PREDICTORS FOR RESPONSE TO PRUCALOPRIDE IN PATIENTS WITH CHRONIC CONSTIPATION

Abstract

Objectives: Prucalopride is a selective 5-HT4 agonist, effective for treatment of chronic constipation, as demonstrated in three pivotal phase III trials and a trial in elderly patients. The aim of this study was to explore if baseline patient characteristics could predict efficacy of prucalopride after four weeks treatment. Two efficacy outcome measures were considered: R1) response defined as an increase from baseline of ≥1 point in patient-reported satisfaction with treatment; R2) response defined as an average of ≥3 spontaneous complete bowel movements (SCBM) per week. Methods: Data from four trials were combined (n=1256). Univariate associations of baseline characteristics with response were investigated. Stepwise logistic regression was used to build a predictive model and classification trees analysis (CTA) models were constructed on data from a random selection of two-thirds of the patients. The predictive value of the constructed model was evaluated on the remaining one-third using ROC-curve analysis. Results: Univariate associations identified 4 covariates strongly associated with R1 (satisfaction with treatment, abdominal pain, laxative use, rectal symptoms) and 10 with R2 (SCBM, CBM, SBM, BM, fewer BM than you like, satisfaction with BM frequency and regularity, abdominal symptoms, laxative use, frequency of BM in preceding six months). The final stepwise logistic regression models contained the same baseline covariates, whereas the CTA-models contained fewer covariates, but had a comparable predictive value (67% positive predictive value, 80% sensitivity and 41% specificity for R1 and 44% positive predictive value, 17% sensitivity and 91% specificity for R2). Conclusions: Self-reported satisfaction with treatment, symptoms, and laxative use at baseline has meaningful predictive value for response to treatment with prucalopride in patients with chronic constipation.