Determining the Optimum Outcome for Trials Evaluating Symptom Improvement in Functional Gastrointestinal Disorders: 1334

Purpose: Functional bowel diseases are common and therapy remains suboptimal. Trials often use how much the patients' symptoms have improved as an outcome but the optimum cut off that defines improvement is unclear. This has hampered the development of treatments for functional dyspepsia (FD) and irritable bowel syndrome (IBS) and an FDA report has highlighted the need for validated, well-defined outcome measures in these disorders. We have evaluated whether the stringency of definition of symptom improvement (e.g. improvement in symptoms versus absence of symptoms) impacts on the outcome of trials in a systematic review. Methods: We identified all placebo controlled randomized controlled trials (RCTs) that evaluated pharmacological therapy in FD or IBS from Cochrane systematic reviews. Eligible studies reported on pain or global symptom improvement as a dichotomous outcome and gave at least two cut-off points with different stringencies for symptom improvement. Interventions where there were four or more eligible RCTs were synthesized using a random effects model with relative risk (RR) of remaining symptomatic as the outcome. Heterogeneity was determined by the I squared statistic. Results: 132 RCTs were identified, 95 studies were excluded and 37 were eligible. There were 5 different intervention/functional disease combinations. In all cases the most stringent definition of symptom improvement resulted in a greater treatment effect and 2/5 comparisons the more stringent definition gave a statistically significant result in favor of active treatment whilst the less stringent definition did not (table). In 3/5 cases the more stringent definition gave less heterogeneity between studies, in one case the heterogeneity remained very similar and in one case the less stringent definition resulted in the least heterogeneity (Table).Table: [1334] Outcomes according to stringency of symptomatic improvementsConclusion: A systematic review of available data suggests that a stringent definition of symptoms, such complete absence of symptoms, is the best at differentiating active treatment from placebo in RCTs of functional gastrointestinal diseases and should be considered as the primary outcome in future trials of therapy. Disclosure: Dr Paul Moayyedi has been on the advisory boards for AstraZeneca, Nycomed and Johnson and Johnson. He has also accepted speaker fees from these companies as well as Wyeth and Takeda. His chair is funded in part by an unrestricted donation given to McMaster University by AstraZeneca. Dr Grigorios Leontiadis has acted as a consultant for AstraZeneca and Axcan Pharma and has received speaking honoraria from AstraZeneca, Sanofi-Aventis, Janssen-Cilag, and GlaxoSmithKline. Alexander Ford - None. Dixuan Zhang - None. Racquel Simpson - None. Karen Dearness - None.