The Combined (Genesis) Technique of Radial Keratotomy

The Combined (Genesis) Technique of Radial Keratotomy A Prospective, Multicenter Study Steven M. Verity, MD/ Jonathan H. Talamo, MD,2 Arturo Chayet, MD,3 Thomas C. Wolf, MD, 4 Peter Rapoza, MD, 5 David]. Schanz lin, MD, l Stephen Lane, MD, 6 Ken Kenyon, MD, 5 Kerry K. Assil, MD,l for the Refractive Keratoplasty Study Group Purpose: A prospective, multicenter, clinical study to evaluate a standardized surgical (Genesis) protocol which includes combined-technique radial incisions in patients seeking reduction of their physiologic myopia. The combined incisions were designed to incorporate the safety of the centrifugal incision technique with the efficacy of the centripetal incision technique. Methods: A total of 375 eyes undergoing radial keratotomy procedures performed in six different clinical centers were analyzed. All procedures were performed in accordance with the Genesis nomograms. The Genesis protocol called for using preoperative screening pachymetry to guide central clear zone size selection, incising the thinnest corneal quadrant first, suturing comeal perforations, and discouraged more than one enhancement procedure, when indicated. Globe fixation technique served as a study variable. Resuns: Mean follow-up was 6.2 months (range, 1.5-12 months). Mean residual cycloplegic refraction was -0.48 ± 0.61 diopters (D) (range, -2.50 to +1.50 D); 92% of eyes were within 1 D of the planned goal of -0.50 D and 85% were within 1 D of emmetropia; 14% were myopic; and 1% was hyperopic by more than 1 D. Uncorrected visual acuity was 20/40 or better in 95% of eyes; the remaining 5% retaining myopic refractive errors. A single procedure was performed in 73% of eyes, and 99% received less than two enhancements. Of eyes with no enhancements, 97% had uncorrected visual acuity of 20/40 or better. One study eye (0.3%) had a two-line loss of spectacle visual acuity. There were no invasions of the central clear zone. Globe fixation was a significant predictor for enhancement incidence (P < 0.001) but not for perforation incidence (P = 0.06). Incision sequence was predictive for perforation incidence (P < 0.0002). Conclusion: The combined-technique of radial keratotomy, coupled with the Genesis surgical protocol, affords centrifugal incision safety with centripetal incision efficacy. The Genesis nomograms, with a built in refractive outcome goal of -0.50 D provide an acceptable degree of accuracy and predictability while guarding against hyperopic overcorrection. Ophthalmology 1995;102:1908-1~17

Originally received: August 30, 1994. Revison accepted: July 18, 1995. I Anheuser-Busch Eye Institute Saint Louis University Health Sciences Center, St. Louis. 2 Massachusetts Eye and Ear Infirmary Harvard Medical School, Boston. 3 Centro Oftalmologico De Tijuana, Tijuana, Mexico. 4 Dean A. McGee Eye Institute University of Oklahoma, Oklahoma City. S Cornea Consultants, Boston.

1908

Associated Eye Physicians, Stillwater, Minnesota. Presented at the American Academy of Ophthalmology Annual Meeting, San Francisco, Oct/Nov 1994. The authors have no proprietary interest in any of the material presented in this article. Dr. Assil serves as an instructor for Alcon Surgical, Fort Worth, Texas, receiving honoraria for teaching courses in refractive keratotomy. Reprint requests to Kerry K. Assil, MD, SINSKEY Eye Institute, 2232 Santa Monica Blvd, Santa Monica, CA 90404. 6

Verity et al . Combined-technique (Genesis) RK Since its development by Russian ophthalmologists and early prospective evaluation in the United States ,' >' radial keratotomy (RK) techniques and technology have evolved to enable a greater degree of safety and predictability. Technologic advances in instrumentation, the use of computer-assisted videokeratography, as well as a greater understanding of corneal biomechanics, including the influences of patient age, wound morphology, and wound healing, all have contributed to the improved level of safety, efficacy, and predictability.v " At the time of its introduction into the United States, the procedure as performed in Russia consisted of radial corneal incisions beginning at the limbus and extending in a centripetal fashion to the central clear zone. In the interest of safety, the Prospective Evaluation of Radial Keratotomy (PERK) Study Group adopted an incisional technique that radiated from the central clear zone to the limbus. Centripetal or "Russian-style" incisions since have been demonstrated to achieve greater uniformity and depth for a given diamond blade setting compared with centrifugal or "American-style" incisions, possibly yielding a greater refractive effect."" A third "combined" technique for performing RK incisions was de veloped in an effort to couple centrifugal incision safet y with centripetal incision

