- Email: [email protected]
Outcome of corneal and laser astigmatic axis alignment in photoastigmatic refractive keratectomy1
articles
Outcome of corneal and laser astigmatic axis alignment in photoastigmatic refractive keratectomy Samir G. Farah, MD, Eric Olafsson, David G. Gwynn, MD, Dimitri T. Azar, MD, Frederick S. Brightbill, MD ABSTRACT Purpose: To compare the refractive results of laser astigmatic treatment in eyes in which the astigmatic axes of the eye and laser are aligned by limbal marking at the 6 o’clock position and in eyes that are not marked. Setting: University Hospital and Clinics, Madison, Wisconsin, USA. Methods: This retrospective study comprised 143 eyes that had photoastigmatic refractive keratectomy with the VISX Star excimer laser. The eyes were divided into marked (G1) and unmarked (G2) groups. Based on the preoperative astigmatism, each group was subdivided into low astigmatism (ⱕ1.00 diopter [D]) and high astigmatism (ⱖ1.25 D). Early postoperative manifest refractions (1.0 to 2.5 months) were analyzed. The Alpins vector analysis method was used to calculate the target induced astigmatism, surgically induced astigmatism, difference vector (DV), magnitude of error (ME), angle of error (AE), and index of success (IS). Results: There was no significant difference between the groups in DV, ME, and IS. When the subgroups were analyzed, the DV and ME were comparable; the IS in the G1 high astigmatism subgroup was significantly better than that in the G2 high astigmatism subgroup (0.22 ⫾ 0.08 and 0.29 ⫾ 0.04, respectively; P ⬍ .0001). There was comparable scatter of AE values; 30% and 36% in G1 and G2, respectively, had an AE of 0. Similar scatter was observed in the subgroups. Of the eyes that had an AE of 0, 90% and 43% in the high astigmatism subgroups of G1 and G2, respectively (P ⬍ .05), had full correction of astigmatism. Conclusion: Limbal marking and subsequent eye and laser astigmatic axis alignment improved the refractive outcome of laser astigmatic treatment of ⱖ1.25 D. A preliminary report of an ongoing prospective randomized study of eyes that had laser in situ keratomileusis is included. J Cataract Refract Surg 2000; 26:1722–1728 © 2000 ASCRS and ESCRS
M
yopic astigmatic laser refractive surgery consists of flattening the steepest meridian of the cornea. In the VISX Star excimer laser system, power is delivered through a contracting-slit apparatus, with
© 2000 ASCRS and ESCRS Published by Elsevier Science Inc.
blade movement along the steepest meridian of the cornea. Astigmatic power and axis are determined while the patient is seated during refraction. However, laser treat0886-3350/00/$–see front matter PII S0886-3350(00)00695-7
CORNEAL AND LASER ASTIGMATIC AXIS ALIGNMENT IN PARK
ment is applied while the patient is supine. Studies1,2 have failed to show statistical differences in cylinder axis measurement in the seated versus the supine position when comparing groups but have documented significant ocular cyclotorsion at the individual level.1–3 If not accounted for, such eye movement would theoretically decrease the refractive outcome of astigmatic surgery.4 Although photoastigmatic refractive keratectomy (PARK) outcomes have been reported to be comparable to those of photorefractive keratectomy (PRK),5– 8 most observers agree that accuracy can be improved.9,10 Even though many refractive surgeons in the United States and abroad place a reference point on the cornea to match the corneal and laser beam astigmatic meridia during treatment, we are unaware of studies that discuss the refractive outcome of astigmatism treatment in marked versus unmarked eyes. In this study, we retrospectively analyzed a series of patients to compare the early results of PARK in a group in which the astigmatic meridia of the cornea and laser were aligned by limbal marking at the 6 o’clock position and in an unmarked group.
Patients and Methods The charts of 1 surgeon’s (F.S.B.) patients who had PARK between May 1997 and May 1998 at the University of Wisconsin Eye Clinics were reviewed. One hundred forty-five patients (211 eyes) were initially identified. The criteria for inclusion in the study were
Accepted for publication August 8, 2000. From the Cornea Service, Department of Ophthalmology and Visual Sciences, University of Wisconsin, Madison, Wisconsin (Farah, Olafsson, Gwynn, Brightbill), and the Cornea Service, Massachusetts Eye and Ear Infirmary, Harvard Medical School, Boston, Massachusetts (Azar), USA. Presented in part at the Symposium on Cataract, IOL and Refractive Surgery, Seattle, Washington, USA, April 1999. Supported by an unrestricted grant from Research to Prevent Blindness, Inc., and the Wisconsin Lion Foundation to the Department of Ophthalmology and Visual Sciences at the University of Wisconsin, Madison. None of the authors has a financial interest in any product mentioned. Reprint requests to Frederick S. Brightbill, MD, 2870 University Avenue, Suite 102, Madison, Wisconsin 53705, USA.
