Using disproportionality analysis to explore the association between periostitis and triazole antifungals in the FDA Adverse Event Reporting System Database

Hailong Li, Miao Zhang, Xue‐Feng Jiao, Yu Zhu, Yan Liu, Linan Zeng, Huiqing Wang, Lei Zhang, Wei Zhang, Lingli ZhangScientific 2023 · DOI

TL;DR

This study leverages the FDA Adverse Event Reporting System (FAERS) to investigate a potential link between periostitis and triazole antifungals. The findings strongly suggest a significant association between voriconazole and periostitis, while other triazole antifungals showed no such signal, highlighting the need for clinical attention and consideration of alternative treatments.

What Problem Does This Paper Address?

Invasive fungal infections are a serious concern, and triazole antifungals are a cornerstone of treatment and prophylaxis, particularly in immunocompromised patients like transplant recipients. While generally effective, these drugs can have side effects. Case reports have emerged describing periostitis (inflammation of the bone membrane) associated with voriconazole use. However, a comprehensive, real-world analysis of this association across a large patient population was lacking. This paper aims to fill that gap by systematically analyzing a large adverse event database to determine if a statistically significant association exists between periostitis and triazole antifungals, and to identify which specific drugs within this class might be implicated.

Methods

The researchers employed disproportionality analysis, a standard pharmacovigilance technique for signal detection in adverse event databases. They utilized the OpenVigil 2.1 software to extract and analyze data from the FDA Adverse Event Reporting System (FAERS) from Q1 2004 to Q2 2022. The core of their analysis involved constructing 2x2 contingency tables to compare the reporting frequency of periostitis with triazole antifungals against the reporting frequency of periostitis with all other drugs. Several statistical metrics were calculated to quantify this association: chi-squared (χ²), relative reporting ratio (RRR), reporting odds ratio (ROR), proportional reporting ratio (PRR), and Bayesian confidence propagation neural networks (BCPNN) of information components (IC). Positive signals were identified based on established criteria (e.g., χ² > 4, PRR > 2). Subgroup analyses were also conducted to explore potential associations based on patient sex and age.

Datasets

The primary dataset used in this study is the FDA Adverse Event Reporting System (FAERS). This is a publicly available database containing spontaneous reports of adverse events and medication errors submitted by healthcare providers, manufacturers, and patients. The data spans from the first quarter of 2004 to the second quarter of 2022. No new datasets were introduced.

Key Insights and Ideas

  • Voriconazole stands out: The most striking finding is the strong and consistent signal linking voriconazole to periostitis, while other triazole antifungals (fluconazole, itraconazole, posaconazole, ravuconazole) did not show similar associations in this dataset. This suggests a drug-specific rather than a class-wide effect.
  • Mechanism hints at fluoride: The paper references existing literature suggesting that voriconazole's trifluoride structure might contribute to periostitis through increased circulating fluoride levels or direct effects of fluoride. This provides a potential biological rationale for the observed association.
  • Clinical relevance is high: The strong statistical signals, even with the inherent limitations of FAERS data, warrant significant clinical attention. The authors emphasize that this association should be considered in clinical practice.
  • Subgroup analyses confirm signal: The association between voriconazole and periostitis was observed across different sexes and age groups, reinforcing the robustness of the signal.

Key Results

The study identified 143 patients who experienced periostitis while using voriconazole. Disproportionality analysis revealed a statistically significant association between voriconazole and periostitis, with metrics like RRR = 583.6, PRR = 1808.9, and ROR = 1831.7 all indicating a strong signal. In contrast, no safety signals were detected for periostitis with other triazole antifungals. Subgroup analyses by sex and age also showed positive signals for voriconazole.

Limitations and Future Work

The authors acknowledge several limitations inherent to using the FAERS database:

  • Spontaneous reporting: Data relies on voluntary reporting, which can lead to under-reporting, over-reporting, and misreporting.
  • Data quality: FAERS data can suffer from non-standard reporting and missing information (e.g., age, sex).
  • Geographic bias: A significant portion of reports originate from the United States, limiting the generalizability of findings to other populations.
  • Causality: FAERS data can identify associations but cannot definitively prove causality.

Future work should focus on verifying causality through prospective studies. The authors also suggest that clinicians should consider alternative triazole antifungals (like posaconazole or itraconazole) if clinically indicated, given the lack of association with periostitis observed for these drugs.

Summary updated

Abstract

Though triazole antifungals are the first choice for preventing and treating invasive fungal infections, periostitis caused by voriconazole has been described in emerging case reports; however, no studies exist on this association in real-world clinical settings. Our study aimed to identify the association between periostitis and triazole antifungals by analyzing data from the FDA Adverse Event Reporting System (FAERS). We extracted and analyzed reports on the association between periostitis and triazole antifungals in FAERS from the first quarter of 2004 to the second quarter of 2022 using OpenVigil 2.1. Disproportionality analysis was performed to evaluate the association between periostitis and triazole antifungals, and chi-squared (χ<sup>2</sup>), relative reporting ratio (RRR), reporting odds ratio (ROR), proportional reporting ratio (PRR), and Bayesian confidence propagation neural networks (BCPNN) of information components (IC) were reported. In total, 143 patients experienced periostitis while using voriconazole. Disproportionality analysis identified an association between periostitis and voriconazole (χ<sup>2</sup> = 82,689.0, RRR = 583.6, 95%CI [472.4, 721.1], PRR = 1808.9, 95%CI [1356.0, 2412.9], ROR = 1831.7, 95%CI [1371.6, 2446.3], IC = 9.2, 95%CI [8.6, 9.8]). However, no safety signals were observed between periostitis and other triazole antifungals. When stratified by sex and age, disproportionality analysis identified positive signals between periostitis and voriconazole. The possible association between periostitis and voriconazole should attract sufficient attention in clinical practice. Alternative treatment with other triazole antifungals can be considered, and causality needs to be verified in further prospective studies.
BibTeX
@inproceedings{li2023using,
  author = {Hailong Li and Miao Zhang and Xue‐Feng Jiao and Yu Zhu and Yan Liu and Linan Zeng and Huiqing Wang and Lei Zhang and Wei Zhang and Lingli Zhang},
  title = {{Using disproportionality analysis to explore the association between periostitis and triazole antifungals in the FDA Adverse Event Reporting System Database}},
  booktitle = {Scientific},
  year = {2023},
  doi = {10.1038/s41598-023-27687-0},
}
Figure