Skip to main content


Publication bias in Clinical Trials on monoclonal antibodies: a cross-sectional study

Date and Location




Sunday 22 September 2013 - 10:30 - 12:00


Presenting author and contact person

Presenting author

Douglas Henrique ...

Contact person

Maria Eduarda S Puga
Abstract text
Background: Publication bias can affect the reliability of systematic reviews. This study investigated the problem from the perspective of biologicals, with high aggregate economic value, in order to assess whether the nature of the medical supply may change the patterns of publication and dissemination of results. Objectives: To evaluate the characteristics (indices and ratios) of publication and dissemination of results of clinical trials on monoclonal antibodies adalimumab, bevacizumab, rituximab, trastuzumab and infliximab registered at Methods: Cross-sectional study. In the first stage of the research, the sample included every protocols of interventional clinical trials, phases III-IV, completed and registered at on the above mentioned monoclonal antibodies (n = 243). Set the initial sample, we evaluated the patterns of publication and result’s disclosure at, through a search strategy that considered Pubmed, Embase, Lilacs, Cochrane Central and Google Scholar. Results: Among the 243 trials that comprised the initial sample, 169 (≈ 69.5%) were published and 61 (≈ 25.1%) had their results disclosed at The industry sponsored, wholly or partially, 169 trials (≈ 69.5%). Considering the subsample of unpublished studies (n = 74), 51 (≈ 69%) were fully or partially funded by industry. The prevalence of placebo controlled or single arm studies is greater in trials funded by industry (≈59.41%) rather than in studies not supported by industry (≈ 40.64%). Conclusions: Publication bias in clinical trials is intense, despite the nature of intervention (in this study, monoclonal antibodies). The source of funding (involving or not the industry) did not change the patterns of publication, suggesting that publication bias occurs similarly in all cases. However, studies involving placebo or single arm studies (therefore, with a poor design) were more common in the sample financed by industry.