- Email: [email protected]
IntraLase-Enabled Astigmatic Keratotomy for Post-Keratoplasty Astigmatism: On-Axis Vector Analysis
IntraLase-Enabled Astigmatic Keratotomy for Post-Keratoplasty Astigmatism: On-Axis Vector Analysis Nikhil L. Kumar, BMed, FRANZCO, Igor Kaiserman, MD, MSc, MHA, Raneen Shehadeh-Mashor, MD, Wiwan Sansanayudh, MD, Rusty Ritenour, MD, FRCSC, David S. Rootman, MD, FRCSC Purpose: To determine the refractive predictability, stability, efficacy, and complication rate of femtosecond laser-enabled astigmatic keratotomy for post-keratoplasty astigmatism. Design: A retrospective case series (pilot study). Participants: Thirty-seven eyes of 34 patients. Methods: All eyes underwent IntraLase-enabled astigmatic keratotomy for high astigmatism (⬎5 diopters [D]) after penetrating keratoplasty. Main Outcome Measures: Uncorrected visual acuity (UCVA), best-corrected visual acuity (BCVA), manifest refraction, higher-order aberrations, and complications. Results: Mean follow-up was for 7.2 months. Uncorrected visual acuity improved from a mean of 1.08⫾0.34 logarithm of the minimum angle of resolution preoperatively to a mean of 0.80⫾0.42 postoperatively (P⫽0.0016). Best-corrected visual acuity improved from a mean of 0.45⫾0.27 preoperatively to 0.37⫾0.27 postoperatively (P⫽0.018). The defocus equivalent was significantly reduced by more than 1 D (P⫽0.025). The value of absolute astigmatism was reduced from 7.46⫾2.70 D preoperatively to 4.77⫾3.29 D postoperatively (P⫽0.0001). Higherorder aberrations were significantly increased. The efficacy index was 0.6⫾0.6. There were no cases of perforation, wound dehiscence, or infectious keratitis. Three eyes (8%) experienced an episode of graft rejection. Overcorrection occurred in 9 eyes (24%). Conclusions: IntraLase-enabled astigmatic keratotomy is an effective treatment for high astigmatism after penetrating keratoplasty with an encouraging refractive predictability. Future studies may help refine the treatment parameters required to achieve reduction of cylinder with greater accuracy. Financial Disclosure(s): Proprietary or commercial disclosure may be found after the references. Ophthalmology 2010;117:1228 –1235 © 2010 by the American Academy of Ophthalmology.
Visual outcomes after penetrating keratoplasty are often limited by significant astigmatism, ametropia, and anisometropia. More than 5 diopters (D) of astigmatism are common after full-thickness corneal transplantation.1– 4 Conservative therapeutic options in the setting of high astigmatism are difficult to use. Spectacle use may lead to intolerable aniseikonia. Contact lens fitting is challenging over a clear graft with topographic alterations;5 thus, a conservative approach does not always yield maximal visual acuity.6 Previously, surgical options for astigmatic reduction were limited to continuous suture adjustment, selective suture removal, relaxing incisions with or without compression sutures, manual astigmatic keratotomy, wedge resections, photorefractive keratectomy, and LASIK.2,4,7–13 Suture manipulation is possible within the first year. Relaxing incisions, manual astigmatic keratotomy, and wedge resections have been associated with unpredictable results, wound gape, and occasionally perforation. LASIK has been reported to give stable and predictable results after pene-
1228
© 2010 by the American Academy of Ophthalmology Published by Elsevier Inc.
trating keratoplasty but has limited capacity for astigmatic reduction.4,6 There has been great enthusiasm for the application of femtosecond laser technology to improve patient outcomes and decrease complications in penetrating keratoplasty.14 –16 The use of femtosecond laser technology to create astigmatic keratotomy incisions has recently been reported.16 –20 This technique has the advantage of greater precision in arc depth, length, and curvature. We previously compared the results of manual astigmatic keratotomy with IntraLase-enabled astigmatic keratotomy (IEAK).18 IntraLase-enabled astigmatic keratotomy significantly improved both uncorrected visual acuity (UCVA) and best-corrected visual acuity (BCVA) and brought the cylinder vector closer to neutral. Manual astigmatic keratotomy exhibited a trend toward an improvement in UCVA and BCVA but overcorrected the cylinder vector. IntraLase-enabled astigmatic keratotomy had a lesser detrimental effect on higher-order aberrations than manual astigmatic keratotomy. ISSN 0161-6420/10/$–see front matter doi:10.1016/j.ophtha.2009.10.041
Kumar et al 䡠 IntraLase-Enabled Astigmatic Keratotomy Vector Analysis To date, there is a paucity of data on the predictability, rate of change, and duration of effect, efficacy, and complication rate of this procedure beyond 1 month. In this retrospective pilot study, we report these in a series of 37 eyes with a mean follow-up of 7 months.
Materials and Methods Study Design A retrospective review of 37 eyes of 35 patients who had undergone IEAK for high astigmatism was completed. There were 21 male and 14 female patients in the study. Approval for the study was obtained from the University Health Network Research Ethics Board to access retrospective data. Patients with at least 5 D of post-keratoplasty regular astigmatism were included. All were at least 18 months post-keratoplasty, and no residual sutures were in place when IEAK was performed. Patients with high irregular astigmatism were excluded. Irregular astigmatism was calibrated by analysis of the wavefront OPD and Zernike decomposition of the higher-order aberrations (NIDEK, OPD Scan II ARK 10000, Gamagori, Japan). Thirty-four eyes of 32 patients had a minimum of 4 months follow-up and were included in the outcome measurement analysis. Three eyes of 3 patients required early resuturing because of overcorrection, within 1 month of the incisions. These were excluded from outcome measurement analysis, because it was impossible to determine the effect of the original incisions. Their data were included in the calculation of complication rates. Outcome measures included UCVA, BCVA, manifest refraction, corneal topography, and higher-order aberrations. Patients were followed at 1 day, 1 week, 1 month, 3 months, 6 months, and 12 months. At each visit pre- and postoperatively, slit-lamp examination and tonometry were performed and outcome measures were recorded. For outcome measurement analysis, both interval and last visit data were used. The efficacy index was calculated as the ratio of the mean postoperative UCVA to the mean preoperative BCVA.21
Surgical Procedure Preoperative planning included assessment of the amount and axis of corneal astigmatism, corneal thickness, and graft size. When there was disagreement between the axis of the manifest refraction and the topography, we favored the topography axis. The amount of topographic cylinder rather than the manifest cylinder was used to determine the length (degrees) of the keratotomy. Most patients also had imaging with the Pentacam rotating Scheimpflug camera (Oculus, Wetzlar, Germany). All treatments were paired symmetric (same length) incisions centered on the steep axis. The depth of the incisions was set at approximately 90% depth, based on either the Pentacam or ultrasound pachymetry at the incision location. Each incision was made 0.5 mm within the graft– host junction, such that the diameter was set at 1 mm less than the graft diameter measured by calipers at the time of surgery. The surgery was performed using the 60 kHz IntraLase FS system (IntraLase Corp., Irvine, CA) under topical anesthesia (proparacaine 0.5%), with the laser adapted to make therapeutic cuts. The eyelids were prepared with Betadine sponges. With the use of a sterile marking pen, the graft– host junction was marked in the steep and flat axis, because this helped to center the incisions on the graft. The IntraLase limbal suction ring was then applied, and the cone was positioned. Applanation was judged as adequate
if the fluid meniscus was at least beyond the graft– host junction. There were no suction breaks during treatment in any case. Once complete, suction was released, and the ring was removed. Treatment parameters were initially based on the topographic location and radial extent of the steep meridian. After experiencing early overcorrections within the first 10 cases, the treatment nomogram was altered. Thereafter, up to 6 D of cylinder was treated with 40 to 60-degree arc length, 6 to 10 D with 65 to 75-degree arc length, and greater than 10 D with 90-degree arc length. After surgery, patients were treated with topical tobramycin and dexamethasone (TobraDex; Alcon, Mississauga, Ontario, Canada) 4 times daily for 1 week. Thereafter, they were placed on a maintenance dose of topical steroid according to their individual requirements.
Analysis of Astigmatism Astigmatism analysis has been outlined in previous publications.18,22 Because astigmatism traverses an entire cycle in 180 degrees, the doubled-angle polar plot was used for analysis. The doubled-angle plot has the 0 and 180-degree axes at the same location (Fig 1). Procedures that are astigmatically neutral have the centroid of the surgically induced cylinder data at the centre of the plot, the “null point.” Surgically induced refractive change (SIRC) was assessed by 2 methods: (1) change in absolute cylinder value and (2) vector analysis, calculating the mean pre- and postoperative centroids.22 For standard descriptive statistics (means, standard deviations) to be applied correctly, each data point must be converted to an x–y coordinate system. Therefore astigmatism data were converted from polar coordinates (axis, cylinder) to Cartesian coordinates (x and y values) using the following equations: x ⫽ CylinderⴱCos 共2ⴱaxis兲 y ⫽ CylinderⴱSin 共2ⴱaxis兲 In the formulas, the angle of the axis of astigmatism is doubled to give the correct x and y values. The centroid, or mean of a set of x and y values, is calculated by independently finding the mean of each variable (xm, ym). The centroid (mean astigmatism) is then converted back to standard polar notation:
Figure 1. Postoperative versus preoperative DEQ. The area above the solid line represents a worsened DEQ, and the area beneath the solid line represents an improved DEQ. Dashed line represents linear regression, and the solid line represents no change in DEQ. As can be seen, the linear regression is shifted toward improved DEQ. D ⫽ diopters; DEQ ⫽ defocus equivalent.
