- Email: [email protected]
Spontaneous Long-Term Changes of Corneal Power and Astigmatism After Suture Removal After Penetrating Keratoplasty Using a Regression Model
Spontaneous Long-Term Changes of Corneal Power and Astigmatism After Suture Removal After Penetrating Keratoplasty Using a Regression Model ACHIM LANGENBUCHER, PHD, GOTTFRIED O.H. NAUMANN, MD, AND BERTHOLD SEITZ, MD, FEBO
● PURPOSE: To assess the diagnosis-based spontaneous long-term changes in corneal power and refraction with a regression model in the all-sutures-out time period following non-mechanical penetrating keratoplasty (PK). ● DESIGN: Retrospective non-randomized clinical trial. ● METHODS: SETTING: Clinical practice. STUDY POPULATION: 147 eyes [47 Fuchs dystrophy (FD); 100 keratoconus (KC)] were studied after suture removal in this retrospective longitudinal study. MAIN OUTCOME MEASURES: Zeiss keratometry [equivalent power (KEQ) and astigmatism (KAST)], corneal topography analysis [equivalent power (TEQ) and astigmatism (TAST)], and subjective refractometry [spherical equivalent (SEQ) and refractive cylinder (RAST)] were assessed in at least three up to 16 ophthalmologic examinations in the all-sutures-out time period. OBSERVATION PROCEDURE: The time course of each target variable was analyzed in a longitudinal manner (time interval > 12 months) separately for each patient with a linear regression model. ● RESULTS: Post-keratoplasty follow-up ranged from 31 months to 10.3 years. In the linear regression model, the annual change in FD/KC showed an increase/a decrease in KEQ (0.29 ⴞ 0.50/ⴚ0.63 ⴞ 0.46 diopters, P ⴝ .02) and an increase/a decrease in TEQ (0.37 ⴞ 0.54/ⴚ0.69 ⴞ 0.49 diopters, P ⴝ .04) corresponding to a decrease/an increase in SEQ (ⴚ0.31 ⴞ 0.47/0.63 ⴞ 0.43 diopters, P ⴝ .02). KAST/TAST/RAST showed a minimal annual decrease (ⴚ0.06 ⴞ 0.41/ⴚ0.05 ⴞ 0.45/ⴚ0.06 ⴞ 0.41 diopters) in FD but an increase in KC (0.46 ⴞ 0.41/0.51 ⴞ 0.43/0.46 ⴞ 0.38 diopters) (P ⴝ .05/0.06/0.12).
Accepted for publication Jan 26, 2005. From the Augenklinik mit Poliklinik der Universität Erlangen-Nürnberg, Erlangen, Germany. Inquires to Achim Langenbucher, PhD, Augenklinik mit Poliklinik der Universität Erlangen-Nürnberg, Schwabachanlage 6, D-91054 Erlangen, Germany; fax: 09131-853-4271; e-mail: achim.langenbucher@augen. imed.uni-erlangen.de 0002-9394/05/$30.00 doi:10.1016/j.ajo.2005.01.038
©
2005 BY
● CONCLUSIONS: In the follow-up after post-keratoplasty suture removal, patients with FD/KC tend to develop a spontaneous myopic shift (steepening of the cornea)/hyperopic shift (flattening of the cornea). In contrast with those with FD, patients with KC should be counseled on the fact that astigmatism may increase again over time after suture removal. (Am J Ophthalmol 2005;140:29 –34. © 2005 by Elsevier Inc. All rights reserved.)
I
N NON– HIGH-RISK KERATOPLASTIES THE PROGNOSIS OF
corneal grafts has improved in the last decades because of enhanced treatment techniques and storage methods and is superior compared with all other organ transplantations. Primary goals after corneal transplantation are the preservation of a clear graft and the rehabilitation of the visual function.1–5 Many studies have been performed to assess the influencing factors for graft survival and the rate of repeat keratoplasty because of graft failure.6 –10 From June 1989 to December 2004, more than 1650 non-mechanical penetrating keratoplasties (PK) have been performed using the 193 nm excimer laser in our Department of Ophthalmology in Erlangen. The most frequent indications were keratoconus (KC) and Fuchs dystrophy (FD). Especially in young patients with KC, non-mechanical trephination offers clear advantages, such as reduction of keratometric astigmatism and refractive cylinder, higher regularity of the corneal topography, less asymmetry in the topographic pattern, and a significantly better visual outcome, which are the crucial criteria for patient satisfaction.4,7,11 In a prospective, randomized, cross-sectional, clinical, single-center study including 179 eyes with primary central PK for FD (diameter 7.5 mm) or KC (diameter 8.0 mm), graft oversize 0.1 mm, 16-bite, double-running diagonal suture, we found a significantly smaller keratometric astigmatism (3.0 diopters) after suture
ELSEVIER INC. ALL
RIGHTS RESERVED.
29.e1
● MAIN OUTCOME MEASURES AND REGRESSION ANALYSIS: Main outcome measures of parameters included: (1)
removal for laser trephination in contrast with 6.1 diopter for motor trephination (P ⬍ .0009). In contrast, Mader and colleagues12 reported a mean keratometric astigmatism of 5.6 diopters with all sutures out after conventional trephination in 162 eyes. Wiffen and associates published a final median keratometric astigmatism of 5.5 diopters using the Hessburg-Barron trephine.10 With sutures in place, the corneal topography is mostly determined by the suture tension, whereas the pattern of corneal topography after suture removal is affected by the wound configuration and the adaptation between the donor and recipient cut.12–16 At this time, very little is known about the long-term behavior of the all-suture-out corneal curvature following PK. The purpose of this study was to assess the spontaneous long-term changes of refraction, corneal power, and astigmatism in the all-suture-out time period following nonmechanical PK comparing eyes with KC and FD in a longitudinal fashion, using a linear regression fit separately for each variable and each patient.
keratometric corneal power (Zeiss Ophthalmometer, Zeiss), (2) keratometric astigmatism, (3) topographic corneal power (TMS-1, release 1.61, Tomey, Erlangen, Germany), (4) topographic astigmatism, (5) spherical equivalent of the refraction (trial glasses in a trial frame), and (6) refractive cylinder. For each patient at least three up to 16 valid measurements of parameters (1)–(6) in the all-sutures-out time period (baseline examination around 6 weeks after final suture removal) in a time interval of at least 1 year were used to formulate an individual regression model separately for each patient and each parameter. The slope of the linear regression yields the absolute change of the target parameters/time interval. From the difference between the measured target parameter and the predicted data given by the regression model, the residuum was calculated and considered the prediction error of the model. This method has been described in detail in previous studies applied to the problem of graft endothelial cell loss after PK.18,19 ● STATISTICAL ANALYSIS AND REGRESSION MODELS:
Data were recruited and collected in a relational database system (Access 97 for Windows) and analyzed using SPSS/PC 11.0 (Windows NT). Comparisons between patient groups (FD vs KC) were performed using non-parametric tests (Mann-Whitney U test for unpaired samples). Linear regression analysis was performed separately for each target variable and each patient to describe the time course of the values after suture removal. For evaluation of the prediction error, we derived the Durbin-Watson residuum between the observed and the predicted data from the fit. For assessment of correlations between variables the Spearman correlation coefficient was calculated. A P-value ⬍.05 was considered statistically significant.
