Screening for cancer reduced targeted cancer mortality without a significant increase in “off-target” mortality, an analysis of 17 large randomized studies showed.

Noncancer (“off-target”) mortality increased by 0.2% in screened versus non-screened patients. Separate analyses of the individual studies yielded relative risk (RR) values of 0.89 to 1.09 for off-target mortality in screened patients, all of which were associated with confidence intervals that included 1.00.

The results address a common criticism of randomized trials of screening that have all-cause mortality as the endpoint, reported Jiayao Lei, PhD, of the Karolinska Institute in Stockholm, and co-authors in JAMA Network Open.

“Our findings support, at most, a minimal increase in off-target mortality caused by cancer screening,” the authors concluded. “Understanding the potential harms related to overdiagnosis and overtreatment remains crucial, but the overall balance of risk appears to favor the continuation of cancer screening practices because screening saves lives.”

“Combining the actual numbers of target-specific deaths with off-target deaths mixes two outcomes, which are strikingly heterogeneous, and conceals important information,” they added. “Looking at all-cause mortality in a meta-analysis of screening trials is problematic because one would not expect any effect in any trial in which there was no effect on targeted mortality … . Our findings highlighted the importance of quantifying the benefits and harms of screening separately and using appropriate methods to combine these to estimate the overall impact on a composite endpoint such as all-cause mortality.”

The results ran counter to another meta-analysis of the same randomized trials, which showed no improvement in lifetime years gained, with the possible exception of screening sigmoidoscopy for colorectal cancer. Nonetheless, the authors of that meta-analysis did not “advocate that all screening should be abandoned,” but suggested that organizations and individuals “who promote cancer screening tests by their effect to save lives may find other ways of encouraging screening.”

In an editorial that accompanied the negative analysis, three of the co-authors asserted that once screening has begun, “it is difficult or indeed impossible to phase out screening programs, even when research has failed to document significant benefits. We believe that transparent, evidence-based discussions about cancer screening tests with a delicate balance of benefits and harms have become a threat to powerful stakeholders.”

In contrast to the earlier meta-analysis, which had a primary endpoint of lifetime years gained with screening, Lei and colleagues focused on off-target mortality.

“If screening does not extend life, it must be associated with increased off-target mortality to balance the reduction in cancer-specific mortality,” they argued. “Here, we assess the difference in off-target mortality between the arms of randomized clinical trials of cancer screening.”

The 17 studies, published between 1989 and 2022, comprised 1.3 million participants and 18.5 million person-years of follow-up. The studies addressed screening for colorectal (eight studies), prostate (three), lung (three), breast (two), and multiple (one) cancers. Lei and co-authors did not address potential biases of each trial, and they collected data from the original publications, not meta-analyses.

The meta-analysis showed virtually identical off-target mortality in the screened and non-screened populations, as reflected in the RR of 1.00 (95% CI 1.00-1.01). The confidence intervals associated with the 0.2% increase in off-target mortality ranged from -0.5% to 0.9%, meaning that off-target mortality in the screened population could be no more than 0.9% greater as compared with the non-screened study participants.

In the non-screened control arms of the trials, targeted cancer deaths, as a proportion of all deaths, ranged from 2.6% to 5.0% in the colorectal cancer studies, 2.3% to 5.8% of the prostate cancer studies, 23.3% to 33.1% of the lung cancer studies, and 3.9% to 10.8% of the breast cancer studies, as well as 13.5% of cancer deaths among female participants and 13.8% of male participants in the multi-cancer screening study.

A 20% reduction in targeted mortality would correspond to an approximately 0.7%-5.1% reduction in all-cause mortality across cancer types.

“For breast and targeted lung screening, the reduction in on-target deaths greatly exceeded the maximal expected increase in off-target deaths,” the authors stated.

“Where screening has been shown to have a beneficial effect on targeted cancer mortality, combining that result with the lack of a substantial association with off-target mortality means that would infer with reasonable confidence that screening must be saving lives,” they added.

Please enable JavaScript to view the

comments powered by Disqus.