Topical nonsteroidal anti-inflammatory drugs (NSAIDs) for management of pain in patients after photorefractive keratectomy (PRK). Pain after PRK is a major concern for both patients and surgeons. Although evidence supports the use of NSAIDs postoperatively, no consensus exists regarding the preferred regimen. The study aimed to compare the efficacy and safety of different topical NSAIDs. This study was prospectively registered with PROSPERO (ID: CRD42023417651). A systematic search of electronic databases was performed, for randomized controlled trials reporting topical NSAIDs' outcomes of corneal re-epithelization, rescue analgesics intake, and pain in days 0 to 3 after PRK (postoperative days [PODs] 0 to 3). Studies were graded for risk of bias. Data were extracted, and standardized mean differences (SMDs) were evaluated in a network meta-analysis in accordance with the Cochrane's guidelines, to which a frequentist approach model was fitted. Transitivity was assessed using the net split method. Treatment effectiveness was ranked using forest plots based on comparison with placebo. P-scores (P) and league tables were used to examine combined direct and indirect comparisons. Of 1540 studies identified, 27 were included. These encompassed 2286 patients across 11 countries, evaluating 7 distinct topical NSAIDs. At POD0, ketorolac (P 0.764), flurbiprofen (P 0.763), and bromfenac (P 0.717) were the most efficient drugs overall and displayed significantly lower pain scores than placebo. Other than that, flurbiprofen held the highest rank for reported pain throughout, significantly outperforming placebo on POD1 (P 0.874, SMD -1.19, 95% CI -1.86 to -0.52), POD2 (P 0.882, SMD -1.05, 95% CI -1.82 to -0.27), and POD3 (P 0.939, SMD -1.14, 95% CI -2.1 to -0.18). Other NSAIDs were significantly better than placebo only on POD1 and POD0. Rescue analgesic intake analysis favored indomethacin (P 0.834, SMD -0.8, 95% CI -1.33 to -0.27), ketorolac, and diclofenac. Compared with placebo, re-epithelization was slowed to different significances with all NSAIDs but flurbiprofen (P 0.991, SMD -0.7, 95% CI -1.38 to -0.03). Flurbiprofen was favorable in pain scores on typically painful postoperative days and re-epithelization times. However, analgesics intake, a more objective outcome, suggested superiority of other NSAIDs. Inconsistencies may be explained by the small sample size. For clinical interpretation, NSAID effect sizes should be taken into consideration.
Read full abstract