ImportanceColorectal cancer (CRC) can compromise the mucosal barrier and subsequently allow a route for bacterial invasion into the portal system or systemic circulation. Despite preliminary data suggesting that patients who experienced pyogenic liver abscess (PLA) have higher CRC incidence rates, data from outside the Southeast Asian population are sparse.ObjectiveTo investigate whether there is an association between PLA and the subsequent incidence of diagnosed CRC using the Veterans Health Administration (VHA) cohort.Design, Setting, and ParticipantsA patient-level matched retrospective cohort study was conducted at 127 VHA hospitals across the US from January 1, 2003, to December 31, 2020. Patients were followed up to 10 years from PLA diagnosis. Data analysis was performed from April 14, 2002, to October 31, 2023. All patients who were admitted to VHA hospitals with a diagnosis of PLA were included. For each patient with PLA, up to 3 controls without diagnosis of PLA, matching age, sex, and health care facility, were selected.ExposurePyogenic liver abscess.Main Outcomes and MeasuresThe primary outcome was CRC diagnosis during the follow-up period. A multivariable Fine-Gray subdistribution regression model with time-dependent coefficient was used to estimate the time-varying hazard ratio (HR) of CRC incidence while accounting for mortality as a competing event.ResultsA total of 8286 patients with PLA (male, 96.5%; mean [SD] age, 65.8 [11.9] years) and 23 201 patient-level matched controls (male, 96.3%; mean age, 65.3 [11.7] years) were included. A diagnosis of CRC was found in a significantly higher proportion of patients with PLA compared with controls (1.9% [159 of 8286] vs 0.8% [196 of 23 201]; P < .001). The incidence of CRC was significantly higher among patients with PLA during the first 3 years from PLA diagnosis (HR, 3.64; 95% CI, 2.70-4.91 at 0.5 years; HR, 2.51; 95% CI, 1.93-3.26 at 1 year; HR, 1.74; 95% CI, 1.33-2.28 at 2 years; and HR, 1.41; 95% CI, 1.05-1.89 at 3 years), but not significant after 3 years. This association was not observed among patients whose PLA was likely secondary to cholangitis or cholecystitis (HR, 1.78; 95% CI, 0.89-3.56 at 0.5 years).Conclusions and RelevanceIn this patient-level matched retrospective cohort study, a significantly higher incidence of CRC was observed up to 3 years from PLA diagnosis. The findings suggest that offering CRC screening for patients with cryptogenic PLA may be useful, especially patients who have not been screened according to guidelines.