In the updated analysis, 71% of trial participants had died as ascertained from the National Death Index.
In intention-to-treat analyses, the risk of colorectal cancer mortality was a significant 32% lower in the group screened annually (relative risk, 0.68) and 22% lower in the group screened biennially (RR, 0.78) as compared with the nonscreened group, according to data reported at the meeting and recently published (N. Engl. J. Med. 2013;369:1106-14).
"We don’t know what happened after 1998," Dr. Shaukat reminded attendees. "At best, the effects that we are seeing might be dilute, and if truly the control group had remained unscreened, we would have seen perhaps larger differences."
The absolute cumulative colorectal cancer mortality rates were 0.02 with annual screening, 0.02 with biennial screening, and 0.03 with no screening. "This separation [in curves] started at about 13 years of follow-up and persisted through 30 years of follow-up," she pointed out.
All-cause mortality was statistically indistinguishable between groups, although the trial was underpowered to assess this outcome.
In subgroup analyses, there was a near-significant interaction of screening with sex (P = .06), whereby the benefit was greater among men (RR, 0.62) than among women (RR, 0.83).
Additionally, among men, there was a significant interaction with age (P = .04), whereby screening was most beneficial among those 60-69 years old at baseline (RR, 0.46). The benefit among women appeared to be restricted to those who started screening at age 60 or later.
"We don’t have information on the incidence of colon cancer, and hence we can’t comment on right- versus left-sided colon cancer mortality," Dr. Shaukat noted.
Dr. Shaukat disclosed no conflicts of interest.