WASHINGTON—With up to 23 years of follow-up data on thousands of participants in the National Polyp Study (NPS), it is clear that colonoscopy with removal of polyps produces "a substantial, long-term reduction in colorectal cancer (CRC) mortality," said Ann Zauber, PhD, a biostatistician at Memorial Sloan-Kettering Cancer Center.
The landmark study had previously shown that colonoscopy with polypectomy markedly reduced the incidence of the disease. The new data on mortality provides "an even more clinically important indicator," said Dr. Zauber, who presented the study at Digestive Disease Week 2007 (abstract 268).
The National Polyp Study enrolled more than 9,000 participants between 1980 and 1990 at seven centers. Each participant had an initial colonoscopy, during which any polyps were removed. About 2,600 of the participants had polyps, and, of these, 1,418 were randomized to one of two surveillance schedules—follow-up at 1 and 3 years, or follow-up at more than 3 years. The other patients with adenomas refused randomization.
The new analysis compared CRC death rates among participants with adenomas to death rates that would have been expected had they not had their polyps removed. Deaths among participants and causes of deaths were obtained by matching the study records with the National Death Index, based on social security number, birth date, and other identifying factors; about 98% of the records were matched, according to a subsample of 161 known deaths in the study.
Dr. Zauber and her colleagues found 12 CRC deaths among those with adenomas, 5 in the group randomized to structured follow-up and 7 among those who had refused randomization. Comparing these "observed" deaths to the deaths that would have been expected without colonoscopy, they found substantial differences. Among the general population, 39 CRC deaths would have been expected, according to SEER data. The 12 actual deaths represented a 69% reduction.
The authors also looked at observed deaths in relation to cumulative CRC deaths that would have been expected among patients with adenomas as predicted by a computer model. Compared to the 150 deaths predicted by the model, which is reliable and well-validated, Dr. Zauber said, colonoscopy with polypectomy reduced deaths by 92%.
Finally, the researchers compared deaths in the adenoma group to deaths among NPS participants who had no neoplasms at the initial colonoscopy. CRC death rates were comparable between the two groups.