METHODS ● STUDY POPULATION/DONOR DETAILS AND SURGICAL INTERVENTION: Inclusion criteria for this retrospective, non-randomized clinical study consisted of: primary central round PK without previous surgery, FD with a trephination diameter 7.5/7.6 mm in recipient/donor (n ⫽ 47; 12 men, 35 women; 27 left eye [OS], 20 right eye [OD], or KC with a trephination diameter 8.0/8.1 mm in recipient/donor (n ⫽ 100; 71 men, 29 women; 51 OS, 49 OD), interval of surgery between June 1992 and March 2002. Exclusion criteria were maculopathies, optic nerve atrophies, amblyopia, complications such as immunologic graft reactions during follow-up and further surgery such as arcuate keratotomies, cataract surgery, and so on. In 147 patients, non-mechanical PK was performed with the 193 nm excimer laser (MEL 60, Zeiss-Meditec, Jena, Germany) along metal aperture masks with eight orientation teeth/notches. In 29 patients with FD, a simultaneous extracapsular cataract extraction and posterior chamber lens implantation (triple procedure) was necessary. After temporary fixation of the donor button in the recipient bed with eight interrupted 10-0 nylon sutures, permanent wound closure was ensured by a 16-bite, double-running, diagonal cross-stitch suture.17 The pre- and postoperative medication was standardized. Pre-operatively, patients with PK only received pilocarpine 2% eye drops; those with triple procedure received atropine 1% eye drops. After surgery, ofloxacin ointment, cycloplegics, and a vitamin D– containing gel was applied. Topical corticosteroids (prednisolone acetate) were started five times a day not before epithelial closure and were tapered by one drop a day every 6 weeks.
29.e2
AMERICAN JOURNAL
RESULTS OVERALL, THE MEAN FOLLOW-UP PERIOD WAS 3.9 ⫾
1.7 years (range 31 months to 10.3 years). In FD, the individual follow-up ranged from 31 months to 9.1 years with a mean of 4.0 ⫾ 1.6 years and in KC, it ranged from 33 months to 10.3 years with a mean of 3.9 ⫾ 1.7 years. Between groups, there was no statistically significant difference. No patient in the FD or KC group developed an immunologic graft rejection in the follow-up period. One of the double-running sutures was removed 11 to 14 months after PK, and the second suture was removed 15 to 21 months after PK. The time interval between PK and suture removal did not differ significantly between FD and KC. The mean time interval for regression analysis, time period between first and last follow-up examination in the all-sutures-out time period, was 2.7 ⫾ 1.5 years (range 12 months to 8.4 years). In FD, it was 2.6 ⫾ 1.4 years (range OF
OPHTHALMOLOGY
JULY 2005
TABLE 1. Linear Regression Analysis Mean ⫾ SD; MinMax; Median in Diopters
Total population Fuchs’ dystrophy (FD, n ⫽ 47) P Keratoconus (KC, n ⫽ 100)
Keratometric Central Power
Keratometric Astigmatism
Topographic Central Power
Topographic Astigmatism
Spherical Equivalent
Refractive Cylinder
⫺0.33 ⫾ 0.64 ⫺2.0–1.1; ⫺0.42 0.29 ⫾ 0.50; ⫺1.0–1.1; 0.29 .02 ⫺0.63 ⫾ 0.46 ⫺2.0–0.5; ⫺0.62
0.29 ⫾ 0.47; ⫺1.4–1.6; 0.27 ⫺0.06 ⫾ 0.41; ⫺1.4–1.1; ⫺0.06 .05 0.46 ⫾ 0.41; ⫺0.4–1.6; 0.47
⫺0.35 ⫾ 0.71; ⫺2.0–1.5; ⫺0.44 0.37 ⫾ 0.54 ⫺1.2–1.5; 0.43 .04 ⫺0.69 ⫾ 0.49; ⫺2.0–0.4; ⫺0.73
0.33 ⫾ 0.51 ⫺1.1–1.9; 0.32 ⫺0.05 ⫾ 0.45 ⫺1.1–1.1; 0.0 .06 0.51 ⫾ 0.43 ⫺0.4–1.9; 0.50
0.32 ⫾ 0.63; ⫺1.1–1.7; 0.46 0.31 ⫾ 0.47; ⫺1.1–0.9; ⫺0.35 .02 0.63 ⫾ 0.43; ⫺0.6–1.7; 0.69
0.30 ⫾ 0.46; ⫺1.3–1.6; 0.28 0.06 ⫾ 0.41; ⫺1.3–1.0; ⫺0.09 .12 0.46 ⫾ 0.38; ⫺0.4–1.6; 0.46
Slope of the linear regression analysis (absolute change per year) of the target variables keratometric central power, keratometric astigmatism, topographic central power, topographic astigmatism, spherical equivalent, and refractive cylinder. The P-values refer to the significance level comparing FD and KC.
12 months to 6.5 years), and in KC, it was 2.7 ⫾ 1.4 years (range 12 months to 8.4 years) without significant differences between groups. The overall number of all-suture-out examinations with valid target variables was 5.9 ⫾ 3.0 (range 3 to 16). In FD, it was 6.0 ⫾ 3.0 (range 3 to 16), and in KC, it was 5.9 ⫾ 3.0 (range 3 to 15). The regression analysis of keratometry yielded a mean decrease of keratometric central power of 0.33 ⫾ 0.64 diopters per year and an increase of the keratometric astigmatism of 0.29 ⫾ 0.47 diopters per year (Table 1). Figure 1 shows the change in spherical equivalent, keratometric power, and topographic power for both groups of patients. The regression analysis of corneal topographic data yielded a decrease of topographic central power of 0.35 ⫾ 0.71 diopters per year, and an increase of the topographic astigmatism of 0.33 ⫾ 0.51 diopters per year (Table 1, Figure 2). The regression analysis of subjective refraction yielded an increase of the spherical equivalent of 0.32 ⫾ 0.62 diopters per year, and an increase of the refractive cylinder of 0.30 ⫾ 0.46 diopters per year. Overall, the change in subjective spherical equivalent correlated inversely with the change in keratometric central power (r2 ⫽ .81, P ⫽ .02; FD: r2 ⫽ .66, P ⫽ .04; KC: r2 ⫽ .86, P ⫽ .01) and with the change in topographic central power (r2 ⫽ .66, P ⫽ .04; FD: r2 ⫽ 0.55, P ⫽ .04; KC: r2 ⫽ .72, P ⫽ .02). Change in refractive cylinder correlated tendentially with the change in keratometric astigmatism (r2 ⫽ .40, P ⫽ .04; FD: r2 ⫽ .36, P ⫽ .06; KC: r2 ⫽ .41, P ⫽ .04) and with the change in topographic astigmatism (r2 ⫽ .38, P ⫽ .06; FD: r2 ⫽ .30, P ⫽ .08; KC: r2 ⫽ .51, P ⫽ .04). The prediction error for keratometry was lower for the keratometric power (0.32 ⫾ 0.24 diopters) compared with the keratometric astigmatism (0.42 ⫾ 0.32 diopters, P ⫽ .02, Table 2). The prediction error for corneal topographic data was tendentially lower for keratometric power (0.36 ⫾ 0.29 diopters) compared with keratometric astigmatism (0.40 ⫾ 0.34 diopters, P ⫽ .05, Table 2). The prediction error of VOL. 140, NO. 1
SPONTANEOUS LONG-TERM CHANGES
FIGURE 1. Change over time in diopters (1 year, annual change) described by the slope of the linear regression line of spherical equivalent (subjective refraction evaluated with trial glasses in a trial frame), keratometric central power, and topographic central power of patients with Fuchs’ dystrophy (FD; n ⴝ 47) and keratoconus (KC; n ⴝ 100) after suture removal (all sutures out) following non-mechanical penetrating keratoplasty using the excimer laser 193 nm. Positive/negative values indicate an increase/ decrease of the spherical equivalent or corneal power over time, whereas values that equal zero indicate a stable spherical equivalent or corneal power over time.
subjective refraction was comparable for the spherical equivalent (0.27 ⫾ 0.20 diopters) and the refractive cylinder (0.29 ⫾ 0.23 diopters, P ⫽ .25, Table 2).