age of at least 21 years, stable refraction on 2 consecutive examinations, best corrected visual acuity of 20/20 or better, preoperative refractive astigmatism ⱕ4.0 diopters (D), absence of connective tissue disease, no history of eye disease or glaucoma, no muscle imbalance, and no haze postoperatively. Forty-one eyes were not included (35 for insufficient follow-up, 5 for presence of corneal haze, and 1 for infectious keratitis). Although a subgroup of patients with preoperative astigmatism of 0.5 D (n ⫽ 27, 18 in G1 and 9 in G2) was initially identified, it was not included in the analysis because patients with less astigmatism may be more tolerant of subtle changes in cylinder power and axis and because if present, most surgeons prefer treating the spherical equivalent. This study comprised 90 patients (143 eyes) who had PARK for myopia and astigmatism. Mean patient age was 44 years ⫾ 13 (SD) (range 25 to 59 years). It was 42 ⫾ 15 years in the marked group and 46 ⫾ 17 years in the unmarked group. Five retreated eyes were included in the study (10% of the total number of eyes). Eyes were divided into marked (G1) and unmarked (G2) groups throughout the study using a table of random variation; G1 comprised 40 right eyes and 36 left eyes and G2, 34 and 33, respectively. Each group was divided into 2 subgroups based on the preoperative astigmatism: low astigmatism (ⱕ1.00 D; G1 ⫽ 40, G2 ⫽ 34) and high astigmatism (ⱖ1.25 D; G1 ⫽ 36, G2 ⫽ 33). Marking Technique at the Slitlamp Patients were seated at the slitlamp. Eye marking was done by F.S.B. or by S.G.F. while observed by F.S.B. After it was confirmed that the slitlamp beam and the patient’s corneal plane were in a position perpendicular to the floor, patients were asked to look straight ahead into the distance as if looking through the operator. The slitlamp light was made coaxial with the binoculars and the light dimmed and narrowed to include only the pupillary margins. The lower lid was retracted and the inferior limbus dried using a Merocel sponge. The 6 o’clock position of the pupil was marked in the middle of the light, just outside the limbus, using a fine-tip sterile marker pen. The lower lid remained retracted for 5 to 10 seconds, allowing time for the mark to dry. The patient was then positioned under the laser and asked to fixate on the blinking red light. After the
J CATARACT REFRACT SURG—VOL 26, DECEMBER 2000
1723
CORNEAL AND LASER ASTIGMATIC AXIS ALIGNMENT IN PARK
Figure 1. (Farah) Limbal mark and 6 o’clock arm of reticle alignment.
lid speculum was inserted, the head was rotated to match the mark with the 6 o’clock arm of the laser reticle (Figure 1). Surgical Technique The VISX Star excimer laser was used in all cases. An epithelial laser scrape technique at a depth of 45 m was used in all eyes. Residual epithelium was scraped using a #64 Beaver blade or a cyclodialysis spatula. The PRK mode was then applied, aiming at a 100% astigmatic correction, using the elliptical ablation profile. Postoperatively, a sterile bandage contact lens was inserted and the patient discharged on tobramycin and dexamethasone 4 times daily and diclofenac 2 times daily. After the epithelium healed completely, patients were switched to fluorometholone 0.1% 4 times daily, tapered over a period of 3 to 4 months. A manifest refraction was performed at each visit starting 1 month postoperatively. Data Analysis Although later refractions were available, early postoperative manifest refractions (earliest refraction giving a 20/20 or better acuity between 1.0 and 2.5 months of surgery) were analyzed using the Alpins vector analysis method.11 The change in astigmatism with each treatment was calculated based on the change in refractive astigmatism at the corneal plane. The VectrAK version 1.0 software (Assort Pty Ltd.) was used to calculate each treatment’s target induced astigmatism (TIA), surgically 1724
induced astigmatism (SIA), difference vector (DV), angle of error (AE), magnitude of error (ME), and index of success (IS). The TIA vector represents the intended treatment (planned astigmatic surgery). It is calculated as 1.20 times the preoperative astigmatism. The SIA vector represents the actual change in astigmatism induced by the surgery. The SIA and TIA vectors can differ in magnitude (diopters) and direction; treatment success occurs when they are equal. The DV indicates the magnitude and direction of change required to achieve the initial goal. It represents the remaining postoperative astigmatism. The ME is the value (diopters) of the difference between the SIA and TIA vectors’ magnitude. It has a positive value if the SIA is larger than the TIA (overcorrection) and a negative value if the SIA is smaller than the TIA (undercorrection). The AE is the angle between the SIA and TIA vectors on a 180 degree vector diagram. It is negative if the SIA vector lies farther clockwise to the TIA vector; it is positive if the SIA vector lies farther counterclockwise. In an optimal treatment, the ME and AE equal zero. The IS is the ratio of the DV to the TIA magnitudes and is optimally equal to zero. The IS represents astigmatism remaining as a portion of the TIA. The geometric means and standard deviations of the various parameters in G1, G2, and the subgroups were calculated using the same program. The 2-tailed t test and the Fisher exact test were used to perform the statistical comparisons.
Results In the first step, the parameters in G1 (n ⫽ 76) and G2 (n ⫽ 67) were calculated and compared. No statistical difference was found between the corresponding parameters. In G1 and G2, the mean TIA was 1.21 ⫾ 0.58 D and 1.22 ⫾ 0.54 D, respectively; the mean SIA, 1.09 ⫾ 0.65 D and 1.08 ⫾ 0.59 D, respectively; the mean DV, 0.46 ⫾ 0.38 D and 0.43 ⫾ 0.35 D, respectively; the mean ME, – 0.12 ⫾ 0.38 D and – 0.13 ⫾ 0.39 D, respectively; and the mean IS, 0.31 ⫾ 0.14 and 0.26 ⫾ 0.11, respectively (Figure 2). In the second step, G1 and G2 were divided into low and high astigmatism subgroups to see whether the results improved with increasing amounts of preoperative astigmatism. In the low astigmatism subgroups of
J CATARACT REFRACT SURG—VOL 26, DECEMBER 2000
CORNEAL AND LASER ASTIGMATIC AXIS ALIGNMENT IN PARK
Figure 2. (Farah) Mean of DV, ME, and IS in the 2 groups. G1 ⫽ marked group; G2 ⫽ unmarked group; DV ⫽ difference vector; ME ⫽ magnitude of error; IS ⫽ index of success.
