n small incision refractive lenticule extraction (SMILE), a refractive lenticule is created within the corneal stro-ma and subsequently extracted through a sstro-mall periph-eral incision.1,2 Currently, the procedure allows correction of myopia and myopic astigmatism. In previous studies, the outcome of SMILE has been found to be on par with femto-second laser-assisted LASIK (FS-LASIK) for myopia,3-5 and we recently reported the safety of the treatment to be high.6
In keratorefractive surgery, correction of astigmatism is generally more difficult than correction of simple spherical errors.7-9 To date, no studies have systematically evaluated the outcome of astigmatic corrections with SMILE. However, in a study of the flap-based femtosecond lenticule extraction (a previous evolutionary step of SMILE), a 10% undercorrec-tion of the attempted astigmatic correcundercorrec-tion was found in eyes with mostly low degrees of astigmatism.10
The current study evaluated the outcome of SMILE in patients with low to high degrees of myopic astigmatism.
PATIENTS AND METHODS
Since January 2011, SMILE has been routinely performed at the Department of Ophthalmology, Aarhus University Hospital, Denmark. By March 2013, more than 1,800 myopic eyes had been treated. All eyes with an attempted cylinder correction of 0.75 diopters (D) or more were identified from the electronic patient records. Patients were divided into two groups: a low astigmatic group with attempted cylinder cor-rection from 0.75 to 2.25 D and a high astigmatic group with an attempted cylinder correction of 2.50 D or more. Patients
I
ABSTRACTPURPOSE: To evaluate the outcome after small incision refractive lenticule extraction (SMILE) in patients with myopic astigmatism.
METHODS: Seven hundred seventy-five eyes from 403 patients with myopia treated with SMILE for a cylinder of 0.75 diopters (D) or more were identified from pa-tient records. Six hundred sixty-nine eyes were defined as receiving low (< 2.5 D) and 106 eyes as receiving high (>2.5 D) astigmatic correction. Patients were ex-amined before and 3 months after surgery. SMILE was performed with a Visumax femtosecond laser (Carl Zeiss Meditec, Jena, Germany). Preoperative and postopera-tive refractions were converted to polar values. Induced torsion and achieved corrections of sphere and cylinder were determined.
RESULTS: In low astigmatism, the mean preoperative spherical equivalent (SE) was -7.57 ±1.67 D and the cylinder was -1.22 ± 0.49 D. Three months after sur-gery, SE was -0.19 ± 0.48 D from target, astigmatism was 0.17 ± 0.42 D undercorrected, and a small but significant torsion of the cylinder axis corresponding to 0.05 ± 0.37 D was found. The astigmatic undercorrec-tion measured 13% per diopter of attempted correcundercorrec-tion. In high astigmatism, preoperative SE was -5.91 ± 2.56 D and cylinder was -3.22 ± 0.67 D. After surgery, the average astigmatic undercorrection was 0.59 ± 0.65 D, equivalent to 16% per diopter of attempted correc-tion. No undercorrection in SE occurred and no torsion was found.
CONCLUSIONS: SMILE in myopic astigmatism offers predictable correction of SE, but a small, significant un-dercorrection of the astigmatic error. The undercorrec-tion increases with the attempted treatment. Only very little rotation of the cylinder axis was found.
[J Refract Surg. 2014;30(4):240-247.]
From the Department of Ophthalmology, Aarhus University Hospital, Aarhus, Denmark.
Submitted: December 5, 2013; Accepted: January 2, 2014; Posted online: April 4, 2014
The authors thank Nicolaj Aagaard, Jens Christian Hedegaard, Henrik Sejersen, and Christina Møller for invaluable aid with clinical controls. The authors have no financial or proprietary interest in the materials pre-sented herein.
Correspondence: Anders Ivarsen, MD, PhD, Department of Ophthalmology, Aarhus University Hospital, Noerrebrogade 44, DK-8000 Aarhus C, Denmark. E-mail: ai@dadlnet.dk
doi:10.3928/1081597X-20140320-02
Correction of Myopic Astigmatism With Small
Incision Lenticule Extraction
were examined before and 3 months after surgery. Ex-aminations included manifest refraction and uncor-rected (UDVA) and coruncor-rected distance visual acuity (CDVA). In a few cases, postoperative anterior segment optical coherence tomography (Heidelberg Spectralis; Heidelberg Engineering GmbH, Heidelberg, Germany) was performed to visualize the intrastromal interface. Specially trained optometrists performed all clinical refractions.
