Abstract
The effect of different sequences of invitations to Faecal Occult Blood Test (FOBT) screening regarding colorectal cancer (CRC) incidence has never been evaluated. In 2008-2012, all residents in Stockholm-Gotland, Sweden, born 1938-1954, were randomly assigned by birth year to different calendar years of invitation to guaiac-based FOBT (g) or Faecal Immunochemical Test (f) screening at 60-69 years (1-5 rounds), or not (0). Linkage was made to the national Cancer- and Cause of Death Registers on CRC diagnosis and mortality 1958-2020, and the Swedish Colorectal Cancer Register regarding stage. Follow-up started age 60 and CRC incidence, calculated per 100,000 person-years, was assessed during screening (age 60-69) and post screening (age 70-73). Stage I-II and III-IV was assessed post screening. 364,668 individuals were included. During screening, incidence rate ratio was significantly higher in sequences (0, g, g, g, g) (RR 1.25, 95% CI 1.09-1.43), (g, g, g, g, f) (RR 1.17, 95% CI 1.01-1.35), and (g, g, f, f, f) (RR 1.14, 95% CI 1.01-1.29). Post screening, the largest decrease was seen in sequences (g, g, g, g, f) and (g, g, g, f, f), RR 0.65, 95%, CI 0.47-0.90, and RR 0.53, 95% CI 0.30-0.94, respectively. There was an overall decreasing trend along sequences from (0, 0, 0, 0, g) to (g, g, f, f, f) post screening and both stages I-II and III-IV (p < 0.001). We could demonstrate a decreased CRC incidence post screening proportional to the number of invitations with implications for future modeling studies and risk-based screening strategies.