Comparative study of small incision refractive lenticule extraction with 5 mm versus 6.5 mm optic zone for myopia

Abstract

Purpose:

To compare residual stromal thickness (RST) in eyes undergoing small incision refractive lenticule extraction (SMILE) using a lenticular diameter of 6.5 mm versus those with a diameter of 5 mm.

Methods:

In this retrospective comparative case series, consecutive patients who underwent SMILE between 2016 and 2021 with at least 6 months of follow-up were included. Preoperative best-corrected distance visual acuity (BCDVA), refractive error, contrast sensitivity, central corneal thickness, keratometry, higher order aberrations, and scotopic pupil size were recorded using a Placido disk topography with Sheimpflug tomography-based system. Patients underwent SMILE with a lenticular diameter of 6.5 mm until 2018 (n = 372 eyes). Thereafter, the lenticular diameter was reduced to 5 mm (n = 318). The RST, postoperative refraction, aberrations, subjective glare, and halos were compared across groups at 1 and 6 months.

Results:

The mean age of participants was 26.8 ± 5.8 years with a mean preoperative spherical equivalent of −4.48 D ± 2.16 D (range: −0.75 to −12.25 D) and mean scotopic pupil of 3.7 ± 0.75 mm. Eyes in the 5 mm group had 30.6 m (95% confidence interval [CI] = 28 to 33 m, P < 0.001) greater RST compared to the 6.5 mm group after adjusting for spherical equivalent and preoperative pachymetry. There were no differences in vision, contrast sensitivity, aberrations (wavefront error of 0.19 ± 0.2 vs. 0.25 ± 0.2, P = 0.19) or glare between the two groups.

Conclusion:

SMILE performed with a lenticular diameter of 5 mm leads to greater RST across the myopic range, but without inducing significant higher-order aberrations.

Refractive surgery has undergone paradigm shifts because of the introduction of laser in situ keratomileusis (LASIK) by Pallikaris et al.[1] Continuous refinements including wavefront-guided LASIK and femtosecond–LASIK (FS-LASIK) have made treatments safer and more efficacious with improved patient satisfaction.[2,3] The main drawback of LASIK is the risk of corneal ectasia with a reported incidence of 0.06 to 0.8%;[4,5] it depends upon the degree of refractive error, preoperative pachymetry, thickness of the residual stromal bed, and the preoperative corneal topography findings.[6] Ectasia was attributed to the creation of the corneal flap that reduced the biomechanical strength of the cornea.[7]

To overcome this, and theoretically reduce the risk of ectasia, refractive lenticule extraction (ReLEx-SMILE) was introduced about a decade ago.[8] The ReLEx-SMILE procedure is effective in treating myopia up to −10 D sphere and −5 D cylinder with no flap-related complications such as in LASIK.[9,10] It also induces lesser higher-order aberrations and has a lower incidence of dry eye, diffuse lamellar keratitis, and epithelial growth, making it, perhaps, the safest corneal refractive procedure available today.[9] Yet, corneal ectasia following ReLEx–SMILE has been reported in past, mostly in eyes with subclinical keratoconus.[11,12] Though one would expect SMILE to be potentially safer than FS-LASIK in eyes with borderline topography due to the lack of a corneal flap, the residual stromal thickness (RST) may be the most important determinant predicting ectasia, irrespective of the type of procedure.[7]

Currently, the majority of femtosecond laser platforms for SMILE recommend using a 6.5 mm lenticule diameter for optimum results. We hypothesized that reducing this lenticular diameter may theoretically lead to the reduction of the lenticule thickness extracted, thereby increasing the RST. In this study, we reduced the lenticular diameter to 5 mm to report on the RST, refractive outcomes, and aberration profile, and subjective scotopic visual quality of these eyes compared to those who underwent SMILE with the conventional 6.5 mm lenticular diameter.

Methods

This was a retrospective study conducted at a tertiary refractive surgical practice in western India. The study was approved by the Institutional Ethics Committee and followed the tenets of the Declaration of Helsinki. All surgical procedures were performed after obtaining expressed written consent from patients.

