ORIGINAL ARTICLE

SMILE for Hyperopia With and Without

Astigmatism: Results of a Prospective
Multicenter 12-Month Study
Dan Z. Reinstein, MD, MA(Cantab), FRCOphth; Walter Sekundo, MD, PhD;
Timothy J. Archer, MA(Oxon), DipCompSci(Cantab), PhD; Pavel Stodulka, MD;
Sri Ganesh, MD; Beatrice Cochener, MD, PhD; Marcus Blum, MD, PhD; Yan Wang, MD, PhD;
Xingtao Zhou, MD, PhD

RESULTS: The preoperative spherical equivalent was +3.20

ABSTRACT ± 1.48 D (range: +0.25 to +6.50 D). At the 12-month follow-
up visit, 81% of eyes treated were within ±0.50 D and 93% of
PURPOSE: To investigate the safety and effectiveness of eyes were within ±1.00 D of intended correction. A total of
small incision lenticule extraction (SMILE) in patients who 1.2% of eyes lost two or more lines of CDVA at the 12-month
have hyperopia with or without astigmatism. follow-up visit, and 83% were at least 20/20, correspond-
ing to a safety index of 1.005 at 12 months. Of the 219 eyes
METHODS: This was a prospective multicenter trial includ- with plano target, 68.8% had an uncorrected distance visual
ing 374 eyes of 199 patients treated by SMILE for hyperopia acuity of 20/20 or better and 88% were at least 20/25 uncor-
using the VisuMax femtosecond laser (Carl Zeiss Meditec rected at 12 months. There were no statistically significant
AG). Inclusion criteria were sphere up to +6.00 diopters (D), changes in contrast sensitivity.
cylinder up to 5.00 D, and maximum hyperopic meridian
up to +7.00 D, with preoperative corrected distance visual CONCLUSIONS: SMILE was found to be an effective treatment
acuity (CDVA) of 20/25 or better. The optical zone was 6.3 method for the correction of compound hyperopic astigma-
mm with a transition zone of 2 mm. The minimum lenticule tism, demonstrating a high level of efficacy, predictability,
thickness was set at 25 µm in the center and at 10 µm at the safety, and stability.
edge. Patients were examined at 1 day, 1 week, and 1, 3, 6,
9, and 12 months after surgery. Standard refractive surgery [J Refract Surg. 2022;38(12):760-769.]
outcomes analysis was performed.

From Reinstein Vision, London, United Kingdom (DZR, TJA); London Vision Clinic, EuroEyes Group, London, United Kingdom (DZR, TJA); the
Department of Ophthalmology, Phillips University of Marburg, Germany (WS); Gemini Eye Clinic, Zlín and Gemini Eye Clinic, Prague, Czech
Republic (PS); Nethradhama Superspeciality Eye Hospital, Bangalore Jayanagar, India (SG); University Brest Hospital, Brest, France (BC); the
Department of Ophthalmology, HELIOS–Hospital Erfurt, Erfurt, Germany (MB); Tianjin Eye Hospital, Heping District, Tianjin, China (YW); and
EYE & ENT Hospital, Fudan University, Shanghai, China (XZ).
© 2022 Reinstein, Sekundo, Archer, et al; licensee SLACK Incorporated. This is an Open Access article distributed under the terms of the
Creative Commons Attribution 4.0 International (https://creativecommons.org/licenses/by/4.0). This license allows users to copy and distribute,
to remix, transform, and build upon the article, for any purpose, even commercially, provided the author is attributed and is not represented
as endorsing the use made of the work.
Submitted: July 26, 2022; Accepted: November 2, 2022
Supported by Carl Zeiss Meditec AG, Jena, Germany. The sponsor participated in the design of the study, data management, data analysis,
interpretation of the data, and review of the manuscript.
Disclosure: Drs. Reinstein, Blum, Ganesh, Sekundo, and Zhou are consultants for Carl Zeiss Meditec AG. Dr. Cochener reports that University
Brest Hospital received funding for the study. Dr. Stodulka received funding for the study from Carl Zeiss Meditect AG. Dr. Reinstein has a
proprietary interest in the Artemis Insight 100 technology (ArcScan, Inc, Golden, Colorado) through patents administered by the Cornell Center
for Technology Enterprise and Commercialization (CCTEC), Ithaca, New York. The remaining authors have no proprietary or financial interest
in the materials presented herein.
Drs. Reinstein and Sekundo contributed equally to this work and should be considered as equal first authors.
Clinical Trial No. NCT03431571.
Correspondence: Dan Z. Reinstein, MD, MA(Cantab), FRCOphth, Reinstein Vision, London Vision Clinic, EuroEyes Group, 138 Harley Street,
London W1G 7LA, United Kingdom. Email: [email protected]
doi:10.3928/1081597X-20221102-02

