Article Text

Download PDFPDF

Original article
Small incision corneal refractive surgery using the small incision lenticule extraction (SMILE) procedure for the correction of myopia and myopic astigmatism: results of a 6 month prospective study
  1. Walter Sekundo1,
  2. Kathleen S Kunert2,
  3. Marcus Blum2
  1. 1Department of Ophthalmology, Phillips University of Marburg, Germany
  2. 2Department of Ophthalmology, Helios Hospital Erfurt, Erfurt, Germany
  1. Correspondence to Professor Dr Walter Sekundo, Department of Ophthalmology, Philipps University of Marburg, Robert-Koch-Strasse 4, 35037 Marburg, Germany; sekundo{at}


Aim This 6 month prospective multi-centre study evaluated the feasibility of performing myopic femtosecond lenticule extraction (FLEx) through a small incision using the small incision lenticule extraction (SMILE) procedure.

Design Prospective, non-randomised clinical trial.

Participants Ninety-one eyes of 48 patients with myopia with and without astigmatism completed the final 6 month follow-up. The patients' mean age was 35.3 years. Their preoperative mean spherical equivalent (SE) was −4.75±1.56 D.

Methods A refractive lenticule of intrastromal corneal tissue was cut utilising a prototype of the Carl Zeiss Meditec AG VisuMax femtosecond laser system. Simultaneously two opposite small ‘pocket’ incisions were created by the laser system. Thereafter, the lenticule was manually dissected with a spatula and removed through one of incisions using modified McPherson forceps.

Main outcome measures Uncorrected visual acuity (UCVA) and best spectacle corrected visual acuity (BSCVA) after 6 months, objective and manifest refraction as well as slit-lamp examination, side effects and a questionnaire.

Results Six months postoperatively the mean SE was −0.01 D±0.49 D. Most treated eyes (95.6%) were within ±1.0 D, and 80.2% were within ±0.5 D of intended correction. Of the eyes treated, 83.5% had an UCVA of 1.0 (20/20) or better, 53% remained unchanged, 32.3% gained one line, 3.3% gained two lines of BSCVA, 8.8% lost one line and 1.1% lost ≥2 lines of BSCVA. When answering a standardised questionnaire, 93.3% of patients were satisfied with the results obtained and would undergo the procedure again.

Conclusion SMILE is a promising new flapless minimally invasive refractive procedure to correct myopia.

  • Cornea
  • treatment lasers
  • clinical trial

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Lamellar refractive surgery developed from the concepts and work of Jose Barraquer. In the original keratomileusis procedure for myopia a disc approximately 300 μm thick was dissected from the anterior cornea in a freehand fashion and reshaped using a cryolathe. In the late 1980s Ruiz developed an automated microkeratome that controlled the speed as it passed across the cornea leading to more consistent results. This procedure has become known as automated lamellar keratoplasty (ALK). In the 1990s the combination of a microkeratome and an excimer laser (for the refractive cut) was developed by Pallikaris, Burrato and others, further increasing the predictability of the refractive procedure. This procedure, known as laser assisted in situ keratomileusis (LASIK) has gained wide acceptance wordwide.1 The limits of this procedure have been already proven in long-term 6 and 10 year follow-up studies related to the induction of aberrations and regression.2 3

Femtosecond lenticule extraction (FLEx) is a new procedure that does not require a microkeratome or an excimer laser. In FLEx both the flap and the refractive lenticule are created in a one-step procedure using a femtosecond laser. Meanwhile, the first peer reviewed publications4 5 as well as the first reports by other investigators (R Shah, unpublished results) awoke an interest in this new technology and procedure. Despite the fact that FLEx does not utilise an excimer laser, this procedure requires a flap that is smaller, but similar to the one for the femtosecond laser-assisted laser in situ keratomileusis (Femtolasik) operation. A substantial part of complications, for example dry eyes and disturbance of corneal biomechanics, caused by or after LASIK surgery is believed to be linked to the creation of the flap regardless of the method of flap cutting. FLEx has a potential to work without lifting the flap. Thus, this prospective study was designed to investigate the possibility of a FLEx procedure through a smaller incision avoiding creation of a flap. This modification was named small incision lenticule extraction (SMILE). To the best of our knowledge this is the first prospective study on this topic.

