Adolescent intake of animal products has been proposed to contribute to prostate cancer (PCa) development because of its potentially carcinogenic constituents and influence on hormone levels during adolescence.
We used data from 159,482 participants in the NIH-AARP Diet and Health Study to investigate associations for recalled adolescent intake of red meat (unprocessed beef and processed red meat), poultry, egg, canned tuna, animal fat and animal protein at ages 12–13 years with subsequent PCa risk and mortality over 14 years of follow-up. Cox proportional hazard regression was used to calculate hazard ratios (HRs) and 95% confidence intervals (CIs) of total (n = 17,349), advanced (n = 2,297) and fatal (n = 804) PCa.
Suggestive inverse trends were observed for adolescent unprocessed beef intake with risks of total, advanced and fatal PCa (multivariable-adjusted P-trends = 0.01, 0.02 and 0.04, respectively). No consistent patterns of association were observed for other animal products by PCa outcome.
We found evidence to suggest that adolescent unprocessed beef intake, or possibly a correlate of beef intake, such as early-life socioeconomic status, may be associated with reduced risk and mortality from PCa. Additional studies with further early-life exposure information are warranted to better understand this association.
Although prostate cancer (PCa) is the most commonly diagnosed cancer and the second leading cause of cancer death among US men , little is known about its aetiology and even less about prevention strategies. A large focus of epidemiologic research over the past few decades has been on the possible role of diet in prostate carcinogenesis. This avenue of research was motivated by positive ecologic correlations between per-capita dietary intake (e.g., total fat, animal protein and milk) and PCa mortality , and by increased PCa incidence and mortality rates among men who migrated from low- to high-PCa risk countries .
Although several migration studies observed higher risks of PCa with longer time since or younger age at migration [4,5,6,7,8], most research on diet to date has focused on mid- to later-life diet and only a few studies have focused on early-life or adolescent diet [9,10,11,12,13,14,15]. Nonetheless, adolescence may be an important sensitive period for PCa because the prostate may be more susceptible to deleterious exposures while it is growing and developing rapidly . This hypothesis is supported by several observations, including the finding of stronger associations for early-life than for adult exposures (e.g., hig-fat diet) with prostate lesion development in animal studies [17, 18]; the mathematical estimation of PCa initiation (i.e., first genomic alteration) as early as puberty in some men , similar to estimated for a growing number of other cancers [20, 21]; the observation of PCa precursor lesions and small PCa foci in men as young as their twenties and thirties in human histologic studies [22,23,24,25,26,27]. Additional supportive observations include the divergence in prostate lesion prevalence , as well as in PCa incidence and mortality , by race in men in their thirties and forties, and the observation of positive associations for characteristics influenced by early-life diet and other early-life exposures (e.g., height and timing of puberty [30, 31]) with PCa incidence and mortality [32,33,34,35,36]. Although genetic factors may also play a role, these findings are consistent with a possible early-life environmental (e.g., dietary) contribution to prostate carcinogenesis.
To inform the possible role of adolescent diet in prostate carcinogenesis, we previously took advantage of data collected in the large, ongoing NIH-AARP Diet and Health (AARP) Study to investigate recalled adolescent dairy product and calcium intake in relation to subsequent PCa risk and mortality . In that study, we observed positive associations for adolescent dairy product and calcium intake, but these associations attenuated after controlling for red meat intake, suggesting that adolescent red meat intake might be related to PCa risk or mortality. Red meat intake has been proposed to contribute to PCa risk by several possible mechanisms, including increasing exposure to mutagenic heterocyclic amines, polycyclic aromatic hydrocarbons, N-nitroso compounds (for processed meat, in particular) and haem iron [38, 39]. Consistent with these possible mechanisms, a recent pooled analysis found a positive association between adult unprocessed red meat intake and PCa risk in North American men , and a recent meta-analysis found a positive association for adult processed, but not total, red meat intake with PCa risk . Several other studies have also observed positive associations for adolescent red meat intake with cancers of the breast, colorectum and pancreas [41,42,43,44].
Besides red meat, foods high in animal fat and protein more broadly have been proposed to contribute to PCa risk by a variety of additional mechanisms. These include increasing steroid hormone levels, decreasing levels of their binding proteins, increasing insulin-like growth factor levels [38, 45,46,47,48] and contributing to the earlier onset and longer duration of puberty, as well as longer lifetime exposure to steroid and other hormones [49,50,51]. However, these mechanisms may not apply equally to all animal products. For example, fish contains nutrients that could potentially protect against PCa (e.g., omega-3 polyunsaturated fatty acids), a hypothesis supported by the observation of inverse associations between adult fish intake and PCa in some epidemiologic studies [52, 53].
To further investigate the relation between adolescent animal product intake and PCa, we examined associations for a wide range of animal products (i.e., unprocessed beef, processed red meat, poultry, canned tuna and eggs) and nutrients (animal fat and protein) consumed in adolescence with later PCa risk and mortality in the AARP Study.
