific finding are unclear, although a difference in food sources of n-3 fatty acids can be a factor; the possibility of a chance finding cannot be ruled out.

Although early studies on meat intake and prostate cancer largely showed a direct association,1, 2, 3 these were almost entirely case–control studies, whereas recent cohort studies yielded inconsistent findings on relation of total, red or processed meat.28, 29, 30, 32, 33, 47 A positive association with meat intake could reflect several constituents other than fat, including heme iron, zinc and compounds produced during the cooking process.2 Indeed, 1 cohort study reported that very well-done meat was associated with higher risk (RR of >10.0 g/day vs. no consumption = 1.42, 95% CI = 1.05–1.92).32 However, when we investigated meat doneness in our study, we found no association with prostate cancer risk.

Few studies have examined racial/ethnic differences in associations between diet and prostate cancer risk. The Cancer Prevention Study II reported an increased risk with higher intake of total red meat among Black men (RR of highest vs. lowest quartile = 2.0, 95% CI = 1.0–4.2), but not among White men.33 One population-based case–control study also observed that increased intake of foods high in animal fat was associated with prostate cancer among American blacks, but not among American whites.27 However, we did not find a stronger relationship between meat intake and risk in African Americans than in Whites in our study population.

There are several possible limitations to our study. Inaccuracy in quantifying fatty acid intakes based on a food frequency questionnaire is a concern, especially as it would tend to deattenuate associations. However, when we examined calibration-corrected fat intakes, which should be closer to truth, albeit possibly modestly, the findings were consistent. Biochemical indicators allow a more objective assessment,21 but such data were not available for this analysis. Carcinogenic compounds, including a variety of heterocyclic aromatic amines and polycyclic aromatic amines, can be formed during the cooking or smoke-preservation of meat. We did not quantify these carcinogenic compounds. However, because these carcinogens are formed under a variety of cooking methods and temperatures, it is difficult to assess their levels in meats.29 Another potential limitation is that the dietary assessment for this analysis was done only once, at baseline, whereas dietary intake may have changed during the follow-up period. However, the follow-up period was relatively short for the present analysis.

On the other hand, our study had several strengths. It has a prospective design, with a large number of subjects of several ethnic origins. The ranges of dietary intake in the study population were generally wide, and the QFFQ was validated in a previous calibration study.35 We were able to control for several potential confounding factors for prostate cancer. Lastly, the number of cases was large enough to permit ethnic-specific analyses, except in Native Hawaiian men.

In summary, while a weak protective effect of n-3 fatty acid intake was suggested, overall our findings from a large multiethnic cohort study gave no indication that intake of fat, fatty acid, cholesterol or various meats affects prostate cancer risk.

Ancillary