2.3 Analysis
We performed disproportionality analysis using the reporting odds ratio (ROR) with relevant Confidential Interval (CI) to indicate the presence of signals of potential increased risk of drug-related AE [9]. The ROR was calculated by dividing the odds of hepatic dysfunction events reporting for voriconazole by the odds of hepatic dysfunction events reporting for the comparison drug. A signal of increased hepatic dysfunction risk was defined as the lower limit of 95% CI exceed one with the number of cases ≥3 [10].
To validate the robustness of the findings, we conducted five specific comparisons: (1) compared voriconazole with the full database; (2) compared voriconazole with the other triazole antifungal drugs: fluconazole, itraconazole, posaconazole and isavuconazole; (3) compared voriconazole with the echinocandins antifungal drugs: caspofungin, micafungin and anidulafungin; (4) compared voriconazole with the polyene antifungal drug amphotericin B; (5) compared voriconazole with the 5-fluorocytosine.
The primary outcomes were overall hepatic-related adverse events compared with the full database and the other antifungal drugs (see details in Supplementary Materials Table S3 ). The secondary outcomes were the high frequency reporting of hepatic dysfunction events (see details in Supplementary Materials Table S4/S5 ): hepatic function abnormal, cholestasis, alanine aminotransferase increased, blood alkaline phosphatase increased, aspartate aminotransferase increased, liver function test abnormal, gamma-glutamyl transferase increased, liver disorder, hepatocellular injury, blood bilirubin increased.
All the data extraction were performed by Microsoft Access 2010, and the statistical analyses were performed using SPSS 26.0.