Study population and design
The AARP Study is a large prospective cohort study of AARP members administered by the US National Cancer Institute. The study began in 1995–1996 when baseline questionnaires were mailed to AARP members 50–71 years of age residing in six states (California, Florida, Louisiana, New Jersey, North Carolina and Pennsylvania) and two metropolitan areas (Atlanta and Detroit). In total, 339,669 men completed the baseline questionnaire, which assessed their demographic and lifestyle characteristics, including their diet in the previous 12 months. In 1996–1997, participants completed a second questionnaire (“risk factor questionnaire”), which assessed additional lifestyle characteristics, including their diet during adolescence (ages 12–13 years). Finally, in 2004, participants completed a third questionnaire (“follow-up questionnaire”) to update and expand their medical and lifestyle information. Participants are followed for cancer incidence through linkage with cancer registries and for mortality through linkage with the US Social Security Administration Death Master File, National Death Index, state cancer registries, questionnaire responses from next of kin and responses to other mailings . This study was approved by the Special Studies Institutional Review Board of the National Cancer Institute.
We limited the present analysis to men who completed both the baseline and risk factor questionnaires (n = 213,090), and then further excluded men who (1) completed questionnaires by proxy (n = 19,184); (2) reported any cancers including PCa before the date of completion of the risk factor questionnaire (n = 15,703, including 2833 PCa cases); (3) reported extreme energy intake during adolescence (<500 or >4500 kcal/day, n = 10,003); (4) had four or more missing adolescent food items (n = 4121) or (5) had any missing adolescent items on meat, canned tuna and egg intake (n = 4597). After these exclusions, 159,482 remained in the final analytic sample.
Assessment and definitions of adolescent animal product, fat and protein intake
Diet from ages 12–13 years was assessed by a 37-item food-frequency questionnaire (FFQ, Supplementary Fig. 1) included on the risk factor questionnaire. This FFQ was designed to capture major dietary contributors to nutrients of greatest interest for cancer development at the time of FFQ development: fat, fibre, beta-carotene and vitamin C . Participants were asked to select from 1 to 8 frequency options for each food item, ranging from never to ≥2/day. The portion size was not queried on the FFQ. Non-dairy animal products assessed included ground beef, roast beef or steak, bacon or sausage, hot dogs or frankfurters, cold cuts or luncheon meats, chicken or turkey, gravy, canned tuna, eggs and butter. These items were used to derive unprocessed beef intake as the sum of ground and roast beef intake; processed red meat intake as the sum of bacon, hot dog and cold cut intake; total red meat intake as the sum of unprocessed beef and processed red meat intake. All food group intakes were modelled in frequency (times per week) for ease of interpretation.
Nutrient composition for food items was estimated using data from the first nationwide individual food consumption survey: the US Department of Agriculture 1965-66 Household Food Consumption Survey . This survey was used to estimate (1) the frequency of intake for the highest intake category (median of two times/day for all food items, except for whole milk for which a value of three times/day was assigned), (2) the median portion size for boys 12–13 years of age and (3) the nutrient content of food items. Animal fat and protein intake were calculated as the product of the frequency of intake, portion size and nutrient value from the above-mentioned animal products, including dairy products (whole milk, cheese and ice cream). Animal fat and protein intakes were estimated as % of total adolescent energy intake and modelled as quintiles.
Prostate cancer ascertainment and definitions
Total PCa was identified through linkage to state cancer registries and was defined as a first primary diagnosis of PCa (ICD-9 185 or ICD-10 C61). In a previous validation study of AARP members, cancer ascertainment was found to be highly valid through state cancer registries (sensitivity = 89.2% and specificity = 99.5% ). Information on cancer grade (histological) and stage (SEER disease extent and summary stage) was also obtained from state registries, and was used to define: (1) advanced PCa as clinical stages T3–T4, lymph node involvement (N1), metastasis (M1) or PCa death (obtained by linkage to the National Death Index Plus) through December, 2011; (2) more advanced PCa as advanced PCa, excluding stage T3 and (3) fatal PCa as PCa death. We also explored early-onset PCa (defined as PCa diagnosed before age 65 ) because we hypothesised that early-life exposures might lead to earlier PCa development, as observed for family history and race [29, 58].
Assessment of covariates
We assessed all covariate information by questionnaire. From the baseline questionnaire, we obtained participants’ demographic and lifestyle information, including race/ethnicity; education; marital status; cigarette smoking history; adolescent and adult physical activity levels; medical history (including diabetes and family history of cancer); current waist circumference and height; weight at ages 18, 35, 50 and at the time of questionnaire completion. From the risk factor questionnaire, we obtained participants’ prostate-specific antigen (PSA) testing and digital rectal examination (DRE) screening history, and from the follow-up questionnaire, we obtained participants’ father’s occupation, which we used as a surrogate measure of participants’ adolescent socioeconomic status.