Participants and methods

The inclusion criteria were spherical myopia up to –10 D and myopic astigmatism up to −5 D cyl. Other inclusion criteria were a minimum age of 21 years, best spectacle corrected visual acuity (BSCVA) ≥0.8 (20/25) and no other ocular conditions except myopia. Moreover, the central corneal thickness as measured by AC-Master (Carl Zeiss Meditec AG, Jena, Germany) had to be more than 500 μm and the calculated residual stromal bed after treatment >250 μm. A regular topographic pattern was verified by Atlas topography. Absence of asymmetric corneal thickness was confirmed by a pachymetry map using the Visante-OCT (Carl Zeiss Meditec AG). The follow-up appointments were 1 day, 1 week, 1 month, and 3 and 6 months postoperatively. In the first five patients the fellow eye treatment was carried out not earlier than 1 week after the treatment of the first eye. All remaining patients had a bilateral simultaneous procedure. Patients' average age at the time of surgery was 35.3 years. The mean preoperative spherical equivalent (SE) was –4.75±1.56 D. The mean preoperative myopic astigmatism of –0.78±0.79 (range 0 to −4.5) D cyl was recorded. The mean scotopic pupil diameter was 5.7±1.0 mm and the mean preoperative pachymetry 538.4±22.6 μm.

The main outcome measures were the predictability, efficacy, stability and safety of the refractive results obtained with the novel procedure described below.

The SMILE procedure

SMILE is performed under topical anaesthesia using three drops of preservative-free oxybuprocaine tetrachloride (Conjucain EDO; Bausch & Lomb, Berlin, Germany) applied 2–3 min before surgery. After standard sterile draping and insertion of the aspirating speculum, the patient's eye is positioned under the VisuMax integrated surgical microscope (Carl Zeiss Meditec AG). Afterwards, the table moves to the laser treatment position under an illuminated and curved suction contact glass (so-called treatment pack). While the patient fixates an internal target light for centration, the cornea is partly applanated by moving the table upward towards the curved contact glass. The surgeon observes this motion through the operating microscope and controls the movement with a joystick. Once an appropriate centration has been achieved, the surgeon initiates the automatic suction. The patient continues to observe the blinking target light even when the suction is being applied. The VisuMax femtosecond laser produces ultra-short pulses of light, at a repetition rate of 200 kHz with a typical pulse energy ≤300 nJ, that are focused at a precise depth in the corneal tissue. A plasma state develops with optical breakdown, and a small gas bubble is formed from the vaporisation of tissue. A series of bubbles are created in a spiral fashion with a typical spot distance of 3–5 μm resulting in cleaving of tissue planes.

A schematic drawing of the procedure is shown in figure 1. Four subsequent femtosecond incisions are performed: the posterior surface of the refractive lenticule, the lenticule border and the anterior surface of the refractive lenticule. The refractive lenticule extends centripetally for 0.5 mm. The side cuts for the access to the lenticule are positioned 180° apart and have an extension of 80° of cord length. One-half of the eyes treated had the opening incisions in the 12 and 6 o'clock position (surgeon WS) (figure 2) and the other half at the 3 and 9 o'clock position (surgeon MB). After the suction has been released, the patient is moved towards the observation position under the VisuMax integrated surgical microscope. A thin spatula is inserted through the side cut over the roof of the refractive lenticule dissecting this plane followed by the bottom of the lenticule. The lenticule is subsequently grasped with modified serrated McPherson forceps (Geuder, GmbH, Heidelberg, Germany; design M. Blum) and removed. After the removal of the lenticule the intrastromal space is flushed assuring an adequate flow from one incision to the opposite incision using a standard Lasik irrigating cannula.

Figure 1

A schematic drawing of the small incision lenticule extraction (SMILE) procedure. The VisuMax femtosecond laser system cuts the back of the refractive lenticule followed by its front surface incision. In the third step two vertical opposite incisions 80° arc length are carried out. When the patient has returned to the observation position the final fourth step is performed manually, with the lenticule being dissected through the side cut opening incision and removed manually using forceps.

Figure 2

A 6 month postoperative view of the cornea with two fine scars corresponding to the opening incisions at the 12 and 6 o'clock position. The dilated pupil (for fundus examination) underlines the perfect arcuate shape of the opening incisions.

The intended thickness of the upper arcade of tissue (equivalent of ‘flap thickness’ in standard FLEx) varied between 110 and 130 μm; its diameter was chosen between 7.0 and 7.5 (mean 7.3±0.2) mm, always at least 0.5 mm larger than the diameter of the refractive lenticule. The lenticule diameter varied between 6.0 and 7.0 mm with a mean of 6.5±0.3 mm according to the patient's scotopic pupil. The same equation as for FLEx1 was used to calculate the geometry and the thickness of the refractive lenticule. The postoperative regimen consisted of preservative-free ofloxacine (Floxal EDO; Bausch & Lomb) dexamethasone (Dexa EDO; Bausch & Lomb) and hyaluronic acid lubricating drops (VisLube, Chemedica, Switzerland) four times per day each for 1 week. After this, only lubricating drops were used up to 3 months as needed. Follow-up intervals were 1 day, 1 week, and 1, 3 and 6 months.

