Beginning surveillance colonoscopy five years after removal of high-risk adenomas was noninferior to beginning at three years for colorectal cancer (CRC) incidence, according to an interim analysis of the ongoing European Polyp Surveillance (EPoS) II trial published in The New England Journal of Medicine. At 5.5 years of follow-up, cumulative CRC incidence was 0.77% in the five-year group vs. 0.82% in the three-year group.
“The results of our trial may encourage updates to the recommendations for surveillance intervals,” wrote corresponding authors Rodrigo Jover, MD, PhD, of Hospital General Universitario Dr. Balmis in Alicante, Spain, and Michael Bretthauer, MD, PhD, of the University of Oslo in Norway, and colleagues.
The noninferiority trial included patients with high-risk adenomas from 49 endoscopy centers across eight European countries. High-risk adenomas were defined as at least one adenoma with a diameter of 10 millimeters or larger, high-grade dysplasia, or villous growth; or three to 10 adenomas of any kind. Patients were randomly assigned to undergo their first surveillance colonoscopy five years after polyp removal or at three years. The authors noted that three years is the interval recommended in current U.S. and international guidelines. Patients in the three-year group also underwent surveillance colonoscopy at five years.
The primary endpoint is the 10-year cumulative incidence of CRC, with a prespecified noninferiority margin of 0.7 percentage points for the upper boundary of the confidence interval for the between-group difference. The current interim analysis was conducted after 5.5 years of follow-up.
Of the 10,799 patients who underwent randomization, 5,398 were assigned to the five-year group and 5,401 to the three-year group. The between-group difference in cumulative CRC incidence was −0.05 percentage points. The upper boundary of the 99.12% confidence interval for that difference was 0.68 percentage points, just under the 0.7-point margin — “narrowly meeting the criterion for noninferiority,” the authors wrote. They said they used a stringent threshold at the interim analysis to reserve most of the trial’s allowable statistical error for the 10-year analysis.
“This finding may trigger a discussion of the categorization of these patients as being at high risk for colorectal cancer,” the authors wrote. The finding they referred to was that more than 99% of patients did not receive a CRC diagnosis within five years after polyp removal.
The trial was randomized and required a complete colonoscopy with adequate bowel cleansing and removal of all identified polyps at enrollment. For those reasons, according to the authors, its estimates “may be more relevant to clinical practice than those from previous studies.”
The distribution of cancer stage at diagnosis did not appear to differ substantially between groups, based on the authors’ descriptive comparison. Five patients died of CRC — three (0.06%) in the five-year group and two (0.04%) in the three-year group. The investigators did not perform statistical testing on deaths because of the small number of events.
Patients in the five-year group underwent 4,180 (48.5%) fewer colonoscopies during follow-up than those in the three-year group (4,434 vs. 8,614); the three-year group’s protocol included an additional surveillance procedure. More-frequent surveillance identified more lesions:
advanced adenomas in 14.55% of the three-year group vs. 11.07% of the five-year group
any adenoma in 71.77% vs. 54.82%
serrated polyps in 41.63% vs. 25.95%
The authors noted that confidence intervals for these secondary outcomes were not adjusted for multiple comparisons and should not be used to infer treatment effects. Colonoscopy-related adverse events occurred in 0.63% of the five-year group and 0.52% of the three-year group. The authors found no evidence of a substantial difference between groups.
“Whether the higher level of detection is beneficial for patients remains to be seen,” the authors wrote. They noted that the between-group difference in advanced adenomas was small and that even advanced adenomas grow slowly. “The 10-year analysis may provide further insight.”
The investigators noted that adherence to the planned five-year surveillance colonoscopy was limited: 70% of patients in the five-year group and 65% of those in the three-year group underwent the procedure. Patients who did not undergo surveillance may have had undiagnosed CRC, so inverse probability weighting was used to account for missing colonoscopy data. There was also no sample-size calculation for this five-year interim analysis, so the investigators used what they described as a “very conservative” confidence interval.
The trial’s primary endpoint will be assessed once 10-year follow-up is complete.
The trial was funded by the Research Council of Norway and others. Dr. Jover reported no disclosures, and Dr. Bretthauer reported research grant support from the Research Council of Norway and the Kristian Gerhard Jebsen Foundation. Several coauthors reported relationships with companies including Olympus, Boston Scientific, Fujifilm, Pentax of America, Medtronic, Norgine, Ipsen, and Exact Sciences.