Figure 3. (Farah) Mean of DV, ME, and IS in the subgroups. G1 ⫽ marked group; G2 ⫽ unmarked group; Low Ast ⫽ low astigmatism subgroup; High Ast ⫽ high astigmatism subgroup; DV ⫽ difference vector; ME ⫽ magnitude of error; IS ⫽ index of success.
G1 (n ⫽ 40) and G2 (n ⫽ 34), the IS was statistically better in G2 (P ⬍ .001). The mean TIA was 0.81 ⫾ 0.13 D and 0.80 ⫾ 0.13 D, respectively; the mean SIA, 0.75 ⫾ 0.39 D and 0.78 ⫾ 0.35 D, respectively; the mean DV, 0.47 ⫾ 0.42 D and 0.33 ⫾ 0.36 D, respectively; the mean ME, – 0.06 ⫾ 0.39 D and – 0.01 ⫾ 0.34 D, respectively; and the mean IS, 0.41 ⫾ 0.18 and 0.24 ⫾ 0.18, respectively (Figure 3). In the high astigmatism subgroups of G1 (n ⫽ 36) and G2 (n ⫽ 33), the IS was statistically better in G1 (P ⬍ .0001). The mean TIA was 1.67 ⫾ 0.54 D and 1.65 ⫾ 0.46 D, respectively; the mean SIA, 1.47 ⫾ 0.68 D and 1.39 ⫾ 0.63 D, respectively; the mean DV, 0.45 ⫾ 0.34 D and 0.53 ⫾ 0.30 D, respectively; the
mean ME, – 0.20 ⫾ 0.36 D and – 0.26 ⫾ 0.41 D, respectively; and the mean IS, 0.22 ⫾ 0.08 and 0.29 ⫾ 0.04, respectively (Figure 3). In the third step, the AE was analyzed. In G1, 28 eyes (37%) had a positive AE and 25 (33%), a negative AE. In G2, 21 eyes (31%) had a positive AE and 22 (33%), a negative AE. Eyes that had identical orientations of the SIA and TIA vectors (AE of 0) were analyzed. In G1, 23 eyes (30%) (13 [33%] in the low astigmatism subgroup and 10 [28%] in the high astigmatism subgroup) had an AE of 0 and in G2, 24 eyes (36%) (17 [50%] in the low astigmatism subgroup and 7 [21%] in the high astigmatism subgroup) had an AE of 0.
J CATARACT REFRACT SURG—VOL 26, DECEMBER 2000
1725
CORNEAL AND LASER ASTIGMATIC AXIS ALIGNMENT IN PARK
Figure 4. (Farah) Distribution by subgroup of eyes that had AE and IS of 0. Dark column ⫽ G1, marked group ( n ⫽ 20); light column ⫽ G2, unmarked group (n ⫽ 17); Low ⫽ low astigmatism subgroup; High ⫽ high astigmatism subgroup.
Eyes that had full correction of the preoperative astigmatism (ie, on axis and full correction) (AE and IS of 0) were analyzed. There were 20/23 (87%) (11/13 [85%] in the low astigmatism subgroup and 9/10 [90%] in the high astigmatism subgroup) in G1 and 17 (70%) (14 [82%] in the low astigmatism subgroup and 3 [43%] in the high astigmatism subgroup) in G2 (Figure 4). There was a statistically better result in the high astigmatism subgroup of G1 (P ⬍ .05).
Discussion In this study, we analyzed and compared the astigmatic change after PARK in a group of eyes in which the astigmatic meridian of the cornea and laser had been aligned after placing a reference point at the limbus and in a group in which there was no placement of a reference point. Our results showed that when all preoperative astigmatic levels were analyzed together, there was no statistical difference between the 2 groups in DV, AE, ME, and IS. When the groups were divided into subgroups of low and high astigmatism, the IS, which reflects the proportion of remaining astigmatism to treated astigmatism, showed a statistically better result in the high astigmatism subgroup of the marked group than in that of the unmarked group. We conclude that limbal marking and subsequent corneal and laser astigmatic axis alignment improve the refractive outcome of PARK for astigmatism ⱖ1.25 D. Early (1.0 to 2.5 month) postoperative manifest refractions were analyzed to minimize the influence of wound healing and 1726
regression of effect occurring with later (6 month) refractions.12 Smith and coauthors1,2 studied position-related astigmatic axis change. They used 3 different methods in their measurements: rocking the cylinder, Jackson cross cylinder, and Maddox double rod. No statistically significant difference in cylinder axis measurement in the seated versus supine positions was observed using either of these techniques.1,2 The Jackson cross-cylinder method1 revealed a difference of 2.3 ⫾ 1.9 degrees (range 0 to 7 degrees); the rocking the cylinder method,1 a difference of 4.3 ⫾ 3.5 degrees (range 0 to 13 degrees); and the Maddox double-rod method,2 a difference of 0.2 ⫾ 1.2 degrees (range 0 to 4 degrees). However, the same authors1 reported that 25% of the studied eyes had a change in axis of 7 to 13 degrees and of the 50 eyes studied, 24 (48%) excyclotorted, 21 (42%) incyclotorted, and 5 (10%) had no change in axis. Using a reference point on the cornea and videokeratography for intraoperative identification of astigmatic axis change, Suzuki et al.3 found a mean axial misalignment of 4.4 ⫾ 2.8 degrees (range 0 to 14 degrees) between the seated and supine positions. Of the 70 eyes analyzed, 43% showed clockwise misalignment and 43%, counterclockwise misalignment; in 14%, misalignment did not occur. In our study, although we did not record the amount and direction of cyclotorsion intraoperatively, we were aware that most G1 eyes were off axis and had to be adjusted by repositioning the head to align the reference point and the 6 o’clock arm of the laser reticle.