760
W
ith more than 6 million treatments worldwide, The inclusion criteria were as follows: (1) patient age
small incision lenticule extraction (SMILE) has of 18 years or older; (2) preoperative CDVA of 20/25 or
been established as a modern, minimally inva- better in both eyes; (3) central corneal thickness of great-
sive, and reliable procedure for the treatment of myopia er than 500 µm and calculated postoperative residual
and myopic astigmatism since the first publications in stromal thickness of 250 µm or greater; (4) normal cor-
2011.1,2 SMILE for hyperopia has also been developed neal topography; (5) no preceding refractive surgery; (6)
and preliminary studies published.3-9 The initial feasi- subjective hyperopic sphere up to +6.00 D; (7) subjective
bility study of hyperopia treatment using femtosecond astigmatism up to +5.00 D; (8) refraction at the maximum
lenticule extraction, published in 2013, showed prom- hyperopic meridian up to +7.00 D; (9) a difference in
ising initial results, but was associated with refractive spherical equivalent refraction (SEQ) between cyclople-
regression and loss of corrected distance visual acuity gic and manifest refractions up to 1.00 D; (10) predicted
(CDVA).3 As a consequence of this investigation, the len- postoperative keratometry less than 51.00 D (calculated
ticule profile was redesigned to have both a larger optical by adding the maximum hyperopia treated to the pre-
zone and a larger transition zone, which was shown to operative keratometry); (11) no contact lens wear before
improve safety and stability, and centration was applied baseline measurements for at least 1 week (soft lenses), 2
to the corneal vertex rather than the pupil center.4,5 weeks (soft toric lenses), or 1 month (rigid gas permeable
This updated large transition zone profile was first lenses); (12) patients able to understand the patient infor-
investigated for hyperopic SMILE treatment by Rein- mation and willing to sign an informed consent; and (13)
stein et al6-8 and Pradhan et al9 at the Tilganga Insti- patients willing to comply with all follow-up visits and
tute of Ophthalmology. This study found that optical the respective examinations.
zone centration was similar to hyperopic LASIK and The exclusion criteria were identical to the strict
that the achieved optical zone diameter after SMILE general exclusion criteria for corneal laser refractive
with a 6.3-mm optical zone (and 2-mm transition surgery, such as autoimmune dermal and/or collagen
zone) was larger than for laser in situ keratomileusis vascular diseases, diabetes, ocular comorbidities (eg,
(LASIK) with a 7-mm optical zone with corresponding glaucoma, keratoconus, pellucid marginal degenera-
improvement in the induction in spherical aberration tion, or cataract), pregnancy/nursing, systemic or topi-
with SMILE compared to the same zone LASIK.6,7 This cal medications with steroids/antimetabolites/retinoids
study also showed promising visual and refractive re- and other drugs, inability to stay still in a supine posi-
sults at 38 and 129 months, demonstrating refractive tion, and/or weight greater than 135 kg (weight limita-
stability despite the high refractive average correction tion of the VisuMax laser bed).
in the patients recruited for study. The preoperative examination was identical to the
Following these early studies, a large scale, multi- previous hyperopic femtosecond lenticule extraction
center, international, prospective study was executed study.4,5 Visual acuity was measured to the nearest let-
for hyperopic SMILE using the VisuMax femtosecond ter (0.02 logMAR) using the CSV-1000 standardized
laser (Carl Zeiss Meditec AG). Early Treatment of Diabetic Retinopathy Study chart
(VectorVision). Manifest refraction and cycloplegic re-
PATIENTS AND METHODS fraction were performed according to a standardized
This prospective, multicenter study was conducted be- protocol to push maximum plus and maximum cylin-
tween July 2017 and October 2020 at the following sites: der.10 The manifest refraction was repeated at a sepa-
Department of Ophthalmology, Phillips University of rate visit before the day of surgery, with the advantage
Marburg, Marburg, Germany (Sekundo); London Vision of being able to refer to the previous manifest, cyclo-
Clinic, London, United Kingdom (Reinstein); Gemini Eye plegic, and aberrometry refractions. This final mani-
Clinic, Zlín and Prague, Czech Republic (Stodulka); Neth- fest refraction was used to plan the treatment. The
radhama Superspeciality Eye Hospital, Bangalore, India CSV-1000 chart with four rows of sine-wave gratings
(Ganesh); Hôpital Morvan, Brest, France (Cochener); De- was used to measure contrast sensitivity at spatial fre-
partment of Ophthalmology, HELIOS–Hospital Erfurt, quencies of 3, 6, 12, and 18 cycles/degree (cpd). Cor-
Erfurt, Germany (Blum); Tianjin Eye Hospital, Heping neal topography was measured using the Atlas 9000
District, Tianjin, China (Wang); and EYE & ENT Hospital, topography system (Carl Zeiss Meditec AG) to monitor
Fudan University, Shanghai, China (Zhou). The study was keratometry and higher order aberrations.
listed on clinicaltrials.gov (NCT03431571), was approved In patients with bilateral ametropia, a bilateral si-
by the respective Ethics Committees/Institutional Review multaneous procedure was performed on the day of
Boards at each participating institution, and adhered to surgery. In cases where only one eye fulfilled the in-
the tenets of the Declaration of Helsinki. clusion criteria, the fellow eye was either left untreat-

• Vol. 38, No. 12, 2022 761

ed or received a standard femtosecond laser–assisted manifest refraction, contrast sensitivity using the CSV-
LASIK treatment. 1000, slit-lamp examination, corneal topography using
Figure A (available in the online version of this arti- the Atlas, corneal tomography using the MS-39 OCT
cle) summarizes the parameters, geometry, and cutting (CSO Italia), corneal aberrations, and intraocular pres-
order of the lenticule, as has already been described in sure using the Ocular Response Analyzer (Reichert).
detail in previous publications.4,5 The standard proto-
col was to use an optical zone of 6.3 mm with a transi- Statistical Analysis
tion zone of 2 mm (corresponding to a total lenticule A sample size calculation was performed using an
diameter of 8.3 mm), although the zone was adjusted exact binomial test including seven approval criteria
between 6 and 7 mm for some cases. A medium-sized for safety, efficacy, predictability, stability, and cyl-
(M) contact glass was used in all cases with sufficient inder correction. Including a drop-out rate of 20%,
white-to-white horizontal diameter (11.7 mm) to ac- the total sample size was calculated as 374 eyes with
commodate the large optical zone and transition zone. an a error probability of 0.05 and a statistical power
The minimum lenticule thickness used was fixed at 25 (1-b error probability) of 0.8. The sample size calcu-
µm centrally and between 10 and 24 µm peripherally. lation also took into account the variance inflation
A cap thickness of 120 µm was used. The femtosecond due to partial correlation between right and left eyes.
laser energy and spot spacing settings were optimized The impact of this correlation was calculated based
for each individual laser.11 The energy was usually ap- on a previous hyperopic data set. Because binomial
proximately 130 nJ (VisuMax software setting 26) with criteria refer to rates, the intraclass correlation (ICC)
a fixed laser spot and tracking distance of 4.5 µm. The was considered.13,14 For all performance criteria that
small incision width ranged from 2 to 4 mm and was allow the calculation of the ICC, a correction factor
placed superotemporally or superiorly. At some sites, was applied. The correction factor was calculated as
a reserve small incision was created superonasally ac- Deff = (2n1 + n2) / [2n1 / (1 + ICC) + n2], where n1 was
cording to the Standard Operating Procedure.11 The ap- the number of patients treated bilaterally and n2 was
proximate suction time for the average case of hyperopic the number of patients treated monocularly. Because
SMILE was 31 seconds (± approximately 3 seconds). the numbers n1 and n2 were not known in advance, it
After suction was applied, the centration achieved was was conservatively assumed that all patients would
confirmed to be on the corneal vertex with reference be treated bilaterally (ie, n2 = 0). This resulted in a
to the topography eye image, as described previously.11 simplified formula: Deff = 1 + ICC.13,14 The correction
For eyes with a cylinder of 2.00 D or greater, marks was applied if a clear correlation was present (ie, if
were placed on the cornea to adjust for cyclotorsion by ICC was greater than 0.3).15 The sample size was mul-
rotating the contact glass after suction was applied.12 tiplied by the correction factor for the respective cri-
teria. The final sample size was chosen as the largest
Surgical Technique sample size required across all criteria. Full details of
Due to the use of the M size contact glass and larg- the sample size calculation are included in the Clini-
er cap diameter, a modified version of the Reinstein cal Investigation Report (data on file, Carl Zeiss Med-
SMILE Separator was used that has a longer 9-mm sep- itec AG).
arating arm (compared to 8 mm for the similar myopic All measured data were entered and collected using
instrument). This enables the safe separation of the an electronic case report form for a centralized analy-
interfaces without the instrument stressing the small sis. Regular monitoring visits were performed by an
incision. The lenticule was separated using the same independent clinical research organization according
technique as each surgeon used for myopic SMILE, to the International Organization for Standardization
but here ensuring that the lenticule interface separa- standard ISO 14155 and clinical guidelines. In addi-
tion was first conducted peripherally followed by the tion, the Marburg site as a Leading Investigator was
central cornea where the lenticule is thinnest (25 µm). monitored by the Government Authority of the Fed-
eral State of Hesse (Germany). A comprehensive and
Postoperative Evaluation detailed 230-page Clinical Investigation Report was
Patients were given antibiotic and steroid eye drops produced and provided to the European authorities.
according to the standard routine protocol at each site. Outcome analysis was performed according to the
Patients were observed for 12 months with postopera- Standard Graphs for Reporting Refractive Surgery16
tive visits at 1 week and 1, 3, 6, 9, and 12 months. The and vector analysis was performed using the Alpins
postoperative examination included measurement of method.17 Data from the 12-month visit were used
uncorrected distance visual acuity (UDVA), CDVA, for analysis. Eyes where the intended postoperative

