Changes of corneal topographic measurements and higher-order aberrations after surgery for exotropia

Purpose To analyze changes in corneal topographic measurements and higher-order aberrations (HOAs) after horizontal muscle surgery for exotropia. Design Retrospective, observational study. Methods A total of 131 eyes of 121 patients who had undergone surgery for exotropia were included. The eyes with unilateral lateral rectus (ULR) or bilateral lateral rectus (BLR) recession(s) were assigned to group A, and those with unilateral lateral rectus recession & medial rectus resection (R&R) to group B. Corneal topographic measurements and HOAs were compared between the preoperative and postoperative periods using a Placido-dual Scheimpflug analyzer (Galilei 2TM, Ziemer, Port., Switzerland) for each group. Results In group A, simulated keratometery (sim K) was significantly changed until 3 months postoperatively relative to the preoperative value (postoperative 1 week, p = 0.017; 1 month, p = 0.037; and 3 months, p = 0.023, respectively). All steep K (steep sim K, steep-Kpost, and TCP-steep K) parameters also were significantly changed at postoperative 1 month (p<0.001, p = 0.015, p<0.001, respectively), but not at 3 months. Among the higher-order aberrations, spherical aberration (Z40) and secondary horizontal astigmatism (Z42) at postoperative 1 week had significantly changed from the preoperative values, as had horizontal quadrafoil (Z44) at 1 month. However, in group B, only vertical quadrafoil (Z4-4) showed statistically significant changes, at postoperative 1 and 3 months. None of the other postoperative parameters was significantly different from the corresponding preoperative value. Conclusion Lateral rectus recession induced changes in both corneal topographic measurements and HOAs, whereas R&R did so only in HOAs. These changes might explain some patients' complaints about poor quality of vision.


Results
In group A, simulated keratometery (sim K) was significantly changed until 3 months postoperatively relative to the preoperative value (postoperative 1 week, p = 0.017; 1 month, p = 0.037; and 3 months, p = 0.023, respectively). All steep K (steep sim K, steep-Kpost, and TCP-steep K) parameters also were significantly changed at postoperative 1 month (p<0.001, p = 0.015, p<0.001, respectively), but not at 3 months. Among the higher-order aberrations, spherical aberration (Z 4 0 ) and secondary horizontal astigmatism (Z 4 2 ) at postoperative 1 week had significantly changed from the preoperative values, as had horizontal quadrafoil (Z 4 4 ) at 1 month. However, in group B, only vertical quadrafoil (Z 4 -4 ) showed statistically significant changes, at postoperative 1 and 3 months. None of the other postoperative parameters was significantly different from the corresponding preoperative value. PLOS

Introduction
Surgical correction of exotropia, to restore the normal ocular position by weakening the lateral rectus muscle with/without strengthening the medial rectus muscle through changing the orientation of their action plan, has been the main treatment for deviations [1]. Many studies have shown that changes of refractive error or astigmatic power are transient and not clinically important in most cases [2][3][4]; meanwhile, it has been reported that such changes remain stable long term in some patients [5][6][7]. Also, there have been many studies on topographic changes following strabismus surgery. Recession of a single rectus muscle typically effects a decrease of corneal curvature in the meridian of the recessed muscle, though paradoxically, it occasionally increases the focusing power along this meridian [4,[6][7][8][9][10]. Incidence of increased with-the-rule astigmatism after horizontal muscle recess-resect procedures also has been reported [11].
Recently, many imaging techniques for anterior corneal assessment have been developed. Orbscan TM (Bausch & Lomb, Rochester, NY), which uses Placido disk technology to display conventional corneal topography, was the first commercially viable hybrid system [12]. A hybrid Placido-Scheimpflug device, Pentacam TM (Oculus, Inc., Wetzlar, Germany), became commercially available in 2002 [13]. Lately, the Placido-dual Scheimpflug analyzer (Galilei TM , Ziemer, Port., Switzerland), which uses two Scheimpflug cameras, has been introduced as well [14]. The dual-camera system derives images from both sides, which minimizes the effect of decentration due to eye movements on corneal pachymetry and posterior corneal curvature measurements [15].
In normal eyes, most of the optical irregularity is caused by second-order aberrations, which are called refractive errors. Higher-order aberrations (HOAs), by contrast, are a relatively small component of optical irregularity. However, many authors believe that HOAs might play a significant role in reducing retinal image quality [16,17]. HOAs, unlike lowerorder aberrations such as myopia, hyperopia, and astigmatism, cannot be corrected with spectacles. HOAs usually have been studied for the purposes of refractive surgery [18,19].
Curiously, patients who have undergone strabismus surgery sometimes complain about their quality of vision even though the visual acuity or refractive error was not changed. We hypothesized that such patients might be influenced by postoperative changes of HOAs, as HOAs have been known to affect vision quality. However, there remains only scant information in the literature on post-strabismus-surgery changes in corneal topographic measurements or HOAs. Therefore, the aim of this study was to analyze surgically induced changes of corneal topographic measurements and HOAs using Galilei after surgery for exotropia.