J CATARACT REFRACT SURG—VOL 26, DECEMBER 2000
CORNEAL AND LASER ASTIGMATIC AXIS ALIGNMENT IN PARK
Figure 5. (Farah) Mean IS in 31 eyes treated by LASIK. G1 ⫽ marked group; G2 ⫽ unmarked group.
When we analyzed the AE, the data showed no difference between G1 and G2 in eyes that had an AE of 0 (30% and 36%, respectively). Among the eyes that had an AE of 0, 87% and 70% in G1 and G2, respectively, had an IS of 0. In the high astigmatism subgroups, 90% in G1 and 43% in G2 had full cylinder correction. It has been shown theoretically4 that a correcting cylinder of equal magnitude to the cylinder present but with axis misalignment leads to loss of effect and undercorrection. For example, an 8 degree misalignment could lead to a 25% loss of effect. Possible causes of axis misalignment include position-induced ocular torsion, head tilting during surgery, unmasking of a cyclophoria, and distortion of the globe by a lid speculum. The limitations of our study include small sample size, retrospective design, possible inaccuracies in the marking method, surgeon bias during the postoperative refractions, and positioning problems of the head. (We attempted to minimize the likelihood of this type of error in our study by maintaining a neutral head position during both seated and supine axis determination.) Photoastigmatic refractive keratectomy has been successful in reducing but often not completely eliminating astigmatism. Several variables have been suggested to decrease astigmatic laser treatment outcome such as axis misalignment,4 posterior corneal surface astigmatism, wound healing,13 laser nomogram power values,14 differences between refractive and topographic axes, centration of the optical zone, and regression of effect.12 Since excimer cylindrical ablation tends to undercorrect the ocular astigmatism, it has been suggested that
the cylinder power applied should be increased to a value greater than that to be corrected.14 How much of this undercorrection is power or wound-healing related and how much is axis alignment related has to be determined. The fact that there was only 1 variable change (marking) in the surgical technique between G1 and G2 in our study leads us to believe that axis alignment is an important variable. We believe that quicker wound healing and earlier stabilization of the refraction15 in eyes treated by laser in situ keratomileusis (LASIK) will help reveal the advantage of limbal marking and subsequent corneal and laser astigmatic axis alignment for all levels of astigmatism. Preliminary data in G1 and G2 groups in 31 LASIKtreated eyes (Figure 5) confirm this. Thus, although the current study shows an advantage of marking in the high astigmatism subgroup only, we recommend that corneal and laser astigmatic axis alignment be performed for levels of preoperative astigmatism of 0.75 D or more.
References 1. Smith EM, Talamo JH, Assil KK, Petashnick DE. Comparison of astigmatic axis in the seated and supine positions. J Refract Corneal Surg 1994; 10:615– 620 2. Smith EM, Talamo JH. Cyclotorsion in the seated and supine patient. J Cataract Refract Surg 1995; 21:402– 403 3. Suzuki A, Maeda N, Watanabe H, et al. Using a reference point and videokeratography for intraoperative identification of astigmatism axis. J Cataract Refract Surg 1997; 23:1491–1495 4. Stevens JD. Astigmatic excimer laser treatment: theoretical effects of axis misalignment. Eur J Implant Refract Surg 1994; 6:310 –318 5. Shieh E, Moreira H, D’Arcy J, et al. Quantitative analysis of wound healing after cylindrical and spherical excimer laser ablations. Ophthalmology 1992; 99:1050 –1055 6. McDonnell PJ, Moreira H, Clapham TN, et al. Photorefractive keratectomy for astigmatism: initial clinical results. Arch Ophthalmol 1991; 109:1370 –1373 7. McDonnell PJ, Moreira H, Garbus J, et al. Photorefractive keratectomy to create toric ablations for correction of astigmatism. Arch Ophthalmol 1991; 109:710 –713 8. Taylor HR, Guest CS, Kelly P, Alpins N. Comparison of excimer laser treatment of astigmatism and myopia. Arch Ophthalmol 1993; 111:1621–1626 9. Seiler T, McDonnell PJ. Excimer laser photorefractive keratectomy. Surv Ophthalmol 1995; 40:89 –118 10. Snibson GR, Carson CA, Aldred GF, Taylor HR. One year evaluation of excimer laser photorefractive keratec-
J CATARACT REFRACT SURG—VOL 26, DECEMBER 2000
1727
CORNEAL AND LASER ASTIGMATIC AXIS ALIGNMENT IN PARK
tomy for myopia and myopic astigmatism. Arch Ophthalmol 1995; 113:994 –1000 11. Alpins NA. A new method of analyzing vectors for changes in astigmatism. J Cataract Refract Surg 1993; 19:524 –533 12. Schipper I, Senn P, Wienecke L, Øyo-Szerenyi KD. Photoastigmatic refractive keratectomy for primary treatment and revision of myopic astigmatism. J Cataract Refract Surg 1997; 23:1465–1471 13. Tabin GC, Alpins N, Aldred GF, et al. Astigmatic change
1728
1 year after excimer laser treatment of myopia and myopic astigmatism. J Cataract Refract Surg 1996; 22:924 – 930 14. Taylor HR, Kelly P, Alpins N. Excimer laser correction of myopic astigmatism. J Cataract Refract Surg 1994; 20: 243–251 15. Hersh PS, Brint SF, Maloney RK, et al. Photorefractive keratectomy versus laser in situ keratomileusis for moderate to high myopia. Ophthalmology 1998; 105:1512– 1523