Surgery
In all patients, surgery was performed using topi-cal anesthesia with two drops of 0.8% oxybuprocain tetrachloride a few minutes before surgery. SMILE was performed using a 500-kHz Visumax femtosecond la-ser (Carl Zeiss Meditec, Jena, Germany) as previously detailed.3,4 The patient was asked to fixate the blinking target, the eye was centered under the laser, and suc-tion was applied. The posterior surface of the lenticule was cut initially, followed by the anterior surface that was extended toward the periphery to allow surgical manipulation of the lenticule. A 30°- to 60°-incision was created at the 12-o’clock position. The diameter of the lenticule ranged from 6.0 to 7.0 mm with a transi-tion zone of 0.0 or 0.1 mm. The diameter of the cap ranged from 7.3 to 7.7 mm and the cap thickness var-ied from 110 to 130 µm.
All surgeries were performed with one of two dif-ferent laser settings: setting 1 with a laser cut energy index of 25 to 27 (approximately 130 nJ) and a spot spacing of 2.5 to 3.0 µm, or setting 2 with a laser cut energy index of 34 (approximately 170 nJ) and a spot spacing of 4.5 µm. Laser setting 1 was used for sur-gery in 244 low astigmatic and 28 high astigmatic eyes, whereas setting 2 was used in the remaining 425 low and 78 high astigmatic procedures. A blunt spatula was used to break any tissue bridges after the laser treatment and the lenticule was removed with a pair of forceps. Postoperative treatment comprised
chlor-amphenicol (Takeda Pharma, Roskilde, Denmark) and fluorometholone (Flurolon; Allergan, Upplands Väsby, Sweden) four times a day tapered over 2 weeks. AnAlySeS
All preoperative and postoperative spherocylindri-cal refractions were converted into spherispherocylindri-cal equivalent (SE) and two polar values as previously reported.11,12
In brief, to fully characterize the refractive cylinders, we calculated the astigmatic polar values at the main meridian (AKP) and at an oblique meridian 45° coun-terclockwise to the main meridian (AKP+45). Surgically induced astigmatism was determined from the change in AKP from before to after surgery (DAKP), and surgi-cally induced torsion or rotation of the cylinder axis was determined from the change in AKP+45 (DAKP+45), with negative values indicating a clockwise change. The errors in treatment, AKPerror and AKP+45,error, were calculated by subtraction of the achieved change from the intended value.
The average polar values were converted to net cyl-inder format as previously shown and the blur due to the postoperative refraction (error of treatment) was estimated by calculation of the amount of defocus as previously reported.12
Data were tested for normality using the D’Agostino– Pearson test. Statistics were performed as unpaired t tests, and as a bivariate analysis of the combined mean polar values with calculation of two-dimensional con-fidence ellipses and determination of Hotelling’s T2 as previously detailed.11,13
RESULTS
A total of 775 eyes with an attempted astigmatic correction of 0.75 D or more were identified. Of these eyes, 669 were treated for astigmatism below 2.50 D, and the remaining 106 eyes were treated for astigma-tism of 2.50 D or more. The preoperative and postop-erative patient characteristics are given in Table 1. TABLE 1
Patient Demographics and Preoperative and Postoperative Refraction by Group
a Parameter Low Astigmatism (669 Eyes) High Astigmatism (106 Eyes)Age (y) 39 ± 8 (19 to 56) 36 ± 9 (21 to 54)
Gender (M/F) 131/213 23/36
Preoperative sphere (D) -6.96 ± 1.69 (-12.50 to 0) -4.30 ± 2.66 (-9.50 to -0.25) Preoperative cylinder (D) -1.22 ± 0.49 (-2.75 to 0) -3.22 ± 0.67 (-5.75 to -2.50) 3-month postoperative sphere (D) -0.09 ± 0.53 (-3.25 to 2.25) 0.25 ± 0.61 (-1.50 to 2.25) 3-month postoperative cylinder (D) -0.44 ± 0.36 (-1.75 to 0) -0.79 ± 0.58 (-2.75 to 0)
D = diopters
low AStigmAtiSm
Three months after surgery, the intended and achieved spherical equivalent corrections were highly correlated (R2 = 0.93; P < .001) (Figure 1A) with 77% and 95% of the eyes within ±0.50 and ±1.00 D of the intended correction, respectively (Figure 2).
