Systematic review and meta-analysis of femtosecond laser–enabled keratoplasty versus conventional penetrating keratoplasty

Abstract

Purpose: The aim of this meta-analysis is to compare femtosecond laser–enabled keratoplasty with conventional penetrating keratoplasty regarding postoperative astigmatism, visual functions, graft rejection, graft failure, and complications. Methods: Eligible studies were retrieved from five mainstream electronic databases, including PubMed, Embase, Ovid MEDLINE, Cochrane Library, and ClinicalTrial.gov. Postoperative topographic astigmatism was set as the primary outcome, and best-corrected visual acuity, spherical equivalent, endothelial cell density, graft rejection, graft failure, and complications were chosen as the secondary outcomes. Standard mean difference and risk ratio were the size effects for continuous data and binomial data, respectively. The data were pooled through either the random-effects model or the fixed-effects model based on data heterogeneity. Moreover, subgroup analyses were conducted when the heterogeneity occurred distinctly (I2 > 50%). Results: A total of seven comparative studies were included. The pooled standard mean difference (−0.32, 95% confidence interval: −0.74 to 0.10) showed that femtosecond laser–enabled keratoplasty was not superior over conventional penetrating keratoplasty to decrease postoperative topographic astigmatism (p = 0.14). However, best-corrected visual acuity after femtosecond laser–enabled keratoplasty was significantly better than that after conventional penetrating keratoplasty (p = 0.00, standard mean difference: −0.23, 95% confidence interval: −0.37 to −0.10). Furthermore, endothelial cell density after femtosecond laser–enabled keratoplasty was preserved significantly better (p = 0.03, standard mean difference: 0.63, 95% confidence interval: 0.07–1.20). Moreover, spherical equivalent, graft rejection, graft failure, and complications represented no distinct differences between femtosecond laser–enabled keratoplasty and conventional penetrating keratoplasty (p > 0.05). Conclusion: Femtosecond laser–enabled keratoplasty may not be superior over conventional penetrating keratoplasty in decreasing postoperative topographic astigmatism, but might have advantages to achieve best-corrected visual acuity and endothelial cell density preservation. In addition, the two techniques seem to be comparable regarding spherical equivalent, graft rejection/failure, and complications.