1229
Ophthalmology
Volume 117, Number 6, June 2010
Table 1. Indications for Penetrating Keratoplasty Diagnosis
No. (%)
Keratoconus Pseudophakic bullous keratopathy Corneal scar Fuchs’ dystrophy Other
19 (51) 8 (22) 6 (16) 3 (8) 1 (3)
values for the DEQ correspond to a smaller blur circle; consequently, a lower postoperative than preoperative DEQ is ideal. It has been used as a measure of surgical success and is thought to correlate well with UCVA.22,23 The DEQ is calculated as follows: DEQ ⫽ magnitude of spherical equivalent ⫹ ½ 共magnitude of cylinder兲 .
Statistical Analysis Cylinder ⫽ ⫹ Angle ⫽ 1/2ⴱArc tan (ym ⁄ xm) (x2m
y2m)0.5
(If xm and ym⬎0 then axis ⫽ angle, if xm⬍0 then axis ⫽ angle⫹90 degrees, if xm⬎0 and ym⬍0 then axis ⫽ angle ⫹180 degrees.) The standard deviation of the mean astigmatism is an elliptical area surrounding the centroid. The axes of the ellipse are twice the standard deviations of the x and y values (sx and sy). On a doubled-angle plot, the y-axis is coincident with the axis of oblique astigmatism and the x-axis is coincident with the axis of orthogonal astigmatism. Populations having with or against the rule astigmatism have ellipses that are oriented horizontally. Populations with a higher proportion of oblique astigmatism have ellipses that are oriented vertically.22 A modification of the doubled-angle plot was made and described as “on-axis analysis.” The horizontal axis is defined as the original axis of each patient. All preoperative data are set at zero and aligned on this axis. The postoperative axis is then calculated in relation to the preoperative axis. In this modified plot, mean pre- and postoperative astigmatism vectors were calculated and plotted as usual. The preoperative standard deviation is a line on the horizontal axis, whereas the postoperative standard deviation is an ellipse. If the surgical procedure affected only the axis that was intended to treat, then the postoperative vectors should be aligned on the horizontal axis. However, if the surgical procedure induced vectors in directions other than the intended one, the postoperative vectors would deviate from the horizontal axis. The defocus equivalent (DEQ) is proportional to the diameter of the blur circle on the retina for a given pupil size. Lower
A paired t test was used to compare continuous variables (SPSS software version 13; SPSS Inc., Chicago, IL). Probabilities of less than 5% were considered significant (P⬍0.05), and probabilities less than 1% were considered to be highly significant (P⬍0.01).
Results Mean follow-up after IEAK was 7.2⫾3.5 months (range 4 –15 months). The indications for the original penetrating keratoplasty are summarized in Table 1. A comparison of pre- and postoperative outcome measures is summarized in Table 2. Detailed data of cases included in outcome measurement analysis are shown in Table 3 (available at http://aaojournal.org). Uncorrected visual acuity expressed in the logarithm of the minimum angle of resolution improved significantly from 1.08⫾0.34 preoperatively to 0.80⫾0.42 postoperatively (P⫽0.0016, paired t test). Best-corrected visual acuity (logarithm of the minimum angle of resolution) improved significantly from 0.45⫾0.27 preoperatively to 0.37⫾0.27 postoperatively (P⫽0.018). Although the mean spherical equivalent did not significantly change (P⫽ 0.99), the DEQ was significantly reduced by more than 1 D (from 8.5⫾3.5 D to 7.2⫾4.59 D, P⫽0.025). Mean absolute astigmatism was reduced from 7.46⫾2.70 preoperatively to 4.77⫾3.29 postoperatively (P⫽0.0001), a highly significant result. The mean astigmatism vector was reduced from 2.52⫻122 degrees ⫾ 5.4 to 0.41⫻126⫾4.0 degrees (P⫽0.07). Both absolute orthogonal and oblique astigmatism were significantly reduced (P⫽0.005 and P⫽0.04, respectively). Total high order aberrations were increased
Table 2. Comparison of Before and After IntraLase-Enabled Astigmatic Keratotomy Visual Parameters (⫾ Standard Deviation) Parameter
Pre-IEAK
Post-IEAK
P Value (Paired t Test)
BCVA (logMAR) UCVA (logMAR) Spherical equivalent DEQ (D) Absolute astigmatism (D) Mean astigmatism vector (D) Absolute orthogonal astigmatism (D) Absolute oblique astigmatism (D) Aberrometry (mm) High order aberrations Tilt aberrations Coma-like aberrations Trefoil aberrations Quatrefoil aberrations Spherical aberrations
0.45⫾0.27 1.08⫾0.34 ⫺4.01⫾3.85 8.49⫾3.52 7.46⫾2.70 2.52⫻122⫾5.4 4.7⫾2.9 4.4⫾3.5
0.37⫾0.27 0.80⫾0.42 ⫺3.95⫾4.69 7.23⫾4.49 4.77⫾3.29 0.41⫻126⫾4.0 2.8⫾2.1 3.2⫾3.4
0.018 0.0016 0.99 0.025 0.0001 0.07 0.005 0.04
2.8⫾1.7 3.6⫾2.6 1.4⫾0.9 1.6⫾1.1 0.8⫾0.8 0.6⫾0.5
4.2⫾3.5 4.0⫾3.0 1.8⫾1.3 2.7⫾2.5 1.4⫾1.4 0.8⫾0.7
0.01 0.41 0.022 0.01 0.038 0.005
BCVA ⫽ best-corrected visual acuity; D ⫽ diopters; DEQ ⫽ defocus equivalent; IEAK ⫽ IntraLase-enabled astigmatic keratotomy; logMAR ⫽ logarithm of the minimum angle of resolution; UCVA ⫽ uncorrected visual acuity.
1230
Kumar et al 䡠 IntraLase-Enabled Astigmatic Keratotomy Vector Analysis
Figure 2. Mean absolute astigmatism plotted against time after IEAK. The mean astigmatism seems to be stable throughout the follow-up period (6 months). IEAK ⫽ IntraLase-enabled astigmatic keratotomy.
from 2.8⫾1.7 to 4.2⫾3.5 (P⫽0.01). When investigated separately, each of the higher-order aberrations was significantly increased with the exception of tilt aberrations. Figure 1 outlines the comparison between pre- and postoperative DEQ for each case. A trend toward an improved, lower value for the DEQ postoperatively is seen with the mean effect (linear regression) depicted as a dotted line. Twenty-two eyes (65%) had a lower DEQ postoperatively, 4 eyes (11%) experienced no change in DEQ, and 8 eyes (24%) had a higher DEQ postoperatively. Of the 8 eyes with a higher DEQ after the procedure, 4 experienced an improvement in UCVA and BCVA, 3 were unchanged, and 1 lost 1 line of UCVA and BCVA. Figure 2 shows the mean reduction in astigmatism in diopters plotted against time after treatment. The effect of the astigmatic keratotomy is seen to reach its maximum effect 1.5 months after treatment, with slight regression and then stabilization between 3 and 6 months.
Figure 3. Achieved orthogonal astigmatic correction plotted against intended orthogonal astigmatic correction. The dotted line represents an exact correlation between intended and achieved correction. The solid line represents the mean effect of all procedures (linear regression) and closely superimposes the dotted line. Dashed lines represent 95% confidence interval. D ⫽ diopters.
Figure 4. Achieved oblique astigmatic correction plotted against intended oblique astigmatic correction. The dotted line represents an exact correlation between intended and achieved correction. The solid line represents the mean effect of all procedures (linear regression). Dashed lines represent 95% confidence interval. D ⫽ diopters.
Figure 3 shows the achieved astigmatic correction and the intended astigmatic correction for orthogonal astigmatism. A precise relationship between achieved and intended correction is established. Figure 4 shows the achieved and intended astigmatic correction for oblique astigmatism. Overall, a slight overcorrection was achieved for high minus cylinder and a slight undercorrection was achieved for high plus cylinder. However, for lower degrees of astigmatism, a precise relationship between the achieved and the intended astigmatism is again established. Figure 5 depicts the doubled-angled polar plot. The centroid of the preoperative data is displaced from the null point. The standard
Figure 5. Double-angled polar plot of the preoperative and postoperative refractive astigmatism. The ellipses signify 1 standard deviation around the mean astigmatism vector (stars). The mean astigmatism vector is brought closer to the null position by IEAK, and the standard deviation ellipse is reduced. IEAK ⫽ IntraLase-enabled astigmatic keratotomy.
1231
Ophthalmology
Volume 117, Number 6, June 2010 rection occurred in 9 eyes (24%). Six of the overcorrected eyes had originally undergone penetrating keratoplasty for keratoconus, 2 for corneal scarring and 1 for pseudophakic bullous keratopathy. Three eyes were resutured because of early overcorrection and were excluded from outcome measurement analysis. They were treated for 5, 9, and 12 D of cylinder with arc sizes of 60, 75, and 90 degrees, respectively.
Discussion
Figure 6. A modified doubled-angled polar plot, termed “on-axis analysis.” The horizontal axis is defined as the original axis of each patient. The postoperative axis is then calculated in relation to the preoperative axis. The ellipse signifies 1 standard deviation around the mean astigmatism vector (star). Although the mean postoperative astigmatism (grey star) is brought closer to the null point, additional vectors seem to be induced in other directions. IEAK ⫽ IntraLase-enabled astigmatic keratotomy; Pre-op ⫽ preoperative.
deviation ellipse is oriented horizontally and to the left of the null point, indicating a population with predominantly with-the-rule astigmatism. Figure 6 is a modified doubled-angled polar plot depicting “on-axis analysis.” By definition, the mean preoperative values are on the horizontal axis. Although the mean postoperative astigmatism vector is brought closer to the null point both horizontally (on-axis correction) and vertically (off-axis–induced astigmatism), the postoperative individual values deviate from the horizontal axis, implying that vectors have been induced in other directions. The mean postoperative astigmatism vector is close to the null point vertically and horizontally; therefore on average the procedure corrected the astigmatism on the intended axis without intentionally inducing cylinder in any specific direction. Figure 7 compares change in total higher-order aberrations with arc size. There is an increased change in higher-order aberrations with increased arc length, but this is not statistically significant (correlation coefficient 0.12).
The femtosecond laser creates arcuate incision length, depth, radius, and symmetry with a precision that was previously unable to attain using manual techniques. Despite this, the effect of the IntraLase-enabled astigmatic keratotomy incisions is still subject to the same vagaries of corneal wound healing that modulate the initial response. Keratotomy healing begins with an epithelial plug that is eventually replaced by hypercellular scar tissue, explaining delayed changes in corneal curvature.24 Therefore, although it is necessary to investigate measurement outcomes, it is equally important to measure the rate of change of effect, duration of effect, stability, and predictability of the intervention. In this study we aimed to look at the results of IEAK after the wound has stabilized (⬃6 months postoperatively). Our results indicate that IEAK is effective in reducing high astigmatism in corneal grafts. Both UCVA and BCVA were significantly improved. The DEQ, a measure of surgical success, and the absolute value of astigmatism were significantly reduced. Such findings are similar in recent studies.19,20 There was no significant change in spherical equivalent. This is because the flattening produced by the incisions in the steep axis is countered by the steepening of the unincised meridian, known as the “coupling effect.” When there is little or no change in spherical equivalent, the corneal steepening/flattening coupling ratio is close to 1.25 This has
Efficacy Twenty-four eyes (71%) gained 1 or more lines of vision, 7 eyes (21%) remained unchanged, and 3 eyes (8%) lost 1 line of vision. The efficacy index, calculated as the ratio of the mean postoperative BCVA to the mean preoperative UCVA, was 0.6⫾0.6.
Complications There were no cases of perforation, wound dehiscence, or infectious keratitis. Three eyes experienced an episode of graft rejection (8%), and each case resolved on topical steroid therapy. Overcor-
1232
Figure 7. Change in total higher-order aberrations (microns) plotted against arc size (degrees). There is a slight correlation between increased arc size and a greater change in total higher-order aberrations.
Kumar et al 䡠 IntraLase-Enabled Astigmatic Keratotomy Vector Analysis been reported in several other studies analyzing the effect of relaxing incisions.26 –32 The mean effect of all procedures was to reduce the DEQ, thereby reducing the size of the blur circle. In Figure 1, those eyes below the solid line represent an improved and reduced postoperative DEQ, and those eyes above the solid line represent a worsened and increased postoperative DEQ. Despite this, those cases above the line experienced a trend toward improved UCVA and BCVA. The mean absolute value of astigmatism decreased from 7.46⫾2.70 D to 4.77⫾3.29 D over the first 3 months postoperatively (Fig 2). The initial reduction to 4.00 D occurred within the first 6 weeks, followed by a slight regression of the effect at 3 months. Between 3 and 6 months, there was no change in effect. This suggests that by 3 months postprocedure, the reduction in astigmatism has stabilized. Further follow-up data are required to establish the duration of effect. An almost identical trend in the change in astigmatism after femtosecond-enabled astigmatic keratotomy for high post-keratoplasty astigmatism has been reported.20 The predictability of the procedure may be defined as how closely the intended correction correlates with achieved correction. For orthogonal astigmatism, the mean achieved reduction in cylinder closely approximates the intended reduction (Fig 3) at low and high levels of astigmatism. For oblique astigmatism, the mean achieved reduction in cylinder closely approximates the intended reduction for lower levels of astigmatism; however, the effect tends toward undercorrection at higher levels (Fig 4). This decrease in refractive predictability with higher corrections has been reported.22 Overall, the achieved reduction correlated well with the intended reduction in astigmatism. However, because 3 eyes were excluded from outcome measurement analysis as the result of early resuturing, our results may be marginally superior when compared with the actual outcome of all cases. Figure 5 shows that the mean astigmatism vector is brought closer to the null point by IEAK. This is consistent with previous findings.18 In addition, the position, size, and orientation of the standard deviation ellipse are all altered by the procedure. The position of the standard deviation ellipse is shifted to the right because there is less with the rule astigmatism. The standard deviation ellipse is smaller because there is a reduction in mean absolute astigmatism. Also, the orientation of the ellipse has changed from horizontal to vertical. This signifies that with the rule astigmatism is predominant preoperatively, whereas oblique astigmatism is predominant postoperatively. By using on-axis analysis (Fig 6), we noted that although the mean astigmatism vector is well corrected by IEAK, new vectors seem to have been created in other directions. These vectors seem to be created at random, because their average is close to the null point. Another possibility is that after penetrating keratoplasty, several vectors are present in the graft (induced by tight sutures and asymmetric healing), and once the main astigmatism vector is reduced by IEAK, the other smaller vectors that were already present become apparent. In such cases,
additional more limited IEAK procedures might help further improve corneal astigmatism. There were no cases of perforation, graft– host junction dehiscence, or infectious keratitis. This compares favorably with prior reports on manual astigmatic keratotomy.2,26 Complications of the procedure included an episode of allograft rejection in 3 cases (8%). Each case was treated with topical corticosteroids and resolved without further incident. The rate of graft rejection after IEAK seems to be higher than that previously found in series of manual astigmatic keratotomy.2,26 Overcorrection occurred in 9 cases (24%). On further analysis of the overcorrected cases, 6 of the 9 occurred during the initial period when arc length was determined by analyzing the location and extent of the steepest meridians on the topographic map. Thereafter the protocol was altered as outlined above, such that the rate of overcorrection decreased dramatically. There were 3 cases of overcorrection in the subsequent 27 eyes. Overall, two thirds of the overcorrected eyes had originally undergone penetrating keratoplasty for keratoconus. It is plausible that ectatic eyes are more at risk for overcorrection. Our data indicate that IEAK significantly reduces the mean absolute astigmatism and improves UCVA, BCVA, and DEQ, and is thus effective. The corrections of orthogonal and oblique astigmatism are achieved with a high degree of refractive predictability. The refractive effect stabilizes and reaches a plateau between 3 and 6 months postoperatively. IntraLase-enabled astigmatic keratotomy adversely affects higher-order aberrations. There is also a greater predominance of oblique astigmatism in the postoperative population. These factors may diminish the beneficial effects of the procedure and may explain the smaller improvement in BCVA when compared with UCVA. Patients may experience greater glare and reduced contrast sensitivity. A similar adverse effect on higher-order aberrations has been reported in a cohort of manual astigmatic keratotomy.18 The present study attempted to determine the effect of IEAK over a mean period of 6 months. Consequently, those eyes with significant early overcorrection that required resuturing within the first 6 weeks could not be included in the outcome measurement analysis. Although only 3 of the eyes initially enrolled in the study were in this category, this represents a selection bias and favorably affects the outcome measurement results. These 3 eyes were included in the calculation of complication rates. We believe that incisions must be performed in the donor because the graft– host junction provides a new functional limbus and has an unpredictable thickness.33 We currently recommend customizing arc length according to the degree of astigmatism that requires correction, as outlined previously. If there is less than 6 D of post-penetrating keratoplasty astigmatism, either photorefractive keratectomy or a limited IEAK with photorefractive keratectomy to reduce the amount of tissue removed by photorefractive keratectomy alone may be the optimal choice. In conclusion, this study showed that IEAK has an encouraging refractive predictability. The effect has been shown to stabilize by 3 months and reach a plateau between
1233
Ophthalmology
Volume 117, Number 6, June 2010
3 and 6 months. The efficacy index was similar to previous findings for IEAK and favorable when compared with a previous cohort of manual astigmatic keratotomy.18 The allograft rejection rate was higher than expected, and although these episodes were readily treated and resolved on topical corticosteroids, this warrants further scrutiny in future studies. The overcorrection rate of IEAK was also high, but decreased significantly on use of a more conservative treatment regimen as outlined. The precise length, curvature, depth, and symmetry of the IEAK incisions may afford a greater effect than a similar arc length produced by a manual technique. However, financial considerations and laser availability must be considered when choosing between IntraLase-enabled and manual astigmatic keratotomy. As more data become available, it is necessary to refine the treatment nomogram for IntraLase-enabled astigmatic keratotomy for high astigmatism after penetrating keratoplasty.
References 1. Williams KA, Roder D, Esterman A, et al. Factors predictive of corneal graft survival: report from the Australian Corneal Graft Registry. Ophthalmology 1992;99:403–14. 2. Poole TR, Ficker LA. Astigmatic keratotomy for post-keratoplasty astigmatism. J Cataract Refract Surg 2006;32:1175–9. 3. Bochmann F, Schipper I. Correction of post-keratoplasty astigmatism with keratotomies in the host cornea. J Cataract Refract Surg 2006;32:923– 8. 4. Vajpayee RB, Sharma N, Sinha R, et al. Laser in-situ keratomileusis after penetrating keratoplasty. Surv Ophthalmol 2003;48:503–14. 5. Genvert GI, Cohen EJ, Arentsen JJ, Laibson PR. Fitting gas permeable contact lenses after penetrating keratoplasty. Am J Ophthalmol 1989;99:511– 4. 6. Forseto AS, Francesconi CM, Nosé RA, Nosé W. Laser in situ keratomileusis to correct refractive errors after keratoplasty. J Cataract Refract Surg 1999;25:479 – 85. 7. Hardten DR, Lindstrom RL. Surgical correction of refractive errors after penetrating keratoplasty. Int Ophthalmol Clin 1997;37:1–35 8. Troutman RC. Corneal wedge resections and relaxing incisions for postkeratoplasty astigmatism. Int Ophthalmol Clin 1983;23:161– 8. 9. Arffa RC. Results of a graded relaxing incision technique for postkeratoplasty astigmatism. Ophthalmic Surg 1988;19: 624 – 8. 10. Lugo M, Donnenfeld ED, Arentson JJ. Corneal wedge resection for high astigmatism following penetrating keratoplasty. Ophthalmic Surg 1987;18:650 –3. 11. Gothard TW, Agapitos PJ, Bowers RA, et al. Four incision radial keratotomy for high myopia after penetrating keratoplasty. J Refract Corneal Surg 1993;9:51–7. 12. John ME, Martines E, Cvintal T, et al. Photorefractive keratectomy following penetrating keratoplasty. J Refract Corneal Surg 1994;10(suppl):S206 –10. 13. Dada T, Vajpayee RB. Laser in situ keratomileusis after PKP [letter]. J Cataract Refract Surg 2002;28:7– 8. 14. Slade SG. Applications of the femtosecond laser in corneal surgery. Curr Opin Ophthalmol 2007;18:338 – 41.
1234
15. Buratto L, Bohm E. The use of the femtosecond laser in penetrating keratoplasty. Am J Ophthalmol 2007;143:737– 42. 16. Harissi-Dagher M, Azar DT. Femtosecond laser astigmatic keratotomy for postkeratoplasty astigmatism. Can J Ophthalmol 2008;43:367–9. 17. Kiraly L, Herrmann C, Amm M, Duncker G. Reduction of astigmatism by arcuate incisions using the femtosecond laser after corneal transplantation [in German]. Klin Monatsbl Augenheilkd 2008;225:70 – 4. 18. Bahar I, Levinger E, Kaiserman I, et al. Intralase-enabled astigmatic keratotomy for postkeratoplasty astigmatism. Am J Ophthalmol 2008;146:897–904. 19. Hoffart L, Proust H, Matonti F, et al. Correction of postkeratoplasty astigmatism by femtosecond laser compared with mechanized astigmatic keratotomy. Am J Ophthalmol 2009; 147:779 – 87. 20. Nubile M, Carpineto P, Lanzini M, et al. Femtosecond laser arcuate keratotomy for the correction of high astigmatism after keratoplasty. Ophthalmology 2009;116:1083–92. 21. Koch DD, Kohnen T, Obstbaum SA, Rosen ES. Format for reporting refractive surgical data. J Cataract Refract Surg 1998;24:285–7. 22. Holladay JT, Moran JR, Kezirian GM. Analysis of aggregate surgically induced refractive change, prediction error, and intraocular astigmatism. J Cataract Refract Surg 2001;27: 61–79. 23. Holladay JT, Lynn MJ, Waring GO III, et al. The relationship of visual acuity, refractive error, and pupil size after radial keratotomy. Arch Ophthalmol 1991;109:70 – 6. 24. Eiferman RA, Schultz GS, Nordquist RE, Waring GO III. Corneal wound healing and its pharmacologic modification after refractive keratotomy. In: Waring GO III, ed. Refractive Keratotomy for Myopia and Astigmatism. St. Louis, MO: Mosby; 1992;749 –79. 25. Wilkins MR, Mehta JS, Larkin FP. Standardized arcuate keratotomy for postkeratoplasty astigmatism. J Cataract Refract Surg 2005;31:297–301. 26. Fronterre A, Portesani GP. Relaxing incisions for postkeratoplasty astigmatism. Cornea 1991;10:305–11. 27. Krachmer JH, Fenzl RE. Surgical correction of high postkeratoplasty astigmatism: relaxing incisions vs wedge resection. Arch Ophthalmol 1980;98:1400 –2. 28. Hjortdal JO, Ehlers N. Paired arcuate keratotomy for congenital and postkeratoplasty astigmatism. Acta Ophthalmol Scand 1998;76:138 – 41. 29. Cohen KL, Tripoli NK, Noecker RJ. Prospective analysis of photokeratoscopy for arcuate keratotomy to reduce postkeratoplasty astigmatism. J Refract Corneal Surg 1989;5:388 –93. 30. Koay YP, McGhee CN, Crawford GJ. Effect of a standard paired arcuate incision and augmentation sutures on postkeratoplasty astigmatism. J Cataract Refract Surg 2000;26: 553– 61. 31. Akura J, Matsuura K, Hatta S, et al. A new concept for the correction of astigmatism: full-arc, depth-dependent astigmatic keratotomy. Ophthalmology 2000;107:95–104. 32. Price FW, Grene RB, Marks RG, Gonzales JS, ARC-T Study Group. Astigmatism reduction clinical trial: a multicenter prospective evaluation of the predictability of arcuate keratotomy; evaluation of surgical nomogram predictability. Arch Ophthalmol 1995;113:277– 82. 33. Kaiserman I, Bahar I, Rootman DS. Corneal wound malapposition after penetrating keratoplasty: an optical coherence tomography study. Br J Ophthalmol 2008;92:1103–7.
Kumar et al 䡠 IntraLase-Enabled Astigmatic Keratotomy Vector Analysis
Footnotes and Financial Disclosures Originally received: March 22, 2009. Final revision: October 24, 2009. Accepted: October 26, 2009. Available online: February 16, 2010.
Supported as provision of material supplies from AMO Inc., Irvine, California. Manuscript no. 2009-398.
Department of Ophthalmology and Vision Sciences, Toronto Western Hospital, University of Toronto, Ontario, Canada. Financial Disclosure(s): The author(s) have made the following disclosure(s): David S. Rootman received funding as stated above.
Correspondence: Nikhil L. Kumar, BMed, FRANZCO, Department of Ophthalmology and Vision Sciences, Toronto Western Hospital, University of Toronto, 399 Bathurst St., Toronto, Ontario, Canada M5T2S8. E-mail: nik_kumar1@ yahoo.com.
1235
Ophthalmology
Volume 117, Number 6, June 2010
Table 3. Detailed Data of Cases Included in Outcome Measurement Analysis No.
Age
Diagnosis
Duration between PKP and IEAK (yrs)
Preoperative Manifest Refraction
Postoperative Manifest Refraction
Post IEAK Follow-up (mos)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
76 37 21 70 62 70 28 58 62 51 40 41 41 58 59 59 70 72 75 48 44 67 43 65 53 44 38 77 62 74 51 21 49 85
FED KC KC PBK SCAR SCAR KC SCAR SCAR KC KC KC KC PBK KC KC SCAR PBK PBK KC KC PBK SCAR FED ICE KC KC PBK FED PBK KC KC KC PBK
9 8 1.5 1.5 12 8 1.5 Data absent 5 30 5 8 8 4 4 Data absent 2 1.5 8 1.5 3.5 Data absent 5 5 1.5 1.5 5 2 5 Data absent 17 1.5 17 3
5.75–7.75*65 1.75–5.0*178 7.0–8.0*60 ⫺0.75 to 7.0*86 ⫺3.00 to 12.5*18 3.50–6.0*55 ⫺4.25 to 12.0*44 2.75–5.5*55 ⫺7.0 to 5.5*32 1.75–8.0*20 7.0–10.0*112 ⫺1.0 to 6.5*134 ⫺2.0 to 6.0*91 1.0–5.5*19 ⫺3.0 to 5.0*157 ⫺2.25 to 6.5*176 ⫺2.0–11.0*36 0.0–12.0*3 5.5–7.0*90 ⫺1.5 to 11.0*84 0.0–6.0*31 ⫺5.0 to 5.0*100 8.75–4.5*30 0.0–6.0*82 ⫺1.5 to 5.5*20 6.0–5.0*13 ⫺4.5 to 5.0*177 0.0–8.25*40 0.5–6.25*167 0.0–10.0*160 ⫺1.75 to 11.25*156 ⫺2.0 to 5.5*15 0.0–13.0*44 ⫺2.25 to 7.0*17
0.00–1.50*70 ⫺1.0 ⫺7.5 to 4.0*95 ⫺1.5 to 2.5*135 10.5–1.75*90 1.0–4.0*55 ⫺10.0 to 5.75*130 2.5–2.5*28 ⫺2.0 to 3.25*51 ⫺7.0 to 6.0*110 ⫺0.5 to 7.5*81 ⫺2.0 to 5.0*171 0.0–4.0*30 ⫺4.5 to 0.75*88 ⫺5.0 to 2.0*46 ⫺7.0 to 6.0*85 4.5–9.0*61 ⫺2.5 to 5.25*139 2.0–4.5*10 ⫺4.5 to 2.5*115 ⫺1.25 to 0.75*9 ⫺6.25 to 6.0*120 ⫺9.5 to 4.5*90 ⫺1.5 to 3.0*79 ⫺1.25 to 5.0*20 2.0–1.5*180 0.5–2.75*78 ⫺0.25 to 4.0*1 ⫺1.0 to 4.75*145 ⫺1.5 to 7.5*155 ⫺9.5 to 6.0*80 ⫺5.0 to 2.0*180 ⫺1.75 to 14.0*40 ⫺1.0 to 0.5*80
6 4 4 7 12 4 4 4 15 5 6 11 11 5 4 4 9 6 9 6 9 4 5 10 6 9 14 10 6 4 8 7 15 4
CX
OC OC
OC
OC OC R OC
R
R
CX ⫽ complications; FED ⫽ Fuchs’ endothelial dystrophy; ICE ⫽ iridocorneal endothelial syndrome; IEAK ⫽ IntraLase-enabled astigmatic keratotomy; KC ⫽ keratoconus; OC ⫽ overcorrected; PBK ⫽ pseudophakic bullous keratopathy; PKP ⫽ penetrating keratoplasty; R ⫽ graft rejection episode; SCAR ⫽ visually significant corneal scar.