DISCUSSION CORNEAL TRANSPLANTATION, BOTH IN FD AND IN KC, ARE
among the main diagnoses for normal-risk keratoplasties IN
CORNEAL POWER
AND
REFRACTION
29.e3
tions with a complete set of target variables were considered for the analysis. In FD, the spherical equivalent decreased, indicating a myopic shift after suture removal, whereas in KC, the spherical equivalent increases, indicating a hyperopic shift. The latter finding was unexpected. These tendencies are in accordance with the steepening of the cornea in FD measured with keratometry and with corneal topography and the flattening of the cornea in KC. The astigmatism remained stable in FD but gained in KC over time. This effect correlated only in part with the objective measurement of the keratometry and topography analyses. In general, the prediction error for all target variables was small, indicating a good representation of the longitudinal data with the linear regression model applied in this study. Unfortunately, in the literature authors rarely differentiate clearly between sutures-in and all-sutures-out results. Concerning the changes attributable to post-keratoplasty suture removal, classic papers include those by Filatov and associates,14 Mader and associates12 and Musch and associates22 From these papers it becomes clear that corneal astigmatism may change unpredictably and by large amounts when all remaining sutures are removed after PK. Mader and colleagues reported that in 62% of eyes the astigmatism changed 2 or more diopters (range ⫹11.94 diopters to ⫺17.87 diopters). The average vectorial change was 6.5 diopters (range 0.56 diopters to 19.8 diopters). There was no decrease in the amount of astigmatic change with increasing time between surgery and suture removal. Graft size and diagnosis had no effect on the amount of astigmatic change. Astigmatic errors became stable with less than 1 diopter of change between successive examinations within 6 months after suture removal. Similar results were obtained for eyes with double-running and interrupted-running sutures.12 In a prospective randomized study, we found that the change of vector-corrected astigmatism was significantly smaller after suture removal following laser trephination (4.3 diopters), in contrast with motor trephination (6.9 diopters) [Seitz and associates, oral paper, presented at the ARVO meeting 2000 in Fort Lauderdale, Florida]. In contrast, the present study looked at spontaneous long-term changes in the all-sutures-out time period after post-keratoplasty suture removal. This question has been studied very rarely in the literature up to now: Tuft and Gregory16 found a trend toward a gradual flattening and increase of corneal cylinder over time in KC. This is in accordance with our present findings. Small amounts of central flattening and increased astigmatism may be explained by the irregular bulging of the keratoconic recipient rim allowing the graft-host-junction to act like a hinge thus leaving the graft center less curved.23 Additionally, Ruhswurm and associates24 studied the long-term clinical outcome after keratoplasty with the guided trephine system in 31 eyes with KC. They found a myopic shift associated with a steepening of the cornea
FIGURE 2. Change over time in diopters (1 year, annual change) described by the slope of the linear regression line of refractive cylinder (subjective refraction evaluated with trial glasses in a trial frame), keratometric astigmatism, and topographic astigmatism of patients with FD (n ⴝ 47) and KC (n ⴝ 100) after suture removal (all sutures out) following nonmechanical penetrating keratoplasty using the excimer laser 193 nm. Positive/negative values indicate an increase/decrease over time, whereas values that equal zero indicate a stable refractive cylinder or corneal astigmatism over time.
with a superior prognosis because of a low rate of immunologic graft rejections, low rate of post-keratoplasty ocular hypertension, and a very low rate of irreversible graft failures.4,20 In the present study, we focused on the spontaneous changes of corneal curvature and subjective refraction after suture removal (all sutures out). Of a sequence of at least three valid data sets after suture removal consisting of spherical equivalent, refractive cylinder, keratometric central corneal power, keratometric astigmatism, topographic central corneal power, and topographic astigmatism, we fitted linear regression lines separately for each patient to evaluate the longitudinal section of these target variables. The characteristic result of each regression analysis fitted to the data sets is (1) the slope of the regression line, which indicates the annual change of the target variable, and (2) the prediction error as the residuum or geometric distance between the samples and the regression line characterizing the quality of the fit.18,19 In general, regression parameters are less sensitive to short-term fluctuations compared with respective parameters extracted from a cross-sectional analysis.18,21 We included only eyes where the time distance between the first and the last valid measurement of the target variable after suture removal was at least 12 months to enhance the predictive value of our results. For the regression analysis, sequences of three up to 16 examination parameters were used. Only follow-up examina29.e4
AMERICAN JOURNAL
OF
OPHTHALMOLOGY
JULY 2005
TABLE 2. Prediction Error for Keratometry Mean ⫾ SD; MinMax; Median in Diopters
Total population Fuchs’ dystrophy (FD, n ⫽ 47) Keratoconus (KC, n ⫽ 100)
Keratometric Power
0.32 ⫾ 0.24 0.0–1.1; 0.25 0.31 ⫾ 0.24; 0.0–1.1; 0.25 0.33 ⫾ 0.25 0.0–1.1; 0.26
Keratometric Astigmatism
0.43 ⫾ 0.32; 0.0–1.6; 0.37 0.40 ⫾ 0.36; 0.0–1.6; 0.29 0.44 ⫾ 0.31; 0.0–1.6; 0.38
Topographic Power
Topographic Astigmatism
Spherical Equivalent
0.36 ⫾ 0.29 0.0–1.3; 0.32 0.35 ⫾ 0.31 ⫺0.0–1.3; 0.26 0.36 ⫾ 0.28; 0.0–1.2; 0.34
0.40 ⫾ 0.34 0.0–1.81; 0.35 0.37 ⫾ 0.34 0.0–1.4; 0.24 0.42 ⫾ 0.35 0.0–1.8; 0.37
0.27 ⫾ 0.20; 0.0–0.96; 0.23 0.29 ⫾ 0.22; 0.0–0.9; 0.25 0.27 ⫾ 0.20; 0.0–1.0; 0.23
Refractive Cylinder
0.29 ⫾ 0.23; 0.0–1.0; 0.25 0.27 ⫾ 0.19; 0.0–0.7; 0.25 0.30 ⫾ 0.24; 0.0–1.0; 0.24
Prediction error (difference between the measured target parameter and the predicted data given by the regression model) of the target variables keratometric central power, keratometric astigmatism, topographic central power, topographic astigmatism, spherical equivalent, and refractive cylinder.
(P ⫽ .008) and a low decrease of astigmatism after suture removal. These results are in contrast with our observations, where the central corneal power decreased associated with a hyperopic shift in keratoconic patients. This discrepancy may be attributable to the different trephination technique and suturing technique. Yi and associates25 studied the impact of the preoperative corneal architecture on the refractive outcome after PK before and after suture removal in 236 eyes with KC. They found that the visual outcome following PK using non-mechanical trephination is independent of the patient’s pre-keratoplasty corneal curvature or irregularity, and that suture removal is responsible for a regularization of the keratometry mires and corneal topography. Many studies support substantial and rapid improvement of visual function after PK for KC.1,3,5,13,26 –28 However, only limited information is provided about functional results after PK for FD.29 In general, functional outcome after KC is superior compared with those after FD, which could be evaluated in a prospective comparative study in our department.4,7,21 Differences may be the because of the age structure of both patient groups and the presumed lower rate of age-related ocular diseases in KC. A further aspect may be the smaller trephination diameter of 7.5 mm in patients with FD in contrast with 8.0 mm in patients with KC, which may predispose eyes with FD to develop a higher degree of corneal irregularities or asymmetries attributable to the suture area being located more centrally.30,31 The latter decision for difference in the trephination diameters in FD and KC was based upon the significantly smaller corneal diameter in eyes with FD in contrast with eyes with KC.32 However, the graft oversize of 0.1 mm, the surgical and suturing technique, as well as the pre- and post-operative medication were standardized in the present study.4 Overall, the trephination technique used has an impact on the required graft oversize. Punching the donor from the endothelial side a nominally 0.25-mm “oversize” typically results in a same-size fit, because the recipient hole is VOL. 140, NO. 1
SPONTANEOUS LONG-TERM CHANGES
larger than the trephine size, especially at the endothelial level, and the donor button is smaller than the trephine size, especially at the endothelial level.33 Applying the Krumeich GTS, the refined Hanna trephine, or excimer laser trephination all of which produce the donor from the epithelial side using an artificial anterior chamber, may eliminate the need for donor oversize completely. On principle, a larger donor oversize produces a steeper graft—at least after suture removal. It is well known that single sutures tend to result in a flatter graft as long as they are in place. In conclusion, in the long-term follow-up after postkeratoplasty suture removal, patients with FD tend to develop a spontaneous myopic shift because of a steepening of the cornea, whereas patients with KC tend to develop a spontaneous hyperopic shift because of a flattening of the cornea. In contrast with those with FD, patients with KC should be counseled on the fact that astigmatism may increase again over time after suture removal. These findings may be important for the determination of the power of a lens implant in case of a subsequent cataract surgery especially if a toric intraocular lens is chosen. Additionally, the need of repeated adjustments of spectacles or contact lenses after PK, especially in case of KC, may be explained, at least in part.