762
refraction was not emmetropia were excluded in the TABLE 1
efficacy analysis. For the calculation of lost/gained Study Demographics
lines of visual acuity, logMAR values were converted Parameter Value
to Snellen lines. Unchanged was considered to be be- Eyes (patients) 374 (199)
tween -0.9 and +0.9 lines, and gained/lost a line +1.0 Age (years) 38.8 ± 12.5 (18 to 69)
to +1.9 and -1.0 to -1.9, respectively. Accordingly,
Gender (M/F % [patients]) 45% (89) / 55% (110)
gain/loss of two lines was considered as +2.0 to +2.9
lines and -2.0 to -2.9 lines. For statistical analysis, the Manifest refractive sphere (D), +2.76 ± 1.48 (-0.25 to +6.00)
mean ± SD (range)
values of the CSV-1000 were converted to logarithmic
Manifest refractive cylinder (D), +0.88 ± 0.87 (0.00 to +4.75)
units. All statistics were calculated using SAS (SAS mean ± SD (range)
Institute) and Microsoft Excel 2019 (Microsoft Corpo-
Manifest spherical equivalent +3.20 ± 1.48 (+0.25 to +6.50)
ration) software. refraction (D), mean ± SD (range)
Attempted hyperopic sphere (D), +3.05 ± 1.41 (0.25 to +6.00)
RESULTS mean ± SD (range)
Patient Population Attempted refractive cylinder +0.88 ± 0.87 (0.00 to +4.75)
A total of 199 patients (374 eyes) were recruited (D), mean ± SD (range)
for the study, as per the sample size calculation, of Attempted spherical equivalent +3.49 ± 1.38 (0.75 to +6.50)
whom 175 patients were treated binocularly and 24 refraction (D), mean ± SD (range)
patients were treated monocularly. Data were avail- Preoperative minimum cor- 558 ± 29 (500 to 649)
able at 3 months in 352 eyes (94.1%), 6 months in 338 neal thickness (µm), mean ± SD
(range)
eyes (90.4%), 9 months in 315 eyes (84.2%), and 12
months in 323 eyes (86.4%). Follow-up percentages Scotopic pupil diameter (mm), 5.30 ± 1.09 (2.50 to 7.90)
mean ± SD (range)
were slightly lower than projected due to the coro-
D = diopters; SD = standard deviation
navirus disease 2019 (COVID-19) pandemic. Table
1 shows demographic data for the study population.
The mean age was 38.8 ± 12.5 years with the oldest
patients treated in London, United Kingdom (mean: 24 µm in 39 eyes (10.4%). All lenticules were created
52.3 ± 11.6 years) and the youngest patients treated with a central minimum thickness of 25 µm.
in Zlín, Czech Republic (mean: 27.3 ± 6.7 years). The
highest hyperopia was treated in Zlín, Czech Republic Efficacy
(SEQ: +4.26 ± 1.31 D) and the lowest in Marburg, Ger- Figure 1 shows the Standard Graphs for Reporting
many (SEQ: +2.56 ± 1.20 D). Table 2 shows the num- Refractive Surgery for the data from the 12-month visit.
ber of eyes by each diopter combination of maximum As shown in Figure 1A, UDVA was 20/20 or better in
hyperopia and cylinder treated. The optical zone used 68% of eyes, relative to 85% with preoperative CDVA
was 6 to 6.2 mm in 21 eyes (5.6%), 6.3 mm in 352 eyes of 20/20 or better. UDVA was 20/16 or better in 28%
(94.1%), and 7 mm in 1 eye (0.27%). The minimum and was 20/40 or better in all eyes at the 12-month visit.
peripheral lenticule thickness was 10 µm in 263 eyes UDVA was within one line of preoperative CDVA in
(70.3%), 11 to 15 µm in 72 eyes (19.3%), and 20 to 93% of eyes (Figure 1B).