We used Cox proportional hazard regression to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for associations of each animal product, and animal fat and protein exposure with PCa risk. Participants contributed time-at-risk from the month of the return of the adolescent FFQ until the month of PCa diagnosis (total and advanced PCa analyses only), death, re-location outside of the study catchment area, loss to follow-up or December 31, 2011, whichever occurred first. To calculate P-trends, we assigned the mid-point value to each frequency category or the median to each quintile and then modelled these values as continuous variables. The shape of all associations (i.e., linear versus non-linear) was explored by spline regression (Supplementary Fig. 2).
We selected potential confounders a priori based on findings from previous studies. Analyses were adjusted first for age as a continuous variable (model 1) and then for variables that could serve solely as confounders or could address detection bias. These included adolescent energy intake; race/ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic and other); family history of PCa (yes/no); education (12 years or less, post high school or some college, college graduate and post graduate); marital status (married, separated and never married); smoking history (never, former and current smoker); adult alcohol intake, waist circumference and body mass index (BMI); adolescent and adult physical activity (never or rarely, 1–3 times per month, 1–2 times per week, 3–4 times per week and 5 or more times per week); screening PSA and DRE history (never, once and more than once), diabetes history (yes/no), father’s occupation [professional or technical, managerial, other non-manual (e.g., bank teller), manual, in a trade (e.g., carpenter), other manual (e.g., farm labourer) and unknown] and adult dietary counterparts of the exposures of interest to control for the correlation between adolescent and adult dietary exposures (model 2). In the third set of models, we additionally adjusted for variables that could serve as either confounders or mediators (adolescent BMI and height), and in a fourth set, we additionally controlled for intake of other adolescent food groups. Finally, to explore the effect of individual animal products/groups independent of other animal foods, we ran the fifth set of models further controlling for other animal products/groups and macronutrients (i.e., mutually adjusting for red meat, poultry, and canned tuna; and for vegetable and/or total fat and protein, as appropriate). As the results from these five sets of models were similar, we present only those from age-adjusted and the fourth set of models in the main tables.
To further investigate the potential for detection bias (i.e., differential PCa screening and detection by exposure ), we performed analyses stratified by the number of PSA tests in the 3 years before baseline. Additional stratified or restricted analyses included (1) analyses restricted to men <55 years of age at the time of completion of the adolescent FFQ because of previously observed greater validity of adolescent diet recall in people <55 years of age [60,61,62,63,64,65]; (2) stratified analyses by father’s occupation as a marker of adolescent socioeconomic status, food availability and expenditures; (3) stratified analyses by calendar period during which participants passed through puberty (completely during the Great Depression or World War II [WWII], partially during WWII and partially after WWII, and completely following WWII) to capture food availability during adolescence and the likelihood that diet recalled for early puberty (ages 12–13 years) also reflected diet later in puberty. As no clear patterns were observed in these stratified/restricted analyses, we limited the presentation of these findings to the supplementary information (Supplementary Table 1). Finally, as a further sensitivity analysis, we added prevalent PCa cases back into the analytic sample to test for any influence that excluding these men might have introduced on our findings (e.g., exclusion of cases with greater adolescent red meat and earlier PCa onset or earlier adoption of PCa screening); this sensitivity analysis also resulted in similar findings. All analyses were performed using SAS version 9.4.
After all relevant exclusions, 159,482 participants were included in the analysis. Between the date of completion of the risk factor questionnaire and December 31, 2011, 17,349 men were diagnosed with any PCa, 2297 with advanced and 804 with fatal PCa. On average, participants were in their early 60 s when they completed the risk factor questionnaire, and most were non-Hispanic White, had completed some post high school education, were married and former or current smokers and had been screened for PCa at least twice (Table 1 and Supplementary Table 2). In general, men tended to have been normal weight and physically active in adolescence and to be overweight and less physically active in adulthood.
Red meat was the most commonly consumed form of animal fat and protein, followed by eggs and poultry. Canned tuna was not commonly consumed during participants’ adolescence. Approximately 7% of participants consumed red meat ≤2 times/week, 15% 3–4 times/week, 20% 5–6 times/week, 33% once/day and 27% ≥2 times/day. For canned tuna, these values were 49% ≤11 times/year, 28% 2 times/month, 19% 1–2 times/week and 4% ≥3 times/week. Overall, animal fat and protein intake were high: the 10th, 50th and 90th percentile values were 27% of total energy intake (%E), 36%E, 43%E for animal fat and 8%E, 12%E, 15%E for animal protein, respectively.
Compared to men with lower intakes during adolescence, men with higher intakes of total red meat, animal fat and protein tended to be younger and were more likely to have entered puberty after the Great Depression/WWII (Table 1). They also had slightly higher BMI and consumed grains, fruits and vegetables with lesser frequency as adolescents; and engaged in less physical activity and consumed a greater amount of red meat and alcohol, and a greater percentage of their energy from animal fat and protein as adults.