REFERENCES 1. Brahma A, Ennis F, Harper R. Visual function after penetrating keratoplasty for keratoconus: a prospective longitudinal evaluation. Br J Ophthalmol 2000;84:60 – 66. 2. Kus MM, Seitz B, Langenbucher A, Naumann GOH. Endothelium and pachymetry of clear grafts 15 to 33 years after penetrating keratoplasty—a report on 20 patients. Am J Ophthalmol 1999;127:600 – 602. 3. Lim L, Pesudovs K, Coster DJ. Penetrating keratoplasty for keratoconus: visual outcome and success. Ophthalmology 2000;107:1125–1131. IN
CORNEAL POWER
AND
REFRACTION
29.e5
4. Seitz B, Langenbucher A, Kus MM, Küchle M, Naumann GOH. Nonmechanical corneal trephination with the excimer laser improves outcome after penetrating keratoplasty. Ophthalmology 1999;106:1156 –1165. 5. Silbiger JS, Cohen EJ, Laibson PR. The rate of visual recovery after penetrating keratoplasty performed on keratoconus patients. CLAO J 1996;22:266 –269. 6. Naumann GOH. Corneal transplantation in anterior segment diseases. The Bowman Lecture (Number 56) Part II. Eye 1995;9:395– 424. 7. Naumann GOH, Seitz B, Lang GK, Langenbucher A, Kus MM. One hundred ninety-three excimer laser trepanation in perforating keratoplasty. Report of 70 patients. Klin Monatsbl Augenheilkd 1993;203:252–261. 8. Price FW, Whitson WE, Johns S, Gonzales JS. Risk factors for corneal graft failure. J Refract Surg 1996;12:134 –147. 9. Price FW, Whitson WE, Marks RG. Graft survival in four common groups of patients undergoing penetrating keratoplasty. Ophthalmology 1991;98:322–328. 10. Wiffen SJ, Maguire LJ, Bourne WM. Keratometric results of penetrating keratoplasty with the Hessburg-Barron and Hanna trephine system using a standard double-running suture technique. Cornea 1997;16:306 –313. 11. Seitz B, Langenbucher A, Nguyen NX, Kus MM, Küchle M, Naumann GOH. Results of the first 1000 consecutive elective nonmechanical keratoplasties using the excimer laser. A prospective study over more than 12 years. Ophthalmologe 2004;101:478 – 488. 12. Mader TH, Yuan R, Lynn MJ. Changes in keratometric astigmatism after suture removal more than one year after penetrating keratoplasty. Ophthalmology 1993;100:119 –127. 13. Brierly SC, Izquierdo L Jr, Mannis MJ. Penetrating keratoplasty for keratoconus. Cornea 2000;19:329 –332. 14. Filatov V, Alexandrakis G, Talamo JH, Steinert RI. Comparison of suture-in and suture-out postkeratoplasty astigmatism with single running suture or combined running and interrupted sutures. Am J Ophthalmol 1996;122:696 –700. 15. Lin DTC, Wilson SE, Reidy JJ. Topographic changes that occur with 10-0 running suture removal following penetrating keratoplasty. Refract Corneal Surg 1990;6:21–25. 16. Tuft SJ, Gregory W. Long-term refraction and keratometry after penetrating keratoplasty for keratoconus. Cornea 1995; 14:614 – 617. 17. Hoffmann F. Suture technique for perforating keratoplasty. Klin Monatsbl Augenheilkd 1976;169:584 –590. 18. Langenbucher A, Nguyen NX, Kus MM, Blüthner K, Küchle M, Seitz B. Regression analysis of corneal endothelium after nonmechanical penetrating keratoplasty. Klin Monatsbl Augenheilkd 2000;216:393–399. 19. Langenbucher A, Seitz B, Nguyen NX, Naumann GOH. Corneal endothelial cell loss after nonmechanical penetrat-
29.e6
AMERICAN JOURNAL
20.
21.
22.
23.
24.
25.
26.
27.
28. 29.
30.
31.
32.
33.
OF
ing keratoplasty depends on diagnosis: a regression analysis. Graefes Arch Clin Exp Ophthalmol 2002;240:387–392. Vail A, Gore SM, Bradley BA. Clinical and surgical factors influencing corneal graft survival, visual acuity, and astigmatism. Ophthalmology 1996;193:41– 49. Seitz B, Langenbucher A, Nguyen NX, Kus MM, Küchle M, Naumann GOH. Graft endothelium and thickness after penetrating keratoplasty, comparing mechanical and nonmechanical excimer laser trephination: a prospective randomized study. Graefes Arch Clin Exp Ophthalmol 2001;239:12– 17. Musch DC, Meyer RF, Sugar A. The effect of removing running sutures on astigmatism after penetrating keratoplasty. Arch Ophthalmol 1988;106:488 – 492. Meyer HJ, Cordesmeyer H. Postoperative astigmatism following keratoplasty in keratoconus. Fortschr Ophthalmol 1987;84:427– 428. Ruhswurm I, Scholz U, Pfleger T, Zehetmayer M, Hanselmayer G, Skorpik C. Three year clinical outcome after penetrating keratoplasty for keratoconus with the guided trephine system. Am J Ophthalmol 1999;127:666 – 673. Yi L, Seitz B, Langenbucher A, Nguyen NX, Naumann GOH. Impact of preoperative corneal curvature on the outcome of penetrating keratoplasty in keratoconus. Cornea 2003;22:409 – 412. Buzard KA, Fundingsland BR. Corneal transplant for keratoconus: results in early and late disease. J Cataract Refract Surg 1997;23:398 – 406. Olsen RJ, Pingree M, Ridges R. Penetrating keratoplasty for keratoconus: a long-term review of results and complications. J Cataract Refract Surg 2000;26:987–991. Troutman RC, Lawless MA. Penetrating keratoplasty for keratoconus. Cornea 1987;6:298 –305. Pineros O, Cohen EJ, Rapuano CJ. Long-term results after penetrating keratoplasty for Fuchs’ endothelial dystrophy. Arch Ophthalmol 1996;114:15–18. Cohen KL, Tripoli NK, Pellom AC. Effect of tissue fit on corneal shape after transplantation. Invest Ophthalmol Vis Sci 1984;25:1226 –1231. Seitz B, Langenbucher A, Küchle M Naumann GOH. Impact of graft diameter on corneal power and the regularity of postkeratoplasty astigmatism before and after suture removal. Ophthalmology 2003;110:2162–2167. Seitz B, Langenbucher A, Zagrada D, Budde W, Kus MM. Corneal dimensions in various types of corneal dystrophies and their effect on penetrating keratoplasty. Klin Monatsbl Augenheilkd 2000;217:152–158. Troutman RC. Consultations. Refractive Corneal Surg 1992; 8:246 –247.
OPHTHALMOLOGY
JULY 2005
Biosketch Achim Langenbucher, PhD, is Assistant Professor in the Department of Ophthalmology, University of Erlangen¨ urnberg, Erlangen, Germany. He studied bioengineering and received his degree in 1992. After completing his doctoral N thesis in 1995, he received his PhD in 2000. Dr. Langenbucher investigates the applications of lasers in ophthalmology in the field of diagnostics and therapy; and he is head of the research group for ophthalmologic lasers and optics in the Department of Ophthalmology. He has received the Achievement Award of the American Academy of Ophthalmology in 2001.