TABLE 2
Distribution of Eyes by Sphere and Cylinder Treated (Positive Cylinder Notation)
Bin Sphere (D)
Distribution
Cylinder 0.00 to +1.00 +1.01 to +2.00 +2.01 to +3.00 +3.01 to +4.00 +4.01 to +5.00 +5.01 to +6.00 Total
0.00 to +1.00 38 (10.16%) 70 (18.72%) 66 (17.65%) 44 (11.76%) 34 (9.09%) 2 (5.88%) 274 (73.26%)
+1.01 to +2.00 2 (0.53%) 14 (3.74%) 7 (1.87%) 21 (5.61%) 8 (2.14%) 5 (1.34%) 57 (15.24%)
+2.01 to +3.00 5 (1.34%) 8 (2.14%) 7 (1.87%) 7 (1.87%) 4 (1.07%) 0 (0%) 31 (8.29%)
+3.01 to +4.00 5 (1.34%) 4 (1.07%) 1 (0.27%) 0 (0%) 0 (0%) 0 (0%) 10 (2.67%)
+4.01 to +5.00 2 (0.53%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 2 (0.53%)
Total 52 (13.90%) 96 (25.67%) 81 (21.66%) 72 (19.25%) 46 (12.30%) 27 (7.22%) 374 (100%)
D = diopters

• Vol. 38, No. 12, 2022 763

Figure 1. Nine standard graphs for reporting refractive surgery showing the visual and refractive outcomes 12 months after hyperopic small inci-
sion lenticule extraction using the VisuMax femtosecond laser (Carl Zeiss Meditec AG). (A) UDVA; (B) UDVA vs CDVA; (C) change in CDVA; (D) SEQ
attempted vs achieved; (E) SEQ accuracy; (F) SEQ stability; (G) refractive astigmatism; (H) TIA vs SIA; (I) refractive astigmatism angle of error.
CDVA = corrected distance visual acuity; D = diopters; Postop = postoperative; Preop = preoperative; SEQ = spherical equivalent refraction; SIA =
surgically induced astigmatism; TIA = target induced astigmatism; UDVA = uncorrected distance visual acuity

764
 TABLE 3
Stability of Spherical Equivalent Refraction and Atlas Keratometry
Mean 3 to 12 Months 3 to 12 Months
3 to 12 Months Change Change Within Change Within
Parameter Preoperative 3 Months 12 Months Change per Month ±0.50 D ±1.00 D
Eyes with refractive 374 352 323 316 316 316 316
data, n
Spherical equiva- +3.49 ± 1.34 -0.02 ± 0.57 +0.12 ± 0.56 +0.12 ± 0.46 0.013 72.15% 93.35%
lent refraction (+0.75 to +6.50) (-2.00 to +3.75) (-1.75 to +3.00) (-1.88 to +1.75);
adjusted for intend- P < .001
ed target (D), mean
± SD (range)
Refractive cylinder +0.88 ± 0.87 +0.42 ± 0.41 +0.40 ± 0.44 -0.02 ± 0.38 -0.002 67.72% 96.84%
(D), mean ± SD (0.00 to +4.75) (0.00 to +2.25) (0.00 to +2.00) (-1.00 to +1.25);
(range) P = .319
Eyes with keratom- 373 352 320 313 313 313 313
etry data, n
Average keratom- 42.79 ± 1.51 45.23 ± 1.93 45.15 ± 1.95 -0.15 ± 0.62 -0.017 75.40% 93.93%
etry (D), mean ± SD (38.79 to 46.59) (40.13 to 50.03) (39.88 to 49.74) (-3.35 to +5.13);
(range) P = .024
Corneal astigma- 1.23 ± 0.88 1.24 ± 0.62 1.23 ± 0.64 -0.002 ± 0.40 -0.0002 81.47% 97.44%
tism (D), mean ± (0.02 to 4.16) (0.03 to 3.45) (0.04 to 3.87) (-1.60 to +1.98);
SD (range) P = .671
D = diopters; SD = standard deviation
The Atlas is manufactured by Carl Zeiss Meditec AG.

Safety Astigmatism
There was a gain of one line of CDVA in 10% of Postoperatively, the cylinder improved in all cases
eyes, whereas there was one line loss in 11% and and was 1.00 D or less in 93% and 0.50 D or less in
two or more lines loss in 1.2% at the 12-month vis- 75% of eyes (Figure 1G). There were no eyes with in-
it (Figure 1C). For eyes with preoperative CDVA of duced cylinder greater than 2.00 D. The scatter plot
20/25 or better, none had a final postoperative CDVA for surgically induced astigmatism vector (SIA) ver-
worse than 20/40. sus target induced astigmatism vector (TIA) shows
Table A (available in the online version of this arti- that the refractive cylinder correction was on target
cle) includes the photopic contrast sensitivity data be- in terms of magnitude (Figure 1H). The angle of error
fore and at the 12-month visit after the surgery. At 12 histogram (Figure 1I) shows that the refractive correc-
months, there was a statistically significant decrease tion was placed accurately on the intended meridian
at 6 cpd (P = .0018), 12 cpd (P < .0001), and 18 cpd (P for the majority of eyes, with 76% within ±15 degrees.
< .0001). There was a decrease by more than 0.25 log Figure B (available in the online version of this arti-
units in more than 30% of eyes for the 12 and 18 cpd cle) shows the vector analysis for refractive cylinder,
frequencies. and the main outcome measures are shown in Table B
(available in the online version of this article).
Predictability
SEQ predictability was within ±0.50 D in 81% and Refractive and Corneal Stability
within ±1.00 D in 93% of eyes at the 12-month visit Figure 1F displays the change in SEQ over time.
(Figure 1E). The mean SEQ improved from +3.49 ± Table 3 shows SEQ, refractive cylinder, and Atlas av-
1.38 D (range: 2.11 to 4.87 D) to +0.12 ± 0.56 D (range: erage keratometry and corneal astigmatism before sur-
-0.44 to +0.68 D) at 12 months. The attempted versus gery and change between 3 and 12 months after sur-
achieved scatter plot (Figure 1D) shows that there gery. As expected for hyperopic treatments, there was
was a trend for undercorrection to a small degree in an initial overcorrection that healed to the final SEQ
low hyperopia, with the undercorrection more signifi- from -0.17 ± 0.53 D at 1 month to -0.02 ± 0.57 D at 3
cant above +4.00 D. The slope of the attempted versus months after surgery. There was a small further aver-
achieved plot was 0.9079, indicating that further no- age hyperopic shift of +0.12 D between 3 months and 1
mogram adjustments would have produced even bet- year (P < .001). This change was deemed to be corneal
ter predictability for higher hyperopia. rather than lenticular as it correlated to the change in

