Background
Colorectal cancer (CRC) screening programmes have been shown effective in reducing CRC incidence and mortality. Risk stratification offers opportunities for further optimisation of screening programmes by better balancing the harms and benefits of screening. To this day, nearly all population screening programmes only differentiate their screening approach by age, resulting in a fixed upper age limit of screening. While further risk stratification based on prior faecal haemoglobin (f-Hb) concentration seems promising, studies are mainly limited to the target population of screening. Risk-stratified screening beyond the upper age limit might be beneficial as well, considering the increasing life expectancy and the harm-to-benefit ratio of screening widely varying amongst elderly.

Aim & Methods
This study assessed interval CRC (IC) risk in individuals who have reached the upper age limit of screening (≥75 years), and whether this risk differs by the f-Hb concentration of their last faecal immunochemical test (FIT). Data of individuals with a negative FIT (<47µg Hb/g faeces) in their last screening round were selected in the national screening database and linked to the Dutch cancer registry to identify CRCs diagnosed within 24 months of the negative FIT. IC risk was compared between subgroups (sex, last f-Hb concentration and screening round) and tested with a Chi-squared test. Survival and multivariable logistic regression analyses assessed whether sex, last f-Hb concentration and screening round were associated with IC risk or stage distribution (early vs late stage), respectively.

Results
A total of 305,761 individuals with a complete follow-up (24 months) after a negative FIT were included. The overall IC risk was 21.6 per 10,000 individuals with a negative FIT. Individuals with detectable f-Hb (>0µg Hb/g faeces) in their last screening round developed IC more often than those without (HR 4.87, 95%CI: 4.19-5.65) (Table 1). Moreover, their ICs were more likely to be diagnosed at a late stage compared to individuals without detectable f-Hb (OR 1.45, 95%CI: 1.06-2.01). Individuals who had participated in three or four screening rounds had a lower IC risk than those who had participated only once (HR 0.49, 95%CI: 0.37-0.63). Screening round was not associated with stage distribution. IC risk and stage distribution did not differ between men and women.

Conclusion
Our results show that IC risk in individuals aged ≥75 differs substantially by the last f-Hb concentration and number of prior screening rounds. A risk-stratified upper age limit may therefore be beneficial. While individuals with detectable f-Hb in their final FIT may benefit from additional screening, screening the elderly is not without risks. Follow-up studies should assess the harm-to-benefit ratio of a risk-stratified upper age limit and identify the optimal screening scenario.

Table 1; Overview of IC risk, including results of a Cox regression analysis with IC as outcome variable

Folkert Van Kemenade

Appearances