The trusted source for
healthcare information and
The View from the Right Side: Effectiveness of Colonoscopy
Abstract & commentary
By Malcolm Robinson, MD, FACP, FACG, AGAF
Synopsis: Although colonoscopy is associated with reduced deaths from colorectal cancer, its effect seems to be limited to deaths from cancer developing in the left colon.
Source: Baxter NN, et al. Association of colonoscopy and death from colorectal cancer: A population-based, case-control study. Ann Intern Med 2009;150:1-8.
Colorectal cancer (CRC) is often deadly, and it remains the second most frequent cause for cancer-related death in North America. Although its utility for this purpose has never been validated in randomized clinical trials, colonoscopy has been strongly recommended by various medical societies as the preferred screening method for CRC. Some indirect evidence for efficacy of colonoscopy comes from randomized trials of fecal occult blood testing (with colonoscopy for positive tests) that seem to indicate a reduction in CRC mortality. Previous case-control studies have suggested that colonoscopy might reduce CRC incidence by 50% and CRC deaths by 60%. However, these studies largely excluded women and they primarily evaluated sigmoidoscopy rather than complete colonoscopy. This exceptionally large and statistically complicated Canadian study attempted to evaluate the association between colonoscopy and subsequent CRC deaths. Ontario provincial patients aged 52-90 who were diagnosed with CRC between January 1996 and December 2001 and died of CRC by December 2003 were each matched with 5 controls that did not die of CRC. There were 10,292 case patients (with CRC) and 51,460 controls identified. A total of 719 case patients (7%) and 5031 controls (9.8%) had undergone colonoscopy. Complete colonoscopy was strongly associated with fewer deaths from left-side CRC (odds ratio [OR], 0.33; confidence interval [CI], 0.28-0.39), but there was not a positive association with death from right-side CRC (OR, 0.99; CI, 0.86-1.14). In this retrospective provincial data analysis, screening colonoscopy could not be differentiated from colon exams designed to evaluate signs or symptoms. The authors hypothesized that colonoscopy might not be effective in the right colon due to the greater likelihood of poor right colon bowel cleansing, the presence of non-pedunculated or even flat or depressed polyps or cancers, or other causes of inadequate colonic mucosal evaluation.
In an accompanying editorial by David Ransohoff, MD, of the University of North Carolina, some potential defects of the study are elaborated. Included is the fact that this was not a randomized controlled trial, that most colonoscopies (70%) were performed by internal medicine physicians and surgeons and family physicians rather than gastroenterologists, and that rapid colonoscope withdrawal could have led to missed lesions (a recently documented phenomenon not necessarily known at the time of this study). It was also noted that any significant symptoms prior to colonoscopy might have preferentially associated the procedure with CRC and death, thus obscuring any favorable effect of colonoscopy on cancer development and mortality. However, this possibility may have been countered by the exclusion of patients who developed cancer within 6 months of the colonoscopy. There seems to be some literature suggesting that right colon cancers may grow more rapidly than those in the left colon, and this could make routine colonoscopy less effective in right-side CRC prevention. The bottom line seems to be that the precise utility of colonoscopy for CRC and CRC death prevention will remain uncertain until large randomized controlled studies are completed. The previous hope that 90% of CRC could be prevented by colonoscopic screening seems quite unrealistic, and patients should be told that the actual reduction of risk is closer to 60% or 70%. As Dr. Ransohoff comments, this reduction in risk is still quite respectable. For example, there is no proven cancer mortality reduction with screening for prostate cancer. Likewise, breast cancer screening reduces cancer mortality by 25% or less. The American Society for Gastrointestinal Endoscopy (ASGE) was quite alarmed by the release of this article. As might have been expected, the ASGE response was that patients should only have colonoscopies done by highly experienced and qualified gastroenterologists and that colonoscopies almost certainly have dramatically improved since the era of the Canadian data acquisition. One hopes that this is true, but only data will answer this question.