Analysis of drug-induced interstitial lung disease caused by herbal medicine using the Japanese Adverse Drug Event Report database

Abstract Background Drug-induced interstitial lung disease (DIILD) is a severe adverse event leading to morbidity and mortality. This study evaluated the adverse event indicators of DIILD and time-to-onset profiles following the daily intake of herbal drugs ( Scutellariae radix “ogon” in Japanese, Bupleuri radix “saiko” in Japanese, and Pinelliae tuber “hange” in Japanese) using the Japanese Adverse Drug Event Report database. DIILD was defined in accordance with the Medical Dictionary for Regulatory Activities. Methods The Japanese Adverse Drug Event Report database contained 830,079 reports published between April 2004 and April 2023. The association between herbal medicines and DILLD was evaluated using the pharmacovigilance index as the reporting odds ratio (ROR), logistic regression models, propensity score-matching techniques, and Weibull shape parameters. Results The adjusted RORs using multivariate logistic regression models for Scutellariae radix (daily intake), Pinelliae tuber (daily intake), sex (male), age (≥ 60 years), Scutellariae radix (daily intake)*age (≥ 60 years), and Scutellariae radix (daily intake)* Pinelliae tuber (daily intake) were 1.47 (1.36 − 1.59), 1.05 (1.01 − 1.10), 1.45 (1.34 − 1.57), 1.92 (1.74 − 2.11), 3.35 (3.12 − 3.60), and 1.49 (1.46 − 1.53), respectively. DIILD onset profiles were evaluated using the Weibull shape parameter. A logistic plot of daily intake and onset of DIILD was drawn. ROR signals were detected in 32 of 54 herbal medicines, including Scutellariae radix , Bupleuri radix , and Pinelliae tuber . The median duration (days) (interquartile range) to DIILD onset was 36.0 (27.0–63.0) for Saikokaryukotsuboreito, 35.0 (21.0–55.0) for Saireito, and 31.0 (13.5–67.5) for Shosaikoto. The Weibull shape parameter beta (95% confidence interval) values for Saikokaryukotsuboreito, Saireito, and Shosaikoto were 1.36 (1.08–1.67), 1.36 (1.20–1.52), and 1.31 (0.98–1.68), respectively. Conclusions DIILD demonstrated a dose-dependent to crude drugs. Clinicians should strive for the early detection of DIILD and avoid the inadvertent administration of herbal medicines.