Objectives Much of the literature on cancer overdiagnosis due to screen detection focuses on the assessment of trends in mortality, incidence, and indolent cancers. Overdiagnosed cases after cancer screening also occur because screen-detected patients die of a non-cancer cause of death (competing cause of death) prior to their hypothetical symptomatic date of cancer diagnosis. The purpose of this study is to characterize cause of death after lung cancer diagnosis for participants in the National Lung Screening Trial (NLST).

Method Participants were 55–74 years at randomization to chest X-ray or low dose computed tomography (LDCT) and were asymptomatic current or former smokers with a 30 pack-year history. The primary endpoint was lung cancer specific mortality verified through death certificate and medical records. Participants lost to follow-up were submitted to the National Death Index. Median follow-up duration was 6.5 years. We focus on LDCT as it is the standard screening modality for lung cancer detection.

We examined cause of death by cancer characteristics and detection method, calculating time from date of diagnosis to date of death (or censor if alive at the end of the trial) by plotting Kaplan-Meier curves. We enumerated the experience (health history, tumor characteristics, treatment and complications) of those who died of a cause other than lung cancer. In a sensitivity analysis we examined cause of death as reported by death certificate compared with endpoint verification.

Results Of the 53 248 participants randomized, 2403 were diagnosed with lung cancer during the trial, 1079 of which were in the LDCT arm. At the end of trial follow-up, 54% did not have a reported death, 41% had a reported death due to lung cancer, and approximately 4% had a death due to a cause other than lung cancer. Among those with death from non-lung cancer causes, 60% were diagnosed with stage I lung cancer and 78% had a positive LDCT screening test. Competing causes of death were primarily other cancers and chronic diseases such as cardiovascular and respiratory illnesses. Overall, those with a cause of death other than lung cancer appear to live longer (median 957 days (IQR 315–1307)) than those with a lung cancer cause of death (median 300 days, (IQR 109–691)).

Conclusions Of lung cancer patients who were deceased at that end of the trial, the majority of deaths were due to lung cancer. Among those who were diagnosed but died of competing causes, it may be of public health importance to characterize their experience to elucidate the harms due to screening as well as possible health interventions for co-existing conditions.