WITH LASER-ASSISTED IN SITU KERATOMILEUSIS

Transcription

1 Corneal Changes After Laser in Situ Keratomileusis: Measurement of Corneal Polarization Magnitude and Axis RAYMUND ANGELES, MD, TERESA ABUNTO, MD, CHRISTOPHER BOWD, PHD, LINDA M. ZANGWILL, PHD, DAVID J. SCHANZLIN, MD, AND ROBERT N. WEINREB, MD PURPOSE: Laser in situ keratomileusis (LASIK) involves ablation of the corneal stroma, which may induce a change in birefringence. The purpose of this study was to determine the effect of LASIK on corneal birefringence by measuring corneal polarization magnitude (CPM) and axis (CPA). STUDY DESIGN: Cohort study. METHODS: In this prospective study, we measured the change in CPM and CPA before and after LASIK with a scanning laser polarimeter ([SLP] GDx-VCC; Laser Diagnostic Technologies, San Diego, California). Scans were completed on 23 subjects before and 3 months after LASIK. 14 normal controls were tested twice during the same time interval. Change in CPM, CPA, corneal thickness, and corneal curvature measurements were compared between LASIK and normal subjects. RESULTS: At baseline, the mean (95% confidence interval) values of CPM, CPA, corneal thickness, and corneal curvature measurements of the total population (n 37) were 41.6 nm (36.6, 46.5); 31.5 degrees (25.7, 37.3); m (540.0, 556.7); and 7.6 mm (7.5, 7.7), respectively. There were no significant differences in baseline values between normal and LASIK subjects. The reproducibility, measured as the average standard deviation of CPM and CPA measurements in 30 normal control eyes, was 1.95 nm (1.43, 2.48) and 1.69 degrees (0.92, 2.46), respectively. Mean CPA, corneal thickness, and corneal curvature measurements were significantly different in patients after LASIK (all P <.0001). Mean absolute values of the change in both CPM and CPA were significantly greater in LASIK patients (4.8 Accepted for publication Nov 3, From the Hamilton Glaucoma Center and the Department of Ophthalmology, University of California, San Diego, California. This study was supported in part by National Institutes of Health Grant EY (L.M.Z.). Doctor Weinreb is a consultant who has received research support from Laser Diagnostic Technologies, San Diego, California. Inquiries to Robert N. Weinreb, MD, Hamilton Glaucoma Center, University of California, San Diego, 9500 Gilman Dr, La Jolla, CA ; fax: (858) ; [email protected] nm [3.3, 6.4], and 10.4 degrees [6.8, 14.1], respectively) than in normal subjects (2.43 nm [1.53, 3.33], and 1.64 degrees [1.15, 2.14], respectively; both P <.05). The absolute value of change in CPA was linearly associated with the absolute value of change in both corneal thickness (R ) and corneal curvature (R ). CONCLUSIONS: LASIK causes a measurable change in corneal birefringence as measured by the CPM and CPA that may be related to loss of corneal tissue. Comparison of SLP measurements before and after LASIK requires eye-specific compensation to adjust for the change in corneal birefringence. (Am J Ophthalmol 2004;137: by Elsevier Inc. All rights reserved.) WITH LASER-ASSISTED IN SITU KERATOMILEUSIS (LASIK), corneal stromal photoablation results in changes in the thickness and curvature of the cornea. The effect of these corneal architectural changes on corneal birefringence have been previously hypothesized but not demonstrated. 1 5 When a light beam is perpendicular to the corneal surface, the cornea can be described as a linear retarder characterized by two orthogonal linear polarizations such that light with one polarization propagates through the material slower than the other, and thus is retarded in phase. The extent of this retardation is referred to as the corneal polarization magnitude (CPM). The orientation of the slower polarized light defines the slow axis of corneal birefringence, or the corneal polarization axis (CPA). 6 The propagation of polarized light through the cornea is affected by the orientations of the corneal lamellae and by the refractive imbalance between the collagen fibrils and the ground substance. 7 Thus, ultrastructural changes in the cornea brought about by LASIK could potentially affect corneal birefringence. 7 Furthermore, subsequent changes in corneal curvature and thickness after LASIK could also affect the birefringent properties of the cornea. 8 This study was designed to ascertain whether LASIK affects CPA and CPM /04/$ BY ELSEVIER INC. ALL RIGHTS RESERVED. 697 doi: /j.ajo