In Table 2, the polar values of the low astigmatic group are detailed. After surgery, the mean reduction in spherical equivalent power was 7.26 ± 1.57 D, and the mean induced surgical flattening at the main me-ridian was 1.04 ± 0.57 D. However, a significant un-dercorrection in both spherical power and astigmatic polar value (AKP) was observed. Furthermore, the observed change in astigmatic power at the oblique
meridian indicated that a small but significant coun-terclockwise torsion of the cylinder axis had occurred (DAKP+45). When examining right and left eyes sepa-rately, the DAKP+45 averaged -0.02 ± 0.35 D in right eyes, and 0.12 ± 0.34 D in left eyes, indicating no tor-sion in right eyes, but a small significant counterclock-wise torsion in left eyes (P < .001).
The intended and achieved astigmatic corrections were significantly correlated (R2 = 0.50, P < .001), but linear regression showed an average undercor-rection of 13% per diopter of attempted corundercor-rection (Figure 3A). As seen in Figures 3B-4A, the error in as-tigmatic correction (AKPerror) showed no correlation to the intended spherical correction (R2 = 0.003 ns) or the intended astigmatic correction (R2 = 0.03; P = < .001).
Figure 5A shows the bivariate analysis of the po-lar values. Confidence ellipses are given for individual samples and for the combined mean, and demonstrate a significant combined error of treatment (Hotelling’s T2 = 108; P < .001). The astigmatic error, given in con-ventional net cylinder format, was 0.17 D at 8° relative to the preoperative stronger meridian, and the mean defocus after surgery was 0.48 ± 0.35 D.
A total of 602 eyes were treated for plano refraction (± 0.25 D), and 487 of these eyes (81%) had a UDVA of 20/25 or better (Figure 6A). By 3 months, 19 eyes had lost two or more lines in CDVA (Figure 6B); however, these patients were offered a late reexamination, and all eyes recovered to within one line of the preopera-tive CDVA.
HigH AStigmAtiSm
In highly astigmatic eyes, the intended and achieved spherical equivalent corrections 3 months after sur-gery were strongly correlated (R2 = 0.97; P < .001)
Figure 2. Error in spherical equivalent refraction given as the percentage of eyes obtaining the specified refraction.
equivalent correction in low astigmatism (A) and high astigmatism (B). Points above and below the dashed lines indi-cate over- and under-correction, respectively. The solid lines indicate the outcome of the lin-ear regression analyses.
Figure 3. (A) Achieved ver-sus intended astigmatic correction for eyes with low (open circles) and high matism. In the high astig-matic eyes, crosses indicate eyes treated with laser set-ting 1 and closed circles indi-cate eyes treated with laser setting 2. Points above and below the dashed lines indi-cate over- and undercorrec-tion, respectively. (B) Error in astigmatic correction at the main meridian (AKPerror) as a function of the intended astigmatic correction for eyes with low (open circles) and high (closed circles) astig-matism.
Figure 4. Error in astigmatic correction at the main merid-ian (AKPerror) as a function of the intended spherical cor-rection for (A) low and (B) high astigmatic eyes.
TABLE 2
Mean ± SD Polar Values (D) in Eyes With High and Low Astigmatism
Parameter Preoperative Postoperative Target Error of treatmentHigh astigmatism SE power -7.57 ± 1.67 -0.31 ± 0.55 -0.12 ± 0.29 -0.19 ± 0.48a AKP 1.22 ± 0.49 0.18 ± 0.40 0.01 ± 0.13 0.17 ± 0.42a AKP+45 0 0.05 ± 0.35 0.00 ± 0.10 0.05 ± 0.37b Low astisgmatism SE power -5.91 ± 2.56 -0.14 ± 0.53 -0.09 ± 0.29 -0.05 ± 0.49 AKP 3.22 ± 0.67 0.61 ± 0.63 0.01 ± 0.16 0.59 ± 0.65a,c AKP+45 0 0.03 ± 0.45 0.00 ± 0.16 0.03 ± 0.47
SD = standard deviation; D = diopters; SE = spherical equivalent; AKP = astigmatic polar value, or the astigmatic power, at the main meridian; AKP+45 = astig-matic power at the reference meridian, 45° counterclockwise from the main meridian.
aSignificant undercorrection (paired t test, P < .001). bSignificant torsion (paired t test, P < .001).