In age- and multivariable-adjusted models, significant inverse trends were observed for greater adolescent intake of unprocessed beef (multivariable-adjusted HR = 0.91, 95% CI: 0.84–0.98; P-trend = 0.01) and eggs (HR = 0.90, 95% CI: 0.85–0.96; P-trend < 0.01) with total PCa (Table 2). Although a suggestive inverse trend was also observed for total red meat (P-trend = 0.06), the category-specific HRs did not clearly indicate a dose–response relationship. No associations were observed for other animal products, including processed red meat, poultry and canned tuna, and for total animal fat and protein. In analyses stratified by the number of PSA tests before baseline, trends for unprocessed beef intake were slightly, but not statistically significantly stronger among lesser screened men (0 PSA tests: multivariable-adjusted HR = 0.86, 95% CI: 0.72–1.03, P-trend = 0.11; 1 PSA test: HR = 0.85, 95% CI: 0.72–0.99, P-trend = 0.04) and weaker among more heavily screened men (≥2 PSA tests: HR = 0.94, 95% CI: 0.84–1.04, P-trend = 0.21; P-interaction = 0.36, comparing ≥2 to 0 PSA tests), whereas no differences in patterns were observed for egg intake.
A generally similar pattern of findings was observed for advanced PCa as for total PCa (Table 3). A significant inverse trend was observed for adolescent unprocessed beef intake and advanced PCa, with the strongest trend observed among men without a history of PSA screening (HR = 0.63, 95% CI: 0.40–0.99, P-trend = 0.04) and the weakest among men with a history of at least two screening tests (HR = 0.84, 95% CI: 0.63–1.14, P-trend = 0.27; P-interaction = 0.06, comparing ≥2 to 0 PSA tests). Suggestive inverse trends were observed for processed (P-trend = 0.07) and total red meat intake (P-trend = 0.06), but category-specific HRs did not indicate clear dose–response relationships. For egg intake, a trend of similar magnitude was observed for advanced as for total PCa, but this trend was no longer statistically significant. No significant trends were observed for poultry, canned tuna and animal fat and protein intake.
Finally, a generally similar pattern of findings was observed for more advanced and fatal PCa as for advanced PCa (Table 4 and Supplementary Table 3). A significant inverse trend was observed for greater adolescent unprocessed beef intake and fatal PCa, with the strongest trend observed among men without a history of PSA screening (HR = 0.55, 95% CI: 0.27–1.12, P-trend = 0.10) and the weakest among men with a history of at least two PSA tests (HR = 0.79, 95% CI: 0.46–1.34, P-trend = 0.38; P-interaction = 0.08, comparing ≥2 to 0 PSA tests). Significant inverse trends were also observed for processed red meat (albeit not a linear trend), and for animal fat and protein intake. Trends for animal fat intake did not vary by screening history, whereas those for animal protein intake were limited to men without a history of PCa screening (0 PSA tests: HR = 0.62, 95% CI: 0.41–0.94, P-trend = 0.02; ≥2 PSA tests: HR = 1.02, 95% CI: 0.73–1.44, P-trend = 0.90). No significant trends were observed for intakes of total red meat, eggs, poultry and canned tuna.
In our large cohort analysis of adolescent diet and PCa, greater adolescent intake of unprocessed beef was associated with reduced risks of total, advanced and fatal PCa, whereas no consistent patterns of association were observed for other animal products by PCa outcome or screening history.
Overall, our findings for adolescent animal product intake are difficult to compare to the small literature on this topic because of minimal overlap with previous studies in terms of foods and nutrients studied, geography, PCa outcome and sample size. Nonetheless, our findings are generally similar to those from most previous studies. Although our inverse results for adolescent unprocessed beef intake with total, advanced and fatal PCa differ in statistical significance from those from a small Swedish case–control study, they are still similar in direction to their previous findings (multivariable-adjusted OR = 0.7, 95% CI: 0.3–1.4, P-trend = 0.71 comparing 30 to 6 times/month for “meat/sausage” intake). Our inverse to null findings for egg and animal fat intake with risks of total and advanced PCa are also similar to their previous findings for intake of eggs (multivariable-adjusted OR = 1.2, 95% CI: 0.7–2.0, P-trend = 0.52 comparing 75 to 2 times/month) and total fat (multivariable-adjusted OR = 0.7, 95% CI: 0.4–1.1, P-trend = 0.38 comparing 226 to 78 g/day). They are also similar to those from another small, American case–control study that compared cases of aggressive PCa to controls by adolescent intake of high saturated fatty acid foods (multivariable-adjusted odds ratio [OR] = 1.1, 95% CI: 0.6–1.9 comparing high to low intake derived from whole milk, cheese, ice cream, eggs and red meat ). Finally, our generally null to inverse findings for animal product intake are consistent with those from another American study that observed a null association for a vegetarian lifestyle before age 15 and PCa among Seventh Day Adventist members (multivariable-adjusted relative risk = 0.80, 95% CI: 0.51–1.26 comparing one or two Adventist and practicing vegetarian parents to none ). Thus, although our findings differ in some respects from those from previous studies, they are consistent, at least, in their lack of support for the hypothesis that greater adolescent intake of meat, eggs and animal fat and protein promotes PCa development.
