This prospective study compared femtosecond LASIK and small-incision lenticule extraction (SMILE) for myopia and myopic astigmatism.
Study design
In this randomized, paired-eye, single-masked trial, 70 patients with bilateral myopia and myopic astigmatism underwent SMILE in 1 eye and LASIK in the other. Each eye was treated with the VisuMax (Carl Zeiss Meditec, Jena, Germany) 500-kHz femtosecond laser system followed by excimer ablation using the WaveLight EX500 excimer laser (Alcon Laboratories, Inc, Fort Worth, TX) in the LASIK eye. The primary outcome was refractive predictability at 3 months, measured as a percentage of eyes achieving a spherical equivalent (SE) within 1.0 D of attempted correction. Secondary outcomes were efficacy and safety at 3 and 12 months.
Outcomes
At 3 months, there was no statistically significant difference in the refractive predictability between LASIK- and SMILE-treated eyes. Nearly all SMILE and LASIK eyes (99% and 97%, respectively) achieved SE within 1.0 D of attempted correction, whereas 87% of SMILE eyes and 92% of LASIK eyes were within 0.5 D of attempted correction (P=0.26).
At 12 months, SMILE outcomes were similar to LASIK in terms of UDVA of 20/40 or better, UDVA of 20/20 or better, efficacy index and safety index. There were no major intraoperative or postoperative complications during the trial, although 3 different participants experienced suction loss during the femtosecond laser procedure (1 eye during LASIK flap rim cut, 2 eyes during SMILE) without impeding the completion of the surgery. There were no significant differences in preoperative or postoperative contrast sensitivity, total aberrations and higher-order aberrations between groups at any time point.
Limitations
In this study, refractive surgery was performed on a homogenous population of Asian patients with an experienced surgeon who had performed over 50 cases of SMILE prior to this study. The results may not be transferrable to different patient populations or inexperienced SMILE surgeons. It was also difficult to mask patients, as LASIK requires a second excimer laser for treatment. In addition, the study only minimally addressed astigmatism and astigmatic correction in SMILE-treated eyes.
An accompanying editorial in the same issue of this journal nicely summarizes another inherent limitation in this noninferiority study design. The authors set a difference of 10% as a predetermined margin for noninferiority; this also led to their target sample size for recruitment. For example, whereas 92% of LASIK patients demonstrated a SE within 0.5 D of target refraction, the lower bound of the estimate for SMILE would have had to be more than 82% to show noninferiority. However, the mean difference in accuracy between SMILE and LASIK (deviation of target SE from achieved SE) was very small (0.02±0.05 D). In essence, failures from either procedure were extremely rare. As the authors state, "the predetermined margins for noninferiority were not meaningful when comparing our final results of 99% of SMILE eyes versus 97% of LASIK eyes achieving ±1.0 D of attempted correction." To put this in context, failures occurred approximately 3% of the time in the LASIK group; as such, SMILE could have resulted in 3 times as many failures and still have been considered noninferior.
Clinical significance
Both SMILE and LASIK appear to be highly efficacious procedures for moderate myopia, with similar safety and predictability. Treating patients with SMILE led to noninferior refractive outcomes, efficacy, safety, and predictability compared with femtosecond LASIK. While LASIK has been studied extensively with overall good visual outcomes and safety profile, SMILE may offer long-term advantages such as increased biomechanical corneal strength as well as the lack of flap-related complications. It will be of interest to the refractive community to follow long-term data.