The aim of this study was to systematically evaluate the efficacy, safety and predictability of implantable collamer lens (ICL) and corneal refractive surgery (CRS) for high myopia.

The Pubmed, Embase, Cochrane Librara, and CNKI databases were searched using the subject terms (phakic intraocular lenses, keratectomy, keratomileusis, laser in situ, photorefractive keratectomy, refractive errors and myopia). The search time was from the self-built database to May 2020. All clinical controlled trials of collect randomized controlled trials and observational studies. Two researchers independently screened the literatures and extracted uncorrected visual acuity (UCVA), best corrected visual acuity (BCVA), average spherical equivalent error (SER), the scope of SER, the type of ICL and CRS. The I² statistics were used to evaluate the heterogeneity of the data. When the heterogeneity was large (I²>50%), the random effects model was used to analysis; otherwise, the fixed effects model test was used. Binary variables was represented by risk ratio (RR) and its 95% confidence interval (CI).

After preliminary search, 476 related studies were obtained, and then the repeated, irrelevant and outcome indicators that did not meet studies were excluded. Finally, 11 studies were included, with a total of 988 eyes. All studies were observational design. All subjects were adults (aged 18 or above) with high myopia or high myopic astigmatism. Patients′ SER ranged from -6.00 D to-25.75 D and -6.00 D to -23.75 D in the ICL group and CRS group, respectively. The follow-up duration of included studies ranged from 1 to 49 months. There were 6 studies with 7 scores, 5 studies with 8 scores according to the evaluation of risk of bias by NOS, which had high quality. During the last follow-up, the proportion of UCVA 20/20 in the ICL group was higher than that in CRS group by the sensitive-analysis. The heterogeneity of two groups was relatively small (I²=0). After the fixed effects model test, the difference between two groups was statistically significant (RR=1.19, 95%CI: 1.06 to 1.34, P<0.05). Compared with CRS group, ICL group had more patients gaining BCVA of 1 lines or more and less patients in losing BCVA of 1 line or more by the sensitive-analysis. The heterogeneity of two groups in gaining BCVA of 1 lines or more was relatively larger (I²=83%). After the random effects model analysis, the difference between two groups was not statistically significant (RR=1.65, 95%CI: 0.94~2.90, P>0.05). The heterogeneity of two groups in losing BCVA of 1 line or more was relatively small (I²=0). After the fixed effects model test, the difference between two groups was statistically significant (RR=0.11, 95%CI: 0.02 to 0.56, P<0.05). The secondary surgeries/complications were 12.6% and 20.8% by the sensitive-analysis. The heterogeneity of two groups was relatively small (I²=45%). After the fixed effects model test, the difference between two groups was not statistically significant (RR=0.56, 95%CI: 0.28 to 1.14, P>0.05). The proportion of predictability within ±0.50 D and ±1.00 D of the ICL group were higher than that in CRS group by the sensitive-analysis. The heterogeneity of two groups was within ±0.50 D relatively small (I²=0). After the fixed effects model test, the difference between two groups was statistically significant (RR=1.31, 95%CI: 1.12 to 1.53, P<0.05). The heterogeneity of two groups was within ±1.00 D relatively large (I²=85%). After the random effects model analysis, the difference between two groups was not statistically significant (RR=1.07, 95%CI: 0.96 to 1.20, P>0.05).

For the treatment of high myopia, ICL shows a better performance in terms of the efficacy, safety and predictability than CRS; while there was not statistical difference in the secondary surgeries or complications between them.