2 METHODS SUBJECTS: Twenty-three consecutive patients undergoing myopic LASIK by a single surgeon (D.J.S.), who consented to participate in this study between June and December 2002, were prospectively enrolled. Fourteen normal subjects, not undergoing LASIK, who consented to participate within the same time period, were enrolled as controls. One eye was randomly selected from all subjects for inclusion in the study. Informed consent, approved by the University of California San Diego Institutional Review Board, was obtained from each participant. All eyes underwent refraction, slit-lamp biomicroscopy, intraocular pressure (IOP) measurement, and dilated fundus examination with an indirect ophthalmoscope using a 20 diopter lens and slit-lamp biomicroscopy using a 78 diopter lens before enrollment as part of the preoperative evaluation for LASIK surgery. Patients having myopic LASIK who were at least 18 years of age were included in the study. Average (95% confidence interval [CI]) sphere and cylinder before LASIK were 4.4 ( 3.5, 5.2) diopters and 1.1 (0.7, 1.5) diopters, respectively. Exclusion criteria were glaucomatous appearing optic discs, intraocular pressure greater than 21 mm Hg, and any corneal or retinal pathology. Glaucomatous-appearing optic discs were defined as having one or a combination of the following findings: disk hemorrhages, pallor, localized notching, rim thinning or excavation, cup:disk ratio asymmetry of greater than 0.2. INSTRUMENTS: The CPM and CPA were measured using a commercial scanning laser polarimeter, the GDx- VCC (Laser Diagnostic Technologies, San Diego, California). Using polarized light, this technology is designed primarily to measure the birefringence of the retinal nerve fiber layer and determine its thickness. This device measures the CPM and CPA of each individual then uses these values to extract retinal nerve fiber layer retardance from the total retardation. Since the anterior segment structures, mainly the cornea and to a lesser extent the lens, also exhibit birefringence; this device has been modified to include a variable corneal compensator which also measures eye-specific corneal birefringence. A detailed description on how CPM and CPA are determined from this data and how the compensator adjusts to minimize corneal birefringence is described elsewhere. 9 In brief, uncompensated macular polarimetry images first are obtained. The macula, because of the radial arrangement of the Henle fibers, exhibits fairly uniform birefringence. The resulting retardation profile reflects the combined retardance of both the cornea and the Henle layer (Figure 1, A). The CPM and CPA are calculated automatically from this signal, and the instrument then adjusts to compensate for them. A fully compensated macular SLP macular image exhibits little birefringence (Figure 1, B). The difference between the total signal and FIGURE 1. (A) Uncompensated scanning laser polarimeter macular image reflecting the combined retardance of the cornea and Henle fiber layer. Red and yellow color indicate high retardance values, blue indicates low retardance. (B) Compensated macular image. Note the uniform blue color. the corneal signal then represents the retinal nerve fiber layer retardation. STUDY DESIGN THE STUDY DESIGN FOLLOWS THE PRINCIPLES OF A COHORT study. Measurements were done on two occasions using the GDx-VCC. For patients, scans were completed on all eyes before and approximately 12 weeks after LASIK. For the control group, the two scanning sessions were obtained at approximately the same interval as with the study group. Multiple scans were taken, and the best quality scan based 698 AMERICAN JOURNAL OF OPHTHALMOLOGY APRIL 2004

3 TABLE 1. Characteristics of Study Population and Normal Controls LASIK Patients n 23 Normal Controls n 14 P Value* Male:female 12:11 3: Age (mean, 95% CI), years 45 ( ) 42.8 ( ).407 Range Number of weeks between 15.1 ( ) 17.2 ( ).376 examinations (mean, 95% CI) CI confidence interval; LASIK laser-assisted in situ keratomileusis. *Student t test. on maximal instrument-provided image quality score and minimum residual retardance measurement was used to report the CPM and CPA. Experienced operators performed all scans (T.A.A., and a nonauthor technician). For scans to be included, software-provided quality scores of 8 or more (of 10) and residual macular retardance measurements of12.5 nm or less were