Background
Despite the fecal immunochemical test (FIT) and colonoscopy both being recommended for colorectal cancer (CRC) screening, the long-term effect of FIT in comparison with colonoscopy on the reduction of CRC incidence and CRC-related mortality remains unknown.
Aims
To compare the long-term effects of CRC incidence, CRC-related mortality and all-cause mortality between subjects who underwent FIT and colonoscopy for CRC screening.
Methods
This is a cohort study of asymptomatic Chinese subjects aged 50 to 70 years recruited for population-based CRC screening in Hong Kong from 2008-2012. They were given the options to choose between a one-off colonoscopy (and surveillance if necessary) and annual FIT screening for a maximum of 4 years. Those tested positive by FIT were offered early colonoscopy for cancer or polyp detection. Letter were sent annually to the FIT group reminding them for the FIT screening. Both FIT and colonoscopy were offered for free. Follow-up was conducted by matching their unique identifier to a territory-wide electronic health record (eHR) linked to the cancer and death registry. The Kaplan–Meier estimator was used to calculate the cumulative risks of CRC incidence, CRC-related mortality and all-cause mortality between the FIT and colonoscopy groups. The risks were compared using risk difference and risk ratio.
Results
A total of 10,309 subjects (mean age [SD]: 58.1 [5.1]; male: 44.3%) received CRC screening, of which 5,690 (55.2%) underwent at least one FIT and 4,619 (44.8%) underwent colonoscopy. During a median follow-up of 13.9 years, 54 subjects (0.9%) in the FIT group and 35 (0.2%) in the colonoscopy group were diagnosed with CRC. 16 and 298 subjects in the FIT group, and 6 and 164 subjects in the colonoscopy group, died from CRC and any cause, respectively. The cumulative risk of CRC incidence, CRC-related mortality and all-cause mortality were 1.03%, 0.31% and 5.77% in the FIT group and 0.79%, 0.15% and 4.09% in the colonoscopy group. There was no statistical difference between both groups in terms of CRC incidence (risk difference: -0.24, 95% confidence interval [CI]: 0.57-1.11; risk ratio: 0.77, 95% CI: 0.48-1.21) and CRC-related mortality (risk difference: -0.24, 95% confidence interval [CI]: 0.57-1.11; risk ratio: 0.77, 95% CI: 0.48-1.21). A risk reduction of 29% in all-cause mortality was observed (risk ratio, 0.71; 95% CI, 0.58-0.85) in subjects who underwent colonoscopy screening when compared to those who underwent FIT screening (table 1 and figure 1).
Conclusions
The long-term risk of CRC incidence and CRC-related mortality was similar for subjects who underwent FIT and colonoscopy screening. Further research is required to investigate the reasons behind the reduction of all-cause mortality in the colonoscopy group.

Table 1. Long term effect of colorectal cancer screening by colonoscopy and fecal immunochemical test.
Figure 1 A. Long-term cumulative risk of colorectal cancer incidence. B. Long-term cumulative risk of colorectal cancer mortality. C. Long-term cumulative risk of all-cause cancer mortality. D. Incidence rate ratios for colorectal cancer in the colonoscopy group as compared with the fecal immunochemical group.