Table 1. Exclusion Criteria Goal of residual myopia > 1.25 0 Age <19 or >65 yrs Best-corrected visual acuity <20/20 Ocular pathology (e.g., retinal disease, ocular hypertension, uveitis , or cataracts) Collagen vascular disease with severe dry eyes Chronic eye rubbing Irregular astigmatism , including corneal ectatic disorders Unstable refraction (> I-D change during previous year) Herpes simplex keratitis Patient refusal to discontinue contant lens wear before procedure Corneal vascularization (except micropannus) Active blepharoconjunctivitis Enrollment in other refractive keratotomy investigational protocols 0= diopter.

We conducted a prospective, multicenter, clinical study in patients undergoing combined-technique radial incisions (combined incisions) using a standardized surgical protocol (Genesis protocol) aimed at maximizing procedure safety while maintaining a high level of efficacy. We compare these results with previously reported data using other RK protocols for the correction of myopia.

ried out at the time of screening. Screening corneal pachymetry data were used to establish the relative thinnest paracentral site and to assess the magnitude of thickness disparity between sites. Because a broad range ofthickness variability may indicate less-achievable correction than predicted by our nomogram, in patients with disparities greater than 75 Jlm, the central clear zone diameter was reduced by 0.25 mm. After providing informed consent, patients underwent RK using the group standard surgical procedure," with a postoperative refractive goal of -0.50 D .

Materials and Methods

Surgical Procedure

Population Studied

Cycloplegic refractions, performed before all primary and enhancement procedures, were used as the basis to form the surgical plan. Determination of central clear zone size and incision number were based on the degree of desired correction, patient age, and the variance of paracentral corneal thickness, using the Genesis nomograms. 10 Patient sex and intraocular pressure were minor modifiers for discerning between adjacent central clear zone alternatives. In general, the nomograms guided incision number to encourage central clear zone size selection between 3.5 and 4.0 mm. The magnitude of refractive astigmatism and topographic pattern served as a secondary guiding variable for radial incision number and central clear zone diameter, and as a primary variable for determining axis of incision placement, respectively. Primary procedure radial incision number varied between two and eight. Two-incision RKs were reserved for patients with low magnitude myopic astigmatism undergoing either primary or enhancement surgery, with the radial incisions placed in the steep refractive axis.!? Maximal primary surgery was defined as eight incisions with optical zones limited to 3.0 mm or larger. The Genesis nomograms historically were established based on modifications of the prior experience of two authors (DJS and KKA), which included

efficacy."

From August 1992 to July 1994, patients seeking RK for reduction of myopia at each of six clinical centers were enrolled in a prospective clinical trial. Patients with myopia beyond the correctable range predicted by the 3.0mm central clear zone of the Genesis eight-incision nomogram 10 were excluded, as were patients not satisfying other study criteria (Table 1). Patients enrolled in other investigational trials also were excluded. All patients who satisfied study entry criteria were enrolled at each center. A total of 407 eyes satisfied the entry criteria, of which 375 eyes (480 procedures) of238 patients had greater than 6 weeks of follow-up, after their most recent procedure. The mean patient age was 36.6 ± 9.4 years (range, 1964 years) . The patients (48% male, 52% female) had preoperative refractive errors, ranging from -9.50 to -1.00 diopters (D) (mean ± standard deviation, -3.83 ± 1.60 D). Preoperative screening included, but was not limited to, visual acuity measurements, slit-lamp biomicroscopy, applanation tonometry, dilated indirect ophthalmoscopy, and computer-assisted corneal topography. Ultrasonic pachymetry measurements in the corneal center as well as eight paracentral locations also were car-