(Figure 1B). A total of 77% of the eyes were within ±0.5 D of the intended correction and 97% were within ±1.0 D (Figure 2).
Table 2 shows the polar values of the high astigmatic eyes. The average reduction in spherical equivalent power was 5.76 ± 2.45 D, and the surgically induced flattening at the main meridian averaged 2.61 ± 0.76 D. A significant undercorrection in AKP was observed. However, in high astigmatic eyes no significant change at the oblique meridian was found (DAKP+45), indicat-ing that no torsion of the cylinder axis had been in-duced. Similarly, there were no significant differences
in DAKP+45 between right and left eyes. Comparing the low and high astigmatic groups, significantly more un-dercorrection was seen after operation for high astig-matism, whereas no difference was observed with re-spect to the error in spherical correction or induced torsion (Table 2).
The intended and achieved astigmatic corrections were weakly correlated (R2 = 0.38, P < .001), with an average undercorrection of 19% per diopter of at-tempted correction (Figure 3A). Of the 20 eyes with an astigmatic undercorrection of more than 1.00 D, sig-nificantly more eyes had been treated with laser setting Figure 5. Bivariate representation of the error of treatment in (A) low and (B) high astigmatic eyes. AKPerror indicates the error of treatment at the main meridian (positive values representing the amount of undercorrection). DAKP+45 indicates the induced astigmatism at a 45° oblique meridian; repre-senting the amount of torsion or rotation of the cylinder axis. Confidence ellipses comprising 95% of the observations are given for individual samples (dashed red) and for the combined mean (solid green); the ellipses being equivalent to the standard variation, respectively the standard error of mean in a univariate analysis. The dimensions of the confidence ellipses indicate the relative inaccuracy of the treatments. The statistical significance may be read directly from the figure, with the combined mean errors being significantly different from zero, as the origin is not included in the confidence ellipses (solid green). All values are given in diopters.
Figure 6. (A) Cumulative per-centage of eyes obtaining the specified uncorrected dis-tance visual acuity in Snellen lines. The figure includes only eyes treated for plano refrac-tion. (B) Percentage of eyes with the specified loss or gain in corrected distance visual acuity 3 months after surgery.
1 (11 of 28 eyes) than with laser setting 2 (9 of 78 eyes) (chi-square test; P < .001), and when considering only eyes treated with laser setting 2, the average undercor-rection in high astigmatic eyes was reduced to 16% per diopter of attempted correction.
In high astigmatic eyes, the error in astigmatic cor-rection (AKPerror) showed only a weak correlation to the intended spherical correction (R2 = 0.05; P = .03) (Figure 4B) or to the intended astigmatic correction (R2 = 0.12; P < .001) (Figure 3B). Considering all eyes (both low and high cylinders), the correlation between AKPerror and the intended astigmatic correction was still weak (R2 = 0.13; P < .001).
The bivariate analysis of the polar values is given in Figure 5B for eyes with high astigmatism. The confi-dence ellipse for the combined means showed a signif-icant combined error of treatment (Hotelling’s T2 = 43;
P < .001). The astigmatic error, given in conventional net cylinder format, was 0.59 D at 1° relative to the preoperative stronger meridian, and the mean defocus after surgery was 0.60 ± 0.36 D.
In the high astigmatic group, 91 eyes were treated for plano refraction (± 0.25 D), and 58 of these eyes (64%) had a UDVA of 20/25 or better (Figure 6A). By 3 months, 1 eye had lost two or more lines of CDVA (Figure 6B).
In Figure 7, an optical coherence tomography image of the changes 3 months after SMILE for high astigma-tism is shown. At the main meridian, an abrupt stro-mal change was observed corresponding to the edge of the removed lenticule. Matching changes in epithelial thickness could be seen in the overlying epithelium (Figure 7A). In contrast, a smooth interface was seen at the 90° orthogonal meridian (Figure 7B).
