Immortal time bias in a retrospective study examining colorectal cancer mortality according to adherence to colonoscopy

Correspondence to Dr Conchubhair Winters, University of Leeds Leeds Institute of Medical Research at St James's, Leeds LS9 7TF, UK; Conchubhair.winters@nhs.net We read with interest the study by Zorzi et al.1 The authors conducted a retrospective cohort study comparing outcomes between individuals with a positive faecal immunochemical test (FIT) as part of an Italian screening programme, dependent on whether they were adherent with a colonoscopy within the programme, reporting an adjusted HR of 2.03 (95% CI 1.68 to 2.44) for mortality from colorectal cancer (CRC) among non-adherent individuals during a median 4.9 years of follow-up. Correct approaches would include using a time dependent analysis,4 studying only ‘survivors’ of the immortal period by moving the start of follow-up to the end of the immortal period,5 or moving the start of follow-up to the date of colonoscopy in adherent individuals, with a date assigned according to the distribution of adherent individuals’ immortal time for the non-adherent group.6 In conclusion, the risk of residual confounding due to inherent differences between these two populations, as well as the high likelihood of immortal time bias, represent major methodological limitations of this study and, therefore, the results should be viewed with caution. Colonoscopy later than 270 days in a fecal immunochemical test-based population screening program is associated with higher prevalence of colorectal cancer.