1909

Ophthalmology

Volume 102, Number 12, December 1995

observations of outcome disparity between combined versus centripetal incisions in eye bank eyes 9 and was further refined based on our personal experience with the combined incisions. Surgery was performed with a miotic pupil under topical anesthesia as previously outlined," The corneal optical center and central clear zone were marked . Intraoperative ultrasonic pachymetry (KMI, Philadelphia, PA) was performed 1.5 mm from the corneal optical center over both the temporal and the thinnest paracentral location as established by preoperative screening pachymetry. A Genesis diamond blade (Alcon Surgical, Ft. Worth, TX) (Fig l) was extended to 100% of the thinnest paracentral corneal measurement using a calibrating microscope (Alcon Surgical). Starting with the thinnest corneal quadrant and proceeding in sequence of graduating corneal thickness, standard combined-technique incisions were initiated by plunging the blade into the stroma adjacent to the central clear zone mark. Holding the blade perpendicular to the corneal surface for several seconds, the surgeon carefully undermined the central clear zone by tilting the knife toward the heel of the foot plates and then radiated in centrifugal fashion, stopping no closer than 1 mm from the corneosclerallimbus. With the Genesis knife, this is accomplished when the heel of the foot plate overlies the corneoscleral limbus. Without removing the diamond blade from the incision groove, the centripetal component of the incision then was performed, again undermining the central clear zone before incision termination. Mechanical globe fixation was left to the discretion of the operating surgeon , who noted fixation technique. Mechanical fixation methods ranged from external pressure with a cellulose sponge or 0.3 forceps across from the incision site to fixation using the Thornton Ring (KMI , Philadelphia, PA). Any intraoperative complications such as corneal perforations, invasion of the central clear zone, crossed incisions, or termination of incisions beyond 1 mm short of the limbus were noted.

Enhancement Procedures Enhancement procedures, when indicated, generally were performed no sooner than 6 weeks after the most recent procedure. The degree of residual myopia, desired refractive outcome, computer-assisted corneal topographic data, and slit-lamp evaluations were used to guide the surgical plan (Table 2). Enhancement procedures were performed with the patient's pupil dilated using the standard protocol .'? Although a maximum of two enhancements was allowed, surgeons were encouraged at the outset to maintain a limit of one enhancement procedure per eye, never exceed ing two.

Dull

Sharp

Figure 1. Schematic diagram of Genes is blade depicts a cutting margin along the angled face and the 2SD-JIm distal portion of the vertical face. Other similar blade designs exist.

the first postoperative week, treatment to modulate the wound-healing response was considered. The pharmocologic regimen in patients suspected of having hyperopia included topical 0.5% pilocarpine and 0.1% dipivefrin four times daily for 6 weeks, In addition, such patients were prescribed oral vitamin C (1000 mg 3 times daily) beginning at the sixth postoperative week for 6 weeks.IO.15.16 Postoperative complications, including infectious keratitis, decreased best-corrected visual acuity, consecutive hyperopia, or incision-groove neovascularization were recorded.

Postoperative Care

Data Analysis

Postoperative medical therapy included a 5-day regimen of topical steroids and prophylactic antibiotics instilled four times daily, followed by 2 days of twice-daily instillation. The most common combinations for topical therapy were either neomycin/polymyxin b/dexamethasone and tobramycin or tobramycin/dexamethasone and ciprofloxacin. If the cycloplegic refraction was greater than + 1.50 D within

All patients were followed from 6 weeks to 1 year postoperatively. Measurements of uncorrected visual acuity, best-corrected visual acuity, manifest and cycloplegic refraction, as well as a slit-lamp evaluation were performed at each postoperative visit. Visual acuity data were collected using standard projected visual acuity charts (Project-a-Chart, Reichert Ophthalmic Instruments, Buffalo ,

1910

Verity et al . Combined-technique (Genesis) RK Table 2. Factors Influencing Radial Keratotomy Surgical Enhancement Options Add New Incisions

Lengthen Incisions

Undercorrection> 1.25 D Optical zone at minimum (3.0 mm)

Undercorrection < 1.25 D Optical zone> 3.0 mm

Limbal zone at minimum (1.0 mm) Fewer than 8 original incisions Corneal topography demonstrates local steep zone

Limbal zone> 1.0 mm Significant astigmatism

D

=

Deepen Incisions Only Original incisions appear shallow at slit lamp Corneal topography demonstrates local steep zone Optical zone at minimum Eight original incisions

diopter.