J CATARACT REFRACT SURG—VOL 26, DECEMBER 2000
Outcome of corneal and laser astigmatic axis alignment in photoastigmatic refractive keratectomy Samir G. Farah, MD, Eric Olafsson, David G. Gwynn, MD, Dimitri T. Azar, MD, Frederick S. Brightbill, MD ABSTRACT Purpose: To compare the refractive results of laser astigmatic treatment in eyes in which the astigmatic axes of the eye and laser are aligned by limbal marking at the 6 o’clock position and in eyes that are not marked. Setting: University Hospital and Clinics, Madison, Wisconsin, USA. Methods: This retrospective study comprised 143 eyes that had photoastigmatic refractive keratectomy with the VISX Star excimer laser. The eyes were divided into marked (G1) and unmarked (G2) groups. Based on the preoperative astigmatism, each group was subdivided into low astigmatism (ⱕ1.00 diopter [D]) and high astigmatism (ⱖ1.25 D). Early postoperative manifest refractions (1.0 to 2.5 months) were analyzed. The Alpins vector analysis method was used to calculate the target induced astigmatism, surgically induced astigmatism, difference vector (DV), magnitude of error (ME), angle of error (AE), and index of success (IS). Results: There was no significant difference between the groups in DV, ME, and IS. When the subgroups were analyzed, the DV and ME were comparable; the IS in the G1 high astigmatism subgroup was significantly better than that in the G2 high astigmatism subgroup (0.22 ⫾ 0.08 and 0.29 ⫾ 0.04, respectively; P ⬍ .0001). There was comparable scatter of AE values; 30% and 36% in G1 and G2, respectively, had an AE of 0. Similar scatter was observed in the subgroups. Of the eyes that had an AE of 0, 90% and 43% in the high astigmatism subgroups of G1 and G2, respectively (P ⬍ .05), had full correction of astigmatism. Conclusion: Limbal marking and subsequent eye and laser astigmatic axis alignment improved the refractive outcome of laser astigmatic treatment of ⱖ1.25 D. A preliminary report of an ongoing prospective randomized study of eyes that had laser in situ keratomileusis is included. J Cataract Refract Surg 2000; 26:1722–1728 © 2000 ASCRS and ESCRS
M
yopic astigmatic laser refractive surgery consists of flattening the steepest meridian of the cornea. In the VISX Star excimer laser system, power is delivered through a contracting-slit apparatus, with
© 2000 ASCRS and ESCRS Published by Elsevier Science Inc.
blade movement along the steepest meridian of the cornea. Astigmatic power and axis are determined while the patient is seated during refraction. However, laser treat0886-3350/00/$–see front matter PII S0886-3350(00)00695-7
CORNEAL AND LASER ASTIGMATIC AXIS ALIGNMENT IN PARK
ment is applied while the patient is supine. Studies1,2 have failed to show statistical differences in cylinder axis measurement in the seated versus the supine position when comparing groups but have documented significant ocular cyclotorsion at the individual level.1–3 If not accounted for, such eye movement would theoretically decrease the refractive outcome of astigmatic surgery.4 Although photoastigmatic refractive keratectomy (PARK) outcomes have been reported to be comparable to those of photorefractive keratectomy (PRK),5– 8 most observers agree that accuracy can be improved.9,10 Even though many refractive surgeons in the United States and abroad place a reference point on the cornea to match the corneal and laser beam astigmatic meridia during treatment, we are unaware of studies that discuss the refractive outcome of astigmatism treatment in marked versus unmarked eyes. In this study, we retrospectively analyzed a series of patients to compare the early results of PARK in a group in which the astigmatic meridia of the cornea and laser were aligned by limbal marking at the 6 o’clock position and in an unmarked group.
Patients and Methods The charts of 1 surgeon’s (F.S.B.) patients who had PARK between May 1997 and May 1998 at the University of Wisconsin Eye Clinics were reviewed. One hundred forty-five patients (211 eyes) were initially identified. The criteria for inclusion in the study were
Accepted for publication August 8, 2000. From the Cornea Service, Department of Ophthalmology and Visual Sciences, University of Wisconsin, Madison, Wisconsin (Farah, Olafsson, Gwynn, Brightbill), and the Cornea Service, Massachusetts Eye and Ear Infirmary, Harvard Medical School, Boston, Massachusetts (Azar), USA. Presented in part at the Symposium on Cataract, IOL and Refractive Surgery, Seattle, Washington, USA, April 1999. Supported by an unrestricted grant from Research to Prevent Blindness, Inc., and the Wisconsin Lion Foundation to the Department of Ophthalmology and Visual Sciences at the University of Wisconsin, Madison. None of the authors has a financial interest in any product mentioned. Reprint requests to Frederick S. Brightbill, MD, 2870 University Avenue, Suite 102, Madison, Wisconsin 53705, USA.
age of at least 21 years, stable refraction on 2 consecutive examinations, best corrected visual acuity of 20/20 or better, preoperative refractive astigmatism ⱕ4.0 diopters (D), absence of connective tissue disease, no history of eye disease or glaucoma, no muscle imbalance, and no haze postoperatively. Forty-one eyes were not included (35 for insufficient follow-up, 5 for presence of corneal haze, and 1 for infectious keratitis). Although a subgroup of patients with preoperative astigmatism of 0.5 D (n ⫽ 27, 18 in G1 and 9 in G2) was initially identified, it was not included in the analysis because patients with less astigmatism may be more tolerant of subtle changes in cylinder power and axis and because if present, most surgeons prefer treating the spherical equivalent. This study comprised 90 patients (143 eyes) who had PARK for myopia and astigmatism. Mean patient age was 44 years ⫾ 13 (SD) (range 25 to 59 years). It was 42 ⫾ 15 years in the marked group and 46 ⫾ 17 years in the unmarked group. Five retreated eyes were included in the study (10% of the total number of eyes). Eyes were divided into marked (G1) and unmarked (G2) groups throughout the study using a table of random variation; G1 comprised 40 right eyes and 36 left eyes and G2, 34 and 33, respectively. Each group was divided into 2 subgroups based on the preoperative astigmatism: low astigmatism (ⱕ1.00 D; G1 ⫽ 40, G2 ⫽ 34) and high astigmatism (ⱖ1.25 D; G1 ⫽ 36, G2 ⫽ 33). Marking Technique at the Slitlamp Patients were seated at the slitlamp. Eye marking was done by F.S.B. or by S.G.F. while observed by F.S.B. After it was confirmed that the slitlamp beam and the patient’s corneal plane were in a position perpendicular to the floor, patients were asked to look straight ahead into the distance as if looking through the operator. The slitlamp light was made coaxial with the binoculars and the light dimmed and narrowed to include only the pupillary margins. The lower lid was retracted and the inferior limbus dried using a Merocel sponge. The 6 o’clock position of the pupil was marked in the middle of the light, just outside the limbus, using a fine-tip sterile marker pen. The lower lid remained retracted for 5 to 10 seconds, allowing time for the mark to dry. The patient was then positioned under the laser and asked to fixate on the blinking red light. After the
J CATARACT REFRACT SURG—VOL 26, DECEMBER 2000
1723
CORNEAL AND LASER ASTIGMATIC AXIS ALIGNMENT IN PARK
Figure 1. (Farah) Limbal mark and 6 o’clock arm of reticle alignment.
lid speculum was inserted, the head was rotated to match the mark with the 6 o’clock arm of the laser reticle (Figure 1). Surgical Technique The VISX Star excimer laser was used in all cases. An epithelial laser scrape technique at a depth of 45 m was used in all eyes. Residual epithelium was scraped using a #64 Beaver blade or a cyclodialysis spatula. The PRK mode was then applied, aiming at a 100% astigmatic correction, using the elliptical ablation profile. Postoperatively, a sterile bandage contact lens was inserted and the patient discharged on tobramycin and dexamethasone 4 times daily and diclofenac 2 times daily. After the epithelium healed completely, patients were switched to fluorometholone 0.1% 4 times daily, tapered over a period of 3 to 4 months. A manifest refraction was performed at each visit starting 1 month postoperatively. Data Analysis Although later refractions were available, early postoperative manifest refractions (earliest refraction giving a 20/20 or better acuity between 1.0 and 2.5 months of surgery) were analyzed using the Alpins vector analysis method.11 The change in astigmatism with each treatment was calculated based on the change in refractive astigmatism at the corneal plane. The VectrAK version 1.0 software (Assort Pty Ltd.) was used to calculate each treatment’s target induced astigmatism (TIA), surgically 1724
induced astigmatism (SIA), difference vector (DV), angle of error (AE), magnitude of error (ME), and index of success (IS). The TIA vector represents the intended treatment (planned astigmatic surgery). It is calculated as 1.20 times the preoperative astigmatism. The SIA vector represents the actual change in astigmatism induced by the surgery. The SIA and TIA vectors can differ in magnitude (diopters) and direction; treatment success occurs when they are equal. The DV indicates the magnitude and direction of change required to achieve the initial goal. It represents the remaining postoperative astigmatism. The ME is the value (diopters) of the difference between the SIA and TIA vectors’ magnitude. It has a positive value if the SIA is larger than the TIA (overcorrection) and a negative value if the SIA is smaller than the TIA (undercorrection). The AE is the angle between the SIA and TIA vectors on a 180 degree vector diagram. It is negative if the SIA vector lies farther clockwise to the TIA vector; it is positive if the SIA vector lies farther counterclockwise. In an optimal treatment, the ME and AE equal zero. The IS is the ratio of the DV to the TIA magnitudes and is optimally equal to zero. The IS represents astigmatism remaining as a portion of the TIA. The geometric means and standard deviations of the various parameters in G1, G2, and the subgroups were calculated using the same program. The 2-tailed t test and the Fisher exact test were used to perform the statistical comparisons.
Results In the first step, the parameters in G1 (n ⫽ 76) and G2 (n ⫽ 67) were calculated and compared. No statistical difference was found between the corresponding parameters. In G1 and G2, the mean TIA was 1.21 ⫾ 0.58 D and 1.22 ⫾ 0.54 D, respectively; the mean SIA, 1.09 ⫾ 0.65 D and 1.08 ⫾ 0.59 D, respectively; the mean DV, 0.46 ⫾ 0.38 D and 0.43 ⫾ 0.35 D, respectively; the mean ME, – 0.12 ⫾ 0.38 D and – 0.13 ⫾ 0.39 D, respectively; and the mean IS, 0.31 ⫾ 0.14 and 0.26 ⫾ 0.11, respectively (Figure 2). In the second step, G1 and G2 were divided into low and high astigmatism subgroups to see whether the results improved with increasing amounts of preoperative astigmatism. In the low astigmatism subgroups of
J CATARACT REFRACT SURG—VOL 26, DECEMBER 2000
CORNEAL AND LASER ASTIGMATIC AXIS ALIGNMENT IN PARK
Figure 2. (Farah) Mean of DV, ME, and IS in the 2 groups. G1 ⫽ marked group; G2 ⫽ unmarked group; DV ⫽ difference vector; ME ⫽ magnitude of error; IS ⫽ index of success.