VOL. 140, NO. 1
SPONTANEOUS LONG-TERM CHANGES
IN
CORNEAL POWER
AND
REFRACTION
29.e7
● PURPOSE: To assess the diagnosis-based spontaneous long-term changes in corneal power and refraction with a regression model in the all-sutures-out time period following non-mechanical penetrating keratoplasty (PK). ● DESIGN: Retrospective non-randomized clinical trial. ● METHODS: SETTING: Clinical practice. STUDY POPULATION: 147 eyes [47 Fuchs dystrophy (FD); 100 keratoconus (KC)] were studied after suture removal in this retrospective longitudinal study. MAIN OUTCOME MEASURES: Zeiss keratometry [equivalent power (KEQ) and astigmatism (KAST)], corneal topography analysis [equivalent power (TEQ) and astigmatism (TAST)], and subjective refractometry [spherical equivalent (SEQ) and refractive cylinder (RAST)] were assessed in at least three up to 16 ophthalmologic examinations in the all-sutures-out time period. OBSERVATION PROCEDURE: The time course of each target variable was analyzed in a longitudinal manner (time interval > 12 months) separately for each patient with a linear regression model. ● RESULTS: Post-keratoplasty follow-up ranged from 31 months to 10.3 years. In the linear regression model, the annual change in FD/KC showed an increase/a decrease in KEQ (0.29 ⴞ 0.50/ⴚ0.63 ⴞ 0.46 diopters, P ⴝ .02) and an increase/a decrease in TEQ (0.37 ⴞ 0.54/ⴚ0.69 ⴞ 0.49 diopters, P ⴝ .04) corresponding to a decrease/an increase in SEQ (ⴚ0.31 ⴞ 0.47/0.63 ⴞ 0.43 diopters, P ⴝ .02). KAST/TAST/RAST showed a minimal annual decrease (ⴚ0.06 ⴞ 0.41/ⴚ0.05 ⴞ 0.45/ⴚ0.06 ⴞ 0.41 diopters) in FD but an increase in KC (0.46 ⴞ 0.41/0.51 ⴞ 0.43/0.46 ⴞ 0.38 diopters) (P ⴝ .05/0.06/0.12).
Accepted for publication Jan 26, 2005. From the Augenklinik mit Poliklinik der Universität Erlangen-Nürnberg, Erlangen, Germany. Inquires to Achim Langenbucher, PhD, Augenklinik mit Poliklinik der Universität Erlangen-Nürnberg, Schwabachanlage 6, D-91054 Erlangen, Germany; fax: 09131-853-4271; e-mail: achim.langenbucher@augen. imed.uni-erlangen.de 0002-9394/05/$30.00 doi:10.1016/j.ajo.2005.01.038
©
2005 BY
● CONCLUSIONS: In the follow-up after post-keratoplasty suture removal, patients with FD/KC tend to develop a spontaneous myopic shift (steepening of the cornea)/hyperopic shift (flattening of the cornea). In contrast with those with FD, patients with KC should be counseled on the fact that astigmatism may increase again over time after suture removal. (Am J Ophthalmol 2005;140:29 –34. © 2005 by Elsevier Inc. All rights reserved.)
I
N NON– HIGH-RISK KERATOPLASTIES THE PROGNOSIS OF
corneal grafts has improved in the last decades because of enhanced treatment techniques and storage methods and is superior compared with all other organ transplantations. Primary goals after corneal transplantation are the preservation of a clear graft and the rehabilitation of the visual function.1–5 Many studies have been performed to assess the influencing factors for graft survival and the rate of repeat keratoplasty because of graft failure.6 –10 From June 1989 to December 2004, more than 1650 non-mechanical penetrating keratoplasties (PK) have been performed using the 193 nm excimer laser in our Department of Ophthalmology in Erlangen. The most frequent indications were keratoconus (KC) and Fuchs dystrophy (FD). Especially in young patients with KC, non-mechanical trephination offers clear advantages, such as reduction of keratometric astigmatism and refractive cylinder, higher regularity of the corneal topography, less asymmetry in the topographic pattern, and a significantly better visual outcome, which are the crucial criteria for patient satisfaction.4,7,11 In a prospective, randomized, cross-sectional, clinical, single-center study including 179 eyes with primary central PK for FD (diameter 7.5 mm) or KC (diameter 8.0 mm), graft oversize 0.1 mm, 16-bite, double-running diagonal suture, we found a significantly smaller keratometric astigmatism (3.0 diopters) after suture
ELSEVIER INC. ALL
RIGHTS RESERVED.
29.e1
● MAIN OUTCOME MEASURES AND REGRESSION ANALYSIS: Main outcome measures of parameters included: (1)
removal for laser trephination in contrast with 6.1 diopter for motor trephination (P ⬍ .0009). In contrast, Mader and colleagues12 reported a mean keratometric astigmatism of 5.6 diopters with all sutures out after conventional trephination in 162 eyes. Wiffen and associates published a final median keratometric astigmatism of 5.5 diopters using the Hessburg-Barron trephine.10 With sutures in place, the corneal topography is mostly determined by the suture tension, whereas the pattern of corneal topography after suture removal is affected by the wound configuration and the adaptation between the donor and recipient cut.12–16 At this time, very little is known about the long-term behavior of the all-suture-out corneal curvature following PK. The purpose of this study was to assess the spontaneous long-term changes of refraction, corneal power, and astigmatism in the all-suture-out time period following nonmechanical PK comparing eyes with KC and FD in a longitudinal fashion, using a linear regression fit separately for each variable and each patient.
keratometric corneal power (Zeiss Ophthalmometer, Zeiss), (2) keratometric astigmatism, (3) topographic corneal power (TMS-1, release 1.61, Tomey, Erlangen, Germany), (4) topographic astigmatism, (5) spherical equivalent of the refraction (trial glasses in a trial frame), and (6) refractive cylinder. For each patient at least three up to 16 valid measurements of parameters (1)–(6) in the all-sutures-out time period (baseline examination around 6 weeks after final suture removal) in a time interval of at least 1 year were used to formulate an individual regression model separately for each patient and each parameter. The slope of the linear regression yields the absolute change of the target parameters/time interval. From the difference between the measured target parameter and the predicted data given by the regression model, the residuum was calculated and considered the prediction error of the model. This method has been described in detail in previous studies applied to the problem of graft endothelial cell loss after PK.18,19 ● STATISTICAL ANALYSIS AND REGRESSION MODELS:
Data were recruited and collected in a relational database system (Access 97 for Windows) and analyzed using SPSS/PC 11.0 (Windows NT). Comparisons between patient groups (FD vs KC) were performed using non-parametric tests (Mann-Whitney U test for unpaired samples). Linear regression analysis was performed separately for each target variable and each patient to describe the time course of the values after suture removal. For evaluation of the prediction error, we derived the Durbin-Watson residuum between the observed and the predicted data from the fit. For assessment of correlations between variables the Spearman correlation coefficient was calculated. A P-value ⬍.05 was considered statistically significant.
METHODS ● STUDY POPULATION/DONOR DETAILS AND SURGICAL INTERVENTION: Inclusion criteria for this retrospective, non-randomized clinical study consisted of: primary central round PK without previous surgery, FD with a trephination diameter 7.5/7.6 mm in recipient/donor (n ⫽ 47; 12 men, 35 women; 27 left eye [OS], 20 right eye [OD], or KC with a trephination diameter 8.0/8.1 mm in recipient/donor (n ⫽ 100; 71 men, 29 women; 51 OS, 49 OD), interval of surgery between June 1992 and March 2002. Exclusion criteria were maculopathies, optic nerve atrophies, amblyopia, complications such as immunologic graft reactions during follow-up and further surgery such as arcuate keratotomies, cataract surgery, and so on. In 147 patients, non-mechanical PK was performed with the 193 nm excimer laser (MEL 60, Zeiss-Meditec, Jena, Germany) along metal aperture masks with eight orientation teeth/notches. In 29 patients with FD, a simultaneous extracapsular cataract extraction and posterior chamber lens implantation (triple procedure) was necessary. After temporary fixation of the donor button in the recipient bed with eight interrupted 10-0 nylon sutures, permanent wound closure was ensured by a 16-bite, double-running, diagonal cross-stitch suture.17 The pre- and postoperative medication was standardized. Pre-operatively, patients with PK only received pilocarpine 2% eye drops; those with triple procedure received atropine 1% eye drops. After surgery, ofloxacin ointment, cycloplegics, and a vitamin D– containing gel was applied. Topical corticosteroids (prednisolone acetate) were started five times a day not before epithelial closure and were tapered by one drop a day every 6 weeks.
