We have concerns regarding the conclusions of Wielders and associates in their manuscript “Prevention of cystoid macular edema after cataract surgery in non-diabetic and diabetic patients: A systematic review and meta-analysis.” The authors include studies of all cystoid macular edema (CME), regardless of whether or not it was associated with decreased vision. Because CME’s clinical importance relates to its effect on vision, visual acuity is a more appropriate primary outcome than any manifestation of CME. Furthermore, there is no standard means to define CME, which confounds systematic review because of inconsistent definitions.

The authors suggest that topical nonsteroidal anti-inflammatory drugs (NSAID) “should always be part of the preventative treatment (of CME) after cataract surgery…” and base this assertion on the results of 3 trials that compared monotherapy with an NSAID vs monotherapy with either fluorometholone 0.1% or betamethasone 0.1%. After a single application, fluorometholone alcohol 0.1% and betamethasone sodium phosphate 0.1% result in peak aqueous concentration of 5.1 ng/mL and 7.7 ng/mL, respectively, vs 669.9 ng/mL after a single application of prednisolone acetate 1%. Both corticosteroids therefore may more closely approximate the effect of placebo. In contrast, a recent prospective randomized trial by Tzelikis and associates using prednisolone acetate 1% reported no significant differences in CME with NSAID treatment.

The authors furthermore provide no mechanism to support their conclusion that NSAIDs are more effective than corticosteroids at reducing CME. The anti-inflammatory properties of NSAIDs are due to their inhibition of prostaglandins, but corticosteroids also inhibit prostaglandins, in addition to downregulating several other inflammatory-mediated events. Consequently, corticosteroids encompass and possess far broader anti-inflammatory properties than NSAIDs, which is evident by their greater clinical effectiveness in treating inflammatory disorders. A greater therapeutic effect of NSAIDs plus corticosteroids can be readily explained by an additive rather than synergistic effect of 2 anti-inflammatory drugs, which may be replicated therefore by increased dosing of a single agent.

Systematic reviews and meta-analyses do not fully account for the variable definitions of CME, unbalanced treatment regimens, conflicts of interest, and differences in corticosteroid formulations that are prevalent among published studies and confound analysis. Furthermore, meta-analyses have a strong tendency to amplify publication bias (tendency to publish positive instead of negative findings). For example, the authors would have excluded the study by Tzelikis and associates because “no patient developed CME in either treatment group.”

The AAO Ophthalmology Technology Assessment Panel recently performed an extensive review on the role of NSAIDs in cataract surgery and concluded that (1) there was no level 1 evidence that supports the long-term visual benefit of NSAID therapy to prevent vision loss from CME after cataract surgery and (2) there was no evidence of a synergistic effect of NSAIDs with corticosteroids. The panel recommended that future trials primarily focus on long-term visual acuity outcomes, fully account for dosing disparities, and avoid conflicts of interest to improve the quality of reporting. Given the considerable limitations and biases in published studies, systematic reviews and meta-analyses will overstate the clinical effectiveness of NSAIDs.