NY). For the purposes of statistical analysis, visual acuity data were converted to the logarithm of the minimum angle of resolution (logMAR) and averaged.F:" Within this system, eyeswith visual acuity of 20/20 or better were assigned a 10gMARvalue of 0, whereas eyes with 20/200 visual acuity were assigned 1.0. Eyes with visual acuity of counting fingersand hand motions were assigned decimal scale values of 0.02 (logMAR value = 1.7) and 0.01 (logMAR value = 2.0), respectively. Data were tabulated for the 1.5-, 3-, 6-, and 12-month visits since the time of the most recent surgical procedure (either primary or enhancement). Additionally, data were stratified in a similar fashion as the PERK study for purposes of comparison. To minimize data bias, first eye-only analysis was additionally carried out, evaluating mean uncorrected visual acuity and refractive outcome in the 238 patients. The enhancement rates in eyes that were mechanically fixated versus those not fixated were compared, as were those among varying ranges of preoperative myopia. The perforation rates in eyes that were mechanically fixated versus those not fixated were compared, as were those among eyes after the protocol's suggested incision sequence versus nonprotocol sequence. The residual cycloplegicrefractiveerror in a subset population was compared at 3 and 12 months postoperatively. Comparisons were made using either a nonpaired or paired samples Student's t test, as appropriate (P < 0.02 significant).

Results Primary procedures alone were performed on 73% (275/ 375) ofthe eyes.Enhancement procedures were performed on the other 27% (100/375) of eyes. Five eyes (1.3%) had more than one enhancement, for a total of 105 enhancement procedures. Tables 3 and 4 present the enhancement data for all patients. The enhancement procedure incidence in the high-myopia group were 3.4 times and 2.4 times greater compared with the low- or moderate-myopia groups, respectively (P = 0.0001) (Table 3). Seventy-one percent (267/375) of eyes underwent mechanical globe fixation. These eyes had a 23% enhancement rate compared with 41% when surgeons did not fixate the globe (P < 0.00l) (Table 4). Follow-up (mean ± standard deviation) was 6.2 ± 3.7 months (range, 1.5-12 months).

Table 5 presents the refraction data over time, with eyes divided by level of baseline myopia. The mean residual cycloplegic refraction for all eyes (Table 5) was -0.48 ± 0.61 D (range, -2.50 to + 1.50 D), whereas (Table 5) this value was -0.49 ± 0.58 D (range, -2.50 to + 1.25 D) for first eyes. The mean uncorrected visual acuity was 20/25 (0.10 ± 0.15; range, 20/200-20/20) (Table 6) for all eyes and 20/ 25 (0.10 ± 0.14; range, 20/200-20/20) for first eyes. Uncorrected visual acuity was 20/40 or better in 95% of eyes. Of the 375 eyes within the study, 275 underwent only the primary procedure and no enhancements, with 97% of these achieving an uncorrected visual acuity of 20/40 or better. Table 7 presents the distribution of eyes corrected to within ± 1.0 D of emmetropia as well as those undercorrected and overcorrected by more than 1 D compared with the PERK Study." No patient was overcorrected by more than 1.50D. The incidence of consecutive hyperopia (spherical equivalent, > 1.00 D) was 1.3% (5/375 eyes). One of these five eyes had undergone an enhancement procedure. At the time of data analysis, 85% of the study eyes were corrected to within 1 D of emmetropia. However, 92% of eyes were corrected to within 1 D of the outcome goal of -0.50 D. The change in cycloplegic refraction from 3 to 12 months was evaluated as a subset in 46 eyes from one center that had no surgical intervention throughout this time interval. The change (mean ± standard deviation) was -0.17 D ± 0.60 (progression of effect) (range, -1.50 to + 1.37 D). In this subset of eyes, the difference between the 3- and 12-month examinations was not significant (P = 0.06, paired samples analysis). Five eyes (1.3%)gained two lines of best-corrected spectacle visual acuity and one patient (0.3%) had a decrease of two lines in the study due to irregular astigmatism, not associated with any intraoperative complication. Ten eyes (2.7%) gained one line of best-corrected spectacle visual acuity after RK, whereas 14 eyes (3.7%) had a decrease of one line. The decrease in best-corrected spectacle visual acuity in 1 of these 14 eyes was associated with an intraoperative perforation, with suture repair. A decrease in another two patients was associated with consecutive hyperopia with cycloplegic refractions of + 1.50 D spherical equivalent in both patients. The remaining 11 eyes were not associated with any complications.

1911

Ophthalmology

Volume 102, Number 12, December 1995

Table 3. Enhancement Data* Baseline Myopia

No.

Enhancements

> 1 Enhancement

Low mvopia es 3.12 D Intermediate myopia 3.25-4.25 D High myopia ~ 4.50 D

150 102 123

22 21 62

1 1 3

Enhancement Rate (%)

15 21 50

diopter. • Rates for low andintermediate myopia were significantly different from high myopia (P = O.Oool).