The following parameters were obtained at each visit:

  • BSCVA and uncorrected visual acuity (UCVA) using different Early Treatment Diabetic Retinopathy Study (ETDRS) charts at each visit

  • Objective and manifest refraction

  • Corneal topography (Atlas; Carl Zeiss Meditec)

  • Wave front measurements (WASCA; Carl Zeiss Meditec)

  • Pachymetry and pachymetry map (AC Master and Visante-OCT; Carl Zeiss Meditec)

  • Goldmann's applanation tonometry (not on day 1 and week 1)

  • Side effects were recorded

  • Patients had to answer a standardised questionnaire consisting of 16 items, such as glare, night driving problems, dryness, pain, which can be found online.

All data were collected on standardised study spreadsheets and entered into Datagraph 3.5b software (Datagraph, Pieger GmbH, Wendelstein, Germany) for analysis. Further statistical analysis was performed using Excel 2003 (Microsoft Inc.) and WINSTAT for Excel 2005.1 (R. Fitch Software Inc, Chicago, Illinois, USA.) The Wilcoxon signed rank test was used to compare mean outcomes.



As shown in figure 3, 49 eyes (53%) had an unchanged BSCVA, 29 eyes (32.3%) gained one line and three eyes (3.3%) two lines. Eight treated eyes (8.8%) lost one line. One eye (1.1%) lost two lines and another eye (1.1%) lost >2 lines of BSCVA.

Figure 3

Safety, shown as a loss/gain in Snellen lines of the best spectacle corrected visual acuity (BSCVA).

Predictability of spherical equivalent and efficacy

At the 6 month follow-up 80.2% of eyes treated were within ±0.5 D and 95.6% within ±1.0 D of the intended refractive target (figure 4). Six months postoperatively the mean SE was −0.01±0.49 D. A UCVA of 0.5 (20/40) or better was obtained in 97.6% of eyes treated and a UCVA of 1.0 (20/20) or better in 83.5% of eyes treated (figure 5). In the latter group, 28.2% had a UCVA of 1.2 and 10.6% of eyes treated had a UCVA of 1.6 (20/12.5)

Figure 4

Predictability of spherical equivalent at 6 months follow-up. 80.2% of eyes are within ±0.5 D and 95.6% within ±1.0 D.

Figure 5

The distribution of uncorrected visual acuity (UCVA) at 6 month follow-up.


On the first few postoperative days the eyes were nearby plano with typical surface-dependent fluctuation in vision as is usual from the LASIK procedure. The refraction stabilised at the 1-week follow-up interval. No further significant change in SE was observed at 1 (0.05 D), 3 (0.14 D) and 6 (0.10 D) months. (figure 6).

Figure 6

Stability of refractive outcome over the 6-month follow-up period. SEQ, spherical equivalent.

Side effects

No cases of transient sensitivity light syndrome or diffuse lamellar keratitis were observed. In one eye, a paracentral perforation of the upper arcade of corneal tissue by the dissecting spatula occurred due to strong adhesion between the roof of the refractive lenticule and the undersurface of the upper corneal layers. This eye (1.1%) lost three lines of BSCVA at 6 months follow-up, but fortunately recovered to only one lost line at 1 year additional follow-up. There was one case of non-progressive epithelial in-growth at the edge of the superior incision. Three cases of a small tear at the incision edge as well as three cases of a localised epithelial defect near the edge were recorded. In one eye the laser incision opening was incomplete. Thus, the incision was opened to a desirable extent using a diamond knife and the lenticule extracted. There were 10.0% of cases with visually insignificant microstriae and one case of superficial punctate corneal staining (1.1%), but no subjective complains of dry eye syndrome recorded at 1 week follow-up.


On the scale of 0 to 100 (0=very poor; 100=best ever known vision) the mean quality of vision was graded 92.3. All patients reported full independence from spectacle correction. The questionnaire was asked in relation to each individual eye treated. Twenty-five (28.4%) reported a marked improvement of their vision and 60 (68.2%) reported an extreme improvement of their vision. The remaining eyes had ‘some improvement’. None complained of worsening. In general, the question ‘Would you have the surgery again?’ was answered by 93.3% with ‘Yes’, 6.7% of the cases answered ‘Not sure’ and there were no ‘No’ answers. However, when asked specifically, 5.5% had some trouble with night vision and 6.6% with driving a car at night. There were no cases of glare, but seven cases (7.7%) felt dryness requiring lubrication compared with two cases (2.2%) preoperatively. The questionnaire used in this study was derived and modified from the questionnaires used in the MEL 80 FDA approval study and has been validated in previous FLEx studies4 5 Further details are displayed in the appendix.