Figure 3. (Farah) Mean of DV, ME, and IS in the subgroups. G1 ⫽ marked group; G2 ⫽ unmarked group; Low Ast ⫽ low astigmatism subgroup; High Ast ⫽ high astigmatism subgroup; DV ⫽ difference vector; ME ⫽ magnitude of error; IS ⫽ index of success.
G1 (n ⫽ 40) and G2 (n ⫽ 34), the IS was statistically better in G2 (P ⬍ .001). The mean TIA was 0.81 ⫾ 0.13 D and 0.80 ⫾ 0.13 D, respectively; the mean SIA, 0.75 ⫾ 0.39 D and 0.78 ⫾ 0.35 D, respectively; the mean DV, 0.47 ⫾ 0.42 D and 0.33 ⫾ 0.36 D, respectively; the mean ME, – 0.06 ⫾ 0.39 D and – 0.01 ⫾ 0.34 D, respectively; and the mean IS, 0.41 ⫾ 0.18 and 0.24 ⫾ 0.18, respectively (Figure 3). In the high astigmatism subgroups of G1 (n ⫽ 36) and G2 (n ⫽ 33), the IS was statistically better in G1 (P ⬍ .0001). The mean TIA was 1.67 ⫾ 0.54 D and 1.65 ⫾ 0.46 D, respectively; the mean SIA, 1.47 ⫾ 0.68 D and 1.39 ⫾ 0.63 D, respectively; the mean DV, 0.45 ⫾ 0.34 D and 0.53 ⫾ 0.30 D, respectively; the
mean ME, – 0.20 ⫾ 0.36 D and – 0.26 ⫾ 0.41 D, respectively; and the mean IS, 0.22 ⫾ 0.08 and 0.29 ⫾ 0.04, respectively (Figure 3). In the third step, the AE was analyzed. In G1, 28 eyes (37%) had a positive AE and 25 (33%), a negative AE. In G2, 21 eyes (31%) had a positive AE and 22 (33%), a negative AE. Eyes that had identical orientations of the SIA and TIA vectors (AE of 0) were analyzed. In G1, 23 eyes (30%) (13 [33%] in the low astigmatism subgroup and 10 [28%] in the high astigmatism subgroup) had an AE of 0 and in G2, 24 eyes (36%) (17 [50%] in the low astigmatism subgroup and 7 [21%] in the high astigmatism subgroup) had an AE of 0.
J CATARACT REFRACT SURG—VOL 26, DECEMBER 2000
1725
CORNEAL AND LASER ASTIGMATIC AXIS ALIGNMENT IN PARK
Figure 4. (Farah) Distribution by subgroup of eyes that had AE and IS of 0. Dark column ⫽ G1, marked group ( n ⫽ 20); light column ⫽ G2, unmarked group (n ⫽ 17); Low ⫽ low astigmatism subgroup; High ⫽ high astigmatism subgroup.
Eyes that had full correction of the preoperative astigmatism (ie, on axis and full correction) (AE and IS of 0) were analyzed. There were 20/23 (87%) (11/13 [85%] in the low astigmatism subgroup and 9/10 [90%] in the high astigmatism subgroup) in G1 and 17 (70%) (14 [82%] in the low astigmatism subgroup and 3 [43%] in the high astigmatism subgroup) in G2 (Figure 4). There was a statistically better result in the high astigmatism subgroup of G1 (P ⬍ .05).
Discussion In this study, we analyzed and compared the astigmatic change after PARK in a group of eyes in which the astigmatic meridian of the cornea and laser had been aligned after placing a reference point at the limbus and in a group in which there was no placement of a reference point. Our results showed that when all preoperative astigmatic levels were analyzed together, there was no statistical difference between the 2 groups in DV, AE, ME, and IS. When the groups were divided into subgroups of low and high astigmatism, the IS, which reflects the proportion of remaining astigmatism to treated astigmatism, showed a statistically better result in the high astigmatism subgroup of the marked group than in that of the unmarked group. We conclude that limbal marking and subsequent corneal and laser astigmatic axis alignment improve the refractive outcome of PARK for astigmatism ⱖ1.25 D. Early (1.0 to 2.5 month) postoperative manifest refractions were analyzed to minimize the influence of wound healing and 1726
regression of effect occurring with later (6 month) refractions.12 Smith and coauthors1,2 studied position-related astigmatic axis change. They used 3 different methods in their measurements: rocking the cylinder, Jackson cross cylinder, and Maddox double rod. No statistically significant difference in cylinder axis measurement in the seated versus supine positions was observed using either of these techniques.1,2 The Jackson cross-cylinder method1 revealed a difference of 2.3 ⫾ 1.9 degrees (range 0 to 7 degrees); the rocking the cylinder method,1 a difference of 4.3 ⫾ 3.5 degrees (range 0 to 13 degrees); and the Maddox double-rod method,2 a difference of 0.2 ⫾ 1.2 degrees (range 0 to 4 degrees). However, the same authors1 reported that 25% of the studied eyes had a change in axis of 7 to 13 degrees and of the 50 eyes studied, 24 (48%) excyclotorted, 21 (42%) incyclotorted, and 5 (10%) had no change in axis. Using a reference point on the cornea and videokeratography for intraoperative identification of astigmatic axis change, Suzuki et al.3 found a mean axial misalignment of 4.4 ⫾ 2.8 degrees (range 0 to 14 degrees) between the seated and supine positions. Of the 70 eyes analyzed, 43% showed clockwise misalignment and 43%, counterclockwise misalignment; in 14%, misalignment did not occur. In our study, although we did not record the amount and direction of cyclotorsion intraoperatively, we were aware that most G1 eyes were off axis and had to be adjusted by repositioning the head to align the reference point and the 6 o’clock arm of the laser reticle.