29.e2
AMERICAN JOURNAL
RESULTS OVERALL, THE MEAN FOLLOW-UP PERIOD WAS 3.9 ⫾
1.7 years (range 31 months to 10.3 years). In FD, the individual follow-up ranged from 31 months to 9.1 years with a mean of 4.0 ⫾ 1.6 years and in KC, it ranged from 33 months to 10.3 years with a mean of 3.9 ⫾ 1.7 years. Between groups, there was no statistically significant difference. No patient in the FD or KC group developed an immunologic graft rejection in the follow-up period. One of the double-running sutures was removed 11 to 14 months after PK, and the second suture was removed 15 to 21 months after PK. The time interval between PK and suture removal did not differ significantly between FD and KC. The mean time interval for regression analysis, time period between first and last follow-up examination in the all-sutures-out time period, was 2.7 ⫾ 1.5 years (range 12 months to 8.4 years). In FD, it was 2.6 ⫾ 1.4 years (range OF
OPHTHALMOLOGY
JULY 2005
TABLE 1. Linear Regression Analysis Mean ⫾ SD; MinMax; Median in Diopters
Total population Fuchs’ dystrophy (FD, n ⫽ 47) P Keratoconus (KC, n ⫽ 100)
Keratometric Central Power
Keratometric Astigmatism
Topographic Central Power
Topographic Astigmatism
Spherical Equivalent
Refractive Cylinder
⫺0.33 ⫾ 0.64 ⫺2.0–1.1; ⫺0.42 0.29 ⫾ 0.50; ⫺1.0–1.1; 0.29 .02 ⫺0.63 ⫾ 0.46 ⫺2.0–0.5; ⫺0.62
0.29 ⫾ 0.47; ⫺1.4–1.6; 0.27 ⫺0.06 ⫾ 0.41; ⫺1.4–1.1; ⫺0.06 .05 0.46 ⫾ 0.41; ⫺0.4–1.6; 0.47
⫺0.35 ⫾ 0.71; ⫺2.0–1.5; ⫺0.44 0.37 ⫾ 0.54 ⫺1.2–1.5; 0.43 .04 ⫺0.69 ⫾ 0.49; ⫺2.0–0.4; ⫺0.73
0.33 ⫾ 0.51 ⫺1.1–1.9; 0.32 ⫺0.05 ⫾ 0.45 ⫺1.1–1.1; 0.0 .06 0.51 ⫾ 0.43 ⫺0.4–1.9; 0.50
0.32 ⫾ 0.63; ⫺1.1–1.7; 0.46 0.31 ⫾ 0.47; ⫺1.1–0.9; ⫺0.35 .02 0.63 ⫾ 0.43; ⫺0.6–1.7; 0.69
0.30 ⫾ 0.46; ⫺1.3–1.6; 0.28 0.06 ⫾ 0.41; ⫺1.3–1.0; ⫺0.09 .12 0.46 ⫾ 0.38; ⫺0.4–1.6; 0.46
Slope of the linear regression analysis (absolute change per year) of the target variables keratometric central power, keratometric astigmatism, topographic central power, topographic astigmatism, spherical equivalent, and refractive cylinder. The P-values refer to the significance level comparing FD and KC.
12 months to 6.5 years), and in KC, it was 2.7 ⫾ 1.4 years (range 12 months to 8.4 years) without significant differences between groups. The overall number of all-suture-out examinations with valid target variables was 5.9 ⫾ 3.0 (range 3 to 16). In FD, it was 6.0 ⫾ 3.0 (range 3 to 16), and in KC, it was 5.9 ⫾ 3.0 (range 3 to 15). The regression analysis of keratometry yielded a mean decrease of keratometric central power of 0.33 ⫾ 0.64 diopters per year and an increase of the keratometric astigmatism of 0.29 ⫾ 0.47 diopters per year (Table 1). Figure 1 shows the change in spherical equivalent, keratometric power, and topographic power for both groups of patients. The regression analysis of corneal topographic data yielded a decrease of topographic central power of 0.35 ⫾ 0.71 diopters per year, and an increase of the topographic astigmatism of 0.33 ⫾ 0.51 diopters per year (Table 1, Figure 2). The regression analysis of subjective refraction yielded an increase of the spherical equivalent of 0.32 ⫾ 0.62 diopters per year, and an increase of the refractive cylinder of 0.30 ⫾ 0.46 diopters per year. Overall, the change in subjective spherical equivalent correlated inversely with the change in keratometric central power (r2 ⫽ .81, P ⫽ .02; FD: r2 ⫽ .66, P ⫽ .04; KC: r2 ⫽ .86, P ⫽ .01) and with the change in topographic central power (r2 ⫽ .66, P ⫽ .04; FD: r2 ⫽ 0.55, P ⫽ .04; KC: r2 ⫽ .72, P ⫽ .02). Change in refractive cylinder correlated tendentially with the change in keratometric astigmatism (r2 ⫽ .40, P ⫽ .04; FD: r2 ⫽ .36, P ⫽ .06; KC: r2 ⫽ .41, P ⫽ .04) and with the change in topographic astigmatism (r2 ⫽ .38, P ⫽ .06; FD: r2 ⫽ .30, P ⫽ .08; KC: r2 ⫽ .51, P ⫽ .04). The prediction error for keratometry was lower for the keratometric power (0.32 ⫾ 0.24 diopters) compared with the keratometric astigmatism (0.42 ⫾ 0.32 diopters, P ⫽ .02, Table 2). The prediction error for corneal topographic data was tendentially lower for keratometric power (0.36 ⫾ 0.29 diopters) compared with keratometric astigmatism (0.40 ⫾ 0.34 diopters, P ⫽ .05, Table 2). The prediction error of VOL. 140, NO. 1
SPONTANEOUS LONG-TERM CHANGES
FIGURE 1. Change over time in diopters (1 year, annual change) described by the slope of the linear regression line of spherical equivalent (subjective refraction evaluated with trial glasses in a trial frame), keratometric central power, and topographic central power of patients with Fuchs’ dystrophy (FD; n ⴝ 47) and keratoconus (KC; n ⴝ 100) after suture removal (all sutures out) following non-mechanical penetrating keratoplasty using the excimer laser 193 nm. Positive/negative values indicate an increase/ decrease of the spherical equivalent or corneal power over time, whereas values that equal zero indicate a stable spherical equivalent or corneal power over time.
subjective refraction was comparable for the spherical equivalent (0.27 ⫾ 0.20 diopters) and the refractive cylinder (0.29 ⫾ 0.23 diopters, P ⫽ .25, Table 2).