Higher order aberrations

The mean total induction of higher order aberrations (HOA) was 0.04±0.07 μm. For the spherical aberration Z4, 0 the mean induction was 0.008±0.07 μm, for the coma Z3, +1 −0.04±0.12 μm and for Z3, −1 −0.09±0.13.


In a recent prospective clinical study on 108 eyes with myopia and myopic astigmatism treated by the FLEx, Blum and Sekundo reported 98.1% of eyes treated being within ±1.0 D, and 74.8% of eyes within ±0.5 D of intended correction.5 They also showed that 97.1% of patients were satisfied with the results and would undergo the FLEx procedure again.5 However, they also reported peripheral visually insignificant microstriae in 15.7%, superficial punctate keratitis in 13.9% and subjective dry eye syndrome in 8.3% at 1 week follow-up (our unpublished data). Microstriae, flap displacement and dry eye syndrome are believed to be a result of flap creation regardless of the type of procedure, whether it is microkeratome LASIK, Femto-LASIK or FLEx.6 7 When FLEx became a feasible option for corneal refractive correction it was unclear whether this procedure could be executed through a smaller incision avoiding a flap creation.

This very first study on SMILE answers this question. First, a successful refractive correction using two incisions of 80° arc lengths is not only possible, but also provided 83% of eyes treated with an UCVA of 20/20 or better: a result close to outcomes known from Femto-LASIK.8 However, it was our experience that the surgical manoeuvres in the SMILE procedure are far more challenging than in FLEx, as it is of utmost importance to find the first dissection plane over the roof of the lenticule first. When the spatula undermines the lenticule first getting to its base, the surgery becomes difficult, as the thin lenticule ‘sticks’ to the superior corneal layers and becomes almost invisible. This problem does not exist with FLEx, where—in a similar case—the flap can be lifted and the lenticule stripped off from the back of the flap. Nevertheless, as we succeeded in all cases performed with this prototype of the 200 kHz VisuMax laser, one can imagine that the second generation 500 kHz machine might provide an even better dissection plane, making the removal of the lenticule easier and the incision width smaller. With regard to the microstriae, we observed a marked reduction of microstriae compared with FLEx.5 However, some microstriae can still be observed after myopic SMILE due to the compression effect within the superior corneal layers after the removal of the underlying lenticule. Another noteworthy fact was the reduction of superficial punctate corneal staining and subjective dry eye syndrome (p<0.001, Fishers exact test) compared with FLEx. These findings support the hypothesis that SMILE reduces the amount of dissected nerve fibres.

Another important finding was the fact that a 110 to 130 μm-deep vertical corneal incision does not induce detectable irregular corneal topographic astigmatism in an arcuate keratotomy (AK)-like fashion. We attribute this to a very low depth of cut compared with the true AK.1 The concept of two opposite, paired incisions proved successful, as we could assure an adequate flow when flushing the interface. Nevertheless, in 8.8% of cases intrastromal debris were recorded in our series.

Despite the more challenging dissection technique of SMILE in comparison to FLEx performed with the same prototype of VisuMax laser, the current study showed slightly better refractive results with regard to the SE and the stability of the achieved refractive change. We assume that this finding reflects the higher internal structural stability in flapless SMILE-treated corneas. Further improvement might be possible when special SMILE nomograms are used instead of the original FLEx equation.1 So far no enhancement was required for any of the eyes treated. However, we are aware that a possible current drawback of the SMILE procedure is an inability for in situ enhancement. Currently, SMILE-treated eyes would either require an add-on surface ablation or a surgical opening (with femtosecond laser or manually) of the side cut incisions, thus creating a flap, for enhancement purposes. Moreover, the early work of Ruiz et al9 on femtosecond laser intrastromal correction for a low degree of refractive error might eliminate this theoretical problem in the future.

In summary, this multicentre prospective study showed for the first time that femtosecond lenticule extraction can be achieved through a small incision. This novel flapless approach, named SMILE, achieved refractive correction slightly better than FLEx and very close to the latest generation excimer laser technology, while the structural integrity of the cornea and corneal nerve fibres after SMILE appears to be superior to that of the flap procedures.

These first results should encourage further research in order to obtain optimised nomograms and easier surgical manipulation. We also believe that 500 kHz technology may contribute to a further reduction of the opening side cut incisions.


View Abstract


  • See Editorial, p 301

  • Linked articles 186429

  • Funding This study was supported by Carl Zeiss Meditec, Germany.

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval This study was conducted with the approval of the Ethics Committee of the Chamber of Physician of Thuringia, Germany, and the Ethics Committee of the Chamber of Physicians of Rheinland-Palatine, Germany.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles