Association Between Systemic Fluoroquinolones and Risk of Peripheral Neuropathy

Fluoroquinolone use has been associated with peripheral neuropathy in case reports, but there have been few noninterventional studies validating this association, and causality is unclear. The aim of this study was to estimate the association between systemic fluoroquinolone prescription and incident peripheral neuropathy relative to a comparator class of antibiotics, cephalosporins, using multiple databases to increase precision and assess the robustness of the results.

Cohort studies were conducted using data from 2 databases of UK electronic primary care records, Clinical Practice Research Datalink Aurum and GOLD, from April 1, 1997, to December 31, 2019. Among adults (age ≥18 years) without diabetes mellitus or previously recorded peripheral neuropathy prescribed systemic fluoroquinolones or cephalosporins, hazard ratios were estimated between fluoroquinolone prescription and incident peripheral neuropathy within 60 days of prescription using stabilized inverse probability of treatment weighted Cox regression. Hazard ratios were similarly estimated for a positive control outcome of incident tendon rupture within 60 days of prescription. Risk differences for peripheral neuropathy were estimated using a weighted Aalen-Johansen estimator. Estimates were pooled across databases using fixed-effects meta-analysis.

In total, 2,933,982 individuals prescribed fluoroquinolones or cephalosporins in Aurum (63.7% [n = 1,868,793] female; median age 49.5 years [interquartile range (IQR) 34.1-67.5]) and 995,457 in GOLD (63.5% [n = 632,033] female, median age 51.5 years [IQR 35.6-68.0]) were included. Of these individuals, at first treatment episode, 1,002,004 (34.2%) in Aurum and 381,734 (38.3%) in GOLD were prescribed fluoroquinolones. The estimated association between fluoroquinolones and peripheral neuropathy was close to the null at first treatment episode (pooled adjusted hazard ratio [aHR] 0.96, 95% CI 0.77-1.20) and over all treatment episodes (pooled aHR 1.05, 95% CI 0.92-1.20). Hazard ratios were elevated for a positive control outcome of tendon rupture at first treatment episode (pooled aHR 1.98, 95% CI 1.55-2.53) and over all treatment episodes (1.76, 95% CI 1.52-2.04).

Findings were not indicative of a substantive difference in risk of peripheral neuropathy between those prescribed fluoroquinolones and cephalosporins. Potential residual confounding is a study limitation. Study results provide some reassurance to clinicians regarding risk of peripheral neuropathy with fluoroquinolones relative to a comparator antibiotic.