Ying Li, Han Xie, Linbo Tan
Voriconazole initiation was temporally associated with subsequent hypomagnesemia in pediatric patients with hematologic diseases. PSSA may complement spontaneous reporting systems for detecting laboratory-defined safety signals. Weekly serum magnesium monitoring should be considered during the first 30 days of voriconazole therapy. These findings support a temporal association but do not establish causality.
OBJECTIVE: To evaluate the temporal association between voriconazole and hypomagnesemia in pediatric patients with hematologic diseases using prescription sequence symmetry analysis (PSSA) and to assess the utility of hospital information system (HIS)-based active drug safety surveillance.
METHODS: Inpatient data were extracted from the Children's Hospital of Chongqing Medical University from November 2013 to December 2024. The first year served as a data integrity verification window and the subsequent approximately 10-year period as the formal analysis phase. Patients with hematologic diseases were identified using ICD-10 codes C81-C96, D45-D47, and D60-D61. Each patient contributed the first voriconazole prescription and first hypomagnesemia event (Mg2+ < 0.6 mmol/L). The primary analysis used a 365-day washout period, 90-day exposure window, and 7-day blackout period. Adjusted sequence ratios (aSRs) with 95% confidence intervals (CIs) were estimated, with 10 parameter sensitivity analyses, monitoring-intensity correction, four control analyses, and stratified analyses by concomitant medications.
RESULTS: Among 536 eligible patients, 364 (67.9%) had serum magnesium measurements. The post-exposure sequence comprised 58 patients and the reverse sequence eight patients, yielding an aSR of 5.76 (95% CI: 2.75-12.06). Median time to hypomagnesemia after voriconazole initiation was 28 days (IQR: 20-43), and 32/58 events (55.2%) occurred within 30 days. Exploratory dose-response analysis showed no association between weight-based voriconazole dose and hypomagnesemia occurrence (P = 0.437). All parameter sensitivity analyses produced aSRs >3.9; after correction for monitoring intensity, the aSR was 3.33 (95% CI: 1.59-6.98); in a sensitivity grid, the association remained statistically significant for monitoring-intensity ratios up to 2.75. All four control analyses agreed with their prespecified expectations. Associations remained significant in the PPI and aminoglycoside strata, whereas the loop-diuretic non-user stratum showed the same direction but was not statistically significant.
CONCLUSION: Voriconazole initiation was temporally associated with subsequent hypomagnesemia in pediatric patients with hematologic diseases. PSSA may complement spontaneous reporting systems for detecting laboratory-defined safety signals. Weekly serum magnesium monitoring should be considered during the first 30 days of voriconazole therapy. These findings support a temporal association but do not establish causality.