required. No participants were excluded because of unacceptable image quality. During the second scanning session, the device was reset to remeasure the CPM and CPA. Other parameters that were measured and recorded during both visits were the following: refraction, corneal thickness using ultrasound pachymetry (Pachette DGH 500; DGH Technology, Philadelphia, PA), corneal curvature using keratometry (Humphrey Automatic Refractor 597; Humphrey Systems, Dublin, California), and applanation intraocular pressure (Haag Streit International, Bern, Switzerland). To determine intrasubject reproducibility of CPM and CPA measurements, we obtained four macular scans from 30 normal control subjects, defined as those with normal eye examinations and normal visual field results using full-threshold or Swedish Interactive Threshold Strategy standard automated perimetry (Humphrey Field Analyzer; Carl Zeiss Meditiec, Dublin, California). We then reported the average standard deviation of CPM and CPA measurements for all of these subjects combined. SURGERY: Laser-assisted in situ keratomileusis was performed with the LADARVision 4000 excimer laser (Alcon Laboratories, Orlando, Florida) and with the VISX Star 3 Laser (VISX, Santa Clara, California). The Hansatome microkeratome was used to create the flap after the intraocular pressure was increased to a minimum of 65 mm Hg as measured with a Barraquer tonometer. STATISTICAL ANALYSIS Statistical analysis was performed using JMP software (SAS Institute, Cary, North Carolina). We used unpaired Student t tests and paired t tests to compare the baseline with the post- LASIK measurements of CPA and CPM, corneal thickness, and curvature. Linear regression analysis also was used to analyze the relationship between change in corneal structure (thickness and curvature) and change in CPM and CPA. A P value of.05 or less was considered statistically significant. RESULTS WE STUDIED 23 EYES OF 23 CONSECUTIVE PATIENTS (12 males, 11 females) who underwent LASIK and compared them with 14 eyes of 14 normal controls (three males, 11 females). The mean (95% CI) age for LASIK patients and normal controls was 45.0 years (41.8, 48.3; range 27 to 64) and 42.8 years (37.9, 47.7; range 26 to 54), respectively. The mean (95% CI) length of time between examinations for LASIK patients was 15.1 (12.4, 17.1) weeks; for normal subjects it was 17.2 (12.4, 22.0) weeks. There were no significant differences in age and length of time between examinations between LASIK patients and controls (Student t test P.4, P.3, respectively; Table 1). In addition, there were no significant differences (all P.05) in mean (95% CI) residual retardance measurements between patients and controls either at baseline (5.9 nm [4.7, 7.1 nm], and 4.4 nm [3.4, 5.3 nm], respectively) and follow-up testing (5.5 nm [4.5, 6.4 nm], and 5.5 nm [3.8, 7.1 nm], respectively). The reproducibility, measured as the average standard deviation (95% CI), of CPM and CPA measurements in 30 normal control eyes (see Methods) was 1.95 nm (1.43, 2.48) and 1.69 degrees (0.92, 2.46), respectively. These values resulted in coefficients of variations of 0.05 for CPM and 0.07 for CPA. At baseline, the mean (95% CI) values of CPM, CPA, corneal thickness, and corneal curvature of the total population (n 37) were 41.6 nm (36.6, 46.5); 31.5 degrees (25.7, 37.3); m (540.0, 556.7), and 7.6 mm (7.5, 7.7), respectively. The mean baseline CPM, CPA, central corneal thickness, and corneal curvature values were similar in the LASIK patients and normal controls (Student t test, all P.2). Comparing baseline with VOL. 137, NO. 4 CORNEAL CHANGES AFTER LASER IN SITU KERATOMILEUSIS 699

4 TABLE 2. Comparison Between Baseline and Follow-Up Means in Both LASIK Patients and Normal Subjects LASIK Subjects (n 23) Normal Controls (n 14) Follow-Up Mean (95% CI) P Value* Baseline Mean (95% CI) Follow-Up Mean (95% CI) P Value* Variable Baseline Mean (95% CI) CPM (nm) 43.1 ( ) 42.8 ( ) ( ) 39.6 ( ).474 CPA (degrees) 31.0 ( ) 21.0 ( ) ( ) 31.4 ( ).064 Corneal curvature (mm) 7.6 ( ) 3.2 ( ) ( ) 7.6 ( ).072 Corneal thickness ( m) ( ) ( ) ( ) ( ).019 CI confidence interval; CPA corneal polarization axis; CPM corneal polarization magnitude; LASIK laser-assisted in situ keratomileusis. *Paired t test. follow-up measurements in LASIK patients, statistically significant differences were found in mean CPA, corneal thickness and corneal curvature (paired t tests, all P.0001), but not in CPM (Table 2). In normal patients, there was a statistically significant change only in corneal thickness measurements (P.02). Because CPM and CPA may not change in a predictable direction after LASIK, we determined the absolute value of the difference between baseline and follow-up for both parameters. The mean (95% CI) absolute value of change in CPA was significantly greater in the LASIK group (10.4 degrees [6.8, 14.1]) than in the control group (1.6 degrees [1.2, 2.1]; P.001), and the absolute value of change in CPM was marginally greater in the LASIK group (4.8 nm [3.3 nm, 6.4 nm]), than in the control group (2.4 nm [1.5 nm, 3.3 nm]; P.05). Regression analysis, including all eyes, showed a linear correlation between absolute value of change in corneal thickness and in the CPA (R , P.0001), and between the absolute value of change in corneal curvature and the in the CPA (R , P.0001). The absolute value of change in CPM was not linearly associated with change in corneal curvature, corneal thickness or CPA (all R 2 P.07; Figure 2). DISCUSSION THE ADVENT OF SCANNING LASER POLARIMETRY HAS INcreased interest in measuring corneal birefringence Scanning laser polarimetry directly measures retardation, a measure that correlates well with the retinal nerve fiber layer thickness. 12 An early version of the SLP assumed fixed values for CPA (nasally downward) and CPM to compensate for corneal birefringence. Greenfield and associates, 13 however, in a study of 113 eyes showed that the distribution of CPA varies over a wide range. In a study of 73 eyes, Knighton and associates 10 confirmed the individual variation of CPA, and also showed that CPM varied among individuals. These observations led to the development by Zhou and Weinreb of a method to individually compensate for corneal birefringence 14 which confirmed the individual variability of both CPM and CPA. Their method was subsequently incorporated into a commercial SLP system (GDx-VCC) that measures the CPA and CPM of each individual, and then uses these values to extract retinal nerve fiber layer retardance from the total retardation. Our baseline results (pre-lasik and normal eyes combined) for CPM and CPA ( nm and degrees, respectively) were similar to those previously reported by Weinreb and associates 11 using a prototype of the GDx-VCC. They reported a mean ( SD) CPM of 41.1 nm ( 14.3 nm) and a CPA of 28.8 degrees ( 15.4 degrees) for healthy eyes. The results of the current study are in agreement with previous ones in finding a large interindividual variability for both the CPM and CPA. 9,11 In addition, this study evaluated the intrasubject reproducibility of CPA and CPM measurements obtained with the commercial instrument. In general, the reproducibility was good with coefficients of variation less than 0.07 for four scans obtained on the same day. In the current study, a significant difference was noted in the CPA but not the CPM when comparing the mean values between baseline and follow up tests in post-lasik patients. As the change in CPM and CPA may be positive or negative, measuring the means of these parameters in a given population may underestimate the magnitude of change. Measurements of the mean absolute value of the change with time were significantly different between the LASIK patients and the normal controls for all parameters, including CPM. However, the change in CPM was twofold, while that of the CPA was sixfold in post-lasik eyes compared with normal eyes. The change in corneal thickness measurements in normal controls was small, approximately 3.5 m, compared with 48 m in the LASIK group, and may be due to measurement variability or chance. Another possibility is that this small change may be due to reported diurnal changes in corneal thickness AMERICAN JOURNAL OF OPHTHALMOLOGY APRIL 2004