With respect to fish consumption, our results are also difficult to compare to those from previous studies because of likely differences in type and frequency of fish consumed (based on geography), PCa outcomes and sample size. Comparing our findings to those from a previous Icelandic case–control study, our null findings for adolescent canned tuna intake are similar to their null association for adolescent intake of fish (including fish as a meal, and fish as a topping on bread or in a salad) with total and advanced PCa (multivariable-adjusted OR = 0.87, 95% CI: 0.66–1.13, P-trend = 0.25 for total PCa comparing >4 to ≤2 portions/week; and multivariable-adjusted OR = 0.81, 95% CI: 0.46–1.42 for advanced PCa comparing >2 to ≤2 portions/week ). Our findings differ, however, from those from another Nordic study—the previously mentioned Swedish case–control study—that observed suggestive positive findings for fish intake and PCa (multivariable-adjusted OR = 1.8, 95% CI: 1.0–3.5, P-trend = 0.09 comparing 30 to 2 times/month) . Despite these differences, none of the small number of studies on adolescent fish intake and PCa to date support a protective role for fish consumption during adolescence and PCa risk.
Overall, our most notable findings were those for adolescent unprocessed beef intake. Our observed inverse trend for this animal product was unexpected and generally difficult to explain. Although our attenuated findings for total PCa in PSA-screened men suggested a possible role for detection bias (i.e., decreased likelihood of screening by exposure status, leading to an inverse trend with total PCa), the persistence of these findings for advanced and fatal PCa, which might be more likely to result from lack of PCa screening, make this explanation less likely. Another unlikely explanation is exclusion of prevalent PCa cases because similar results were observed in sensitivity analyses adding these men back into the analytic sample as in the main analysis. A further possible explanation for our inverse findings is extensive misclassification of adolescent unprocessed beef intake; however, expected positive associations were observed for adolescent red meat intake and colorectal cancer risk in a previous AARP Study analysis . Finally, another possible explanation for our inverse findings is confounding by greater adolescent socioeconomic status, leading to a lower lifetime risk of PCa by as-yet undefined mechanisms. We attempted to control for adolescent socioeconomic status by adjusting for father’s occupation as a crude measure of socioeconomic status, but the potential for residual confounding remains, particularly as data on father’s occupation were missing for a large proportion of the cohort. This possibility may also explain our previous inverse findings for adolescent intakes of cheese and ice cream, two food items correlated with adolescent red meat intake, and PCa risk , as well as inverse findings for adolescent red meat intake and postmenopausal breast cancer risk in the AARP Study (unpublished data).
Compared to the previous literature, our analysis brings several methodologic strengths to the study of adolescent diet and PCa. These include use of adolescent diet data collected before PCa diagnosis to avoid concerns of differential recall bias; an extremely large sample size to allow for evaluation of advanced and fatal PCa, in addition to total PCa; and collection of extensive demographic, lifestyle, and dietary data to control for numerous factors related to adolescent diet and PCa. The limitations of our analysis include its use of a shorter adolescent FFQ (37 food items), which assessed unprocessed beef but not other unprocessed red meat products, such as pork and lamb. However, as this FFQ was designed to capture foods commonly consumed in the 1940s–1960s , and as fat was one of the original nutrients around which this FFQ was developed, its shorter length may be less of a concern. Another limitation of the adolescent FFQ is its lack of information on portion size. However, imputation of portion size using data collected near in time to participants’ adolescence should have helped to improve exposure estimation. A further limitation of our analysis is its use of recalled adolescent diet data. Although we observed similar results among younger (<55) and older participants (≥55 years of age), we cannot rule out the possibility that non-differential misclassification of exposure due to the long recall time between adolescence and FFQ completion may have contributed to the attenuation of some of our findings to the null. In addition, as our study population was mainly white, our findings may not generalise to men of other race/ethnicities. Finally, minimal information on PCa screening (only before baseline) limited our ability to control for the influence of PCa screening in the analyses.
To conclude, in one of the few studies to investigate adolescent animal product intake and PCa risk, we found evidence to suggest that adolescent unprocessed beef intake, or possibly a correlate of beef intake, such as early-life socioeconomic status, may be associated with reduced risk and mortality from PCa. As this study is the first to observe such an association, additional studies with further early-life exposure information are warranted to replicate and better understand this association.
AARP Study data are available through the National Cancer Institute.
Siegel RL, Miller KD, Jemal A. Cancer statistics, 2020. CA: Cancer J Clin. 2020;70:7–30.
Armstrong B, Doll R. Environmental factors and cancer incidence and mortality in different countries, with special reference to dietary practices. Int J Cancer. 1975;15:617–631.
