Signal detection of adverse drug reactions of cephalosporins using data from a national pharmacovigilance database

Jung Yoon Choi, Jae Hee Choi, Myeong Gyu Kim, Sandy Jeong Rhie

Research output: Contribution to journalArticlepeer-review

2 Scopus citations


This case-non-case study aims to detect signals not currently listed on cephalosporin drug labels. From 2009 to 2018, adverse event (AE) reports concerning antibacterial drugs (anatomical therapeutic chemical (ATC) code J01) in the Korea Adverse Events Reporting System (KAERS) database were examined. For signal detection, three indices of disproportionality, proportional reporting ratio (PRR), reporting odds ratio (ROR), and information component (IC), were calculated. The list of signals was compared with ADRs on the drug labels from the United States, United Kingdom, Japan, and South Korea. A total of 163,800 cephalosporin–AE combinations and 72,265 all other J01–AE combinations were analyzed. This study detected 472 signals and 114 new signals that are not included on the drug labels. Cefatrizine–corneal edema (PRR, 440.64; ROR, 481.67; IC, 3.84) and cefatrizine–corneal ulceration (PRR, 346.22; ROR, 399.70; IC, 4.40) had the highest PRR, ROR, and IC among all signals. Additionally, six serious AEs that were not listed on drug labels such as cefaclor-induced stupor (ten cases) and cefaclor-induced respiratory depression (four cases) were found. Detecting signals using a national pharmacovigilance database is useful for identifying unknown ADRs. This study identified signals of cephalosporins that warrant further investigation.

Original languageEnglish
Article number425
Issue number5
StatePublished - May 2021

Bibliographical note

Publisher Copyright:
© 2021 by the authors. Licensee MDPI, Basel, Switzerland.


  • Adverse drug reaction
  • Cephalosporin
  • KIDS KAERS database (KIDS-KD)
  • Pharmacovigi-lance
  • Signal


Dive into the research topics of 'Signal detection of adverse drug reactions of cephalosporins using data from a national pharmacovigilance database'. Together they form a unique fingerprint.

Cite this