Article : Chemoprevention of Recurrent Colorectal Neoplasia

Paul S. Mueller, MD, MPH, FACP reviewing Dulai PS et al. BMJ 2016 Dec 5.


Nonaspirin nonsteroidal anti-inflammatory drugs conferred the most protection, but were most likely to cause adverse events.

Evidence suggests that, in people with previous colorectal neoplasia, aspirin and nonsteroidal anti-inflammatory drugs (NSAIDs) lower risk for metachronous neoplasia (i.e., occurring after resection of initial neoplasia). In this network meta-analysis of 14 randomized trials that involved 12,000 adults with previous colorectal neoplasia (either polyps or cancer with curative resection), researchers determined the comparative effectiveness and safety of candidate chemopreventive agents in preventing advanced colorectal neoplasia (villous component, high-grade dysplasia, or cancer) within 5 years of the colonoscopy that identified the initial polyps or cancer. Candidate agents were compared head-to-head or with placebo.

Nonaspirin NSAIDs (i.e., celecoxib and sulindac) were superior to placebo and all other candidate agents (i.e., aspirin, calcium, vitamin D, or folic acid, alone or in combination) for preventing recurrent neoplasia. Low-dose (≤160 mg/day) aspirin ranked second in efficacy. Compared with controls, nonaspirin NSAIDs reduced risk for advanced metachronous neoplasia from 7.4% to 3.9% in lower-risk patients (those with 1 or 2 previous small adenomas). Among higher-risk patients (those with more advanced previous neoplasia), nonaspirin NSAIDs conferred a greater risk reduction — from 16.3% to 6.7%. Risk reduction with low-dose aspirin was about half that of nonaspirin NSAIDs.

Regarding safety, nonaspirin NSAIDs were least safe and low-dose aspirin was safest. Serious adverse events occurred in 15.2% of low-dose aspirin recipients, 18.7% of placebo recipients, and 22.1% of nonaspirin NSAID recipients.


CITATION(S):

Dulai PS et al. Chemoprevention of colorectal cancer in individuals with previous colorectal neoplasia: Systematic review and network meta-analysis. BMJ 2016 Dec 5; 355:i6188. 


JWatch

BACK