Background: The interpretation of reports on adverse events of drug compared with non-drug intervention in meta-analyses remains challenging. The grading of adverse events is stratified as mild, severe and fatal events. The reporting bias of the adverse events in meta-analyses can lead to an unbalanced judgment, and the likelihood of an misleading overall effect.

Objectives: i. To illustrate the reporting on adverse events with stratification or non-stratification in the selected published meta-analyses. ii. To explore the impact of interpretation of the adverse events of such interventions in meta-analysis.

Methods: An illustration was carried out using the three meta-analyses of beta-blockers (i.e. drug intervention) versus esophageal banding ligations (i.e. non-drug intervention) for primary prophylaxis of esophageal variceal bleeding. We compared the reporting of adverse events of drugs with or without stratification. We identify the difference in interpretation of the effect estimations of drugs versus non-drug intervention in the overall context of evidence.

Results: Of the three meta-analyses selected, only one has reported the adverse events of drugs according to the stratified grading of mild, severe and fatal events. When the adverse event of non-drug intervention was compared with those of drug without stratification, the overall context of interpretation varies from those of drug with stratification.

Conclusions: The findings highlight the justification of the requirement for stratification of the adverse events of drugs when it was compared with non-drug intervention for balancing the outcomes of meta-analyses. The stratification of adverse events as well as the different underlying mechanisms attributed to the cause of adverse events play a role in providing the evidence of patient safety.