Study population
In this retrospective study, the medical records of 121 patients who had undergone surgery for exotropia between August 2012 and February 2017 with a postoperative follow-up period of 3 months or more were reviewed. Patients were excluded if they had a history of any eye surgery including strabismus surgery and corneal surgery, ocular disease, or neurologic disorder such as Down syndrome or cerebral palsy. Data from any patient who did not cooperate with the testing were excluded. The study was performed in accordance with the tenets of the Declaration of Helsinki, and was reviewed and approved by the Institutional Review Board/Ethics Committee of Hallym University Medical Center with an understanding on exemption from the informed consent for the study of retrospective collection of the clinical data. All data were anonymized before we accessed them.

Ophthalmologic evaluation
All of the patients underwent a complete ophthalmologic examination prior to their exotropia surgery. We noted preoperative characteristics including age at surgery, sex, mean angle of exodeviation at distance and near, best-corrected visual acuity (BCVA), refractive error, and the slit-lamp examination result. Deviation angles were measured by the alternate prism cover test for distance (6 m) and near (33 cm) in all 9 positions of gaze using accommodative targets and the patients' best optical correction. Refractive error was measured by streak retinoscopy after cycloplegia using 1% cyclopentolate hydrochloride (Cyclogyl1, Alcon Lab. Inc., Fort Worth, TX, USA) and 1% tropicamide (Mydriacyl1, Alcon Lab. Inc.). Amblyopia was defined as a between-eye difference of 2 lines or more in visual acuity.

Measurement of corneal topographic measurements and HOAs
We measured the corneal topography and HOAs using the Placido-dual Scheimpflug analyzer (Galilei 2, Ziemer, Port., Switzerland). All of the measurements were performed by three experienced examiners, one examiner having been assigned randomly for each case. The patients were instructed to blink completely just before each measurement.
The HOAs were shown in the wavefront report of Galilei. The wavefront aberration data were analysed with a 6 mm pupil size. The elevation data from the Scheimpflug images were combined to form a 3-dimensional reconstruction of the corneal structure. Internal software automatically converted the corneal elevation profile to corneal wavefront data using Zernicke polynomials. The data analysis included the third-order Zernike components (Z 3 -3 to Z 3 3 ) and fourth-order Zernike components (Z 4 −4 to Z 4 4 ).

Surgery
All of the patients underwent exotropia surgery under general anesthesia by one surgeon (D. G.C.). Unilateral lateral rectus (ULR) recession, bilateral lateral rectus (BLR) recession, or unilateral lateral rectus recession & medial rectus resection (R&R) was performed. The surgical method was selected by the surgeon, who had no preference for BLR recession or R&R. When we performed lateral rectus recession (ULR or BLR), we performed the recession procedure using the scleral suture method, not the hang-back technique. Lateral rectus (LR) recession was performed through a temporal limbal conjunctival incision. After the muscle tendon was dissected from the sclera using curved Stevens tenotomy scissors, the needles were inserted through the sclera, entering the tissue at the mark and emerging slightly lateral and parallel to the limbus. The surgical dosages of each group were based on the largest deviation angles at distance deviation, as indicated in Table 1. Some cases with exotropia of < 25 prism diopters (PD) both at distant and near fixation underwent ULR recession.

Grouping
The eyes with lateral rectus recession in patients who had undergone ULR or BLR recession(s) were assigned to group A, and those with R&R were assigned to group B.

Main outcome measures
The main outcome measures were the postoperative changes of corneal topographic measurements and of HOA in each group. Additionally, the correlation between the recession amount and the postoperative change of HOAs (which showed significant postoperative changes in group A) was investigated.

Statistical analysis
Data were analyzed using SPSS software version 24 (SPSS Inc., Chicago, IL). Continuous variables were expressed as mean ± standard deviations. The Pearson chi-square test and Mann-Whitney U-test were used to compare the preoperative characteristics between the groups. The mixed model was used to compare the pre-and postoperative corneal topographic measurements and HOAs in each group. The correlation between the recession amount and the postoperative change of HOAs was evaluated by Pearson correlation analysis. P values less than 0.05 were considered statistically significant. Table 2 shows the preoperative demographic data on each group. The mean preoperative exodeviations were 20.9 ± 5.6 PD at distance and 21.5 ± 5.3 PD at near in group A, and 29.1 ± 10.0 PD and 31.3 ± 10.1 PD in group B (p <0.001 at both distance and near, Mann-Whitney Utest). The reason for the significantly smaller preoperative angles in Group A is that group A included unilateral lateral muscle recession (one-muscle surgery) for small-angle exotropia as well as bilateral lateral rectus recession (two-muscle surgery), while group B included only the recess-resect procedure (two-muscle surgery).  Postoperative changes of corneal topographic measurement Table 3 shows the results of corneal topographic measurements in group A before and after surgery. The parameter sim K had significantly changed by postoperative 1 week (p = 0.017, mixed model), 1 month (p = 0.037) and 3 months (p = 0.023) compared with the preoperative value. TCP-steep K had significantly changed at postoperative 1 week (p = 0.017). All steep K (steep sim K, steep-Kpost, TCP-steep K) had changed at postoperative 1 month (p<0.001, p = 0.015, p<0.001, respectively), but had reverted to unchanged status at 3 months. However, for each parameter in group B, there was no significant pre-to-postoperative difference in the mean variance (all p>0.05) ( Table 4).  ) at 1 month), none of the values showed a significant association (Fig 1). ) showed a statistically significant change at postoperative 1 and 3 months (p = 0.023, 0.010, respectively). The changes of the other fourthorder aberration parameters did not attain statistical significance.

