With advances in corrective refractive technology, postoperative visual quality has become a top priority for research. In this study, both objective and subjective visual parameters were assessed including the safety, efficacy, HOAs, the OSI, and patient-reported questionnaire in an effort to give a comprehensive comparison.
In this study, the safety index was 1.22 ± 0.21 in the ICL group and 1.15 ± 0.14 in the SMILE group with no intraoperative or postoperative complications. No eyes from either group lost two or more lines of CDVA over 4 years, demonstrating the safety of both procedures. In addition, the efficacy index was comparable between the SMILE and ICL groups. Yan et al. reported the two-year safety and efficacy indexes as 1.24 ± 0.26 and 1.03 ± 0.23, respectively, after EVO-ICL implantation. Yang et al. reported that the safety index was 1.23 ± 0.22, and the efficacy index was 1.04 ± 0.16 4-year postoperatively. Blum et al. reported SMILE provides good long-term safety and efficacy of refraction in a 10-year observation.
SMILE exhibited a slightly superior predictability profile in this study. Eighty-one percent and 72% of eyes were within ± 0.50 D of the attempted spherical equivalent correction after SMILE and EVO-ICL implantation, respectively (P < 0.01). The postoperative mean residual cylinder after SMILE was slightly better than after EVO-ICL implantation. This is partly because in the ICL group, eyes with non-TICL might increase the mean residual SE owing to the uncorrected cylinder refraction. Despite the slight differences between the two procedures in terms of predictability, this study results still had a good performance comparable with previous study findings. In long-term studies of ICL implantation, Yang et al.  observed that 79% eyes had within ± 0.50 D of the attempted correction at 4 years. Igarashi et al. observed 82.9% at 4 years and 68.3% at 8 years, and Alfonso et al. observed 38.0% at 5 years. After SMILE, Han et al. reported 80% eyes within ± 0.50 D of the attempted correction at 3 years, and Li et al. reported 90% at 5 years. Our results were consistent with those of previous studies and demonstrated the safety and effectiveness of the two techniques.
There are concerns about the long-term endothelial cell loss postoperatively. In an FDA clinical study, Sanders et al. found a cumulative endothelial cell loss of between 8.4% and 8.9% over the first 3 years after ICL implantation. Igarashi et al. reported 6.2% of endothelial cell loss 8 years after surgery. Particularly focused on EVO-ICL, Lisa et al. reported a 1.7% reduction in ECD at 1 year. Yang et al. reported a 4.03% reduction at 4 years and demonstrated that the anterior segment biometric parameters, especially the vault, induced a decline in ECD. Our results showed a similar degree of endothelial cell loss (5.6% ± 10.7%) with the previous studies.
In this study, the postoperative ocular HOAs and OSI were assessed for the objective optical performance after the two procedures. Eyes treated with SMILE showed significantly higher amount of spherical aberrations and coma than those treated with ICL. ICL implantation for high myopia has shown excellent aberrometric control compared with other cornea refractive surgeries such as FS-LASIK or PRK.[25, 26] It also showed superiority to SMILE. Siedlecki et al. reported that SMILE-treated eyes showed significantly higher degree of spherical aberration, coma, and total HOAs than ICL-treated eyes at 1 year. Our previous study based on myopic patients has documented that SMILE-treated eyes induced significantly higher amounts of coma (both vertical coma and horizontal coma) than ICL-treated eyes. Few surgically induced HOAs after ICL implantation may be attributed to the retention of unchanged corneal biomechanics and a negative spherical aberration of ICLs. In contrast, SMILE may change the natural asphericity of the cornea into a relatively oblate surface, especially for high-myopic eyes, resulting in more induced HOAs.
There were no differences in the OSI and other parameters between the groups. OSI in both groups (1.3 ± 0.5 in the SMILE group and 1.3 ± 0.7 in the ICL group) was comparable with that reported in previous studies. For ICL-treated eyes, the short-term OSI was 1.08 in Huseynova’s study and was 1.16 in Kamiya’s study. Our previous study found that the mean OSI was 0.71 ± 0.38 at 18 months after SMILE with lower results than that reported in the present study. An OSI > 1.5 indicates significant light scattering.. The mean postoperative value for each group was not over 1.5, which suggested satisfying retinal image quality. Regarding the MTFcutoff value, Kamiya et al. and Qin et al. reported 26.21 cpd and 48.96 cpd, respectively, which were lower and higher, respectively, than that reported in the current study (30.46 cpd). In SMILE, our group found that the MTFcutoff was 38.20 cpd 12-month postoperatively and 37.81 cpd 18-month postoperatively, which were higher that that reported in this study (27.95 cpd). The difference may be due to the longer follow-up period, which needs to be discussed further.
Halo was the most commonly reported night vision disturbance in the ICL group (91%), although it was considered to be non-distressing in more than 90% of cases in both groups. Halos after conventional ICL implantation have been reported in previous studies[10, 11] and were found to be associated with differences between the mesopic pupil size and ICL optic zone diameter, and white-to-white diameter of the cornea. Regarding EVO-ICL, apart from the edge of the optical zone, the cylindrical inner wall of the central hole can be an additional optical surface causing light phenomena.[31, 32] Siedlecki et al. reported a prevalence of 80% for halos at 2 years after EVO-ICL implantation, and Wei et al. reported it as 93.5% at 3 months, which were consistent with our findings.
Previous studies found only very weak and probably clinically irrelevant associations between HOAs and subjectively perceived visual quality after SMILE.[12, 33] This finding may explain the same high-quality subjective visual quality obtained even though the HOA was higher than ICL after SMILE.
This study has some limitations. First, the sample size of both groups was relatively small owing to the long-term follow-up period. Second, only eyes with SE ranging from − 10.00 D to -6.00 D were included owing to the restriction in the application of SMILE. These findings cannot be extrapolated to patients with SE over − 10.00 D. Third, the lack of preoperative data in OQAS and HOAs analysis limited us from exploring the change between pre- and postoperative optical system and finding the associations between the changes and the visual complaints. A strength of the study was its comprehensive report, in both objective and subjective manner, of the four-year long-term visual outcomes after SMILE and ICL implantation in correcting high myopia. This study may help surgeons make clinical decisions for their patients with high myopia.