All consecutive patients who underwent SMILE between April 2016 and June 2021 with at least 6 months of follow-up were included in the analysis. Preoperative workup included best-corrected distance visual acuity (BCDVA), detailed slit-lamp and dilated fundus evaluation, refractive error measurement using objective (Topcon Auto refractometer, Tokyo, Japan), and subjective methods to obtain the best spherical and cylindrical correction and contrast sensitivity assessment using the Pelli–Robson chart at 1 m distance. The central corneal thickness, keratometry values for the steep and flat corneal meridians, scotopic pupil size, and corneal aberrometry were measured using Placido disk topography with a Sheimpflug tomography-based system (Sirius; Costruzione Strumenti Oftalmici, Florence, Italy).

Patients underwent SMILE surgery using standard surgical techniques described previously using the VisuMax (Carl Zeiss Meditec AG; Jena, Germany) laser platform.[9] The spherical and cylindrical correction values were used in the nomogram provided by the manufacturer before performing SMILE. Patients undergoing SMILE up to December 2018 were subjected to a 6.5 mm optic zone (OZ) for lenticule extraction. From January 2019 onward, using the same nomogram, the OZ was reduced to 5 mm. All surgeries were performed by one surgeon using standardized techniques described before.[9] The superficial dissection was carried out first followed by the separation of the deeper lenticule plane.

Postoperatively, patients were prescribed topical antibiotics for 1 week and topical steroids for 10 days. Patients underwent detailed slit-lamp evaluation, and determination of uncorrected distance visual acuity (UCDVA) and contrast sensitivity at postoperative day 1, at 1 month, and 6 months follow-up. At each visit, patients were also shown standard images to demonstrate glare and halos and were asked whether they experienced any glare/haloes in each eye individually during the postoperative period. The residual stromal thickness was obtained from the Sirius tomography scans at the 1-month follow-up. Higher order aberrations were also assessed at the 1-month time point.

Statistical analysis

All continuous variables are presented as mean with standard deviation or median with interquartile (IQR) range and categorical variables are presented as proportions (n, %). Snellen’s visual acuity measurements were converted to logarithm of minimal angle of resolution (logMAR), and the spherical equivalent was calculated as sphere + ½ cylinder for analysis. Group differences between continuous variables were analyzed using the analysis of variance (ANOVA) or the Kruskall– Wallis test for non-parametric variables. Differences in categorical groups between variables were analyzed using the Chi-square or Fischer’s exact test. Changes in parameters at various time points during the study were analyzed using repeated measures ANOVA. Correlation between continuous variables was graphically plotted using the locally weighted scatterplot smoothing (LOWESS) curves.

Univariate and multivariable linear regression was employed using generalized estimating equations with mixed effects models (“xtgee” command in STATA) to identify factors predictive of residual stromal thickness, and results are presented as β coefficient with 95% confidence intervals (CIs). In the multivariable models, covariates were chosen if they showed an association with residual stromal thickness with a P < 0.1. The R² value and regression diagnostics such as multicollinearity, variance inflation, and influence of leverage points were used to check the robustness of the regression equations derived.

All data were entered in Microsoft Excel and analyzed using STATA 12.1 I/C (Fort Worth, Texas, USA). All P values <0.05 were considered statistically significant.

Results

We included 690 eyes of 345 patients with a mean age of 26.8 ± 5.8 years (range 18 to 48 years) and where 221 were women (64%). The mean preoperative BCVA was 0.03 ± 0.06 logMAR (range = 0 to 0.5 logMAR) and the mean preoperative spherical equivalent was −4.48D ± 2.16D (median = −4.25D, interquartile range [IQR] = −5.8 to −2.7D, range = −0.75 to −12.25D). The mean keratometry was 44D ± 1.2D and the mean central corneal thickness was 533 ± 29 m (median = 535 m, IQR = 512 to 555 m, range = 452 to 629 m). The mean scotopic pupillary size in the study group was 3.7 ± 0.75 mm.

