Drug-induced liver injury associated with atypical generation antipsychotics from the FDA Adverse Event Reporting System (FAERS) | BMC Pharmacology and Toxicology

Data acquisition

The data of this study was obtained from the FAERS, one of the most comprehensive spontaneous reporting system databases. The FDA publishes FAERS files every quarter. In this study we used FAERS quarterly data files, reports submitted between the first quarter of 2017 and the first quarter of 2022 were extracted. The AAPs we analyzed included: olanzapine, paliperidone, risperidone, amisulpride, clozapine, aripiprazole, quetiapine and ziprasidone.

This study used the standardized MedDRA queries (SMQ) to retrieve all cases of drug-related liver disease from the FAERS database (SMQ code 20000006), including: liver injury, hepatitis, transaminases increased, hepatic failure, hepatic steatosis, hepatitic cancer, hepatic congestion, hepatic pain, hepatocellular injury, hepatic cirrhosis, hepatomegaly, liver sarcoidosis, hepatic cytolysis, hepatic encephalopathy, hepatic fibrosis, hepatotoxicity, hepatic necrosis. We used the Medical Dictionary for Regulatory Activities (MedDRA version 25.0). Serious adverse events (SAE) in this study were classified as patients whose treatment resulted in hospitalization, death, disability, or other life-threatening consequences.

Inclusion and exclusion criteria

Inclusion criteria for this study: (1). Data from the first quarter of 2017–2022 in the FAERS database; (2). The drug names were the eight AAPs specified in this study; (3). The “Role cod” field was recorded as data for Primary suspect (PS) and Secondary suspect (SS).

Exclusion criteria for this study: (1). Gender, age and other personal information had significant missing data.

Processing

Each quarterly FAERS data file contains seven data tables: demographic characteristics (DEMO), details of medication (DRUG), adverse event (REAC), outcomes of patients treatment (OUTC), source of the report (RPSR), medication start and end date (THER), indications of drugs (INDI). In this study, three tables (“DEMO”, “DRUG”, “REAC”, “OUTC”) were used for analysis. Adverse events were identified by preferred terms (PTs), as coded by the Medical Dictionary for Regulatory Activities (MedDRA). Each table used “$” as the separator to divide the file into several fields. To prevent drug name irregularities, we used The Drugbank database (https://go.drugbank.com/drugs), which contains comprehensive drug names that can be used as a reference for pharmacovigilance analysis.

Since the FAERS database contains reports from numerous sources, there may be multiple reports for the same adverse event, so the data in this study was cleaned and only one report was retained for the same adverse event. According to the method of eliminating duplicate reports recommended by FDA, the “Primaryid”, “Caseid”, and “FDA_DT” fields of DEMO table were selected and sorted in the order of “Caseid”, “FDA_DT”, and “primaryid”. For reports with the same caseid, the one with the largest “FDA_DT” value was reserved in this study. Secondly, for reports with the same “Caseid” and “FDA_DT”, the one with the largest “Primaryid” value was retained. As of 2019 Quarter one there is a new text file that lists deleted files. FDA or Manufacturers may delete cases for various reasons including combiningcases. According to this file, we deleted the reports according to Caseid in the deleted reports list to ensure the accuracy of the data.

The field “Role cod” represents the degree of relationship between adverse events and the drug, including Primary suspect (PS), Secondary suspect (SS), Concomitant (C), Interaction (I). To ensure the accuracy of the study, reports of AAPs recorded as “PS” and “SS” were included in the analysis. The data included in this study were age, gender, adverse reactions, weight, event date, initial FDA receipt date, latest FDA receipt date, reported country, reporter, drug name, outcomes.

Statistical analysis

IBM SPSS V.26.0 and StataCorp Stata 12.0 software were used for analysis. Normally distributed data were presented in the form of mean (SD), and non- normally distributed data were represented by median (Q1, Q3). In this study, a case/non-case approach was used to calculate reporting odds ratio (RORs) and 95% confidence intervals (CIs) [18, 19]. The ROR represents the ratio of the odds of an adverse event for a specific drug against the odds of the same adverse event reported for all other drugs, p-value < 0.05 was considered significant.