Predictive Algorithm for Thiopurine-Induced Hepatotoxicity in Inflammatory Bowel Disease Patients.

Area of Science:

• Hepatology
• Pharmacogenomics
• Gastroenterology

Background:

• Azathioprine (AZA) and mercaptopurine (MP) are thiopurine therapies for inflammatory bowel disease (IBD).
• Hepatotoxicity is a significant adverse drug reaction, leading to treatment discontinuation in approximately 25% of IBD patients within 3 months.
• Optimizing predictive algorithms for thiopurine-induced hepatotoxicity is crucial for patient benefit.

Purpose of the Study:

• To validate and optimize a predictive algorithm for thiopurine-associated hepatotoxicity in IBD patients.
• To improve the identification of patients who may benefit from conventional thiopurine therapy.

Main Methods:

• A multicenter observational study of thiopurine-naive IBD patients receiving AZA or MP.
• Primary outcome: hepatotoxicity within 12 weeks.
• Validation and optimization of two algorithms using age, sex, BMI, and 6-methylmercaptopurine ribonucleotide concentrations at T=1.

Main Results:

• 21 out of 229 patients (9%) developed hepatotoxicity.
• A significant difference in BMI was observed between patients with and without hepatotoxicity (P=0.022).
• Algorithm 1 showed 56% sensitivity and 68% specificity; Algorithm 2 showed 50% sensitivity and 77% specificity.

Conclusions:

• Existing algorithms demonstrated limited predictive accuracy for thiopurine-induced hepatotoxicity.
• Changes in thiopurine prescription patterns, including lower MP dosages, may have influenced algorithm performance.
• Further refinement of predictive models is necessary for clinical application.