D =

There was an intraoperative perforation rate of 2.7% (12/ 435 procedures) among patients when surgeons followed

the recommended incision sequence coinciding with graduating corneal thickness. An intraoperative perforation rate of 15.5% (7/45 procedures) occurred in patients in whom the protocol incision sequence was not followed. This almost sixfold difference in perforation risks was significant (P = 0.0002). Among patients after the recommended incision sequence, there was a 3.6% perforation rate with mechanical globe fixation compared with 1.3% when the globe was not mechanically fixated (P> 0.05).

astigmatism 10 (Table 2); (7) use of ultrathin diamond knives, aiding in improved surgical control and improvement in blade design'v':"; (8) improvements in diamond blade tip calibration accuracy; (9) incorporation of patient age as a variable in the predictive nomograms'v"; (10) total number of radial incisions, ranging from 2 to 12; (II) prohibiting contact-lens wear for up to I year after RK; and (12) early postoperative pharmacologic treatment of suspected overcorrections. It is not currently possible to determine which, if any, ofthese protocol modifications accounts for our apparently improved outcomes. Ofthe 375 eyes studied, 275 underwent only the primary procedure, with 97% of these achieving an uncorrected

Discussion Radial keratotomy surgery as it was performed in this study represents a procedure that has undergone change since it was first developed and evaluated in the United States within the PERK Study. Due to insights gained from that study and from subsequent studies, as well as the incorporation of improved technology, the procedure has become safer and more predictable. Compared with the PERK Study, surgical protocol modifications incorporated into the current study include (I) combined-technique incision style"; (2) termination of incisions l-mm short of the corneoscleral Iimbus'<'"; (3) titration of incision number to encourage large central clear zones; (4) placement of incisions in the thinnest corneal quadrant first to minimize perforation risk as the cornea normally thins during the procedure'P'P; (5) screening pachymetry data to guide central clear zone size selection and facilitate intraoperative pachymetry!"; (6) computer-assisted videokeratography for guiding the surgical plan, including localization of the steep corneal hemimeridia in the patient with two-incision RK and for correction of postoperative

Table 5. Spherical Equivalent of the Cycloplegic Refraction after Radial Keratotomy* Follow-up (mos) 1.5 3.0 6.0 12.0 1.5 3.0

Table

Fixation No fixation • P = 0.0005.

1912

4.

Mechanical Globe Fixation versus

Nonfixation*

No.

Enhancements

267 108

61 44

Enhancement Rate (%)

23 41

6.0 12.0

Baseline Myopia (D) :0;3.12

3.25-4.25 ~4.50

:0;3.12

3.25-4.25 ~4.50

:0;3.12

3.25-4.25 ~4.50

:0;3.12

3.25-4.25 ~4.50

:0;3.12

3.25-4.25 ~4.50

:0;3.12

3.25-4.25 ~4.50

:0;3.12

3.25-4.25 ~4.50

:0;3.12

3.25-4.25 ~4.50

Spherical Equivalent Refraction (D) Mean ± SD

Range

-0.28 ± 0.71 -0.50 ± 0.61 -0.50 ± 0.55 -0.44 ± 0.49 -0.57 ± 0.72 -0.85 ± 0.63 -0.34 ± 0.49 -0.32 ± 0.55 -0.75 ± 0.61 -0.40 ± 0.59 -0.23 ± 0.49 -0.51 ± 0.84 -0.38 ± 0.54 -0.50 ± 0.21 -0.34 ± 0.46 -0.46 ± 0.41 -0.40 ± 0.67 -0.87 ± 0.62 -0.25 ± 0.53 -0.44 ± 0.56 -0.76 ± 0.48 -0.35 ± 0.50 -0.37 ± 0.53 -0.79 ± 0.74

-1.87 to + 1.50 -1.25 to +0.12 -1.25 to +0.37 -1.12 to + 1.00 -1.87 to +0.50 -2.25 to +0.12 -1.50 to +0.88 -1.12 to + 1.00 -1.75 to + 1.50 -2.00 to +0.25 -1.25 to + 1.25 -2.50 to + 1.50 -1.87 to + 1.25 -0.75 to -0.25 -1.00 to +0.25 -1.12 to +0.50 -1.50 to +0.50 - 2.25 to plano -1.50 to +0.88 -1.12 to + 1.00 -1.75 to plano -2.00 to +0.25 -1.25 to + 1.25 - 2.50 to +0.12

D = diopter; SD = standard deviation. • First eye ofeach patient (n = 238).