DISCUSSION
The current study demonstrates that SMILE offers precise and safe correction of low to high degrees of myopic astigmatism. Overall, the correction of spheri-cal equivalent power was very precise, although eyes with low cylinders showed a slight undercorrection in contrast to high astigmatic eyes. With respect to the cylinder correction, a significant undercorrection of astigmatism was observed and the error of treatment increased with higher attempted astigmatic correction. On average, the undercorrection was 13% per diop-ter of attempted cylinder correction in low astigmatic and 16% per diopter in highly astigmatic eyes. In a recent study, we evaluated the outcome of FS-LASIK for high astigmatism in myopic eyes using the Carl Zeiss Meditec MEL-80 excimer laser.9 The patients in this previous study were recruited from the same population as those in the current study and showed
similar preoperative characteristics. After FS-LASIK, we found an error in AKPerror of 0.77 ± 0.62 D, which was close to our current observations of 0.59 ± 0.65 D after SMILE. However, in FS-LASIK, an undertion of 21% per diopter of attempted cylinder correc-tion in highly astigmatic eyes was found, which was somewhat higher than in the current study. In contrast, the U.S. Food and Drug Administration approval stud-ies for the MEL-80 excimer laser14 found a correction ratio (induced/intended astigmatic correction) of up to 1.42 in eyes with low astigmatism, indicating con-siderable overcorrection. Similarly, the U.S. Food and Drug Administration approval of the Wavelight Alle-gretto15 reported a correction ratio of 1.16, again indi-cating overcorrection of the astigmatic correction after LASIK in predominantly low astigmatic eyes. From a clinical point of view, however, slight undercorrection would be preferred to overcorrection because a change in the direction of the cylinder axis probably is poorly accepted by patients.
No other studies on SMILE have performed a rigor-ous evaluation of astigmatic corrections. However, one study evaluated the outcome for up to 6 months after 200-kHz femtosecond lenticule extraction in myopic Figure 7. Anterior segment optical coherence tomography at the lenticule edge 3 months after a 3.5 diopters astigmatic correction with SMILE. (A) Image taken at the main meridian. An abrupt stromal change is observed at the edge of the removed lenticule (arrow). The slightly sloped edge is most likely due to the 0.1-mm transition zone. The overlying epithelium shows matching changes (arrowheads) with thin epithelium at the edge (1) and an increased thickness more centrally (2). (B) Image taken 90° orthogonal to the main meridian. The interface is seen as smooth with no abrupt changes.
eyes with primarily lower degrees of astigmatism. An undercorrection of approximately 10% was reported, which was similar to the observations in this study. Furthermore, the authors found no indication of rota-tion of the cylinder axis. The current study demon-strated a minor induced torsion in left eyes corrected for low astigmatism. However, induced torsion would be expected to be more prominent in highly astig-matic eyes, and to be independent of side because the surgical approach is identical for left and right eyes. Thus, the observed torsion by 3 months is most likely a chance finding. Furthermore, the torsion was small and of little clinical significance, as shown by the net cylinder values of 0.17 D at 8° and 0.59 D at 1° in low and high astigmatic eyes, respectively.
In SMILE, decentration of the lenticule may con-tribute to postoperative astigmatism and coma. In this study, induced higher-order aberrations and decentra-tion of the lenticule were not evaluated. However, in a previous study, the average decentration after SMILE was found to be small.4 In other studies, femtosecond lenticule extraction and FS-LASIK have been demon-strated to induce similar amounts of coma.16 Further-more, astigmatism induced by decentration would most likely influence the main cylinder axis, and in the current study the observed torsion was small, which is why it seems unlikely that decentration had any major impact on the postoperative cylinder. Nev-ertheless, docking of the eye in the SMILE procedure may be more challenging in highly astigmatic eyes, and it would be of interest to evaluate the induction of higher-order aberrations in a future study.
Interestingly, we found the laser settings to influ-ence the outcome after SMILE in highly astigmatic eyes. Thus, the initial laser setting (with low energy and close spot spacing) had significantly more eyes with an undercorrection of at least 1 D than the laser setting with higher energy and larger spot spacing. A similar pattern could not be observed in eyes corrected for low astigmatism. Why laser settings should have an impact on the precision of the astigmatic correction is not obvious because the shape of the lenticule should be identical. However, we recently reported that visual recovery is slower after setting 1 than setting 2.6 Simi-larly, other studies have reported that the laser scan-ning trajectory and energy levels influence the postop-erative outcome.17,18 Whether additional adjustments in the laser parameters can further improve the astig-matic corrections remain to be elucidated.