1235.e1
Visual outcomes after penetrating keratoplasty are often limited by significant astigmatism, ametropia, and anisometropia. More than 5 diopters (D) of astigmatism are common after full-thickness corneal transplantation.1– 4 Conservative therapeutic options in the setting of high astigmatism are difficult to use. Spectacle use may lead to intolerable aniseikonia. Contact lens fitting is challenging over a clear graft with topographic alterations;5 thus, a conservative approach does not always yield maximal visual acuity.6 Previously, surgical options for astigmatic reduction were limited to continuous suture adjustment, selective suture removal, relaxing incisions with or without compression sutures, manual astigmatic keratotomy, wedge resections, photorefractive keratectomy, and LASIK.2,4,7–13 Suture manipulation is possible within the first year. Relaxing incisions, manual astigmatic keratotomy, and wedge resections have been associated with unpredictable results, wound gape, and occasionally perforation. LASIK has been reported to give stable and predictable results after pene-
1228
© 2010 by the American Academy of Ophthalmology Published by Elsevier Inc.
trating keratoplasty but has limited capacity for astigmatic reduction.4,6 There has been great enthusiasm for the application of femtosecond laser technology to improve patient outcomes and decrease complications in penetrating keratoplasty.14 –16 The use of femtosecond laser technology to create astigmatic keratotomy incisions has recently been reported.16 –20 This technique has the advantage of greater precision in arc depth, length, and curvature. We previously compared the results of manual astigmatic keratotomy with IntraLase-enabled astigmatic keratotomy (IEAK).18 IntraLase-enabled astigmatic keratotomy significantly improved both uncorrected visual acuity (UCVA) and best-corrected visual acuity (BCVA) and brought the cylinder vector closer to neutral. Manual astigmatic keratotomy exhibited a trend toward an improvement in UCVA and BCVA but overcorrected the cylinder vector. IntraLase-enabled astigmatic keratotomy had a lesser detrimental effect on higher-order aberrations than manual astigmatic keratotomy. ISSN 0161-6420/10/$–see front matter doi:10.1016/j.ophtha.2009.10.041
Kumar et al 䡠 IntraLase-Enabled Astigmatic Keratotomy Vector Analysis To date, there is a paucity of data on the predictability, rate of change, and duration of effect, efficacy, and complication rate of this procedure beyond 1 month. In this retrospective pilot study, we report these in a series of 37 eyes with a mean follow-up of 7 months.
Materials and Methods Study Design A retrospective review of 37 eyes of 35 patients who had undergone IEAK for high astigmatism was completed. There were 21 male and 14 female patients in the study. Approval for the study was obtained from the University Health Network Research Ethics Board to access retrospective data. Patients with at least 5 D of post-keratoplasty regular astigmatism were included. All were at least 18 months post-keratoplasty, and no residual sutures were in place when IEAK was performed. Patients with high irregular astigmatism were excluded. Irregular astigmatism was calibrated by analysis of the wavefront OPD and Zernike decomposition of the higher-order aberrations (NIDEK, OPD Scan II ARK 10000, Gamagori, Japan). Thirty-four eyes of 32 patients had a minimum of 4 months follow-up and were included in the outcome measurement analysis. Three eyes of 3 patients required early resuturing because of overcorrection, within 1 month of the incisions. These were excluded from outcome measurement analysis, because it was impossible to determine the effect of the original incisions. Their data were included in the calculation of complication rates. Outcome measures included UCVA, BCVA, manifest refraction, corneal topography, and higher-order aberrations. Patients were followed at 1 day, 1 week, 1 month, 3 months, 6 months, and 12 months. At each visit pre- and postoperatively, slit-lamp examination and tonometry were performed and outcome measures were recorded. For outcome measurement analysis, both interval and last visit data were used. The efficacy index was calculated as the ratio of the mean postoperative UCVA to the mean preoperative BCVA.21
Surgical Procedure Preoperative planning included assessment of the amount and axis of corneal astigmatism, corneal thickness, and graft size. When there was disagreement between the axis of the manifest refraction and the topography, we favored the topography axis. The amount of topographic cylinder rather than the manifest cylinder was used to determine the length (degrees) of the keratotomy. Most patients also had imaging with the Pentacam rotating Scheimpflug camera (Oculus, Wetzlar, Germany). All treatments were paired symmetric (same length) incisions centered on the steep axis. The depth of the incisions was set at approximately 90% depth, based on either the Pentacam or ultrasound pachymetry at the incision location. Each incision was made 0.5 mm within the graft– host junction, such that the diameter was set at 1 mm less than the graft diameter measured by calipers at the time of surgery. The surgery was performed using the 60 kHz IntraLase FS system (IntraLase Corp., Irvine, CA) under topical anesthesia (proparacaine 0.5%), with the laser adapted to make therapeutic cuts. The eyelids were prepared with Betadine sponges. With the use of a sterile marking pen, the graft– host junction was marked in the steep and flat axis, because this helped to center the incisions on the graft. The IntraLase limbal suction ring was then applied, and the cone was positioned. Applanation was judged as adequate
if the fluid meniscus was at least beyond the graft– host junction. There were no suction breaks during treatment in any case. Once complete, suction was released, and the ring was removed. Treatment parameters were initially based on the topographic location and radial extent of the steep meridian. After experiencing early overcorrections within the first 10 cases, the treatment nomogram was altered. Thereafter, up to 6 D of cylinder was treated with 40 to 60-degree arc length, 6 to 10 D with 65 to 75-degree arc length, and greater than 10 D with 90-degree arc length. After surgery, patients were treated with topical tobramycin and dexamethasone (TobraDex; Alcon, Mississauga, Ontario, Canada) 4 times daily for 1 week. Thereafter, they were placed on a maintenance dose of topical steroid according to their individual requirements.
Analysis of Astigmatism Astigmatism analysis has been outlined in previous publications.18,22 Because astigmatism traverses an entire cycle in 180 degrees, the doubled-angle polar plot was used for analysis. The doubled-angle plot has the 0 and 180-degree axes at the same location (Fig 1). Procedures that are astigmatically neutral have the centroid of the surgically induced cylinder data at the centre of the plot, the “null point.” Surgically induced refractive change (SIRC) was assessed by 2 methods: (1) change in absolute cylinder value and (2) vector analysis, calculating the mean pre- and postoperative centroids.22 For standard descriptive statistics (means, standard deviations) to be applied correctly, each data point must be converted to an x–y coordinate system. Therefore astigmatism data were converted from polar coordinates (axis, cylinder) to Cartesian coordinates (x and y values) using the following equations: x ⫽ CylinderⴱCos 共2ⴱaxis兲 y ⫽ CylinderⴱSin 共2ⴱaxis兲 In the formulas, the angle of the axis of astigmatism is doubled to give the correct x and y values. The centroid, or mean of a set of x and y values, is calculated by independently finding the mean of each variable (xm, ym). The centroid (mean astigmatism) is then converted back to standard polar notation:
Figure 1. Postoperative versus preoperative DEQ. The area above the solid line represents a worsened DEQ, and the area beneath the solid line represents an improved DEQ. Dashed line represents linear regression, and the solid line represents no change in DEQ. As can be seen, the linear regression is shifted toward improved DEQ. D ⫽ diopters; DEQ ⫽ defocus equivalent.
1229
Ophthalmology
Volume 117, Number 6, June 2010
Table 1. Indications for Penetrating Keratoplasty Diagnosis
No. (%)
Keratoconus Pseudophakic bullous keratopathy Corneal scar Fuchs’ dystrophy Other
19 (51) 8 (22) 6 (16) 3 (8) 1 (3)
values for the DEQ correspond to a smaller blur circle; consequently, a lower postoperative than preoperative DEQ is ideal. It has been used as a measure of surgical success and is thought to correlate well with UCVA.22,23 The DEQ is calculated as follows: DEQ ⫽ magnitude of spherical equivalent ⫹ ½ 共magnitude of cylinder兲 .