DISCUSSION CORNEAL TRANSPLANTATION, BOTH IN FD AND IN KC, ARE
among the main diagnoses for normal-risk keratoplasties IN
CORNEAL POWER
AND
REFRACTION
29.e3
tions with a complete set of target variables were considered for the analysis. In FD, the spherical equivalent decreased, indicating a myopic shift after suture removal, whereas in KC, the spherical equivalent increases, indicating a hyperopic shift. The latter finding was unexpected. These tendencies are in accordance with the steepening of the cornea in FD measured with keratometry and with corneal topography and the flattening of the cornea in KC. The astigmatism remained stable in FD but gained in KC over time. This effect correlated only in part with the objective measurement of the keratometry and topography analyses. In general, the prediction error for all target variables was small, indicating a good representation of the longitudinal data with the linear regression model applied in this study. Unfortunately, in the literature authors rarely differentiate clearly between sutures-in and all-sutures-out results. Concerning the changes attributable to post-keratoplasty suture removal, classic papers include those by Filatov and associates,14 Mader and associates12 and Musch and associates22 From these papers it becomes clear that corneal astigmatism may change unpredictably and by large amounts when all remaining sutures are removed after PK. Mader and colleagues reported that in 62% of eyes the astigmatism changed 2 or more diopters (range ⫹11.94 diopters to ⫺17.87 diopters). The average vectorial change was 6.5 diopters (range 0.56 diopters to 19.8 diopters). There was no decrease in the amount of astigmatic change with increasing time between surgery and suture removal. Graft size and diagnosis had no effect on the amount of astigmatic change. Astigmatic errors became stable with less than 1 diopter of change between successive examinations within 6 months after suture removal. Similar results were obtained for eyes with double-running and interrupted-running sutures.12 In a prospective randomized study, we found that the change of vector-corrected astigmatism was significantly smaller after suture removal following laser trephination (4.3 diopters), in contrast with motor trephination (6.9 diopters) [Seitz and associates, oral paper, presented at the ARVO meeting 2000 in Fort Lauderdale, Florida]. In contrast, the present study looked at spontaneous long-term changes in the all-sutures-out time period after post-keratoplasty suture removal. This question has been studied very rarely in the literature up to now: Tuft and Gregory16 found a trend toward a gradual flattening and increase of corneal cylinder over time in KC. This is in accordance with our present findings. Small amounts of central flattening and increased astigmatism may be explained by the irregular bulging of the keratoconic recipient rim allowing the graft-host-junction to act like a hinge thus leaving the graft center less curved.23 Additionally, Ruhswurm and associates24 studied the long-term clinical outcome after keratoplasty with the guided trephine system in 31 eyes with KC. They found a myopic shift associated with a steepening of the cornea
FIGURE 2. Change over time in diopters (1 year, annual change) described by the slope of the linear regression line of refractive cylinder (subjective refraction evaluated with trial glasses in a trial frame), keratometric astigmatism, and topographic astigmatism of patients with FD (n ⴝ 47) and KC (n ⴝ 100) after suture removal (all sutures out) following nonmechanical penetrating keratoplasty using the excimer laser 193 nm. Positive/negative values indicate an increase/decrease over time, whereas values that equal zero indicate a stable refractive cylinder or corneal astigmatism over time.
with a superior prognosis because of a low rate of immunologic graft rejections, low rate of post-keratoplasty ocular hypertension, and a very low rate of irreversible graft failures.4,20 In the present study, we focused on the spontaneous changes of corneal curvature and subjective refraction after suture removal (all sutures out). Of a sequence of at least three valid data sets after suture removal consisting of spherical equivalent, refractive cylinder, keratometric central corneal power, keratometric astigmatism, topographic central corneal power, and topographic astigmatism, we fitted linear regression lines separately for each patient to evaluate the longitudinal section of these target variables. The characteristic result of each regression analysis fitted to the data sets is (1) the slope of the regression line, which indicates the annual change of the target variable, and (2) the prediction error as the residuum or geometric distance between the samples and the regression line characterizing the quality of the fit.18,19 In general, regression parameters are less sensitive to short-term fluctuations compared with respective parameters extracted from a cross-sectional analysis.18,21 We included only eyes where the time distance between the first and the last valid measurement of the target variable after suture removal was at least 12 months to enhance the predictive value of our results. For the regression analysis, sequences of three up to 16 examination parameters were used. Only follow-up examina29.e4
AMERICAN JOURNAL
OF
OPHTHALMOLOGY
JULY 2005
TABLE 2. Prediction Error for Keratometry Mean ⫾ SD; MinMax; Median in Diopters
Total population Fuchs’ dystrophy (FD, n ⫽ 47) Keratoconus (KC, n ⫽ 100)
Keratometric Power
0.32 ⫾ 0.24 0.0–1.1; 0.25 0.31 ⫾ 0.24; 0.0–1.1; 0.25 0.33 ⫾ 0.25 0.0–1.1; 0.26
Keratometric Astigmatism
0.43 ⫾ 0.32; 0.0–1.6; 0.37 0.40 ⫾ 0.36; 0.0–1.6; 0.29 0.44 ⫾ 0.31; 0.0–1.6; 0.38
Topographic Power
Topographic Astigmatism
Spherical Equivalent
0.36 ⫾ 0.29 0.0–1.3; 0.32 0.35 ⫾ 0.31 ⫺0.0–1.3; 0.26 0.36 ⫾ 0.28; 0.0–1.2; 0.34
0.40 ⫾ 0.34 0.0–1.81; 0.35 0.37 ⫾ 0.34 0.0–1.4; 0.24 0.42 ⫾ 0.35 0.0–1.8; 0.37
0.27 ⫾ 0.20; 0.0–0.96; 0.23 0.29 ⫾ 0.22; 0.0–0.9; 0.25 0.27 ⫾ 0.20; 0.0–1.0; 0.23
Refractive Cylinder
0.29 ⫾ 0.23; 0.0–1.0; 0.25 0.27 ⫾ 0.19; 0.0–0.7; 0.25 0.30 ⫾ 0.24; 0.0–1.0; 0.24
Prediction error (difference between the measured target parameter and the predicted data given by the regression model) of the target variables keratometric central power, keratometric astigmatism, topographic central power, topographic astigmatism, spherical equivalent, and refractive cylinder.
(P ⫽ .008) and a low decrease of astigmatism after suture removal. These results are in contrast with our observations, where the central corneal power decreased associated with a hyperopic shift in keratoconic patients. This discrepancy may be attributable to the different trephination technique and suturing technique. Yi and associates25 studied the impact of the preoperative corneal architecture on the refractive outcome after PK before and after suture removal in 236 eyes with KC. They found that the visual outcome following PK using non-mechanical trephination is independent of the patient’s pre-keratoplasty corneal curvature or irregularity, and that suture removal is responsible for a regularization of the keratometry mires and corneal topography. Many studies support substantial and rapid improvement of visual function after PK for KC.1,3,5,13,26 –28 However, only limited information is provided about functional results after PK for FD.29 In general, functional outcome after KC is superior compared with those after FD, which could be evaluated in a prospective comparative study in our department.4,7,21 Differences may be the because of the age structure of both patient groups and the presumed lower rate of age-related ocular diseases in KC. A further aspect may be the smaller trephination diameter of 7.5 mm in patients with FD in contrast with 8.0 mm in patients with KC, which may predispose eyes with FD to develop a higher degree of corneal irregularities or asymmetries attributable to the suture area being located more centrally.30,31 The latter decision for difference in the trephination diameters in FD and KC was based upon the significantly smaller corneal diameter in eyes with FD in contrast with eyes with KC.32 However, the graft oversize of 0.1 mm, the surgical and suturing technique, as well as the pre- and post-operative medication were standardized in the present study.4 Overall, the trephination technique used has an impact on the required graft oversize. Punching the donor from the endothelial side a nominally 0.25-mm “oversize” typically results in a same-size fit, because the recipient hole is VOL. 140, NO. 1
SPONTANEOUS LONG-TERM CHANGES
larger than the trephine size, especially at the endothelial level, and the donor button is smaller than the trephine size, especially at the endothelial level.33 Applying the Krumeich GTS, the refined Hanna trephine, or excimer laser trephination all of which produce the donor from the epithelial side using an artificial anterior chamber, may eliminate the need for donor oversize completely. On principle, a larger donor oversize produces a steeper graft—at least after suture removal. It is well known that single sutures tend to result in a flatter graft as long as they are in place. In conclusion, in the long-term follow-up after postkeratoplasty suture removal, patients with FD tend to develop a spontaneous myopic shift because of a steepening of the cornea, whereas patients with KC tend to develop a spontaneous hyperopic shift because of a flattening of the cornea. In contrast with those with FD, patients with KC should be counseled on the fact that astigmatism may increase again over time after suture removal. These findings may be important for the determination of the power of a lens implant in case of a subsequent cataract surgery especially if a toric intraocular lens is chosen. Additionally, the need of repeated adjustments of spectacles or contact lenses after PK, especially in case of KC, may be explained, at least in part.