In this study, patients were only evaluated after 3 months. However, the refraction may change during the first months after surgery; probably more so after astig-matic corrections than after pure spherical corrections.
Regression after refractive surgery including photore-fractive keratectomy and LASIK may occur due to an increase in epithelial thickness after myopic refractive surgery19 or due to wound repair with deposition of new stromal tissue.19,20 In SMILE, the wound repair has been demonstrated to be limited,21 but an increase in central epithelial thickness has been noted.22 Most stud-ies have reported little to no regression after SMILE2,3; yet, epithelial hyperplasia could influence the postop-erative refraction and contribute to the observed astig-matic undercorrection in this study. In contrast to ex-cimer laser treatments, the peripheral transition zone in SMILE is small. Whereas this probably has little consequence in myopia where the lenticule becomes thinner toward the edge, it could be significant in high astigmatic corrections where the lenticule edge in the main meridian may become very steep. Because abrupt changes in corneal thickness are usually mitigated by compensatory epithelial hyperplasia, this could con-tribute to undercorrection or postoperative regression of the obtained refractive change at the specific merid-ian. In optical coherence tomography images, we found a rather prominent stromal change corresponding to the lenticule edge with matching changes in epithe-lial thickness. In previous studies of keratorefractive procedures, changes in epithelial thickness have been demonstrated to occur within the first few weeks after surgery.19 Unfortunately, the temporal changes could not be evaluated in the current study, and it remains to be determined if the changes in epithelial thickness at the main meridian contributes to the observed as-tigmatic undercorrection and may change further with increasing time after surgery.
In this study, a loss of two or more lines in CDVA was found in 2.8% of low astigmatic eyes and 0.9% of high astigmatic eyes 3 months after surgery. In comparison to LASIK this is a relatively high loss in CDVA.23 However, patients with a significant visual loss were seen at an extended follow-up, where all eyes were found to have returned to within one line of their preoperative CDVA. Furthermore, in a recent study we evaluated the complications and safety of the first 1,574 eyes treated with SMILE at our department and found late recovery of CDVA in all 24 eyes with a significant loss by 3 months.6
Patients with low astigmatism generally fared bet-ter with respect to UDVA than patients with high cyl-inders. This was to be expected due to less residual astigmatism in the low astigmatic eyes, although the spherical equivalent refraction was slightly less pre-cise. The observed difference in UDVA between low and high astigmatic eyes matched the average defocus value that was marginally better in low astigmatism.
The current study demonstrates that predictable and safe correction of myopic astigmatism can be per-formed with SMILE. However, a small but significant undercorrection was found, in particular for high de-grees of astigmatism. In contrast, minimal torsion was induced by the treatment. Overall, the results were on par with previous evaluations of FS-LASIK and femto-second lenticule extraction.
AUTHOR CONTRIBUTIONS
Study concept and design (AI, JH); analysis and interpretation of data (AI); writing the manuscript (AI); critical revision of the manu-script (JH); administrative, technical, or material support (JH)
REFERENCES
1. Shah R, Shah S, Sengupta S. Results of small incision lenticule extraction: all-in-one femtosecond laser refractive surgery. J Cataract Refract Surg. 2011;37:127-137.
2. Sekundo W, Kunert KS, Blum M. Small incision corneal re-fractive surgery using the small incision lenticule extraction (SMILE) procedure for the correction of myopia and myopic astigmatism: results of a 6 month prospective study. Br J Oph-thalmol. 2011;95:335-359.
3. Vestergaard A, Ivarsen AR, Asp S, Hjortdal JØ. Small-incision lenticule extraction for moderate to high myopia: predictabil-ity, safety, and patient satisfaction. J Cataract Refract Surg. 2012;38:2003-2010.
4. Hjortdal JØ, Vestergaard AH, Ivarsen A, Ragunathan S, Asp S. Predictors for the outcome of small-incision lenticule extrac-tion for myopia. J Refract Surg. 2012;28:865-871.