Gronberg H. Prostate cancer epidemiology. Lancet. 2003;361:859–64.
Parkin DM, Steinitz R, Khlat M, Kaldor J, Katz L, Young J. Cancer in Jewish migrants to Israel. Int J Cancer. 1990;45:614–21.
Tyczynski J, Tarkowski W, Parkin DM, Zatonski, W. Cancer mortality among Polish migrants to Australia. Eur J Cancer. 1994;30A:478–84.
Whittemore AS, Kolonel LN, Wu AH, John EM, Gallagher RP, Howe GR, et al. Prostate cancer in relation to diet, physical activity, and body size in blacks, whites, and Asians in the United States and Canada. J Natl Cancer Inst. 1995;87:652–61.
McCredie M, Williams S, Coates M. Cancer mortality in East and Southeast Asian migrants to New South Wales, Australia, 1975-95. Br J Cancer. 1999;79:1277–82.
Beiki O, Ekbom A, Allebeck P, Moradi T. Risk of prostate cancer among Swedish-born and foreign-born men in Sweden, 1961-2004. Int J Cancer. 2009;124:1941–53.
Andersson SO, Baron J, Wolk A, Lindgren C, Bergstrom R, Adami HO. Early life risk factors for prostate cancer: a population-based case-control study in Sweden. Cancer Epidemiol Biomark Prev. 1995;4:187–92.
Torfadottir JE, Steingrimsdottir L, Mucci L, Aspelund T, Kasperzyk JL, Olafsson O, et al. Milk intake in early life and risk of advanced prostate cancer. Am J Epidemiol. 2012;175:144–53.
Mills PK, Beeson WL, Phillips RL, Fraser GE. Cohort study of diet, lifestyle, and prostate cancer in Adventist men. Cancer. 1989;64:598–604.
Slattery ML, Schumacher MC, West DW, Robison LM, French TK. Food-consumption trends between adolescent and adult years and subsequent risk of prostate cancer. Am J Clin Nutr. 1990;52:752–7.
Torfadottir JE, Valdimarsdottir UA, Mucci LA, Kasperzyk JL, Fall K, Tryggvadottir L, et al. Consumption of fish products across the lifespan and prostate cancer risk. PLoS ONE. 2013;8:e59799.
Whittemore AS, Paffenbarger RS Jr., Anderson K, Lee JE. Early precursors of site-specific cancers in college men and women. J Natl Cancer Inst. 1985;74:43–51.
Dirx MJ, van den Brandt PA, Goldbohm RA, Lumey, LJAJOE. Energy restriction in childhood and adolescence and risk of prostate cancer: results from the Netherlands Cohort Study. Am J Epidemiol. 2001;154:530–7.
Sutcliffe S, Colditz GA. Prostate cancer: is it time to expand the research focus to early-life exposures? Nat Rev Cancer. 2013;13:208–518.
Suttie AW, Dinse GE, Nyska A, Moser GJ, Goldsworthy TL, Maronpot RR. An investigation of the effects of late-onset dietary restriction on prostate cancer development in the TRAMP mouse. Toxicol Pathol. 2005;33:386–97.
Cai LQ, Imperato-McGinley J, Zhu YS. Regulation of prostate 5alpha-reductase-2 gene expression and prostate weight by dietary fat and caloric intake in the rat. Prostate. 2006;66:738–48.
Diamandis EP, Yu H. Does prostate cancer start at puberty? J Clin Lab Anal. 1996;10:468–9.
Van Egeren D, Escabi J, Nguyen M, Liu S, Reilly CR, Patel S, et al. Reconstructing the lineage histories and differentiation trajectories of individual cancer cells in myeloproliferative neoplasms. Cell Stem Cell. 2021;28:514–23. e519
Mitchell TJ, Turajlic S, Rowan A, Nicol D, Farmery J, O’brien T, et al. Timing the landmark events in the evolution of clear cell renal cell cancer: TRACERx renal. Cell. 2018;173:611–23. e617
Sakr WA, Grignon DJ, Crissman JD, Heilbrun LK, Cassin BJ, Pontes JJ, et al. High grade prostatic intraepithelial neoplasia (HGPIN) and prostatic adenocarcinoma between the ages of 20-69: an autopsy study of 249 cases. In Vivo. 1994;8:439–43.
Gu FL, Xia TL, Kong XT. Preliminary study of the frequency of benign prostatic hyperplasia and prostatic cancer in China. Urology. 1994;44:688–91.
Jackson MA, Kovi J, Heshmat MY, Jones GW, Rao MS, Ahluwalia BS. Factors involved in the high incidence of prostatic cancer among American blacks. Prog Clin Biol Res. 1981;53:111–32.
Sakr WA, Haas GP, Cassin BF, Pontes JE, Crissman JD. The frequency of carcinoma and intraepithelial neoplasia of the prostate in young male patients. J Urol. 1993;150:379–85.