• Vol. 38, No. 12, 2022 765

average corneal keratometry of -0.14 D between 3 and and/or reduced tear meniscus at the slit lamp were seen
12 months (P = .024). in 1.24% of eyes at the end of the follow-up period.
Epithelial ingrowth occurred in 1 eye (0.27%) at the in-
Corneal Higher Order Aberrations cision site in an eye with a 4-mm temporal incision,
Table C (available in the online version of this in which a significant epithelial defect occurred at the
article) shows the Atlas corneal higher order aberra- time of the primary surgery. This was managed conclu-
tions, assessed based on the vertex of the topography sively by opening the incision, epithelial removal, and
in a 6-mm zone using the Optical Society of America wound closure by a 10.0 cross-suture, removed 2 weeks
notation, for RMS, spherical aberration, and coma later.
before surgery and change between 3 and 12 months
after surgery. The average increases in corneal aberra- DISCUSSION
tions were within expected ranges: higher order RMS This was a prospective multicenter trial of hyperopic
by +0.36 µm, spherical aberration by +0.50 µm, and SMILE with and without astigmatism and achieved re-
coma by 0.32 µm. fractive results comparable to modern hyperopic cor-
neal refractive surgery outcomes. Complications were
Complications minor and manageable, most of which were expected
Intraoperative complications that occurred during from myopic SMILE experience, with the exception of
this study were all previously described and at rates ob- a higher rate of interface haze. As we know from pho-
served for routine myopic SMILE.11 Suction loss lead- torefractive keratectomy,18 haze prevention by using
ing to the abortion of procedure occurred in 5 of 374 topical corticosteroids for longer than the standard of
eyes (1.34%), and incomplete removal of the lenticule 1 week used in myopic SMILE may mitigate this and
in 7 eyes (1.87%), where all were small and peripheral, is the subject of further investigation.
without adversely affecting CDVA. The procedure was This study built on the first hyperopic SMILE stud-
aborted due to the inability to remove the lenticule in ies by Reinstein et al6-8 and Pradhan et al,9 which dem-
one eye (0.27%). The occurrence of an epithelial de- onstrated similar centration accuracy to hyperopic
fect or corneal erosion was reported in 22 eyes (5.88%), LASIK,6 improved aberration control compared to hy-
mostly at the incision site. These were managed by the peropic LASIK,7 and comparable refractive accuracy.8
application of a bandaged contact lens until the defect The Pradhan et al study,9 for considerably higher hy-
was healed, usually within 24 hours. peropia (mean: +5.61 ± 1.21 D), achieved 76% of eyes
The least anticipated postoperative complication within ±1.00 D and 53% within ±0.50 D of intended
was interface haze, seen more frequently than expected SEQ at 12 months. In a population with lower hypero-
for myopic SMILE. Interface haze occurred with a peak pia (mean: +3.49 ± 1.38 D), the current study achieved
of 17% at the 6-month follow-up visit (Table D, avail- comparable accuracy with 93% of eyes within ±1.00 D
able in the online version of this article), of which 66% and 81% within ±0.50 D of intended correction at 12
were grade 0.5 (trace or faint haze seen only by indirect months.
broad tangential illumination), 25% were grade 1 (haze Both the current and previous SMILE studies dem-
of minimal density seen with difficulty with direct and onstrated superiority over the previous hyperopic fem-
diffuse illumination), and 9% were grade 2 (a mild haze tosecond lenticule extraction study,4 despite an even
easily visible with direct focal slit illumination) and as- lower hyperopic population (mean: +1.96 ± 1.04 D);
sociated with reduction in contrast sensitivity. By the 90% of eyes within ±1.00 D and 69% within ±0.50 D
12-month follow-up visit, the overall rate of interface of intended correction. This improvement was partly
haze reported decreased to 6.5%, which is still higher due to the use of nomogram adjustments, based on the
than seen in myopic SMILE.11 A VisuMax CIRCLE re- aggregate data from the previous femtosecond lenticule
treatment procedure (converting the cap to a flap) was extraction4 and SMILE studies.9
performed at 9 months for 1 eye with recalcitrant vi- In the current study, there was a loss of one line of
sually significant grade 1 haze despite topical cortico- CDVA in 11% of eyes and two lines in 1.2% eyes at 12
steroid therapy. The small residual refraction of -0.50 months of 374 eyes. Table E (available in the online ver-
-0.25 × 113 was simultaneously corrected by excimer sion of this article) includes a summary of the cause and
laser ablation (MEL 90; Carl Zeiss Meditec AG) and mi- clinical course for each of the eyes that lost two lines
tomycin C was applied to the underside of the flap and of CDVA. This safety was comparable to the previous
stromal bed, which subsequently resolved all haze. SMILE study where 16% lost one line of CDVA and no
In other expected postoperative complications, dry eyes lost two lines of CDVA in 93 eyes.9 In the lower hy-
eye signs such as corneal punctate epithelial staining peropic femtosecond lenticule extraction study, there

