Monitoring for 5-aminosalicylate nephrotoxicity in adults with inflammatory bowel disease: prognostic model development and validation using data from the Clinical Practice Research Datalink

Objective To develop and validate a prognostic model for risk-stratified monitoring of 5-aminosalicylate nephrotoxicity.

Methods This UK retrospective cohort study used data from the Clinical Practice Research Datalink Aurum and Gold for model development and validation respectively. It included adults newly diagnosed with inflammatory bowel disease and established on 5-aminosalicylic acid (5-ASA) treatment between 1 January 2007 and 31 December 2019. Drug discontinuation associated with 5-ASA nephrotoxicity defined as a prescription gap of ≥90 days with decline in kidney function was the outcome. Patients prescribed 5-ASAs for ≥6 months were followed-up for up to 5 years. Penalised Cox regression was used to develop the risk equation with bootstrapping for internal validation and optimism adjustment. Model performance was assessed in terms of calibration and discrimination.

Results 13 728 and 7318 participants who contributed 40 378 and 20 679 person-years follow-up formed the development and validation cohorts with 170 (1.2%) and 98 (1.3%) outcome events respectively. Nine predictors were included in the final model, including chronic kidney disease stage 3 and hazardous alcohol use as strong predictors. Age and Body Mass Index were weak predictors. The optimism-adjusted calibration slope, C and D statistics in the development and validation data were 0.90, 0.64 and 0.98, and 1.01, 0.66 and 0.94 respectively.

Conclusion This prognostic model used information from routine clinical care and performed well in an independent validation cohort. It can be used to risk-stratify blood test monitoring during established 5-ASA treatment. A key limitation is that the decline in kidney function could have been due to factors other than 5-ASA nephrotoxicity.