Although diet and prostate cancer has been investigated in numerous epidemiologic studies, few foods have been identified as potentially contributing to increasing or decreasing the risk of this malignancy. High intake of foods containing lycopene and selenium may decrease risk of prostate cancer while diets high in calcium may increase risk [3,13], although the epidemiologic evidence is not entirely clear. Some early studies have suggested that fat intake may be associated positively with prostate cancer [7,14,15], but recent prospective studies have reported no associations with fat consumption [16,17].

Qualitative information and quantitative data were extracted from each study that met the criteria for inclusion. Specifically, information was extracted pertaining to: the year of the study, the study population (i.e., name and nature of the cohort), geographic location of the study, years of follow-up, methods of dietary exposure ascertainment, red meat and processed meat dietary variables and how these variables were defined, the analytical comparison (i.e., the exposure contrast), the number of exposed cases, the relative risk estimates and 95% confidence intervals, and the factors that were adjusted or controlled for in the analyses. A thorough review of each article was conducted to identify cohorts that may have been analyzed in multiple publications. If results were reported in multiple publications, the inclusion of data was based on 1) the size of the study population, 2) duration of follow-up with an emphasis on the most recent publication with the longest follow-up, 3) classification and analytical categorization of red or processed meat, and 4) level of control for potential confounding factors.

Random effects models were used to calculate summary relative risk estimates (SRRE), 95% confidence intervals, and corresponding p-values for heterogeneity. This type of model assumes that the study-specific effect sizes come from a random distribution of effect sizes according to a specific mean and variance [18]. Red meat is commonly defined as beef, pork, lamb, or a combination thereof [13,19], and processed meat is generally defined as meat made largely from pork, beef, or poultry that undergoes methods of preservation, such as curing, smoking, or drying [13,19]. The definitions of red meat and processed meat varied across studies; while some studies explicitly defined these classifications other studies reported no description. Most studies reported data for variables labeled as 'red meat,' 'processed' or 'preserved' meat, although some studies reported data for single meat items, such as beef, pork, liver, or bacon. Meta-analysis models were created for high vs. low red meat and processed meat intake. In addition to high vs. low intake analyses, categorical dose-response regression meta-analyses were conducted using the method proposed by Greenland and Longnecker [20], in which the linear dose-response slope is calculated for each study while accounting for the correlation across intake categories within a study [21]. If the number of cases and person-time data were not available for each intake strata, variance weighted least squares regression was utilized to estimate the slope coefficient. Different intake units were reported across studies; therefore, we used 100 grams as the approximate serving size for red meat and 50 grams for processed meat. Sensitivity analyses were conducted for a variety of methodological quality factors, such as level of adjustment for confounding factors, specification of red/processed meat variable, and number of food items ascertained in the FFQ.

Publication bias was assessed by generating funnel plots for a visual examination, conducting correlation and regression tests for significance, and using a 'trim and fill' procedure to evaluate symmetry around the summary effect [18]. All analyses were performed using STATA [22] and Comprehensive Meta-Analysis [23].

The summary associations across the meta-analysis models of red meat intake ranged between 0.97 and 1.01, with the exception of one model (i.e., the SRRE was 1.13 for the five studies published prior to 2000), and none of the associations were statistically significant (Table ​(Table2).2). Furthermore, there was little heterogeneity across the red meat models. Analyses of high vs. low red meat intake and 100 g increment dose-response regression produced similar results; both SRREs were 1.0.

On average, summary associations for processed meat were slightly stronger in magnitude compared with red meat. However, greater heterogeneity was present in the analyses of processed meat. In sub-group analyses of the eight studies that adjusted simultaneously for at least three potentially important confounding factors, the SRRE was closer to the null compared with the three studies that did not adjust for these factors (1.04 vs. 1.25). Furthermore, the three studies that were not as fully-adjusted were published prior to the year 2000 and were not as large as the other studies. In addition, there was evidence of publication bias in the meta-analysis of processed meat. Theoretically, if all relevant studies were included in the meta-analysis, it would be expected that the funnel plot would be symmetric (i.e., even distribution of point estimates on either side of the mean effect), which was not the case for processed meat (Figure ​(Figure4).4). If the funnel plot is asymmetric, and a higher number of smaller studies are dispersed on the right side of the summary effect, there may be additional small studies for which processed meat data was not reported [23]. Using the "trim and fill" method proposed by Duval and Tweedie, these potentially missing studies are imputed on the other side of the summary effect, and the overall summary association is recalculated. This method trims the asymmetric studies from the right-hand side to identify the unbiased effect (in an iterative procedure), and then fills the plot by re-inserting the trimmed studies on the right as well as their imputed counterparts to the left the mean effect [18,23]. In the analysis of processed meat, this method indicated that four studies may be missing. After imputing data from these potential studies, the SRRE changed from 1.05 (95% CI: 0.99-1.12) to 1.02 (95% CI: 0.94-1.10). It should be noted; however, that this method is merely estimating unpublished data, rather than relying upon actual data.

