Network meta-analysis comparing efficacy and safety of different protocols of corneal cross-linking for the treatment of progressive keratoconus

PurposeThis study aimed to determine the preferred protocol of corneal collagen cross-linking (CXL) in the treatment of progressive keratoconus.MethodsRelevant studies were retrieved in PubMed, EMBASE and Cochrane Central Register of Controlled Trials (CENTRAL). Maximum keratometry value (K-max), best spectacle-corrected visual acuity (BSCVA), manifest refraction spherical equivalent (MRSE), and endothelial cell density (ECD) were evaluated in network meta-analysis.ResultsEight randomized controlled trials (RCTs) were included. Low-level evidence suggested that aCXL with 30mW/cm(2) for 3 min (aCXL-3) might be the best protocol for reducing BSCVA (65.22%) but worst protocol for reducing MRSE (51.53%). aCXL with 18mW/cm(2) for 5 min (aCXL-5) might be the best protocol for reducing K-max (39.58%) and MRSE (77.85%) but might be the worst for preserving ECD (50.98%). aCXL with 9mW/cm(2) for 10 min (aCXL-10) might be the best protocol for preserving ECD (31.53%).ConclusionOverall, three protocols of aCXL are comparable in therapeutic efficacy and safety for treating progressive keratoconus. Despite no direct data comparing the efficacy of each technique according to different patients' profiles, it is reasonable to state that aCXL-5 may be the best for patients at early-stage to reduce K-max and MRSE, aCXL-3 may be the best for patients at mid-stage to improve BSCVA, and aCXL-10 may be the best for patients at late-stage to preserve DEC.