Consumption of total protein and protein sources and colorectal cancer risk in men

B61

Background: While many studies show a positive association between intakes of red and processed meat and colorectal cancer risk, few have examined intakes of total protein or types of protein in relation to colorectal cancer incidence or have examined these associations with colorectal cancer subsites.
 >Methods: We comprehensively analyzed protein intake and colorectal cancer risk in the Health Professionals’ Follow-up Study. Intakes of total protein and protein sources were assessed by a 131-item food frequency questionnaire in 1986, 1990, 1994, 1998, and 2002. Between 1986 and 2004, 969 colorectal cancer cases were ascertained from self-reports on biannual follow-up questionnaires and confirmed by medical records. We used Cox proportional hazards models with time-varying variables to estimate multivariate relative risks (MVRRs) based on baseline or cumulative updated dietary intakes adjusting for age, total energy intake, body mass index, physical activity, family history of colorectal cancer, history of endoscopy, pack-years smoked before age 30, dietary calcium, dietary folate, aspirin use, and alcohol intake. We also used multivariate nutrient density models to examine the substitution of total protein for other macronutrients in relation to risk of colorectal cancer.
 >Results: Overall, total protein intake was not associated with colorectal cancer risk; MVRR (95% CI) for highest vs. lowest quintile (q5 vs. q1) of cumulative updated intake was 0.94 (0.75-1.16) for an isoenergetic substitution of protein for carbohydrate. Results were similar when we used only baseline total protein intake and when we modeled the substitution of total protein for total fat or for specific types of dietary fat. However, among men consuming > 30 grams of alcohol per day, total protein intake was inversely associated with risk; MVRR_{q5 vs. q1} (95% CI) was 0.29 (0.07-1.21), P_trend = 0.04. For cumulative updated sources of protein, the MVRRs _{q5 vs. q1} (95% CI) of colorectal cancer were 1.14 (0.90-1.43), P_trend = 0.35 for red meat; 1.30 (1.02-1.64), P_trend = 0.02 for processed meat; 1.10 (0.89-1.36), P_trend = 0.55 for chicken; and 0.77 (0.62-0.97), P_trend = 0.10 for fish. For subsites of the disease, the positive association with processed meat intake did not vary substantially. However, men in the highest quintile of red meat intake had an increased, although not statistically significant, risk of both proximal colon (MVRR _{q5 vs. q1} (95% CI) = 1.44 (0.97-2.16), P_trend = 0.12) and rectal (MVRR _{q5 vs. q1} (95% CI) = 1.45 (0.86-2.45), P_trend = 0.18) cancers. In addition, men in the highest quintile of fish intake had a 50% lower risk of rectal cancer (but not colon cancer) compared to men in the lowest quintile. Similar RRs were observed when we examined the substitution of red meat protein, processed meat protein, chicken protein, or fish protein for isoenergetic amounts of “other protein” (total protein minus the protein of interest).
 >Conclusions: Our findings suggest that total protein intake is not associated with the risk of colorectal cancer overall. Specific sources of protein, particularly processed meat, may increase the risk, whereas fish may be beneficial.
 >