The summary association for processed meat and prostate cancer in the current assessment was relatively similar to the summary association reported in the aforementioned WCRF/AICR report on diet and cancer (summary RR per serving/week = 1.11, 95% CI: 0.99-1.25) [13]. However, their analysis included data from only four cohorts, whereas, in the current assessment, data from 11 prospective studies of processed meat were analyzed. Their analysis of case-control studies produced a summary effect estimate of 1.01 (95% CI 0.98-1.04) per each processed meat serving/week and prostate cancer [13]. WCRF/AICR judged that the epidemiologic evidence regarding processed meat intake and prostate cancer was "limited-suggestive" and was based on sparse and inconsistent data [13]. The epidemiologic data for red meat and prostate cancer were not summarized in their report.

Although an evaluation of correlates of meat consumption and prostate cancer is beyond the scope of the current assessment, a few factors thought to contribute to positive associations are worth mentioning. Few studies examined fat intake from animal sources, particularly red meat sources, and prostate cancer. Le Marchand et al. [37] reported that intake of "high fat animal products" was associated positively with prostate cancer (RR = 1.6, 95% CI: 1.1-2.4), although the source of animal fat was not limited to meat, as milk and eggs were included with red meat, processed meat, and poultry. Furthermore, diet was ascertained via a small 13-item food frequency questionnaire, thus, the authors could not adjust for total energy intake. In a 1993 study, Giovannucci and colleagues [7] reported that high intake of red meat fat was associated with a greater than two-fold risk of advanced prostate cancer (RR = 2.64, 95% CI: 1.21-5.77). In contrast, in a recent analysis of the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort, Crowe et al. [16] observed inverse associations of 0.94, 0.83, and 0.84 for fat from red and processed meat and total prostate cancer, advanced prostate cancer, and high-grade prostate cancer, respectively.

Investigations of cooking practices, meat doneness, and dietary mutagens have not produced patterns of associations consistent with increasing the risk of prostate cancer, although the available epidemiologic data are limited to few studies. Barbequed and pan-fried meat has been associated inversely with prostate cancer in three large prospective studies [24,29,41]. However, in a sub-group analysis of the aforementioned NIH-AARP cohort, significant positive associations were reported for grilled/barbequed meat and total and advanced prostate cancer but a non-significant inverse association was observed for fatal prostate cancer [26]. In the same study, no associations were observed for pan-fried, microwaved, or broiled meat and total, advanced, or fatal prostate cancer [26]. The relationship between doneness of meat intake and prostate cancer has been inconsistent as two studies reported significant positive associations between consumption of well and very well done meat and prostate cancer risk [24,28] , and two studies observed no associations for well or very well done meat and prostate cancer [17,26]. Few studies have evaluated dietary mutagens and prostate cancer, and no statistically significant associations have been observed for total mutagenic activity, 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline (MeIQx), 2-amino-3,4,8-trimethylimidazo[4,5-f]quinoxaline (DiMeIQx), or Benzo[a]pyrene (BaP), with most RRs slightly above or below the null value [24,28], although a marginally significant RR of 1.28 (95% CI: 1.00-1.65) for BaP has been reported for advanced prostate cancer [26]. A statistically significant positive association between 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP) and total prostate cancer and incident prostate cancer was reported in one study [28], although null or inverse associations were observed between the highest quintile of PhIP and total, advanced, and fatal prostate cancer in another study [26]. The highest quintiles of heme iron, nitrite from meat, and nitrate from meat were associated positively and significantly with advanced prostate cancer among participants in the NIH-AARP cohort, however, no significant associations were observed for total or fatal prostate cancer with the exception of heme iron and total prostate cancer [26]. Additional studies are necessary to fully evaluate any potential associations between consumption preferences, dietary mutagens, heme iron, nitrite/nitrate and prostate cancer.