Three hundred and seventy-two eyes (54%) eyes underwent SMILE using the 6.5 OZ, whereas the remaining 318 (46%) had SMILE with a 5 mm OZ. There were no differences in any baseline parameters between groups [Table 1]. Eyes in the 6.5 mm group had a marginally lower spherical equivalent of about 0.5 D compared to those in the 5 mm group; however, this was not statistically significant. On postoperative day 1, eyes in both groups had UCDVA (mean = 0.09 ± 0.11 logMAR) of 20/25 (i.e., 6/7.5) Senllen’s equivalent and 10% of patients reported glare and halos. There were no differences in UCDVA, glare, and contrast sensitivity between eyes in the 6.5 mm and 5 mm OZ groups across all time points [Table 2]. The RST was significantly higher in the 5 mm OZ group [Table 2] at 6 months time point, and this was uniformly seen across the range of refractive error in the cohort [Fig. 1]. The higher order aberrations induced due to surgery [Fig. 2] were comparable between 6.5 mm and 5 mm OZ [Table 2].

Table 1.

Comparison of preoperative parameters between lenticular diameters used for SMILE during the study

Variable	6.5 mm LD (n=372)	5 mm LD (n=318)	P
Age (years)	27.1±5.8	26.5±5.8	0.35
Gender (% men)	75 (40%)	51 (32%)	0.11
Spherical equivalent	-4.21±1.8	-4.79±2.4	0.08
Mean keratometry	43.9±1.4	44.1±1.4	0.42
BCDVA preop (logMAR)	0.01±0.04	0.04±0.08	0.21
CS preop (log CS)	1.99±0.02	1.98±0.02	0.71
Pachymetry (microns)	539±29	528±29	0.18
Pupil size (mm)	3.63±0.7	3.93±0.6	0.22
Mean lenticule thickness	193±33	167±23	<0.001

LD: Lenticular diameter, BCDVA: Best corrected distance visual acuity, CS: Contrast sensitivity, preop: Preoperatively

Table 2.

Comparison of postoperative parameters between lenticular diameters used for SMILE during the study

Variable	6.5 mm LD (n=372)	5 mm LD (n=318)	P
Postop day 1
UCDVA	0.09±0.04	0.12±0.1	0.36
Glare (n, %)	32 (9%)	35 (11%)	0.30
CS POD1 (log CS)	1.99±0.08	1.98±0.11	0.72
Postop 1 month
UCDVA	0.01±0.06	0.008±0.04	0.64
Glare (n, %)	2 (0.5%)	0	0.50
CS 1 month (log CS)	1.99±0.01	1.98±0.08	0.70
Postop 6 months
Residual stromal thickness	346±39	360±33	<0.001
UCDVA	0.006±0.03	0.008±0.04	0.61
Glare (n, %)	2 (0.5%)	0	0.50
CS 1 (log CS)	1.99±0.02	1.98±0.08	0.70
Change in optical aberrations at 1 month
∆ Spherical aberrations	0.08±0.1	0.14±0.1	0.32
∆ Coma	0.10±0.1	0.19±0.4	0.17
∆ Higher order aberrations	0.12±0.1	0.19±0.2	0.20
∆ Wavefront error	0.19±0.2	0.25±0.2	0.19

UCDVA: Uncorrected distance visual acuity, POD: Postoperative day, CS: Contrast sensitivity

Figure 1.

A locally weighted scatterplot smoothing (LOWESS) curve showing a linear relationship between spherical equivalent and residual stromal thickness in the 6.5 mm versus 5 mm optic zone groups

Figure 2.

(a and b) Wave front error of 0.62 m preoperatively and 0.5 m postoperatively at 5 mm pupil size with OZ of 6.5 mm. (c and d) Wave front error of 0.48 m preoperatively and 0.57 m postoperatively at 5 mm pupil size with OZ of 5 mm

On comparing eyes with low, moderate, and high myopia [Table 3], we found no differences in the preoperative parameters between groups. A significantly higher proportion of patients in the high myopia group reported glare and halos on the first postoperative day [Table 3]; however, this did not persist at 1 and 6 months time points. All other parameters were comparable between eyes with different degrees of myopia at all time points during follow-up, including higher-order aberrations at 1 month.

