Publication bias in studies on biologic therapy for children with inflammatory bowel disease

Objective: The utilization of biologic drugs in children with inflammatory bowel disease (IBD) surged following the publications of positive results in randomized controlled trials and real-world studies. We aimed to explore the extent of publication bias associated with these findings.

Methods: Two reviewers assessed all abstracts evaluating the efficacy or safety of biologics presented at the annual European Society for Pediatric Gastroenterology Hepatology and Nutrition and North American Society for Pediatric Gastroenterology conferences from 2015 to 2019. Abstracts were classified as "positive" or "negative." Time to publication was analyzed using Kaplan-Meier curve and groups were compared using the log-rank test. A Cox proportional model was utilized to determine the likelihood of publication.

Results: Out of 209 included abstracts, only 130 (62%) were published as full manuscripts. The median time to publication was 2.8 years (interquartile range = 0-8.2). In the univariate Cox model, the likelihood of publication was four times higher for abstracts reporting positive results (hazard ratio = 4.4 [95% confidence interval, CI = 2.3-8.5]). The probabilities for publication at 1, 3, and 5 years after the conference were 32%, 59%, and 66% for abstracts with significantly positive results in favor of biologic treatment compared to 10%, 22%, and 25% for those with negative results (p < 0.001). In multivariable model, positive results (odds ratio = 6.4 [95% CI = 2.5-16.4]) were significant associated with publication rate.

Conclusion: Only 62% of abstracts presented in medical conferences regarding biologics in pediatric IBD are eventually published as full manuscripts, and those reporting positive results were more likely to be published and at an earlier time. Clinicians, guideline groups, and medical authorities dealing with drug approval, need to be aware of potential publication bias of published studies when employing evidence-based management strategies.