5 FIGURE 2. (A) Correlation between absolute value of change in corneal polarization axis (CPA) and change in corneal thickness (n 37). (B) Correlation between absolute value of change in CPA and change in corneal curvature (n 37). (C) Correlation between absolute value of change in corneal polarization magnitude (CPM) and change in corneal thickness (n 37). (D) Correlation between absolute value of change in CPM and change in corneal curvature (n 37). (E) Correlation between absolute value of change in CPM and absolute value of change in CPA (n 37). In the current study we did not standardize the time of day at which measurements were made. The effect of LASIK on SLP measurements has been investigated in previous studies, 1 3,5,16 18 but its direct effects on CPM and CPA have not been reported. In this study, we observed that there was a significant difference, compared with baseline, in CPA after LASIK. In addition, there was a significant change in absolute value of both CPM and CPA after LASIK. This agrees with our hypothesis that structural changes in the cornea brought about by LASIK may lead to changes in the corneal birefringence. Linear regression analysis confirmed that the change in CPA was correlated with changes in the corneal thickness and curvature. However, the change in CPM does not appear to be influenced by the change in corneal thickness and curvature. Changes in CPM and CPA did not appear to be correlated. The morphologic changes that occur in the cornea after LASIK have been described previously. 7,19 Vesaluoma and coworkers 19 in a study of 62 eyes using confocal microscopy, showed microfolds in the Bowman layer and the anterior stroma in 96.8% of eyes after LASIK. The clinical significance of slight microfolding appears negligible. However, deeper and more extensive folding might affect the topography of the corneal surface resulting in irregular astigmatism. Whether these changes affect corneal birefringence can only be speculative. Vesaluoma and colleaugues 19 also observed particles of variable size and reflectivity, most probably composed of metallic and cellular debris and inflammatory cells at the flap interface. Their presence coincides with the need for a higher intensity of illumination during scanning laser polarimetry, as noted by Hollo and associates 2 and raises the possibility that the debris significantly increases the reflection of polarized light. This observation was observed during the early postoperative period although it diminishes with time and reaches a stable level at 3 months. The presence of persistent interface particles cannot be discounted. Our follow-up period of approximately 12 weeks post-lasik was designed to eliminate the VOL. 137, NO. 4 CORNEAL CHANGES AFTER LASER IN SITU KERATOMILEUSIS 701

6 FIGURE 3. Case 1. (A) Baseline retinal nerve fiber layer scan: corneal polarization magnitude (CPM) 52, corneal polarization axis (CPA) 55.1 (9/1/2002). Note the hourglass pattern denoting thicker retinal nerve fiber layer in the superior and inferior quadrants (yellow/red color) (B) Post- LASIK (laser-assisted in situ keratomileusis) scan without resetting of corneal compensation. Note reversal of hourglass pattern (12/13/2002) (C) Post-LASIK scan after resetting corneal compensation: CPM 46, CPA 43.3 (12/13/2002). short-term effects of corneal wound healing on SLP measurements. We hypothesize that the loss and disruption of the spatial order of collagen fibrils after photoablation causes the change in the birefringent properties of the cornea. Retardation from stacked layers of fibrils composing the corneal stroma, each at a different angle relative to the others and acting as a single linear retarder, is combined to represent the overall retardation of all of the layers. 20,21 Loss of stromal layers would thus change the net orientation of the fibrils and lead to a change in the corneal birefringence. Because all of our patients received treatment to correct myopia, the effect of corneal steepening (for correction of hyperopia) by LASIK on CPM and CPA was not investigated and remains untested. Other limitations of the current study include the limited sample size and the possibility of interoperator differences in image acquisition. However, the sample size was sufficient to detect significant change in CPM and CPA as a result of LASIK. The change in CPM and CPA after LASIK, and possibly after other surgical procedures involving the cornea, could have an impact in the use of imaging technologies dependent on birefringence such as SLP. The accuracy of the results can be affected if erroneous corneal values are used to compensate for corneal birefringence. As an example, we show the scan results of one patient (Figure 3). Compared with baseline (Figure 3, A), even a small change in the CPM and CPA may lead to a different pattern of retinal nerve fiber layer retardance (Figure 3, B) because of inadequate compensation. These measurements would incorrectly estimate