766
was loss of one line in 10% of eyes and no eyes lost two This suction loss rate was slightly higher than for the
lines in 39 eyes.4 Although the incidence of one line majority of myopic SMILE studies, where suction loss
loss of CDVA was similar between studies, the current is usually in the region of 0.50%,23 but this may be ex-
study produced a small but measurable 1.2% loss of pected to be the case given the longer lenticule cutting
two lines of CDVA (95% confidence interval: 0.34% to time (31 seconds). The newly CE-marked VISUMAX
3.14%). For comparison, the 95% confidence interval 800 laser creates the same lenticular cuts in hyperopic
for two lines loss was 0.00% to 3.97% for the SMILE SMILE in a much shorter time, and can be expected to
study9 and 0.00% to 8.97% for the femtosecond lenti- greatly reduce this suction rate to even less than that of
cule extraction study,4 implying that the current study current VisuMax myopic SMILE. The retained lenticule
was no worse. The inclusion criteria for the current rate is usually in the region of 0.18%.11,24
study was preoperative CDVA of 20/25 or better, and Centration represents one of the more unique po-
CDVA was 20/16 or better in 43% of eyes. In compari- tential differences regarding patient selection for hy-
son, the inclusion criteria in the previous SMILE study peropic SMILE versus LASIK because this involves
was preoperative CDVA of 20/40 or better, and CDVA the interplay between the corneal white-to-white hori-
was 20/25 or better in only 35% of eyes. Therefore, the zontal diameter and the angle kappa of the eye. In the
safety statistics must be considered in this context and prior hyperopic SMILE study, the mean optical zone
are probably not directly comparable. decentration was found to be 0.23 mm from the cor-
The most commonly observed postoperative com- neal vertex,6 compared to 0.20 mm for myopic SMILE
plication was interface haze, which peaked in 17% of using an identical docking technique and centration
eyes (1.5% with grade 2) at 6 months and persisted protocol.25 Although ideally the patient interface con-
in 6.5% of eyes (0.31% with grade 2) at 12 months. tact glass applies suction to the corneolimbal junction,
This was similar to the previous femtosecond lenti- smaller eyes may result in suction ports being applied
cule extraction study where trace haze was reported onto the conjunctiva. In the case of a smaller eye and
for 20% of eyes at 9 months.4 On the other hand, in- small angle kappa, this may not be an issue. How-
terface haze was not reported in the previous SMILE ever, even an eye with an acceptable white-to-white
study.9 This difference may be related to the age of pa- horizontal diameter for an M glass (eg, 11.7 mm), in
tients treated because the average age in the previous the presence of a large angle kappa, may still lead to
SMILE study was 27 years, compared to 38 years in significant conjunctival application of suction ports.
the current study. Another contributing factor might In a series of 404 consecutive myopic eyes treated by
be an observer-dependent bias in recording these trace LASIK, in which the M contact glass was used for eyes
levels of haze; a higher rate of transient haze was re- with a white-to-white horizontal diameter of 11.7 mm
ported at the sites in Marburg, Germany, and London, or less, there were 3 (0.74%) suction losses observed
United Kingdom. (Dan Z. Reinstein, MD, MA(Cantab), FRCOphth, per-
It is known that interface haze is related to the de- sonal communication, April 2022). The issue with
gree of interface dissection trauma, which in turn is conjunctival encroachment of suction ports also re-
related to the optimization of SMILE energy setting pa- lates to centration once suction is achieved because
rameters. The potential influence of energy and spot the application of a spherical contact glass to an asym-
spacing was discussed for the previous femtosecond metric aspheric cornea (tilted cornea by angle kappa)
lenticule extraction study4 and it was suggested that can sometimes result in the inability to achieve suc-
wound healing might differ from myopic SMILE.19 En- tion on the corneal vertex of the coaxially fixating eye
ergy setting optimization has been extensively stud- as intended. This means that patient selection criteria
ied11,20-22 and is the subject of future investigation for for hyperopic SMILE will need to be studied based on
hyperopic SMILE. more than just white-to-white horizontal diameter, but
It might be expected for the incidence of suction also including angle kappa, to ensure that suction is
loss in hyperopic SMILE to be higher than for myopic sufficiently close to the visual axis. This problem does
SMILE because of the longer suction time required. For not arise as much with LASIK because a slightly tem-
example, a myopic lenticule with a 6.5-mm optical zone porally decentered flap due to this scenario has little
within an 8-mm cap takes approximately 28 seconds of impact on centration of the excimer laser ablation,
cutting time, whereas a hyperopic lenticule with a 6.3- because this is locked by an eye tracker based on the
mm optical zone and transition of 2 mm takes approxi- coaxially fixating corneal vertex.26
mately 31 seconds of cutting time. In the current study, Most of the participating surgeons (both highly ex-
suction loss occurred in 5 eyes (1.3%), and incomplete perienced and less experienced in myopic SMILE) re-
removal of the lenticule occurred in 7 eyes (1.87%). ported a distinct learning curve to the surgical process