J CATARACT REFRACT SURG—VOL 26, DECEMBER 2000
CORNEAL AND LASER ASTIGMATIC AXIS ALIGNMENT IN PARK
Figure 5. (Farah) Mean IS in 31 eyes treated by LASIK. G1 ⫽ marked group; G2 ⫽ unmarked group.
When we analyzed the AE, the data showed no difference between G1 and G2 in eyes that had an AE of 0 (30% and 36%, respectively). Among the eyes that had an AE of 0, 87% and 70% in G1 and G2, respectively, had an IS of 0. In the high astigmatism subgroups, 90% in G1 and 43% in G2 had full cylinder correction. It has been shown theoretically4 that a correcting cylinder of equal magnitude to the cylinder present but with axis misalignment leads to loss of effect and undercorrection. For example, an 8 degree misalignment could lead to a 25% loss of effect. Possible causes of axis misalignment include position-induced ocular torsion, head tilting during surgery, unmasking of a cyclophoria, and distortion of the globe by a lid speculum. The limitations of our study include small sample size, retrospective design, possible inaccuracies in the marking method, surgeon bias during the postoperative refractions, and positioning problems of the head. (We attempted to minimize the likelihood of this type of error in our study by maintaining a neutral head position during both seated and supine axis determination.) Photoastigmatic refractive keratectomy has been successful in reducing but often not completely eliminating astigmatism. Several variables have been suggested to decrease astigmatic laser treatment outcome such as axis misalignment,4 posterior corneal surface astigmatism, wound healing,13 laser nomogram power values,14 differences between refractive and topographic axes, centration of the optical zone, and regression of effect.12 Since excimer cylindrical ablation tends to undercorrect the ocular astigmatism, it has been suggested that
the cylinder power applied should be increased to a value greater than that to be corrected.14 How much of this undercorrection is power or wound-healing related and how much is axis alignment related has to be determined. The fact that there was only 1 variable change (marking) in the surgical technique between G1 and G2 in our study leads us to believe that axis alignment is an important variable. We believe that quicker wound healing and earlier stabilization of the refraction15 in eyes treated by laser in situ keratomileusis (LASIK) will help reveal the advantage of limbal marking and subsequent corneal and laser astigmatic axis alignment for all levels of astigmatism. Preliminary data in G1 and G2 groups in 31 LASIKtreated eyes (Figure 5) confirm this. Thus, although the current study shows an advantage of marking in the high astigmatism subgroup only, we recommend that corneal and laser astigmatic axis alignment be performed for levels of preoperative astigmatism of 0.75 D or more.
References 1. Smith EM, Talamo JH, Assil KK, Petashnick DE. Comparison of astigmatic axis in the seated and supine positions. J Refract Corneal Surg 1994; 10:615– 620 2. Smith EM, Talamo JH. Cyclotorsion in the seated and supine patient. J Cataract Refract Surg 1995; 21:402– 403 3. Suzuki A, Maeda N, Watanabe H, et al. Using a reference point and videokeratography for intraoperative identification of astigmatism axis. J Cataract Refract Surg 1997; 23:1491–1495 4. Stevens JD. Astigmatic excimer laser treatment: theoretical effects of axis misalignment. Eur J Implant Refract Surg 1994; 6:310 –318 5. Shieh E, Moreira H, D’Arcy J, et al. Quantitative analysis of wound healing after cylindrical and spherical excimer laser ablations. Ophthalmology 1992; 99:1050 –1055 6. McDonnell PJ, Moreira H, Clapham TN, et al. Photorefractive keratectomy for astigmatism: initial clinical results. Arch Ophthalmol 1991; 109:1370 –1373 7. McDonnell PJ, Moreira H, Garbus J, et al. Photorefractive keratectomy to create toric ablations for correction of astigmatism. Arch Ophthalmol 1991; 109:710 –713 8. Taylor HR, Guest CS, Kelly P, Alpins N. Comparison of excimer laser treatment of astigmatism and myopia. Arch Ophthalmol 1993; 111:1621–1626 9. Seiler T, McDonnell PJ. Excimer laser photorefractive keratectomy. Surv Ophthalmol 1995; 40:89 –118 10. Snibson GR, Carson CA, Aldred GF, Taylor HR. One year evaluation of excimer laser photorefractive keratec-
J CATARACT REFRACT SURG—VOL 26, DECEMBER 2000
1727
CORNEAL AND LASER ASTIGMATIC AXIS ALIGNMENT IN PARK
tomy for myopia and myopic astigmatism. Arch Ophthalmol 1995; 113:994 –1000 11. Alpins NA. A new method of analyzing vectors for changes in astigmatism. J Cataract Refract Surg 1993; 19:524 –533 12. Schipper I, Senn P, Wienecke L, Øyo-Szerenyi KD. Photoastigmatic refractive keratectomy for primary treatment and revision of myopic astigmatism. J Cataract Refract Surg 1997; 23:1465–1471 13. Tabin GC, Alpins N, Aldred GF, et al. Astigmatic change
1728
1 year after excimer laser treatment of myopia and myopic astigmatism. J Cataract Refract Surg 1996; 22:924 – 930 14. Taylor HR, Kelly P, Alpins N. Excimer laser correction of myopic astigmatism. J Cataract Refract Surg 1994; 20: 243–251 15. Hersh PS, Brint SF, Maloney RK, et al. Photorefractive keratectomy versus laser in situ keratomileusis for moderate to high myopia. Ophthalmology 1998; 105:1512– 1523
J CATARACT REFRACT SURG—VOL 26, DECEMBER 2000