Sanchez-Chapado M, Olmedilla G, Cabeza M, Donat E, Ruiz A. Prevalence of prostate cancer and prostatic intraepithelial neoplasia in Caucasian Mediterranean males: an autopsy study. Prostate. 2003;54:238–47.
Yin M, Bastacky S, Chandran U, Becich MJ, Dhir R. Prevalence of incidental prostate cancer in the general population: a study of healthy organ donors. J Urol. 2008;179:892–5.
Jahn JL, Giovannucci EL, Stampfer MJ. The high prevalence of undiagnosed prostate cancer at autopsy: implications for epidemiology and treatment of prostate cancer in the prostate-specific antigen-era. Int J Cancer. 2015;137:2795–802.
Howlader N, Noone A, Krapcho M, Noone A, Neyman N, Aminou R et al. SEER cancer statistics review. 1975–2009 (vintage 2009 populations). Bethesda, MD: National Cancer Institute. 2012. https://seer.cancer.gov/archive/csr/1975_2009_pops09/. Accessed 11 June 2021.
Alimujiang A, Colditz GA, Gardner JD, Park Y, Berkey CS, Sutcliffe S. Childhood diet and growth in boys in relation to timing of puberty and adult height: the Longitudinal Studies of Child Health and Development. Cancer Causes Control. 2018;29:915–26.
Gunnell D, Okasha M, Smith GD, Oliver SE, Sandhu J, Holly JM. Height, leg length, and cancer risk: a systematic review. Epidemiol. Rev. 2001;23:313–42.
Cook MB, Gamborg M, Aarestrup J, Sorensen TI, Baker JL. Childhood height and birth weight in relation to future prostate cancer risk: a cohort study based on the copenhagen school health records register. Cancer Epidemiol Biomark Prev. 2013;22:2232–40.
Aarestrup J, Gamborg M, Cook MB, Baker JL. Childhood height increases the risk of prostate cancer mortality. Eur J Cancer. 2015;51:1340–5.
Bonilla C, Lewis SJ, Martin RM, Donovan JL, Hamdy FC, Neal DE, et al. Pubertal development and prostate cancer risk: Mendelian randomization study in a population-based cohort. BMC Med. 2016;14:66.
Zuccolo L, Harris R, Gunnell D, Oliver S, Lane JA, Davis M, et al. Height and prostate cancer risk: a large nested case-control study (ProtecT) and meta-analysis. Cancer Epidemiol Biomark Prev. 2008;17:2325–36.
Emerging Risk Factors, C. Adult height and the risk of cause-specific death and vascular morbidity in 1 million people: individual participant meta-analysis. Int J Epidemiol. 2012;41:1419–33.
Lan T, Park Y, Colditz GA, Liu J, Wang M, Wu K, et al. Adolescent dairy product and calcium intake in relation to later prostate cancer risk and mortality in the NIH-AARP Diet and Health Study. Cancer Causes Control. 2020;31:891–904.
McCullough ML, Giovannucci EL. Diet and cancer prevention. Oncogene. 2004;23:6349–64.
Bylsma LC, Alexander DD. A review and meta-analysis of prospective studies of red and processed meat, meat cooking methods, heme iron, heterocyclic amines and prostate cancer. Nutr J. 2015;14:125.
Wu K, Spiegelman D, Hou T, Albanes D, Allen NE, Berndt SI, et al. Associations between unprocessed red and processed meat, poultry, seafood and egg intake and the risk of prostate cancer: a pooled analysis of 15 prospective cohort studies. Int J Cancer. 2016;138:2368–82.
Ruder EH, Thiébaut AC, Thompson FE, Potischman N, Subar AF, Park Y, et al. Adolescent and mid-life diet: risk of colorectal cancer in the NIH-AARP Diet and Health Study. Am J Clin Nutr. 2011;94:1607–19.
Gordon-Dseagu V, Thompson FE, Subar AF, Ruder EH, Thiébaut A, Potischman N, et al. A cohort study of adolescent and midlife diet and pancreatic cancer risk in the NIH-AARP Diet and Health Study. Am J Epidemiol. 2017;186:305–17.
Linos E, Willett WC, Cho E, Frazier L. Adolescent diet in relation to breast cancer risk among premenopausal women. Cancer Epidemiol Biomark Prev. 2010;19:689–96.
Linos E, Willett WC, Cho E, Colditz G, Frazier LA. Red meat consumption during adolescence among premenopausal women and risk of breast cancer. Cancer Epidemiol Biomark Prev. 2008;17:2146–51.
Chan JM, Gann PH, Giovannucci EL. Role of diet in prostate cancer development and progression. J Clin Oncol. 2005;23:8152–60.
Allen NE, Appleby PN, Davey GK, Key TJ. Hormones and diet: low insulin-like growth factor-I but normal bioavailable androgens in vegan men. Br J Cancer. 2000;83:95–97.
Barnard RJ, Ngo TH, Leung PS, Aronson WJ, Golding LA. A low-fat diet and/or strenuous exercise alters the IGF axis in vivo and reduces prostate tumor cell growth in vitro. Prostate. 2003;56:201–6.