Table 3.

Comparison of parameters between eyes with low, moderate, and high myopia

Variable	<−3DSph (n=205)	3-6 DSph (n=317)	>-6 DSph (n=168)	P
Baseline
Age (years)	27.5±5.1	26±5.1	27.4±6.6	0.85
Gender (% men)	42 (41%)	56 (34%)	28 (34%)	0.47
Spherical equivalent	−2.1±0.5	−4.4±0.8	−7.5±1.3	<0.001
Mean keratometry	43.7±1.4	44.1±1.4	44.3±1.4	0.22
BCDVA (logMAR)	0.01±0.03	0.02±0.05	0.06±0.09	0.09
CS (log CS)	2±0	1.99±0.02	1.98±0.07	0.14
Pachymetry (microns)	530±32	534±28	535±28	0.27
Pupil size (mm)	3.83±0.6	3.75±0.7	3.72±0.7	0.26
% eyes in 5 mm optic zone	92 (45%)	131 (41%)	95 (56%)	0.50
Lenticule thickness	154±16	182±27	210±25	<0.001
Postop day 1
UCDVA	0.01±0.03	0.01±0.05	0.02±0.06	0.19
Glare (n, %)	14 (7%)	23 (7%)	30 (18%)	0.001
CS (log CS)	2±0	1.99±0.05	1.98±0.08	0.51
Postop 1 month
UCDVA	0±0	0.04±0.09	0.007±0.03	0.10
Glare (n, %)	0	0	2 (1%)	0.06
CS (log CS)	2±0	1.99±0.05	1.98±0.08	0.12
Postop 6 months
Residual stromal thickness	375±35	352±33	325±26	<0.001
UCDVA	0±0.03	0±0.03	0.01±0.04	0.27
Glare (n, %)	0	0	2 (1%)	0.06
CS 1 month (log CS)	2±0	1.99±0.05	1.98±0.08	0.12
Change in optical aberrations at 1 month
□ Spherical aberrations	0.08±0.1	0.10±0.12	0.15±0.14	0.21
□ Coma	0.10±0.11	0.11±0.13	0.24±0.9	0.16
□ Higher order aberrations	0.13±0.13	0.12±0.12	0.22±0.19	0.23
□ Wave front error	0.20±0.17	0.20±0.19	0.25±0.22	0.45

Univariate linear regression analysis with generalized estimating equations showed that spherical equivalent, preoperative pachymetry, and lenticular diameter size were significantly associated with residual stromal thickness. On multivariable analysis, after adjusting for the spherical equivalent, the 5 mm OZ led to 30.6 μ higher RST (95% CI = 28 to 33 microns, P < 0.001) compared to the 6.5 mm group. In separate multivariable models, after adjusting for spherical equivalent as a categorical variable, the 5 mm OZ group still had 27 μ higher residual thickness (95%CI = 24–30 μ, P < 0.001) eyes compared to the 6.5 mm group. The following regression formula was found to accurately predict the RST (R² = 0.82):

RST = −95 +30 (OZ size) +10.9*(Spherical Equivalent) +0.9* pachymetry, where RST = residual stromal thickness, −95 was the constant, OZ size was 0 for 6.5 mm and 1 for 5 mm groups, the spherical equivalent was in diopters and pachymetry was in microns. None of the eyes experienced intraoperative or postoperative complications including lenticule cap adhesions and separation problems.

Discussion

In this study, we found that a lenticular diameter of 5 mm led to excellent refractive correction at 1-month postop with the majority of patients gaining uncorrected vision of 20/20. After adjusting for baseline spherical equivalent and pachymetry, these eyes had 30 m greater RST compared to eyes with 6.5 mm lenticular diameter. Though the induced higher-order aberrations were marginally higher in the 5 mm group, these differences were not statistically significant, and patients did not report greater glare and halos subjectively.