Statistical Analysis Cylinder ⫽ ⫹ Angle ⫽ 1/2ⴱArc tan (ym ⁄ xm) (x2m
y2m)0.5
(If xm and ym⬎0 then axis ⫽ angle, if xm⬍0 then axis ⫽ angle⫹90 degrees, if xm⬎0 and ym⬍0 then axis ⫽ angle ⫹180 degrees.) The standard deviation of the mean astigmatism is an elliptical area surrounding the centroid. The axes of the ellipse are twice the standard deviations of the x and y values (sx and sy). On a doubled-angle plot, the y-axis is coincident with the axis of oblique astigmatism and the x-axis is coincident with the axis of orthogonal astigmatism. Populations having with or against the rule astigmatism have ellipses that are oriented horizontally. Populations with a higher proportion of oblique astigmatism have ellipses that are oriented vertically.22 A modification of the doubled-angle plot was made and described as “on-axis analysis.” The horizontal axis is defined as the original axis of each patient. All preoperative data are set at zero and aligned on this axis. The postoperative axis is then calculated in relation to the preoperative axis. In this modified plot, mean pre- and postoperative astigmatism vectors were calculated and plotted as usual. The preoperative standard deviation is a line on the horizontal axis, whereas the postoperative standard deviation is an ellipse. If the surgical procedure affected only the axis that was intended to treat, then the postoperative vectors should be aligned on the horizontal axis. However, if the surgical procedure induced vectors in directions other than the intended one, the postoperative vectors would deviate from the horizontal axis. The defocus equivalent (DEQ) is proportional to the diameter of the blur circle on the retina for a given pupil size. Lower
A paired t test was used to compare continuous variables (SPSS software version 13; SPSS Inc., Chicago, IL). Probabilities of less than 5% were considered significant (P⬍0.05), and probabilities less than 1% were considered to be highly significant (P⬍0.01).
Results Mean follow-up after IEAK was 7.2⫾3.5 months (range 4 –15 months). The indications for the original penetrating keratoplasty are summarized in Table 1. A comparison of pre- and postoperative outcome measures is summarized in Table 2. Detailed data of cases included in outcome measurement analysis are shown in Table 3 (available at http://aaojournal.org). Uncorrected visual acuity expressed in the logarithm of the minimum angle of resolution improved significantly from 1.08⫾0.34 preoperatively to 0.80⫾0.42 postoperatively (P⫽0.0016, paired t test). Best-corrected visual acuity (logarithm of the minimum angle of resolution) improved significantly from 0.45⫾0.27 preoperatively to 0.37⫾0.27 postoperatively (P⫽0.018). Although the mean spherical equivalent did not significantly change (P⫽ 0.99), the DEQ was significantly reduced by more than 1 D (from 8.5⫾3.5 D to 7.2⫾4.59 D, P⫽0.025). Mean absolute astigmatism was reduced from 7.46⫾2.70 preoperatively to 4.77⫾3.29 postoperatively (P⫽0.0001), a highly significant result. The mean astigmatism vector was reduced from 2.52⫻122 degrees ⫾ 5.4 to 0.41⫻126⫾4.0 degrees (P⫽0.07). Both absolute orthogonal and oblique astigmatism were significantly reduced (P⫽0.005 and P⫽0.04, respectively). Total high order aberrations were increased
Table 2. Comparison of Before and After IntraLase-Enabled Astigmatic Keratotomy Visual Parameters (⫾ Standard Deviation) Parameter
Pre-IEAK
Post-IEAK
P Value (Paired t Test)
BCVA (logMAR) UCVA (logMAR) Spherical equivalent DEQ (D) Absolute astigmatism (D) Mean astigmatism vector (D) Absolute orthogonal astigmatism (D) Absolute oblique astigmatism (D) Aberrometry (mm) High order aberrations Tilt aberrations Coma-like aberrations Trefoil aberrations Quatrefoil aberrations Spherical aberrations
0.45⫾0.27 1.08⫾0.34 ⫺4.01⫾3.85 8.49⫾3.52 7.46⫾2.70 2.52⫻122⫾5.4 4.7⫾2.9 4.4⫾3.5
0.37⫾0.27 0.80⫾0.42 ⫺3.95⫾4.69 7.23⫾4.49 4.77⫾3.29 0.41⫻126⫾4.0 2.8⫾2.1 3.2⫾3.4
0.018 0.0016 0.99 0.025 0.0001 0.07 0.005 0.04
2.8⫾1.7 3.6⫾2.6 1.4⫾0.9 1.6⫾1.1 0.8⫾0.8 0.6⫾0.5
4.2⫾3.5 4.0⫾3.0 1.8⫾1.3 2.7⫾2.5 1.4⫾1.4 0.8⫾0.7
0.01 0.41 0.022 0.01 0.038 0.005
BCVA ⫽ best-corrected visual acuity; D ⫽ diopters; DEQ ⫽ defocus equivalent; IEAK ⫽ IntraLase-enabled astigmatic keratotomy; logMAR ⫽ logarithm of the minimum angle of resolution; UCVA ⫽ uncorrected visual acuity.
1230
Kumar et al 䡠 IntraLase-Enabled Astigmatic Keratotomy Vector Analysis
Figure 2. Mean absolute astigmatism plotted against time after IEAK. The mean astigmatism seems to be stable throughout the follow-up period (6 months). IEAK ⫽ IntraLase-enabled astigmatic keratotomy.
from 2.8⫾1.7 to 4.2⫾3.5 (P⫽0.01). When investigated separately, each of the higher-order aberrations was significantly increased with the exception of tilt aberrations. Figure 1 outlines the comparison between pre- and postoperative DEQ for each case. A trend toward an improved, lower value for the DEQ postoperatively is seen with the mean effect (linear regression) depicted as a dotted line. Twenty-two eyes (65%) had a lower DEQ postoperatively, 4 eyes (11%) experienced no change in DEQ, and 8 eyes (24%) had a higher DEQ postoperatively. Of the 8 eyes with a higher DEQ after the procedure, 4 experienced an improvement in UCVA and BCVA, 3 were unchanged, and 1 lost 1 line of UCVA and BCVA. Figure 2 shows the mean reduction in astigmatism in diopters plotted against time after treatment. The effect of the astigmatic keratotomy is seen to reach its maximum effect 1.5 months after treatment, with slight regression and then stabilization between 3 and 6 months.
Figure 3. Achieved orthogonal astigmatic correction plotted against intended orthogonal astigmatic correction. The dotted line represents an exact correlation between intended and achieved correction. The solid line represents the mean effect of all procedures (linear regression) and closely superimposes the dotted line. Dashed lines represent 95% confidence interval. D ⫽ diopters.
Figure 4. Achieved oblique astigmatic correction plotted against intended oblique astigmatic correction. The dotted line represents an exact correlation between intended and achieved correction. The solid line represents the mean effect of all procedures (linear regression). Dashed lines represent 95% confidence interval. D ⫽ diopters.
Figure 3 shows the achieved astigmatic correction and the intended astigmatic correction for orthogonal astigmatism. A precise relationship between achieved and intended correction is established. Figure 4 shows the achieved and intended astigmatic correction for oblique astigmatism. Overall, a slight overcorrection was achieved for high minus cylinder and a slight undercorrection was achieved for high plus cylinder. However, for lower degrees of astigmatism, a precise relationship between the achieved and the intended astigmatism is again established. Figure 5 depicts the doubled-angled polar plot. The centroid of the preoperative data is displaced from the null point. The standard
Figure 5. Double-angled polar plot of the preoperative and postoperative refractive astigmatism. The ellipses signify 1 standard deviation around the mean astigmatism vector (stars). The mean astigmatism vector is brought closer to the null position by IEAK, and the standard deviation ellipse is reduced. IEAK ⫽ IntraLase-enabled astigmatic keratotomy.
1231
Ophthalmology
Volume 117, Number 6, June 2010 rection occurred in 9 eyes (24%). Six of the overcorrected eyes had originally undergone penetrating keratoplasty for keratoconus, 2 for corneal scarring and 1 for pseudophakic bullous keratopathy. Three eyes were resutured because of early overcorrection and were excluded from outcome measurement analysis. They were treated for 5, 9, and 12 D of cylinder with arc sizes of 60, 75, and 90 degrees, respectively.
Discussion
Figure 6. A modified doubled-angled polar plot, termed “on-axis analysis.” The horizontal axis is defined as the original axis of each patient. The postoperative axis is then calculated in relation to the preoperative axis. The ellipse signifies 1 standard deviation around the mean astigmatism vector (star). Although the mean postoperative astigmatism (grey star) is brought closer to the null point, additional vectors seem to be induced in other directions. IEAK ⫽ IntraLase-enabled astigmatic keratotomy; Pre-op ⫽ preoperative.
deviation ellipse is oriented horizontally and to the left of the null point, indicating a population with predominantly with-the-rule astigmatism. Figure 6 is a modified doubled-angled polar plot depicting “on-axis analysis.” By definition, the mean preoperative values are on the horizontal axis. Although the mean postoperative astigmatism vector is brought closer to the null point both horizontally (on-axis correction) and vertically (off-axis–induced astigmatism), the postoperative individual values deviate from the horizontal axis, implying that vectors have been induced in other directions. The mean postoperative astigmatism vector is close to the null point vertically and horizontally; therefore on average the procedure corrected the astigmatism on the intended axis without intentionally inducing cylinder in any specific direction. Figure 7 compares change in total higher-order aberrations with arc size. There is an increased change in higher-order aberrations with increased arc length, but this is not statistically significant (correlation coefficient 0.12).
