Standardized retrieval of side effects data for meta-analysis of safety outcomes. A feasibility study in acute...

Accurate and complete safety data are indispensable for the proper evaluation of the benefit-to-harm ratio of medical interventions. We evaluated whether a systematic review and meta-analysis of standardized safety data is feasible by requesting information on side effects directly from the investigators of all 38 antibiotic trials on acute sinusitis published in the last decade. We requested standardized information on gastrointestinal toxicity outcomes, including hospitalizations, discontinuations, and days with nausea/vomiting, diarrhea, or both. Responses were received only for 16 trials (42%), and safety data were contributed only for 9 trials (24%). In some trials, safety data had not been collected, had been lost, or had been transferred to other companies. The odds of data retrieval was higher in general medical journals (P =.024) and independently improved with an increase in sample size (P =.064). The available information suggested side effects may equal or exceed in severity the marginal treatment benefits. Interpretation of safety data was further complicated by heterogeneity or lack of information of use of concomitant drugs, mode of collection of safety information, use of blinding, and other study design parameters. Availability of standardized information for performing meta-analysis of safety data may be limited. Standardized reporting, prospective collection, and long-term availability of safety information should be improved.