• Vol. 38, No. 12, 2022 767

of hyperopic SMILE. Lenticule removal was reported nization monitoring, that included multiple surgeons
to be initially more surgically challenging. Because and laser devices, located across the world, incorpo-
no set surgical technique was defined in the study rating a wide range of ethnicity. One potential weak-
protocol, we may expect this to be improved had the ness of the study was that there was unequal recruit-
Standard Operating Protocol been applied across all ment between the nine sites; 69% of patients were
sites.11 For example, the use of the longer 9-mm instru- treated at three sites. There was also a range of previ-
ment shaft meant that those using a superior incision ous experience in myopic SMILE between surgeons,
would find it more difficult to separate the lenticule which could have led to centration, cyclotorsion, and
in deep-set eyes without stressing the wound. A su- lenticule extraction variability and thus may negative-
perotemporal incision provided a more ergonomic lo- ly impact visual recovery and quality of vision param-
cation to access the lenticule. Additionally, the clear- eters. It was not possible to statistically evaluate this
ance zone between the edge of the cap and the edge due to the small sample size at some sites.
of the lenticule is smaller (down to 0.2 mm in some This multicenter study revealed that hyperopic
cases) than for myopic treatments to accommodate the SMILE performed in a broader refractive surgical com-
larger transition zone. This created more challenges to munity with refractive outcomes that match the current
locate the lenticule interface and increased the chance state of the art in hyperopic LASIK. SMILE and LASIK
of false plane dissection. Following the Standard Op- each possess inherent advantages and disadvantages.
erating Procedures employing specific instrumenta- The benefits of SMILE over LASIK with reduced aberra-
tion and angulation, as well as partial tunnelling and tion induction30,31 and dry eye symptoms32-34 have been
the “double-edge confirmation” technique, reduces demonstrated for myopic SMILE. The current study
the probability of false plane creation. Finally, imple- adds to the evidence that these benefits carry over to hy-
menting the “cap recovery technique”27 eliminates the peropic SMILE, providing improved treatment options
difficulties encountered when the lenticule interface for the treatment of hyperopic refractive error.
is inadvertently separated first.
The results of early hyperopic corneal subtractive AUTHOR CONTRIBUTIONS
procedures were less than ideal, but modern pub- Study concept and design (DZR, WS, TJA); data
lished LASIK outcomes rival those of phacoemulsifi- collection (DZR, WS, TJA, PS, SG, BC, MB, YW, XZ);
cation and intraocular lens implantation.28,29 To com- analysis and interpretation of data (DZR, WS, TJA, SG,
pare the current study to modern studies in the past 5 MB); writing the manuscript (DZR, WS, TJA); critical
years (excluding presbyopic profiles), a literature re- revision of the manuscript (WS, PS, SG, BC, MB, YW,
view was performed using the search terms “LASIK” XZ); statistical expertise (DZR, TJA)
and “hyperopia” or “hyperopic” in PubMed, as well as
extracting the U.S. Food and Drug Administration pre- REFERENCES
market approval study data for reference. It is impor- 1. Sekundo W, Kunert KS, Blum M. Small incision corneal re-
fractive surgery using the small incision lenticule extraction
tant to note that studies reporting hyperopic LASIK (SMILE) procedure for the correction of myopia and myo-
outcomes for the MEL 80 and MEL 90 lasers are most pic astigmatism: results of a 6 month prospective study. Br
comparable because the ablation profile is similar to J Ophthalmol. 2011;95(3):335-339. https://doi.org/10.1136/
bjo.2009.174284 PMID:20601657
the hyperopic SMILE lenticule profile used in the cur-
2. Shah R, Shah S, Sengupta S. Results of small incision lenti-
rent study. Table F (available in the online version of cule extraction: all-in-one femtosecond laser refractive sur-
this article) summarizes the key outcome measures of gery. J Cataract Refract Surg. 2011;37(1):127-137. https://doi.
the studies that were identified (see reference list with org/10.1016/j.jcrs.2010.07.033 PMID:21183108
the table) The UDVA results in the current study were 3. Blum M, Kunert KS, Voßmerbäumer U, Sekundo W. Femtosec-
ond lenticule extraction (ReLEx) for correction of hyperopia: first
similar to the results of modern hyperopic LASIK results. Graefes Arch Clin Exp Ophthalmol. 2013;251(1):349-355.
treatments. For studies with at least 75% of eyes with https://doi.org/10.1007/s00417-012-2064-y PMID:22695934
CDVA of 20/20 or better, UDVA varied between 67% 4. Sekundo W, Messerschmidt-Roth A, Reinstein DZ, Ar-
and 90% for UDVA of 20/20 or better and 84% and cher TJ, Blum M. Femtosecond lenticule extraction (FLEx)
100% for UDVA of 20/40 or better compared to 68% for spherocylindrical hyperopia using new profiles. J Re-
fract Surg. 2018;34(1):6-10. https://doi.org/10.3928/108159
and 100% in the current study. Predictability of spher- 7X-20171031-01 PMID:29315435
ical equivalent refraction within ±0.50 D varied be- 5. Sekundo W, Reinstein DZ, Blum M. Improved lenticule shape
tween 47% (for high hyperopic populations) and 95% for hyperopic femtosecond lenticule extraction (ReLEx FLEx):
compared to 81% in the current study. a pilot study. Lasers Med Sci. 2016;31(4):659-664. https://doi.
org/10.1007/s10103-016-1902-2 PMID:26868029
The strength of the study was that it was a prospec-
6. Reinstein DZ, Pradhan KR, Carp GI, et al. Small incision lenti-
tive controlled trial, with full clinical research orga-