The femtosecond laser creates arcuate incision length, depth, radius, and symmetry with a precision that was previously unable to attain using manual techniques. Despite this, the effect of the IntraLase-enabled astigmatic keratotomy incisions is still subject to the same vagaries of corneal wound healing that modulate the initial response. Keratotomy healing begins with an epithelial plug that is eventually replaced by hypercellular scar tissue, explaining delayed changes in corneal curvature.24 Therefore, although it is necessary to investigate measurement outcomes, it is equally important to measure the rate of change of effect, duration of effect, stability, and predictability of the intervention. In this study we aimed to look at the results of IEAK after the wound has stabilized (⬃6 months postoperatively). Our results indicate that IEAK is effective in reducing high astigmatism in corneal grafts. Both UCVA and BCVA were significantly improved. The DEQ, a measure of surgical success, and the absolute value of astigmatism were significantly reduced. Such findings are similar in recent studies.19,20 There was no significant change in spherical equivalent. This is because the flattening produced by the incisions in the steep axis is countered by the steepening of the unincised meridian, known as the “coupling effect.” When there is little or no change in spherical equivalent, the corneal steepening/flattening coupling ratio is close to 1.25 This has
Efficacy Twenty-four eyes (71%) gained 1 or more lines of vision, 7 eyes (21%) remained unchanged, and 3 eyes (8%) lost 1 line of vision. The efficacy index, calculated as the ratio of the mean postoperative BCVA to the mean preoperative UCVA, was 0.6⫾0.6.
Complications There were no cases of perforation, wound dehiscence, or infectious keratitis. Three eyes experienced an episode of graft rejection (8%), and each case resolved on topical steroid therapy. Overcor-
1232
Figure 7. Change in total higher-order aberrations (microns) plotted against arc size (degrees). There is a slight correlation between increased arc size and a greater change in total higher-order aberrations.
Kumar et al 䡠 IntraLase-Enabled Astigmatic Keratotomy Vector Analysis been reported in several other studies analyzing the effect of relaxing incisions.26 –32 The mean effect of all procedures was to reduce the DEQ, thereby reducing the size of the blur circle. In Figure 1, those eyes below the solid line represent an improved and reduced postoperative DEQ, and those eyes above the solid line represent a worsened and increased postoperative DEQ. Despite this, those cases above the line experienced a trend toward improved UCVA and BCVA. The mean absolute value of astigmatism decreased from 7.46⫾2.70 D to 4.77⫾3.29 D over the first 3 months postoperatively (Fig 2). The initial reduction to 4.00 D occurred within the first 6 weeks, followed by a slight regression of the effect at 3 months. Between 3 and 6 months, there was no change in effect. This suggests that by 3 months postprocedure, the reduction in astigmatism has stabilized. Further follow-up data are required to establish the duration of effect. An almost identical trend in the change in astigmatism after femtosecond-enabled astigmatic keratotomy for high post-keratoplasty astigmatism has been reported.20 The predictability of the procedure may be defined as how closely the intended correction correlates with achieved correction. For orthogonal astigmatism, the mean achieved reduction in cylinder closely approximates the intended reduction (Fig 3) at low and high levels of astigmatism. For oblique astigmatism, the mean achieved reduction in cylinder closely approximates the intended reduction for lower levels of astigmatism; however, the effect tends toward undercorrection at higher levels (Fig 4). This decrease in refractive predictability with higher corrections has been reported.22 Overall, the achieved reduction correlated well with the intended reduction in astigmatism. However, because 3 eyes were excluded from outcome measurement analysis as the result of early resuturing, our results may be marginally superior when compared with the actual outcome of all cases. Figure 5 shows that the mean astigmatism vector is brought closer to the null point by IEAK. This is consistent with previous findings.18 In addition, the position, size, and orientation of the standard deviation ellipse are all altered by the procedure. The position of the standard deviation ellipse is shifted to the right because there is less with the rule astigmatism. The standard deviation ellipse is smaller because there is a reduction in mean absolute astigmatism. Also, the orientation of the ellipse has changed from horizontal to vertical. This signifies that with the rule astigmatism is predominant preoperatively, whereas oblique astigmatism is predominant postoperatively. By using on-axis analysis (Fig 6), we noted that although the mean astigmatism vector is well corrected by IEAK, new vectors seem to have been created in other directions. These vectors seem to be created at random, because their average is close to the null point. Another possibility is that after penetrating keratoplasty, several vectors are present in the graft (induced by tight sutures and asymmetric healing), and once the main astigmatism vector is reduced by IEAK, the other smaller vectors that were already present become apparent. In such cases,
additional more limited IEAK procedures might help further improve corneal astigmatism. There were no cases of perforation, graft– host junction dehiscence, or infectious keratitis. This compares favorably with prior reports on manual astigmatic keratotomy.2,26 Complications of the procedure included an episode of allograft rejection in 3 cases (8%). Each case was treated with topical corticosteroids and resolved without further incident. The rate of graft rejection after IEAK seems to be higher than that previously found in series of manual astigmatic keratotomy.2,26 Overcorrection occurred in 9 cases (24%). On further analysis of the overcorrected cases, 6 of the 9 occurred during the initial period when arc length was determined by analyzing the location and extent of the steepest meridians on the topographic map. Thereafter the protocol was altered as outlined above, such that the rate of overcorrection decreased dramatically. There were 3 cases of overcorrection in the subsequent 27 eyes. Overall, two thirds of the overcorrected eyes had originally undergone penetrating keratoplasty for keratoconus. It is plausible that ectatic eyes are more at risk for overcorrection. Our data indicate that IEAK significantly reduces the mean absolute astigmatism and improves UCVA, BCVA, and DEQ, and is thus effective. The corrections of orthogonal and oblique astigmatism are achieved with a high degree of refractive predictability. The refractive effect stabilizes and reaches a plateau between 3 and 6 months postoperatively. IntraLase-enabled astigmatic keratotomy adversely affects higher-order aberrations. There is also a greater predominance of oblique astigmatism in the postoperative population. These factors may diminish the beneficial effects of the procedure and may explain the smaller improvement in BCVA when compared with UCVA. Patients may experience greater glare and reduced contrast sensitivity. A similar adverse effect on higher-order aberrations has been reported in a cohort of manual astigmatic keratotomy.18 The present study attempted to determine the effect of IEAK over a mean period of 6 months. Consequently, those eyes with significant early overcorrection that required resuturing within the first 6 weeks could not be included in the outcome measurement analysis. Although only 3 of the eyes initially enrolled in the study were in this category, this represents a selection bias and favorably affects the outcome measurement results. These 3 eyes were included in the calculation of complication rates. We believe that incisions must be performed in the donor because the graft– host junction provides a new functional limbus and has an unpredictable thickness.33 We currently recommend customizing arc length according to the degree of astigmatism that requires correction, as outlined previously. If there is less than 6 D of post-penetrating keratoplasty astigmatism, either photorefractive keratectomy or a limited IEAK with photorefractive keratectomy to reduce the amount of tissue removed by photorefractive keratectomy alone may be the optimal choice. In conclusion, this study showed that IEAK has an encouraging refractive predictability. The effect has been shown to stabilize by 3 months and reach a plateau between
1233
Ophthalmology
Volume 117, Number 6, June 2010
3 and 6 months. The efficacy index was similar to previous findings for IEAK and favorable when compared with a previous cohort of manual astigmatic keratotomy.18 The allograft rejection rate was higher than expected, and although these episodes were readily treated and resolved on topical corticosteroids, this warrants further scrutiny in future studies. The overcorrection rate of IEAK was also high, but decreased significantly on use of a more conservative treatment regimen as outlined. The precise length, curvature, depth, and symmetry of the IEAK incisions may afford a greater effect than a similar arc length produced by a manual technique. However, financial considerations and laser availability must be considered when choosing between IntraLase-enabled and manual astigmatic keratotomy. As more data become available, it is necessary to refine the treatment nomogram for IntraLase-enabled astigmatic keratotomy for high astigmatism after penetrating keratoplasty.
References 1. Williams KA, Roder D, Esterman A, et al. Factors predictive of corneal graft survival: report from the Australian Corneal Graft Registry. Ophthalmology 1992;99:403–14. 2. Poole TR, Ficker LA. Astigmatic keratotomy for post-keratoplasty astigmatism. J Cataract Refract Surg 2006;32:1175–9. 3. Bochmann F, Schipper I. Correction of post-keratoplasty astigmatism with keratotomies in the host cornea. J Cataract Refract Surg 2006;32:923– 8. 4. Vajpayee RB, Sharma N, Sinha R, et al. Laser in-situ keratomileusis after penetrating keratoplasty. Surv Ophthalmol 2003;48:503–14. 5. Genvert GI, Cohen EJ, Arentsen JJ, Laibson PR. Fitting gas permeable contact lenses after penetrating keratoplasty. Am J Ophthalmol 1989;99:511– 4. 6. Forseto AS, Francesconi CM, Nosé RA, Nosé W. Laser in situ keratomileusis to correct refractive errors after keratoplasty. J Cataract Refract Surg 1999;25:479 – 85. 7. Hardten DR, Lindstrom RL. Surgical correction of refractive errors after penetrating keratoplasty. Int Ophthalmol Clin 1997;37:1–35 8. Troutman RC. Corneal wedge resections and relaxing incisions for postkeratoplasty astigmatism. Int Ophthalmol Clin 1983;23:161– 8. 9. Arffa RC. Results of a graded relaxing incision technique for postkeratoplasty astigmatism. Ophthalmic Surg 1988;19: 624 – 8. 10. Lugo M, Donnenfeld ED, Arentson JJ. Corneal wedge resection for high astigmatism following penetrating keratoplasty. Ophthalmic Surg 1987;18:650 –3. 11. Gothard TW, Agapitos PJ, Bowers RA, et al. Four incision radial keratotomy for high myopia after penetrating keratoplasty. J Refract Corneal Surg 1993;9:51–7. 12. John ME, Martines E, Cvintal T, et al. Photorefractive keratectomy following penetrating keratoplasty. J Refract Corneal Surg 1994;10(suppl):S206 –10. 13. Dada T, Vajpayee RB. Laser in situ keratomileusis after PKP [letter]. J Cataract Refract Surg 2002;28:7– 8. 14. Slade SG. Applications of the femtosecond laser in corneal surgery. Curr Opin Ophthalmol 2007;18:338 – 41.