the retinal nerve fiber layer thickness. After remeasuring the CPA and CPM, however, the resulting retardance map (Figure 3, C) is almost identical to baseline. Therefore it is imperative that the CPM and CPA be remeasured after LASIK for the accurate measurement of retinal nerve fiber layer thickness using this device. In conclusion, we were able to demonstrate that there was a measurable change in CPM and CPA after LASIK. The change in CPA is linearly correlated with the change in corneal curvature and thickness. Whether these changes can significantly affect the discriminating power of SLP in post-lasik eyes remains to be investigated. Certainly these changes will affect the ability to detect change in the RNFL when comparing pre and post-lasik SLP images. Therefore, individual eye-specific corneal compensation is mandatory for detecting retinal nerve fiber layer change with SLP in these circumstances. REFERENCES 1. Gurses-Ozden R, Liebmann JM, Schuffner D, Buxton DF, Soloway BD, Ritch R. Retinal nerve fiber layer thickness remains unchanged following laser-assisted in situ keratomileusis. Am J Ophthalmol 2001;132: Hollo G, Nagy ZZ, Vargha P, Suveges I. Influence of post-lasik corneal healing on scanning laser polarimetric measurement of the retinal nerve fibre layer thickness. Br J Ophthalmol 2002;86: Hollo G, Katsanos A, Kothy P, Kerek A, Suveges I. Influence of LASIK on scanning laser polarimetric measurement of the retinal nerve fibre layer with fixed angle and customised corneal polarisation compensation. Br J Ophthalmol 2003; 87: AMERICAN JOURNAL OF OPHTHALMOLOGY APRIL 2004

7 4. Nevyas JY, Nevyas HJ, Nevyas-Wallace A. Change in retinal nerve fiber layer thickness after laser in situ keratomileusis. J Cataract Refract Surg 2002;28: Roberts TV, Lawless MA, Rogers CM, Sutton GL, Domniz Y. The effect of laser-assisted in situ keratomileusis on retinal nerve fiber layer measurements obtained with scanning laser polarimetry. J Glaucoma 2002;11: Farrell RA, Wharam JF, Kim D, McCally RL. Polarized light propagation in corneal lamellae. J Refract Surg 1999;15: Rumelt S, Cohen I, Skandarani P, Delarea Y, Ben Shaul Y, Rehany U. Ultrastructure of the lamellar corneal wound after laser in situ keratomileusis in human eye. J Cataract Refract Surg 2001;27: Bueno JM, Vargas-Martin F. Measurements of the corneal birefringence with a liquid-crystal imaging polariscope. Appl Opt 2002;41: Greenfield DS, Knighton RW, Feuer WJ, Schiffman JC, Zangwill L, Weinreb RN. Correction for corneal polarization axis improves the discriminating power of scanning laser polarimetry. Am J Ophthalmol 2002;134: Knighton RW, Huang XR. Linear birefringence of the central human cornea. Invest Ophthalmol Vis Sci 2002;43: Weinreb RN, Bowd C, Greenfield DS, Zangwill LM. Measurement of the magnitude and axis of corneal polarization with scanning laser polarimetry. Arch Ophthalmol 2002; 120: Weinreb RN. Evaluating the retinal nerve fiber layer in glaucoma with scanning laser polarimetry. Arch Ophthalmol 1999;117: Greenfield DS, Knighton RW, Huang XR. Effect of corneal polarization axis on assessment of retinal nerve fiber layer thickness by scanning laser polarimetry. Am J Ophthalmol 2000;129: Zhou Q, Weinreb RN. Individualized compensation of anterior segment birefringence during scanning laser polarimetry. Invest Ophthalmol Vis Sci 2002;43: du Toit R, Vega JA, Fonn D, Simpson T. Diurnal variation of corneal sensitivity and thickness. Cornea 2003;22: Gurses-Ozden R, Pons ME, Barbieri C, et al. Scanning laser polarimetry measurements after laser-assisted in situ keratomileusis. Am J Ophthalmol 2000;129: Kook MS, Lee S, Tchah H, Sung K, Park R, Kim K. Effect of laser in situ keratomileusis on retinal nerve fiber layer thickness measurements by scanning laser polarimetry. J Cataract Refract Surg 2002;28: Tsai YY, Lin JM. Effect of laser-assisted in situ keratomileusis on the retinal nerve fiber layer. Retina 2000;20: Vesaluoma M, Perez-Santonja J, Petroll WM, Linna T, Alio J, Tervo T. Corneal stromal changes induced by myopic LASIK. Invest Ophthalmol Vis Sci 2000;41: Radner W, Zehetmayer M, Aufreiter R, Mallinger R. Interlacing and cross-angle distribution of collagen lamellae in the human cornea. Cornea 1998;17: Radner W, Mallinger R. Interlacing of collagen lamellae in the midstroma of the human cornea. Cornea 2002;21: VOL. 137, NO. 4 CORNEAL CHANGES AFTER LASER IN SITU KERATOMILEUSIS 703