Verity et al . Combined-technique (Genesis) RK

Table 6. Uncorrected Visual Acuity after Radial Keratotomy (logMAR)* Mean Uncorrected Acuity Follow-up (mos)

Baseline Myopia (D)

1.5

::5:3.12 3.25-4.25

3.0

::5:3.12 3.25-4.25

6.0

::5:3.12 3.25-4.25

12.0

::5:3.12 3.25-4.25

1.5

::5:3.12 3.25-4.25

3.0

::5:3.12 3.25-4.25

6.0

::5:3.12 3.25-4.25

12.0

::5:3.12 3.25-4.25

LogMAR Scale Snellen Conversion

Mean ± SD

Range (Snellen)

20/22 20/22 20/ 23 20/25 20/25 20/32 20/24 20/26 20/29 20/22 20/23 20/28 20/24 20/26 20/22 20/25 20/24 20/31 20/24 20/28 20/28 20/22 20/22 20/29

0.059 ± 0.097 0.048 ± 0.069 0.069 ± 0.089 0.104 ± 0.120 0.102 ± 0.101 0.208 ± 0.270 0.077 ± 0.134 0.117 ± 0.121 0.161 ± 0.146 0.053 ± 0.071 0.060 ± 0.075 0.145 ± 0.203 0.070 ± 0.105 0.119 ± 0.147 0.048 ± 0.052 0.091 ± 0.111 0.079 ± 0.092 0.198 ± 0.261 0.082 ± 0.156 0.143 ± 0.121 0.141 ± 0.135 0.043 ± 0.066 0.052 ± 0.076 0.157 ± 0.19

20/20-20/40 20/20-20/40 20/20-20/40 20/20-20/40 20/20-20/40 20/20-20/200 20/20-20/80 20/20-20/80 20/20-20/80 20/20-20/30 20/20-20/30 20/20-20/100 20/20-20/40 20/20-20/40 20/20-20/25 20/20-20/40 20/20-20/40 20/20-20/200 20/20-20/80 20/20-20/50 20/20-20/80 20/20-20/30 20/20-20/30 20/20-20/100

~4.50

~4.50

~4.50

~4.50

~4.50

~4.50

~4.50

~4.50

D

= diopter; SD = standard deviat ion . = 238).

• First eye of each patient (n

visual acuity of 20/40 or better. This reflectssome benefits of the above-described insights and technologic advances, coupled with a good predictive value of the combinedtechnique (Genesis) nomogram, because there was no preselection of patients with low or moderate myopia within the study. The study design limited the number of enhancements to no greater than two procedures, and only 5 of the 100 eyes undergoing enhancements received this second procedure . The imposed limitations on enhancement number and the prospective, multicentered

nature of our study lends itself to direct comparison with the PERK Study Group data. Despite a mean residual myopic error of -0.50 D, the mean uncorrected visual acuity for all patients was 20/ 25 after primary and enhancement procedures. Greater than 95% of eyes achieved 20/40 or better visual acuity, with the remaining 5% all retaining residual myopic errors. These results reflect a higher degree of accuracy compared with the l-year PERK Study data, including a standard deviation of 0.6 versus 1.1 D, respectively." The increase

Table 7. Distribution of Cycloplegic Refraction aiter Radial Keratotomy in Comparison with Prospective Evaluation of Radial Keratotomy" Baseline Myopia (D)

::5:3.12 3.25-4.25 ~4.50

D

Undercorrected «-l.OOD)

Emmetropia (±1 D)

Overcorrected (>l.00 D)

Genesis

PERK

Genesis

PERK

Genesis

PERK

93 88 74

84

6 10 25

5 26 56

1 2 1

11 12 6

62 38

= diopter; PERK = Pro spect ive Evaluat ion of Rad ial Keratotom y.

• Spherical equivalent. Data are pre sented as percentage of eyes from baseline myopia groups.