Ngo TH, Barnard RJ, Cohen P, Freedland S, Tran C, deGregorio F, et al. Effect of isocaloric low-fat diet on human LAPC-4 prostate cancer xenografts in severe combined immunodeficient mice and the insulin-like growth factor axis. Clin Cancer Res. 2003;9:2734–43.
Chen C, Chen Y, Zhang Y, Sun W, Jiang Y, Song Y, et al. Association between dietary patterns and precocious puberty in children: a population-based study. Int J Endocrinol. 2018;2018:4528704.
Gunther AL, Karaolis-Danckert N, Kroke A, Remer T, Buyken AE. Dietary protein intake throughout childhood is associated with the timing of puberty. J Nutr. 2010;140:565–71.
Hur J, Giovannucci E. Racial differences in prostate cancer: does timing of puberty play a role? Br J Cancer. 2020;123:349–54.
Szymanski KM, Wheeler DC, Mucci LA. Fish consumption and prostate cancer risk: a review and meta-analysis. Am J Clin Nutr. 2010;92:1223–33.
Reese AC, Fradet V, Witte JS. Omega-3 fatty acids, genetic variants in COX-2 and prostate cancer. J Nutrigenetics Nutrigenomics. 2009;2:149–58.
Schatzkin A, Subar AF, Thompson FE, Harlan LC, Tangrea J, Hollenbeck AR, et al. Design and serendipity in establishing a large cohort with wide dietary intake distributions: the National Institutes of Health-American Association of Retired Persons Diet and Health Study. Am J Epidemiol. 2001;154:1119–25.
USDA. Food and nutrient intake of individuals in the US, Spring 1965. Washington, DC: US Dept of Agriculture; 1972.
Michaud D, Midthune D, Hermansen S, Leitzmann M, Harlan L, Kipnis V, et al. Comparison of cancer registry case ascertainment with SEER estimates and self-reporting in a subset of the NIH-AARP Diet and Health Study. J Registry Manag. 2005;32:70–5.
Rosenblatt KA, Wicklund KG, Stanford JL. Sexual factors and the risk of prostate cancer. Am J Epidemiol . 2001;153:1152–8.
Xu Y, Huang D, Wu Y, Ye D, Zhang N, Gao Y, et al. Family history is significantly associated with prostate cancer and its early onset in Chinese population. Prostate. 2019;79:1762–6.
Potosky AL, Miller BA, Albertsen PC, Kramer BS. The role of increasing detection in the rising incidence of prostate cancer. J Am Med Assoc. 1995;273:548–52.
Maruti SS, Feskanich D, Colditz GA, Frazier AL, Sampson LA, Michels KB, et al. Adult recall of adolescent diet: reproducibility and comparison with maternal reporting. Am J Epidemiol. 2005;161:89–97.
Maruti SS, Feskanich D, Rockett HR, Colditz GA, Sampson LA, Willett WC. Validation of adolescent diet recalled by adults. Epidemiology. 2006;17:226–9.
Potischman N, Weiss HA, Swanson CA, Coates RJ, Gammon MD, Malone KE, et al. Diet during adolescence and risk of breast cancer among young women. J Natl Cancer Inst. 1998;90:226–33.
Dwyer JT, Gardner J, Halvorsen K, Krall EA, Cohen A, Valadian I. Memory of food intake in the distant past. Am J Epidemiol. 1989;130:1033–46.
Chavarro JE, Rosner BA, Sampson L, Willey C, Tocco P, Willett WC, et al. Validity of adolescent diet recall 48 years later. Am J Epidemiol. 2009;170:1563–70.
Wolk A, Bergstrom R, Hansson LE, Nyren O. Reliability of retrospective information on diet 20 years ago and consistency of independent measurements of remote adolescent diet. Nutr Cancer. 1997;29:234–41.
Stiebeling H. Are we well fed?: a report on the diets of families in the United States. U.S. Department of Agriculture. Miscellaneous Publication Number 430. 1941;41–278.
We thank Dr. Linda Liao for assistance with acquiring AARP Study data, and Dr. Stephanie Smith-Warner, Sherry Yaun and Tao Hou for assistance defining prostate cancer outcomes.
This analysis was funded by the Barnes-Jewish Hospital Foundation, the Alvin J. Siteman Cancer Center and the Institute for Clinical and Translational Sciences.
The authors declare no competing interests.
Ethics approval and consent to participate
This study was approved by the Special Studies Institutional Review Board of the National Cancer Institute. Consent from participants was obtained from investigators for the NIH-AARP Diet and Health Study at enrolment.
Consent to publish
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Lan, T., Park, Y., Colditz, G.A. et al. Adolescent animal product intake in relation to later prostate cancer risk and mortality in the NIH-AARP Diet and Health Study. Br J Cancer (2021). https://doi.org/10.1038/s41416-021-01463-1