In recent times, SMILE is becoming the refractive procedure of choice for treating myopia and myopic astigmatism and is being widely adopted by refractive surgeons globally. The main advantage of SMILE over LASIK is that it is flap free, leading to better corneal bio-integrity, intact corneal nerves, and rapid wound healing. Many modifications of SMILE have been presented to improve outcomes and enhance patient satisfaction, many of which have been enumerated by Titiyal et al.[13] One of these modifications is to reduce the size of the lenticular diameter to below the recommended 6.5 mm in an attempt to increase the RST and improve the biomechanical properties of the cornea.

Qian et al.[14] reported retrospective outcomes from 128 eyes undergoing SMILE using either a 6.5 mm or a 6.2 mm optical zone. Their main outcome measures were corneal power distribution measured by a ray tracing method and functional optical zone obtained via a Scheimpflug camera. The authors reported that the change in the total corneal power at 5 mm and 4 mm from the center of the pupil, led to the most accurate target refraction at 1 month in the 6.5 mm and 6.2 mm groups, respectively. These findings prompted us to attempt a reduction in the lenticular diameter to 5 mm. In another study, Fu et al.[15] reported outcomes from SMILE performed across three optic zones, that is, >6.5 mm, 6.3 to <6.5, and <6.3 mm. Authors reported good refractive outcomes in all three groups but patients with high myopia (>−6 D) had a larger difference in planned and achieved functional zone compared to those with lower myopia (<−3D). Similarly, we found good refractive outcomes in all three groups at 1 month.

Moshirfar et al.[11] reported corneal ectasia following SMILE in seven eyes of four patients. In retrospect, all these eyes were seen to have a high Randleman ectasia risk score and the percent tissue altered (PTA) was almost 40%, suggesting that these eyes had subclinical keratoconus at baseline. The theoretical advantage of SMILE over femto–LASIK in improving the biomechanical strength of the cornea has not been confirmed by many studies.[16,17] Perhaps, it is the RST that is the most important predictor of ectasia than the procedure itself.[7] In our study, we found the RST to be significantly better in eyes with a 5 mm lenticule extracted during SMILE, ranging between 28 and 33 μ better, compared to those with a 6.5 mm lenticule, after adjusting for the refractive error. Based on these findings, we believe that it may be prudent to extract a lenticule with a smaller diameter, especially in eyes with higher refractive error and borderline corneal topography readings.

The increased safety of SMILE afforded by a smaller lenticular diameter may be offset by the greater proportion of patients complaining of glare and halos. However, we did not find this in our study where almost none of our participants reported glare and halos subjectively at 1 and 6 months. Additionally, the change in higher-order aberrations was not significantly different in the 5 mm versus 6.5 mm groups. In our study, glare was mainly reported on postoperative day 1, and that too mostly in the high myopia group. Because glare and halos are also influenced by the scotopic pupil size, we believe that our patients, with a mean scotopic pupil size of 3.7 mm, may better tolerate extraction of a smaller lenticular diameter, especially as this may improve long-term safety and prevent ectasia. However, other ethnic populations with larger pupils may experience such phenomenon more frequently and the 5 mm optic zone may be reserved for patients with higher degrees of myopia where the RST is the primary concern.

The main drawback of the study is its retrospective nature. To the best of our knowledge, this is the first study reporting on outcomes from SMILE performed with a 5 mm lenticular diameter with a relatively large sample size.

Conclusion

In conclusion, we found that SMILE performed with a lenticular diameter of 5 mm was effective in delivering excellent refractive correction and improving the corneal biomechanical strength by virtue of significantly greater RST. This did not induce greater higher-order aberrations and was well tolerated by our patients. However, our results cannot be generalized to populations with larger scotopic pupils in whom this modification may be best suited for eyes with higher refractive error, borderline topography, and scotopic pupillary diameters of less than 4 mm. Further studies with a prospective and randomized design are necessary to confirm our findings and increase the widespread adoption of a smaller optical zone in SMILE.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.