Discussion
Several reports have noted changes in refraction caused by surgery on extraocular muscles [2,3,[20][21][22]. Marshall reported that 60% of patients undergoing strabismus surgery showed a change in astigmatism [20]. Most changes of this kind have been thought to be related to changes in corneal curvature secondary to the reduction in the tension of the recessed extraocular muscle transmitted via the sclera to the cornea [4,[6][7][8][9][10]. A high incidence of increased with-the-rule astigmatism-up to 2 D of change-after horizontal muscle recession-resection procedures also has been reported [11].
Complete consideration of corneal topographic measurements, including analysis of the anterior, posterior, and total corneal curvature and HOAs, is important for better understanding of visual performance after strabismus surgery. One of the purposes of the current study was to evaluate and characterize the corneal structure by the Placido-dual Scheimpflug analyzer in patients who had undergone horizontal muscle surgery. In our study, there were significant changes in sim K, steep sim K, steep-Kpost, and TCP-steep K compared with the preoperative values. Sim K showed a significant change throughout the 3-month postoperative period. All of the steep K values showed a significant change at postoperative 1 month; however, they had reverted to the preoperative values by postoperative 3 months. We assumed that the mechanism of the temporal change of steep K is similar to that of the decreased tendency to regress to surgically induced astigmatism. Based on the above results, our study suggests that patients who have undergone ULR or BLR recession, relative to those who have undergone R&R, might need to change their glasses prescription later. In Group B, we did not find any topographic changes, not even at postoperative 1 week, indicating that none of the parameters of the corneal structure had, according to the Placido-dual Scheimpflug analyzer, changed after R&R. When the muscle was detached and placed behind its original insertion, the tension decreased. We considered the possibility that the resected muscle might compensate for the decreased tension of the recessed muscle, thus allowing patients to maintain a regular corneal topography after the R&R procedure. The HOAs have recently been proven to affect vision quality [23,24]. Important information on optical vision quality is provided by corneal wavefront profiles. Seo et al. reported that whereas the root mean square (RMS) of HOA was transiently increased after lateral rectus recession surgery, it returned to the baseline level after one month in children [25]. Among the individual Zernike coefficients, secondary astigmatism, quadrafoil, secondary coma, secondary trefoil, and pentafoil showed a tendency similar to that of the RMS of HOA [25]. Although we did not investigate more than fourth-order aberrations, the third-order aberrations were not changed significantly, whereas some of the fourth-order aberrations (spherical aberration (Z 4 0 ), secondary horizontal astigmatism (Z 4 2 ), horizontal quadrafoil (Z 4 4 )) changed temporally after lateral rectus recession. As the presumed cause of increasing HOAs is considered to be decreasing tension on the sclera, we investigated the correlation between the recession amount relative to cornea proximity and postoperative change of HOAs in group A. However, the recession amount did not show significant correlations with any HOA values. Moreover, we found meaningful changes in the vertical quadrafoil (Z 4 -4 ) after R&R. On these bases, we can conclude that some variables of fourth-order aberrations changed after strabismus surgery, which might be one of the reasons why patients experience discomfort despite good visual acuity postoperatively.
There are some limitations to this study. First, as it was a retrospective study, evaluations of the repeatability and reproducibility of the examinations were not performed. Second, although the Placido ring images used by the device can be affected by tear-film irregularities, patients with dry eye were not excluded [26]. However, as a means of avoiding this effect, we made an effort to make patients blink completely immediately prior to each measurement. Third, we did not investigate data for the late postoperative period (more than 3 months). Because we analyzed data only until postoperative 3 months, we could not determine whether sim K would have maintained significant change after that or not. Further prospective clinical studies need to be conducted to clarify the long-term outcomes.
In conclusion, we showed that lateral rectus recession surgery induced partial changes in both corneal topographic measurements and HOAs. R&R induced change only in the vertical quadrafoil. Clinicians need to consider functional visual outcomes as they relate to surgically induced changes of corneal topographic measurements and HOAs.