Body Composition and Aspirin Dose for Colorectal Adenoma Prevention in a Randomized Clinical Trial

The U. S. Preventive Services Task Force recommends low-dose (81 mg/day) aspirin for the primary prevention of colorectal cancer for those 50–69 years old with 10-year risk of cardiovascular disease ≥10%, and advocates more research on dosing strategies (1). In randomized clinical trials (RCT), aspirin reduces metachronous colorectal adenoma risk by approximately 20% (2). A meta-analysis of RCTs found body weight modified the association between aspirin dose and colorectal cancer risk: 75–100 mg/day aspirin reduced risk among those weighing < 70 kg, and ≥325 mg/day among those weighing < 80 kg (3). We assessed aspirin treatment effects according to height, weight, and body composition measurements predicted from height and weight, in the Aspirin/Folate Polyp Prevention Study (AFPPS), a double-blind placebo-controlled RCT of 81 or 325 mg/day aspirin for the prevention of metachronous adenomas (4).

AFPPS is 3 × 2 factorial RCT comparing 81 and 325 mg/day aspirin with placebo and 1 mg/day of folic acid with placebo (4). Participants recruited from nine clinical centers between 1994 and 1998 were 21–80 years old and recently diagnosed with ≥1 colorectal adenoma. Aspirin treatment ended at a 3-year follow-up colonoscopy. We evaluated the incidence of any colorectal neoplasia ≥1 year following randomization determined by a standardized pathology review, including conventional adenomas, sessile serrated adenomas/polyps (SSA/P), or colorectal cancer. High-risk adenoma (HRA) findings were defined as ≥1 advanced adenoma (tubulovillous/villous adenomas, ≥1 cm in diameter, with high-grade dysplasia, or colorectal cancer) or ≥3 adenomas, without considering SSA/P.

Body mass index (BMI), body surface area (BSA; ref. 5), and fat-free mass (FFM; ref. 6) were calculated from self-reported baseline height and weight. Fat mass (FM) was calculated as weight minus FFM. Risk ratios (RR) with 95% confidence intervals (CI) were estimated from log-linear regression with robust SEs, adjusting for age, sex, center, and treatment assignment. Heterogeneity of RRs for aspirin according to body composition was tested using 2-df Wald tests for interactions with each dose versus placebo in intention-to-treat analyses. Secondary analyses were stratified by sex using different categories for body composition variables when appropriate. Two-sided P ≤ 0.05 was considered statistically significant.

Overall, 1,121 participants underwent randomization. Baseline characteristics by treatment assignment were previously reported, along with summaries of excellent compliance and avoidance of nonprotocol NSAIDs (4). Among 712 men and 409 women, mean (SD) height was 1.8 (0.1) m and 1.6 (0.1) m, respectively. The mean weight was 87.5 (14.6) kg and 71.9 (14.3) kg, respectively. A total of 1,084 participants (97%) completed the follow-up colonoscopy; among these, 470 (43%) had incident colorectal neoplasia, including 140 (13%) with HRA.

For each body composition measurement, risk of colorectal neoplasia and HRA increased comparing the highest with lowest category (Table 1). No statistically significant treatment effect modification was found (Table 2). Treatment effects for low-dose aspirin did not differ according to whether participants weighed <70 kg or ≥70 kg. High-dose aspirin did not, in general, result in statistically significant treatment effects, and there was no evidence to suggest greater risk reduction in those weighing < 80 kg compared with ≥80 kg. Sex-specific effects were of similar magnitudes as overall sex-adjusted effects, but estimates were less precise (Supplementary Tables S1–S4).

Our findings did not show the dose–response pattern by weight identified by Rothwell and colleagues (3). We found low-dose aspirin reduced risk of colorectal neoplasia and HRA by 25% and 48%, respectively, among those weighing ≥ 80 kg, findings not meaningfully different from those weighing < 70 kg. No effect modification was observed for other measurements of body composition.

Overall, AFPPS identified a statistically significant 19% reduced risk of adenomas for low-dose aspirin relative to placebo, and a nonsignificant 4% reduced risk for high-dose aspirin (4). A previous analysis from AFPPS reported treatment effects according to BMI, considering conventional and advanced adenomas as endpoints, but did not evaluate weight alone or other measurements of adiposity (7).

Abdominal visceral adiposity is a known risk factor for colorectal adenomas (8), but measurement requires medical imaging. We calculated FFM and FM using validated prediction equations trained on data from dual-energy x-ray absorptiometry and bioelectrical impedance analysis (6). FFM is often regarded as equivalent to lean body mass (LBM), although the former excludes lipids in the central nervous system, bone marrow, and cell membranes (accounting for ∼4% difference between FFM and LBM; ref. 6). We also calculated classical BSA, as it remains in use to dose some chemotherapies.

AFPPS had excellent follow-up and compliance, avoidance of nonprotocol NSAIDs, standardized pathology, and evaluation of two different aspirin doses by design. Predictions from height and weight may be subject to error, but the randomized design helped balance any inaccuracies by treatment assignment. Weight-based aspirin dosing aims to maximize bioavailability while minimizing side-effects such as bleeding (3), but data from AFPPS does not support the effectiveness of this approach for preventing metachronous colorectal adenomas.

D.J. Ahnen has received speakers bureau honoraria from Ambry Genetics and is a consultant/advisory board member for Cancer Prevention Pharmaceuticals. Together with the Trustees of Dartmouth College, J. A. Baron holds a use patent, not licensed, for the chemopreventive use of aspirin for colorectal cancer. No potential conflicts of interest were disclosed by the other authors.

Conception and design: M.N. Passarelli, E.L. Barry, D.J. Ahnen, J.A. Baron

Development of methodology: M.N. Passarelli

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J.R. Rees, D.J. Ahnen, J.A. Baron

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M.N. Passarelli, E.L. Barry, J.R. Rees, J.A. Baron

Writing, review, and/or revision of the manuscript: M.N. Passarelli, E.L. Barry, J.R. Rees, L.A. Mott, D.J. Ahnen, J.A. Baron

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): L.A. Mott

Study supervision: J.A. Baron

On behalf of the Polyp Prevention Study Group, the authors express their appreciation to the study participants, investigators, and staff. The authors thank Dr. Gail McKeown-Eyssen for comments on the article. Study tablets were provided by Wyeth (Madison, NJ). This research was funded by the NCI at the NIH (R01CA059005 to J.A. Baron), and the National Institute of General Medical Sciences at the NIH (P20GM104416 to M.N. Passarelli).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.