1913

Ophthalmology

Volume 102, Number 12, December 1995

in out com e accuracy ma y reflect technologic advances since the PERK era, the insights concerning patient age and corneal thickness variability as predictive factors, and the relati vely greater number ofenhancement procedures performed in this study. Our 20/40 threshold visual acuit y data are comparable to another contemporary study . I I In contradistinction, our results were achieved in association with a mean residual myopic refracti ve error (versus a hyperopic mean outcome), and a lower enhancement incidence (Table 8). No patient's vision was overcorrected by more than + 1.50 D; therefore, no suture enhancements (Table 8) were required for patients in this study. The mean residual cycloplegic refraction of -0.48 0 is very close to the outcome goal suggested by the Genesis nomogram. Approximately 85% of the study eyes were corrected to within ± 1.00 0 of emmetropia, where 92% were within 1.00 0 of the targeted outcome goal of -0.50 D. This represents a greater degree of accuracy compared with the PERK Study data (60% within ±1.00 D ).4 Consecutive hype ropia occurred in 1.3% of the study eyes. Only one of these five eyes had an enhancement procedure, the remainder being over-responders from the primary procedure. None of these patients with hyperopia lost two lines of best-corrected visual acuity and all achieved uncorrected visual acuity of20/30 or better, underscoring the limitations of defining successful outcome based primarily on an uncorrected visual threshold (20/ 40) value. We speculate that our low incidence of con secutive hyperopia compared with others"! ' ma y be, in part, due to the use of relati vely few incisions, suturing

of perforations, an outcome goal of - 0.50 0 , and the limitation of enhancement number. The termination of incisions short of the corneoscleral limbus and incision number titration favoring relatively large (range, 3.5-4.0 mm) central clear zones also ma y have been beneficial as the 1O-year PERK Study demonstrated an association between incision length and hyperopic drift." In addition, the prohibition of contact-lens wear for I year postoperative ly and the early recognition of potential over-responders coupled with institution of treatment to pharmacologically mo dulate wound healing' Y" may contribute to the low incidence of consecutive hyperopia. The selected pharmacologic protocol is based on the authors' collective anecdotal experience. There has been extensi ve recent discussions concerning the role of pilocarpine on wound healing after keratotomy. In addition, animal models have suggested a role for adrenergic agonists in modulating corn eal response after keratotomy. 16 Clinical studies evaluating the efficacy of these drugs are currently under way; however, their value remains to be established . Our low incidence ofconsecutive patients with hyperopia also ma y reflect the inherently decreased variability in corneal flattening associated with combined-style incisions, We previously demonstrated a 40% variability in corneal flattening associated with both centrifugal and centripetal incision techniques compared with a 20% variability with combined-style incisions in an eye bank eye model. " Previous series have noted a progression of effect with time after RK surgery. v'" The small level of progression

Table 8, Comparison of Keratotomy Techniques

Study design Radial incision number" Range Preoperative refraction Mean ± SD Postoperative refraction Mean ± SD % of eyes ~ 20/40 after enhancements % of eyes ~ 20/40 after 1 procedure % of eyes ± 1.0 D of goal Mean procedure (no. per eye) % enhancements % > 1 enhancement Maximum no . of enhancements % overcorrection > 1 D % sutured for consecutive hyperopia Perforation rate (%) % loss of ~ 2 lines D

=

diopter; PERK

=

American Technique PERK Study4 .5

Russian Technique Werblin Studyll

Combined Technique Genesis System

Prospective

Retrospective

Prospective

8-16

4- 16

2-12

-4.1 ± 1.4

Not reported

-3.8 ± 1.6

-0.2 ± l.lt 88 76 64 1.09 12 0 1 17 0 2.2 3

+0.3 ± 0.6 99t

- 0.5 ± 0.6 95 75§ 92 1.28 23§ 1 2 1 0 2.7 0.3

Prospective Evaluation of Radial Keratotomy.

* Primary procedure maximum incision s numbered eight in each study.

t t

Extrapolated data. Patients with residual myopia anticipating further enhancement were excluded.

§ In globes with mechanical fixation.