5. Zhao J, Yao P, Li M, et al. The morphology of corneal cap and its relation to refractive outcomes in femtosecond laser small inci-sion lenticule extraction (SMILE) with anterior segment optical coherence tomography observation. PLoS One. 2013;8:e70208. 6. Ivarsen A, Asp S, Hjortdal, J. Safety and complications of more
than 1500 small-incision lenticule extraction procedures (pub-lished online ahead of print December 20, 2013). doi: 10.1016/j. ophtha.2013.11.006 Ophthalmology.
7. Abolhassani A, Shojaei A, Baradaran-Rafiee AR, Eslani M, Ela-hi B, Noorizadeh F. Vector analysis of cross cylinder LASIK with the NIDEK EC-5000 excimer laser for high astigmatism. J Refract Surg. 2009;25:1075-1082.
8. Sugar A, Rapuano CJ, Culbertson WW, et al. Laser in situ ker-atomileusis for myopia and astigmatism: safety and efficacy: a report by the American Academy of Ophthalmology. Ophthal-mology. 2002;109:175-187.
9. Ivarsen A, Næser K, Hjortdal J. Laser in situ keratomileusis for high astigmatism in myopic and hyperopic eyes. J Cataract Re-fract Surg. 2013;39:74-80.
10. Kunert KS, Russmann C, Blum M, Sluyterman V L G. Vec-tor analysis of myopic astigmatism corrected by femtosec-ond refractive lenticule extraction. J Cataract Refract Surg. 2013;39:759-769.
11. Næser K, Hjortdal J. Polar value analysis of refractive data. J Cataract Refract Surg. 2001;27:86-94.
12. Næser K. Assessment and statistics of surgically induced astig-matism. Acta Ophthalmol. 2008;86:5-28.
13. Næser K, Hjortdal J. Multivariate analysis of refractive data: mathematics and statistics of spherocylinders. J Cataract Re-fract Surg. 2001;27:129-142.
14. U.S. Food and Drug Administration. Medical Devices. Carl Zeiss Meditec MEL 80 Excimer Laser System - P060004. Available at: http://www.accessdata.fda.gov/cdrh_docs/pdf6/ P060004b.pdf. Accessed October 11, 2013.
15. U.S. Food and Drug Administration. Medical Devices. Wave-Light ALLEGRETTO WAVE™ Excimer Laser System - P020050. Available at: http://www.accessdata.fda.gov/cdrh_docs/pdf2/ P020050b.pdf. Accessed October 11, 2013.
16. Vestergaard A, Ivarsen A, Asp S, Hjortdal JØ. Femtosecond (FS) laser vision correction procedure for moderate to high myopia: a prospective study of ReLEx® flex and comparison with a
ret-rospective study of FS-laser in situ keratomileusis. Acta Oph-thalmol. 2013;91:355-362.
17. Shah R, Shah S. Effect of scanning patterns on the results of femtosecond laser lenticule extraction refractive surgery. J Cat-aract Refract Surg. 2011;37:1636-1647.
18. Riau AK, Ang HP, Lwin NC, Chaurasia SS, Tan DT, Mehta JS. Comparison of four different VisuMax circle patterns for flap creation after small incision lenticule extraction. J Refract Surg. 2013;29:236-244.
19. Ivarsen A, Fledelius W, Hjortdal JØ. Three-year changes in epi-thelial and stromal thickness after PRK or LASIK for high myo-pia. Invest Ophthalmol Vis Sci. 2009;50:2061-2066.
20. Moller-Pedersen T, Cavanagh HD, Petroll WM, Jester JV. Stro-mal wound healing explains refractive instability and haze de-velopment after photorefractive keratectomy: a 1-year confocal microscopic study. Ophthalmology. 2000;107:1235-1245. 21. Riau AK, Angunawela RI, Chaurasia SS, Lee WS, Tan DT,
Meh-ta JS. Early corneal wound healing and inflammatory responses after refractive lenticule extraction (ReLEx). Invest Ophthalmol Vis Sci. 2011;52:6213-6221.
22. Vestergaard AH, Grauslund J, Ivarsen AR, Hjordal JØ. Central corneal sublayer pachymetry and biomechanical properties af-ter refractive femtosecond laser lenticule extraction. J Refract Surg. 2014;30:102-108.
23. Farjo AA, Sugar A, Schallhorn SC, et al. Femtosecond lasers for LASIK flap creation: a report by the American Academy of Ophthalmology. Ophthalmology. 2013;120:e5-e20.
