4. DISCUSSION
The main objective of this retrospective pharmacovigilance study is to investigate the hepatic dysfunction events of voriconazole, based on the real-world data records involved 73 quarters of the FAERS. Among our study we found that all the top ten frequency ADEs are strong correlation signal when compared with the full database.
In our study, 62.20% of patients developed hepatic dysfunction in 15 days of voriconazole use. The median time to event of the hepatic dysfunction events was 8 (IQR 2-28) days. We suggest health professionals be aware of the potentially risk of voriconazole and monitor hepatic function during the first 15 days of voriconazole use.
Voriconazole remains the drug of choice for primary therapy of IA, with liposomal amphotericin B and the newly licensed agent isavuconazole considered alternatives [1]. Combination therapy with mold-active triazoles and echinocandins has been used with the hope of improving outcomes over monotherapy, especially in the setting of refractory disease [11]. Therefore, our study also did a comparison of voriconazole with the other antifungal drugs. In analyses stratified on the triazole antifungal drugs, an increased hepatic dysfunction event reporting was found in voriconazole when compared with fluconazole and isavuconazole, with the RORs (the lower limit of ROR 95% CI > 1) 2.19 (95% CI 1.94-2.47), 2.31 (95% CI 1.66-3.33), respectively. A randomized, non-inferiority trial found that compared to voriconazole, isavuconazole was associated with a lower hepatotoxicity [12]. When comparing voriconazole with amphotericin B, there was a signal of high frequency of reporting hepatic dysfunction event with ROR (95% CI) 1.26 (1.08, 1.48). An observational study found that compared to amphotericin B, voriconazole was significantly associated with hepatotoxicity [p <0.0002; Odds Ratio=22.2 for voriconazole (95% CI, 4.3-116)] [13].
In analyses stratified on comparing to echinocandins, no signal of increased hepatic dysfunction event reporting was found (ROR<1). In previous studies, the echinocandins were considered to have a lower hepatotoxicity when compared with triazoles [14]. A network meta-analysis and systematic review on adverse reactions of antifungal drugs also showed that caspofungin was more significantly associated with a lower incidence of hepatobiliary disorders [15]. Whereas in the real word study, the results showed that echinocandins exhibited a higher hepatotoxicity which are consistent with our findings [16,17]. The risk of developing liver injury and possible hepatic dysfunction by an antifungal agent depends on several factors including underlying hepatic disease, echinocandins were considered as second-line or salvage therapy or antifungal combined therapy when monotherapy is not effective [18,19]. Since echinocandins can be used in patients with liver damage themselves, their related liver damage will be higher than the actual liver damage caused by voriconazole, which is also a shortcoming of this study.
In our hospital, the patient population is mainly liver transplantation and HIV infection, which are the main risk factors of invasive fungal infection. Therefore, precise treatment of invasive fungal infections is crucial in reducing mortality. Since voriconazole is the primary therapy for invasive aspergillosis, especially in immunocompromised patients, some clinical trials showed that supratherapeutic concentrations are correlated with an increased risk of hepatic toxicity [20,21]. Thus, voriconazole TDM is of paramount importance in patients with pre-existing liver disease, since the drug is extensively metabolized by the liver and this population is more difficult to tolerate a deterioration of hepatic function due to voriconazole-induced liver injury [22,23]. In our real word study based on the FAERS, we can’t get the concentration of voriconazole, which is another limitation of this study.
There are inherent limitations with disproportionality analysis methods based on FAERS data. Firstly, the inherent nature of the FAERS which does not allow causality assessment between drug and AE. Secondly, concomitant drugs have not been considered, we will perform relevant drug interaction research in the future. Lastly, due to lack of the patient’s underlying disease status, we did not consider the effect of disease factor which could be an important factor in hepatic abnormal events. Despite these drawbacks, the sample size of is enormous, the prevalence and risk of hepatic dysfunction and other AEs caused by voriconazole need to be further explored in prospective RCT studies.