1234
15. Buratto L, Bohm E. The use of the femtosecond laser in penetrating keratoplasty. Am J Ophthalmol 2007;143:737– 42. 16. Harissi-Dagher M, Azar DT. Femtosecond laser astigmatic keratotomy for postkeratoplasty astigmatism. Can J Ophthalmol 2008;43:367–9. 17. Kiraly L, Herrmann C, Amm M, Duncker G. Reduction of astigmatism by arcuate incisions using the femtosecond laser after corneal transplantation [in German]. Klin Monatsbl Augenheilkd 2008;225:70 – 4. 18. Bahar I, Levinger E, Kaiserman I, et al. Intralase-enabled astigmatic keratotomy for postkeratoplasty astigmatism. Am J Ophthalmol 2008;146:897–904. 19. Hoffart L, Proust H, Matonti F, et al. Correction of postkeratoplasty astigmatism by femtosecond laser compared with mechanized astigmatic keratotomy. Am J Ophthalmol 2009; 147:779 – 87. 20. Nubile M, Carpineto P, Lanzini M, et al. Femtosecond laser arcuate keratotomy for the correction of high astigmatism after keratoplasty. Ophthalmology 2009;116:1083–92. 21. Koch DD, Kohnen T, Obstbaum SA, Rosen ES. Format for reporting refractive surgical data. J Cataract Refract Surg 1998;24:285–7. 22. Holladay JT, Moran JR, Kezirian GM. Analysis of aggregate surgically induced refractive change, prediction error, and intraocular astigmatism. J Cataract Refract Surg 2001;27: 61–79. 23. Holladay JT, Lynn MJ, Waring GO III, et al. The relationship of visual acuity, refractive error, and pupil size after radial keratotomy. Arch Ophthalmol 1991;109:70 – 6. 24. Eiferman RA, Schultz GS, Nordquist RE, Waring GO III. Corneal wound healing and its pharmacologic modification after refractive keratotomy. In: Waring GO III, ed. Refractive Keratotomy for Myopia and Astigmatism. St. Louis, MO: Mosby; 1992;749 –79. 25. Wilkins MR, Mehta JS, Larkin FP. Standardized arcuate keratotomy for postkeratoplasty astigmatism. J Cataract Refract Surg 2005;31:297–301. 26. Fronterre A, Portesani GP. Relaxing incisions for postkeratoplasty astigmatism. Cornea 1991;10:305–11. 27. Krachmer JH, Fenzl RE. Surgical correction of high postkeratoplasty astigmatism: relaxing incisions vs wedge resection. Arch Ophthalmol 1980;98:1400 –2. 28. Hjortdal JO, Ehlers N. Paired arcuate keratotomy for congenital and postkeratoplasty astigmatism. Acta Ophthalmol Scand 1998;76:138 – 41. 29. Cohen KL, Tripoli NK, Noecker RJ. Prospective analysis of photokeratoscopy for arcuate keratotomy to reduce postkeratoplasty astigmatism. J Refract Corneal Surg 1989;5:388 –93. 30. Koay YP, McGhee CN, Crawford GJ. Effect of a standard paired arcuate incision and augmentation sutures on postkeratoplasty astigmatism. J Cataract Refract Surg 2000;26: 553– 61. 31. Akura J, Matsuura K, Hatta S, et al. A new concept for the correction of astigmatism: full-arc, depth-dependent astigmatic keratotomy. Ophthalmology 2000;107:95–104. 32. Price FW, Grene RB, Marks RG, Gonzales JS, ARC-T Study Group. Astigmatism reduction clinical trial: a multicenter prospective evaluation of the predictability of arcuate keratotomy; evaluation of surgical nomogram predictability. Arch Ophthalmol 1995;113:277– 82. 33. Kaiserman I, Bahar I, Rootman DS. Corneal wound malapposition after penetrating keratoplasty: an optical coherence tomography study. Br J Ophthalmol 2008;92:1103–7.
Kumar et al 䡠 IntraLase-Enabled Astigmatic Keratotomy Vector Analysis
Footnotes and Financial Disclosures Originally received: March 22, 2009. Final revision: October 24, 2009. Accepted: October 26, 2009. Available online: February 16, 2010.
Supported as provision of material supplies from AMO Inc., Irvine, California. Manuscript no. 2009-398.
Department of Ophthalmology and Vision Sciences, Toronto Western Hospital, University of Toronto, Ontario, Canada. Financial Disclosure(s): The author(s) have made the following disclosure(s): David S. Rootman received funding as stated above.
Correspondence: Nikhil L. Kumar, BMed, FRANZCO, Department of Ophthalmology and Vision Sciences, Toronto Western Hospital, University of Toronto, 399 Bathurst St., Toronto, Ontario, Canada M5T2S8. E-mail: nik_kumar1@ yahoo.com.
1235
Ophthalmology
Volume 117, Number 6, June 2010
Table 3. Detailed Data of Cases Included in Outcome Measurement Analysis No.
Age
Diagnosis
Duration between PKP and IEAK (yrs)
Preoperative Manifest Refraction
Postoperative Manifest Refraction
Post IEAK Follow-up (mos)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
76 37 21 70 62 70 28 58 62 51 40 41 41 58 59 59 70 72 75 48 44 67 43 65 53 44 38 77 62 74 51 21 49 85
FED KC KC PBK SCAR SCAR KC SCAR SCAR KC KC KC KC PBK KC KC SCAR PBK PBK KC KC PBK SCAR FED ICE KC KC PBK FED PBK KC KC KC PBK
9 8 1.5 1.5 12 8 1.5 Data absent 5 30 5 8 8 4 4 Data absent 2 1.5 8 1.5 3.5 Data absent 5 5 1.5 1.5 5 2 5 Data absent 17 1.5 17 3
5.75–7.75*65 1.75–5.0*178 7.0–8.0*60 ⫺0.75 to 7.0*86 ⫺3.00 to 12.5*18 3.50–6.0*55 ⫺4.25 to 12.0*44 2.75–5.5*55 ⫺7.0 to 5.5*32 1.75–8.0*20 7.0–10.0*112 ⫺1.0 to 6.5*134 ⫺2.0 to 6.0*91 1.0–5.5*19 ⫺3.0 to 5.0*157 ⫺2.25 to 6.5*176 ⫺2.0–11.0*36 0.0–12.0*3 5.5–7.0*90 ⫺1.5 to 11.0*84 0.0–6.0*31 ⫺5.0 to 5.0*100 8.75–4.5*30 0.0–6.0*82 ⫺1.5 to 5.5*20 6.0–5.0*13 ⫺4.5 to 5.0*177 0.0–8.25*40 0.5–6.25*167 0.0–10.0*160 ⫺1.75 to 11.25*156 ⫺2.0 to 5.5*15 0.0–13.0*44 ⫺2.25 to 7.0*17
0.00–1.50*70 ⫺1.0 ⫺7.5 to 4.0*95 ⫺1.5 to 2.5*135 10.5–1.75*90 1.0–4.0*55 ⫺10.0 to 5.75*130 2.5–2.5*28 ⫺2.0 to 3.25*51 ⫺7.0 to 6.0*110 ⫺0.5 to 7.5*81 ⫺2.0 to 5.0*171 0.0–4.0*30 ⫺4.5 to 0.75*88 ⫺5.0 to 2.0*46 ⫺7.0 to 6.0*85 4.5–9.0*61 ⫺2.5 to 5.25*139 2.0–4.5*10 ⫺4.5 to 2.5*115 ⫺1.25 to 0.75*9 ⫺6.25 to 6.0*120 ⫺9.5 to 4.5*90 ⫺1.5 to 3.0*79 ⫺1.25 to 5.0*20 2.0–1.5*180 0.5–2.75*78 ⫺0.25 to 4.0*1 ⫺1.0 to 4.75*145 ⫺1.5 to 7.5*155 ⫺9.5 to 6.0*80 ⫺5.0 to 2.0*180 ⫺1.75 to 14.0*40 ⫺1.0 to 0.5*80
6 4 4 7 12 4 4 4 15 5 6 11 11 5 4 4 9 6 9 6 9 4 5 10 6 9 14 10 6 4 8 7 15 4
CX
OC OC
OC
OC OC R OC
R
R
CX ⫽ complications; FED ⫽ Fuchs’ endothelial dystrophy; ICE ⫽ iridocorneal endothelial syndrome; IEAK ⫽ IntraLase-enabled astigmatic keratotomy; KC ⫽ keratoconus; OC ⫽ overcorrected; PBK ⫽ pseudophakic bullous keratopathy; PKP ⫽ penetrating keratoplasty; R ⫽ graft rejection episode; SCAR ⫽ visually significant corneal scar.
1235.e1