Ketorolac versus nepafenac in cataract surgery

The prospective trial conducted by Duong et al.1 examining the use of ketorolac 0.4% and nepafenac 0.1% following cataract surgery raises some serious concerns. Several design flaws compromise the integrity of the study and therefore call into question the results. Perhaps most important is that masking was virtually impossible because the 2 test agents had different dosing frequencies (4 times a day for ketorolac and 3 times a day for nepafenac), inviting investigator bias. Moreover, although the authors describe the study as randomized, it appears that the surgical coordinator decided the treatment group of each patient, further biasing the results. In addition, the unconventional practice of switching the colored labeling of the masked groups each month increased the possibility of unintentionally mixing data from the 2 groups. The planning and justification of the trial were not explained in the text, raising questions about the primary outcome variable, study size and power, and outcome analysis; thus, it is not possible to discern whether this study was adequately designed and powered to support its conclusions. If not, the study must be considered exploratory and the conclusions can be used only for hypothesis generation. While the reasons for using distinct surgical kits (in which each nonsteroidal antiinflammatory drug [NSAID] is paired with a different antibiotic and corticosteroid) and different NSAID dosing frequencies are understandable, these decisions introduce undue variations into the trial design, making it difficult to assign outcome differences solely to the type of NSAID administered. Finally, bias may have been introduced through analgesic use; although the authors do not comment on patient use of analgesics, it is reasonable to assume that some patients would have taken them, thereby confounding the analgesic-related outcomes of the study. With respect to the outcomes reported, several issues should be raised. First, the authors report a significantly higher posterior capsule opacification (PCO) rate in the nepafenac group, but neglect to describe how they classified PCO with respect to grade and coverage of the posterior capsule. Because PCO is not simply present or absent, but rather is graded on a continuum of severity and extent, the reported percentages provide little information. In addition, the authors fail to address numerous factors that affect PCO rate, including surgical technique (eg, size of capsulorhexis) and relevant patient characteristics such as diabetes mellitus (without diabetic retinopathy) and refractive error.2 Furthermore, the increased rate of PCO development with nepafenac after only 1 month is not consistent with published literature citing a neutral or a protective effect by NSAIDs on PCO development.3,4 Thus, factors independent of NSAID treatment were likely responsible for the between-group difference in PCO rate. The reporting of visual acuity results at day 1 also poses a concern. The numerical advantage of ketorolac 0.4% (0.54 logMAR versus 0.63 logMAR) was described in the text as being “slightly better” than nepafenac. Given the lack of statistical significance, this difference is likely spurious. Finally, of major concern is the interpretation of the patient-reported outcomes. Aside from the patient questionnaire being administered only on day 1, when responses about compliance cannot be expected to be reliable, the authors overstate the importance of the between-group differences in patient satisfaction, compliance, and pain control. In addition to the already stated concerns regarding compliance and analgesic-related outcomes, further reservations about the results should be noted: The small between-group differences in all 3 of these categories (0.25- to 0.40-point differences on a 5-point scale) with such large standard deviations, while questionably statistically significant, can hardly be considered clinically relevant. Thus, the study conclusion discussing the statistical advantage of ketorolac with respect to patient satisfaction, compliance, and pain control is misleading. A more appropriate conclusion would be “nepafenac 3 times a day and ketorolac 4 times a day were both effective and clinically comparable in anterior segment ophthalmic surgery and patient perception.” In conclusion, when one considers the lack of true masking and randomization, the confounding use of surgical kits, as well as the questionable interpretation of much of the results, the validity of the entire study is called into question.