REFERENCES 1. Brahma A, Ennis F, Harper R. Visual function after penetrating keratoplasty for keratoconus: a prospective longitudinal evaluation. Br J Ophthalmol 2000;84:60 – 66. 2. Kus MM, Seitz B, Langenbucher A, Naumann GOH. Endothelium and pachymetry of clear grafts 15 to 33 years after penetrating keratoplasty—a report on 20 patients. Am J Ophthalmol 1999;127:600 – 602. 3. Lim L, Pesudovs K, Coster DJ. Penetrating keratoplasty for keratoconus: visual outcome and success. Ophthalmology 2000;107:1125–1131. IN
CORNEAL POWER
AND
REFRACTION
29.e5
4. Seitz B, Langenbucher A, Kus MM, Küchle M, Naumann GOH. Nonmechanical corneal trephination with the excimer laser improves outcome after penetrating keratoplasty. Ophthalmology 1999;106:1156 –1165. 5. Silbiger JS, Cohen EJ, Laibson PR. The rate of visual recovery after penetrating keratoplasty performed on keratoconus patients. CLAO J 1996;22:266 –269. 6. Naumann GOH. Corneal transplantation in anterior segment diseases. The Bowman Lecture (Number 56) Part II. Eye 1995;9:395– 424. 7. Naumann GOH, Seitz B, Lang GK, Langenbucher A, Kus MM. One hundred ninety-three excimer laser trepanation in perforating keratoplasty. Report of 70 patients. Klin Monatsbl Augenheilkd 1993;203:252–261. 8. Price FW, Whitson WE, Johns S, Gonzales JS. Risk factors for corneal graft failure. J Refract Surg 1996;12:134 –147. 9. Price FW, Whitson WE, Marks RG. Graft survival in four common groups of patients undergoing penetrating keratoplasty. Ophthalmology 1991;98:322–328. 10. Wiffen SJ, Maguire LJ, Bourne WM. Keratometric results of penetrating keratoplasty with the Hessburg-Barron and Hanna trephine system using a standard double-running suture technique. Cornea 1997;16:306 –313. 11. Seitz B, Langenbucher A, Nguyen NX, Kus MM, Küchle M, Naumann GOH. Results of the first 1000 consecutive elective nonmechanical keratoplasties using the excimer laser. A prospective study over more than 12 years. Ophthalmologe 2004;101:478 – 488. 12. Mader TH, Yuan R, Lynn MJ. Changes in keratometric astigmatism after suture removal more than one year after penetrating keratoplasty. Ophthalmology 1993;100:119 –127. 13. Brierly SC, Izquierdo L Jr, Mannis MJ. Penetrating keratoplasty for keratoconus. Cornea 2000;19:329 –332. 14. Filatov V, Alexandrakis G, Talamo JH, Steinert RI. Comparison of suture-in and suture-out postkeratoplasty astigmatism with single running suture or combined running and interrupted sutures. Am J Ophthalmol 1996;122:696 –700. 15. Lin DTC, Wilson SE, Reidy JJ. Topographic changes that occur with 10-0 running suture removal following penetrating keratoplasty. Refract Corneal Surg 1990;6:21–25. 16. Tuft SJ, Gregory W. Long-term refraction and keratometry after penetrating keratoplasty for keratoconus. Cornea 1995; 14:614 – 617. 17. Hoffmann F. Suture technique for perforating keratoplasty. Klin Monatsbl Augenheilkd 1976;169:584 –590. 18. Langenbucher A, Nguyen NX, Kus MM, Blüthner K, Küchle M, Seitz B. Regression analysis of corneal endothelium after nonmechanical penetrating keratoplasty. Klin Monatsbl Augenheilkd 2000;216:393–399. 19. Langenbucher A, Seitz B, Nguyen NX, Naumann GOH. Corneal endothelial cell loss after nonmechanical penetrat-
29.e6
AMERICAN JOURNAL
20.
21.
22.
23.
24.
25.
26.
27.
28. 29.
30.
31.
32.
33.
OF
ing keratoplasty depends on diagnosis: a regression analysis. Graefes Arch Clin Exp Ophthalmol 2002;240:387–392. Vail A, Gore SM, Bradley BA. Clinical and surgical factors influencing corneal graft survival, visual acuity, and astigmatism. Ophthalmology 1996;193:41– 49. Seitz B, Langenbucher A, Nguyen NX, Kus MM, Küchle M, Naumann GOH. Graft endothelium and thickness after penetrating keratoplasty, comparing mechanical and nonmechanical excimer laser trephination: a prospective randomized study. Graefes Arch Clin Exp Ophthalmol 2001;239:12– 17. Musch DC, Meyer RF, Sugar A. The effect of removing running sutures on astigmatism after penetrating keratoplasty. Arch Ophthalmol 1988;106:488 – 492. Meyer HJ, Cordesmeyer H. Postoperative astigmatism following keratoplasty in keratoconus. Fortschr Ophthalmol 1987;84:427– 428. Ruhswurm I, Scholz U, Pfleger T, Zehetmayer M, Hanselmayer G, Skorpik C. Three year clinical outcome after penetrating keratoplasty for keratoconus with the guided trephine system. Am J Ophthalmol 1999;127:666 – 673. Yi L, Seitz B, Langenbucher A, Nguyen NX, Naumann GOH. Impact of preoperative corneal curvature on the outcome of penetrating keratoplasty in keratoconus. Cornea 2003;22:409 – 412. Buzard KA, Fundingsland BR. Corneal transplant for keratoconus: results in early and late disease. J Cataract Refract Surg 1997;23:398 – 406. Olsen RJ, Pingree M, Ridges R. Penetrating keratoplasty for keratoconus: a long-term review of results and complications. J Cataract Refract Surg 2000;26:987–991. Troutman RC, Lawless MA. Penetrating keratoplasty for keratoconus. Cornea 1987;6:298 –305. Pineros O, Cohen EJ, Rapuano CJ. Long-term results after penetrating keratoplasty for Fuchs’ endothelial dystrophy. Arch Ophthalmol 1996;114:15–18. Cohen KL, Tripoli NK, Pellom AC. Effect of tissue fit on corneal shape after transplantation. Invest Ophthalmol Vis Sci 1984;25:1226 –1231. Seitz B, Langenbucher A, Küchle M Naumann GOH. Impact of graft diameter on corneal power and the regularity of postkeratoplasty astigmatism before and after suture removal. Ophthalmology 2003;110:2162–2167. Seitz B, Langenbucher A, Zagrada D, Budde W, Kus MM. Corneal dimensions in various types of corneal dystrophies and their effect on penetrating keratoplasty. Klin Monatsbl Augenheilkd 2000;217:152–158. Troutman RC. Consultations. Refractive Corneal Surg 1992; 8:246 –247.
OPHTHALMOLOGY
JULY 2005
Biosketch Achim Langenbucher, PhD, is Assistant Professor in the Department of Ophthalmology, University of Erlangen¨ urnberg, Erlangen, Germany. He studied bioengineering and received his degree in 1992. After completing his doctoral N thesis in 1995, he received his PhD in 2000. Dr. Langenbucher investigates the applications of lasers in ophthalmology in the field of diagnostics and therapy; and he is head of the research group for ophthalmologic lasers and optics in the Department of Ophthalmology. He has received the Achievement Award of the American Academy of Ophthalmology in 2001.
VOL. 140, NO. 1
SPONTANEOUS LONG-TERM CHANGES
IN
CORNEAL POWER
AND
REFRACTION
29.e7