Signal Detection Between Fluoroquinolone Use and the Risk of Rhegmatogenous Retinal Detachment: Sequence Symmetry Analysis Using Nationwide South Korean Healthcare Database Between 2004 and 2015

Background and Objectives The association between fluoroquinolone and rhegmatogenous retinal detachment (RRD) has been controversial as a result of inconsistent findings. We set out to detect a possible association of fluoroquinolone use and risk of RRD, using sequence symmetry analysis (SSA). Methods We conducted an SSA, case-only design, using a Korean nationwide healthcare database between 2004 and 2015. Exposure was defined as new fluoroquinolone use and outcome as an incident RRD, defined by a diagnosis of RRD (ICD-10: "H33.0") or surgery for RRD. Pairs of exposure and RRD within a 1-year time-window were included. The sequence ratio (SR) was calculated by the ratio of the number of patients prescribed with exposure first and diagnosed with RRD second divided by the number of patients diagnosed with RRD first and prescribed with exposure second. SR was adjusted (aSR) for underlying trends and 95% confidence intervals (CIs) were calculated. In order to observe whether the estimated ratio stabilized over time, we did repeated time-sequential analyses with the cumulative data starting from the 3-year period 2004-2006 to 2015. Results Fluoroquinolone use had a greater association with RRD as compared with other antibiotics [fluoroquinolone: 5234 pairs; aSR = 1.70 (95% CI 1.61-1.80), first-generation cephalosporin: 4139 pairs; aSR = 1.39 (95% CI 1.31-1.80), second-generation cephalosporin: 5914 pairs; aSR = 1.31 (95% CI 1.24-1.38), third-generation cephalosporin: 3650 pairs; aSR = 0.88 (95% CI 0.83-0.95), extended-spectrum penicillin: 4823 pairs; aSR = 1.29 (95% CI 1.31-1.47), macrolides: 4115 pairs; aSR = 1.31 (95% CI 1.24-1.39)]. Time-sequential analyses supported the association between fluoroquinolone and RRD. Conclusions Our detection suggests a possible association between fluoroquinolone use and RRD. However, possible overestimation and reverse causality bias may have influenced our findings due to the limitation of an SSA design.