768
 cule extraction (SMILE) for hyperopia: optical zone centration. spective randomized clinical trial. Am J Ophthalmol. 2017;179:159-
J Refract Surg. 2017;33(3):150-156. https://doi.org/10.3928/108 170. https://doi.org/10.1016/j.ajo.2017.05.005 PMID:28499707
1597X-20161220-01 PMID:28264128 22. Li L, Schallhorn JM, Ma J, Cui T, Wang Y. Energy setting and
7. Reinstein DZ, Pradhan KR, Carp GI, et al. Small incision lenticule visual outcomes in SMILE: a retrospective cohort study. J Re-
extraction (smile) for hyperopia: optical zone diameter and spher- fract Surg. 2018;34(1):11-16. https://doi.org/10.3928/108159
ical aberration induction. J Refract Surg. 2017;33(6):370-376. 7X-20171115-01 PMID:29315436
https://doi.org/10.3928/1081597X-20170331-01 PMID:28586496 23. Reinstein DZ, Archer TJ, Vida RS, Carp GI. Suction stability
8. Reinstein DZ, Pradhan KR, Carp GI, et al. Small incision len- management in small incision lenticule extraction: incidence
ticule extraction for hyperopia: 3-month refractive and visual and outcomes of suction loss in 4000 consecutive procedures.
outcomes. J Refract Surg. 2019;35(1):24-30. https://doi.org/10. Acta Ophthalmol. 2020;98(1):e72-e80. PMID:31448878
3928/1081597X-20181025-01 PMID:30633784 24. Ganesh S, Brar S, Lazaridis A. Management and outcomes of re-
9. Pradhan KR, Reinstein DZ, Carp GI, Archer TJ, Dhungana P. Small tained lenticules and lenticule fragments removal after failed pri-
incision lenticule extraction (smile) for hyperopia: 12-month re- mary SMILE: a case series. J Refract Surg. 2017;33(12):848-853.
fractive and visual outcomes. J Refract Surg. 2019;35(7):442-450. https://doi.org/10.3928/1081597X-20171004-01 PMID:29227514
https://doi.org/10.3928/1081597X-20190529-01 PMID:31298724 25. Reinstein DZ, Gobbe M, Gobbe L, Archer TJ, Carp GI. Optical
10. Reinstein DZ, Yap TE, Carp GI, et al. Reproducibility of zone centration accuracy using corneal fixation-based SMILE
manifest refraction between surgeons and optometrists in a compared to eye tracker-based femtosecond laser-assisted
clinical refractive surgery practice. J Cataract Refract Surg. LASIK for myopia. J Refract Surg. 2015;31(9):586-592. https://
2014;40(3):450-459. https://doi.org/10.1016/j.jcrs.2013.08.053 doi.org/10.3928/1081597X-20150820-03 PMID:26352563
PMID:24581774 26. Reinstein DZ, Gobbe M, Archer TJ. Coaxially sighted corneal
11. Reinstein DZ, Archer TJ, Carp GI. The Surgeon’s Guide to light reflex versus entrance pupil center centration of moder-
Small Incision Lenticule Extraction (SMILE). SLACK Incorpo- ate to high hyperopic corneal ablations in eyes with small and
rated; 2018. large angle kappa. J Refract Surg. 2013;29(8):518-525. https://
doi.org/10.3928/1081597X-20130719-08 PMID:23909778
12. Ganesh S, Brar S, Pawar A. Results of intraoperative manual
cyclotorsion compensation for myopic astigmatism in patients 27. Carp GI, Reinstein DZ, Stuart A, Vida RS, Archer TJ, Finkel JN. Cap
undergoing small incision lenticule extraction (SMILE). J Re- recovery technique and double-edge sign during small-incision
fract Surg. 2017;33(8):506-512. https://doi.org/10.3928/108159 lenticule extraction. J Cataract Refract Surg. 2021;47(9):1191-1195.
7X-20170328-01 PMID:28787514 https://doi.org/10.1097/j.jcrs.0000000000000593 PMID:34468457
13. Karakosta A, Vassilaki M, Plainis S, Elfadl NH, Tsilimbaris M, 28. Reinstein DZ, Carp GI, Archer TJ, et al. LASIK for the correc-
Moschandreas J. Choice of analytic approach for eye-specific out- tion of high hyperopic astigmatism with epithelial thickness
comes: one eye or two? Am J Ophthalmol. 2012;153(3):571-579. monitoring. J Refract Surg. 2017;33(5):314-321. https://doi.org/
e1. https://doi.org/10.1016/j.ajo.2011.08.032 PMID:22078901 10.3928/1081597X-20170111-04 PMID:28486722
14. Bland JM. Cluster randomised trials in the medical literature: 29. Reinstein DZ, Carp GI, Archer TJ, Day AC, Vida RS. Outcomes
two bibliometric surveys. BMC Med Res Methodol. 2004;4(1):21. for hyperopic LASIK with the MEL 90® excimer laser. J Re-
https://doi.org/10.1186/1471-2288-4-21 PMID:15310402 fract Surg. 2018;34(12):799-808. https://doi.org/10.3928/10815
97X-20181019-01 PMID:30540362
15. Murdoch IE, Morris SS, Cousens SN. People and eyes: statistical
approaches in ophthalmology. Br J Ophthalmol. 1998;82(8):971- 30. Ganesh S, Gupta R. Comparison of visual and refractive out-
973. https://doi.org/10.1136/bjo.82.8.971 PMID:9828786 comes following femtosecond laser-assisted LASIK with
SMILE in patients with myopia or myopic astigmatism. J Re-
16. Reinstein DZ, Archer TJ, Randleman JB. JRS standard for re-
fract Surg. 2014;30(9):590-596. https://doi.org/10.3928/108159
porting astigmatism outcomes of refractive surgery. J Refract
7X-20140814-02 PMID:25250415
Surg. 2014;30(10):654-659. https://doi.org/10.3928/108159
7X-20140903-01 PMID:25291747 31. Miraftab M, Hashemi H, Aghamirsalim M, Fayyaz S, Asgari S.
Matched comparison of corneal higher order aberrations in-
17. Alpins N. Astigmatism analysis by the Alpins method. J Cata-
ract Refract Surg. 2001;27(1):31-49. https://doi.org/10.1016/ duced by SMILE to femtosecond assisted LASIK and to PRK
in correcting moderate and high myopia: 3.00mm vs. 6.00mm.
S0886-3350(00)00798-7 PMID:11165856
BMC Ophthalmol. 2021;21(1):216. https://doi.org/10.1186/
18. Baldwin HC, Marshall J. Growth factors in corneal wound heal- s12886-021-01987-3 PMID:33992084
ing following refractive surgery: a review. Acta Ophthalmol
32. Sambhi RS, Sambhi GDS, Mather R, Malvankar-Mehta MS. Dry
Scand. 2002;80(3):238-247. https://doi.org/10.1034/j.1600-
eye after refractive surgery: a meta-analysis. Can J Ophthalmol.
0420.2002.800303.x PMID:12059860
2020;55(2):99-106. https://doi.org/10.1016/j.jcjo.2019.07.005
19. Liu YC, Teo EP, Lwin NC, Yam GH, Mehta JS. Early corneal PMID:31712000
wound healing and inflammatory responses after smile: com-
33. Kobashi H, Kamiya K, Shimizu K. Dry eye after small incision
parison of the effects of different refractive corrections and sur-
lenticule extraction and femtosecond laser-assisted LASIK: me-
gical experiences. J Refract Surg. 2016;32(5):346-353. https://
ta-analysis. Cornea. 2017;36(1):85-91. https://doi.org/10.1097/
doi.org/10.3928/1081597X-20160217-05 PMID:27163621
ICO.0000000000000999 PMID:27560032
20. Donate D, Thaëron R. Lower energy levels improve visual
34. Recchioni A, Sisó-Fuertes I, Hartwig A, et al. Short-term im-
recovery in small incision lenticule extraction (SMILE). J Re-
pact of FS-LASIK and SMILE on dry eye metrics and corneal
fract Surg. 2016;32(9):636-642. https://doi.org/10.3928/108159
nerve morphology. Cornea. 2020;39(7):851-857. https://doi.
7X-20160602-01 PMID:27598734
org/10.1097/ICO.0000000000002312 PMID:32243424
21. Ji YW, Kim M, Kang DSY, et al. Lower laser energy levels lead to
better visual recovery after small-incision lenticule extraction: pro-

• Vol. 38, No. 12, 2022 769

Figure A. Hyperopic lenticule geometry and the cutting order shown by arrows: the lenticule interface is created first (from out-to-in), followed by the lenticule
side cut, the cap interface (from in-to-out), and finally the small incision of 2 to 4 mm.
Figure B. Vector analysis of refractive cylinder displayed as polar plots for target induced astigmatism vector (TIA), surgically induced astigmatism vector
(SIA), difference vector (DV), and correction index (CI).