Controversy Surrounding the Proposed Ocular Adverse Events of Fluoroquinolones.

Abstract

In the current issue of JAMA Ophthalmology, Sandhu et al1 report a retrospectivecohort studyofmedical claimsdata inorder todeterminetheassociationbetweenfluoroquinolonesand uveitis.Theauthorsshouldbe commended on their excellent analyses and interpretationof the results.Their study didnot findanassociationbetween fluoroquinolonesanduveitis. Their conclusions are both in keeping and in contrastwith the findings of recent similar studies.2,3 The differing conclusionsderivedfromthesestudieshighlightsometopics thatneed tobeconsideredwheninterpretingtheresultsofmedicalclaims database research: the importanceofusingproper controls, accounting for confoundingbias, replicationandvalidationof results, and the inherent limitations of this type of research. Uveitis is a heterogeneousdisease, and canbe causedby a variety of infectious andnoninfectious (inflammatory) etiologies. Patients canalsodevelopnoninfectiousuveitis following anunrelatedsystemic infection (eg,poststreptococcaluveitis). Furthermore,patientswithearly signsof a systemic inflammatory disease that is associated with uveitis may present with nonspecific systemic symptoms that canbemisdiagnosedas a systemic infectious process. Thus, in any study that attempts to examine the association between fluoroquinolones and uveitis, thepotential confoundingbiasof this systemicprocess needs tobeproperlyaccountedfor.Forooghianetal3previously published a large, retrospective case-control study examining theassociationbetweenfluoroquinolonesanduveitis.Thestudy usedβ-lactamandmacrolideantibioticsascontrolsandexcluded patients with 9 different systemic diseases known to be associatedwithuveitis. The investigators found that all 3 classes of antibiotics (fluoroquinolones,β-lactams,andmacrolides)were associatedwithuveitis. Theyconcluded that the systemicprocesses forwhich fluoroquinolones are prescribed, and not the fluoroquinolones themselves, are the likely inciting factors for uveitis.Followingthepublicationoftheseresults,anothergroup2 publisheda report in JAMAOphthalmology (usingadatabaseof older men) claiming an association between the fluoroquinolonesmoxifloxacin and ciprofloxacin and uveitis. The control used by Eadie et al in this report was finasteride, and they adjusted for only 3different systemicdiseases known tobe associated with uveitis. It is unclear why they did not use other antibiotics as controls or adjust for amore comprehensive list of systemic inflammatory diseases. The current report by Sandhu et al1 appropriately addresses the issue of potential confounding bias in their analyses. They used β-lactams as a control group and excluded patients with 18 different systemic diseases that are known to beassociatedwithuveitis. Furthermore, theyperformedasecondary analysis using incident diagnosis of a systemic disease known to be associated with uveitis as an outcome measure. When they examined fluoroquinolones as a group, Sandhu and colleagues1 found no increased hazard of uveitis within 30-, 60-, or 90-day treatment periods. A specific exRelated article page 38 Oral Fluoroquinolones and the Risk of Uveitis Original Investigation Research