1914

71

Not reported 1.48 33 11 7 Not reported 3 Not reported Not reported

Verity et al . Combined-technique (Genesis) RK from 3 to 12 months within a sampled subset of our data nearly did achieve significance (P = 0.06). The clinical significance and longer-term implications cannot be determined at this time. The short follow-up time in the current study does not readily enable direct comparison to other reported series. Nevertheless, concern over the development of potential hyperopic drift coupled with the desirability of myopic residual refractive errors with advancing age would suggest that patients should be left slightly undercorrected. Gi ven these considerations, the mean residual refractive error of -0.48 D (after all procedures) in this study would appear to be beneficial. Only one patient (0.3%) had a decrease of two lines, and best-corrected visual acuity improved two lines in five patients after surgery. This compares favorably with 3% of patients who lost and seven patients who gained two lines of best-corrected visual acuity at I year in the PERK Study. Decreased visual acuity after refractive procedures often is associated with irregular astigmatism. It is possible that our low incidence of decreased best-corrected visualacuity is associated with preferential selection of relatively large central clear zones (range, 3.5-4.0 mm). The PERK Study Group recently demonstrated an increased incidence of induced astigmatism associated with small (3.0 mm) central clear zones and variable incision depth (ARVO abstract , 1994). Other factors contributing to a lack of decreased visual acuity in our study thus may include reliable incision depth with associated uniform flattening," in addition to corneal topographic guidance of enhancement plan, suturing of perforations, and imposed limitations on incision and enhancement number. Variables that appeared to influence enhancement incidence included degree of preoperati ve myopia (Tables 2 and 3) and mechanical globe fixation (Table 4). A significantly higher enhancement incidence was noted among patients with high myopia (P = 0.000 I), whereas the enhancement rate difference among low and moderately myopic eyes was not significant (P = O. I3). Globe fixation was associated with a significantly lower enhancement incidence (P < 0.00 I). Globe fixation may elevate intraocular pressure, which may enable reliably deeper incisions.' Although a difference in the perforation rate between mechanically fixated versus nonfixated eyes was noted, it did not achieve significance (P = 0.06). Selected incision sequence, the most common protocol violation in this study, appeared to be a major risk for globe perforation. Two surgeons within the study group reported perforation rates much greater than did other surgeons in the group. A follow-up survey showed that one of these surgeons incised the thinnest corneal quadrant last (instead of the protocol-recommended incision sequence of thinnest quadrant first, graduating to greater corneal thickness) for the first ten patients in his series, experiencing two perforations in these ten eyes. Another surgeon routinely incised the cornea in random sequence, experiencing a perforation rate of 14% (5/35). After controlling for globe fixation method, the observed difference in perforation rate between procedures performed in compliance with protocol-recommended incision sequence (thinnest site first, thickest site last) compared with

any other incision sequence remained significant (P = 0.0002). Additionally, we noted a trend of increased perforation incidence among patients in whom diamondblade extension exceeded the thinnest measured corneal thickness (unpublished data). The intraoperative perforation rate (both micro- and macroperforations) of 2.7% (12/435 procedures) compares favorably with the perforation rates reported in other studies, ranging from 2.2% to II % .4,22-24 In addition, these studies used only centrifugal incisions, which since have been demonstrated to produce shallower depth for a given diamond extension."? It would appear that the increased efficacy of the combined incisions (coupled with the Genesis protocol) is afforded without diminishing the safety level of the PERK-style incisions. Although six of the perforated eyes had persistent aqueous leakage at time of surgery and required suture repair, the visual outcome was still good in these patients (mean uncorrected visual acuity, 20/23; postoperative spherical equivalent [mean ± standard deviation], -0.68 ± 0.94 D). Our data suggest that intraoperative corneal suturing does not adversely affect visual outcome. Media opacities, infectious keratitis, or endophthalmitis did not develop in any patients during this study. While 97% of patients with only primary procedures had an uncorrected visual acuity of 20/40 or better, only 75% of all mechanically fixated eyes had an uncorrected visual acuity at this level after the primary procedure (Table 8). This reflects an intended undercorrection factor in the Genesis combined-technique system. This conservative approach to RK allows the surgeon to titrate the level of surgical effect more accurately through the use of enhancement procedures, also evident in the low number of overcorrected eyes. The combined-technique ofRK used within the Genesis protocol appears to offer the safety of low microperforation and over-correction incidence, in addition to predictably high levels of refractive efficacy. These results may, in part, be afforded by uniformly deep incisions associated with the combined-technique incisions and the described protocol. Our data suggest, therefore, that, in addition, to limiting incision length and number, surgeons may want to fixate the globe, incising the thinnest corneal quadrant first and suturing corneal perforations. In conclusion, the application of principles learned from previously conducted studies of RK, as well as insights into corneal responses to radial incisions and wound healing since the PERK Study, may enable surgeons to achieve acceptable levels of safety and efficacy, with high levels of patient satisfaction.

References I. Fyodorov SN, Durnev VV. Operation ofdosaged dissection of corneal circular ligament in cases of myopia of mild degree. Ann Ophthalmol 1979; II :1885-90. 2. Bores LD , Myers W, Cowden J. Radial keratotomy: an analysis of the American experience. Ann Ophthalmol 1981;13:941-8.

1915

Ophthalmology

Volume 102, Number 12, December 1995

3. Bores LD. Historical review and clinical results of radial keratotomy. Int Ophthalmol Clin 1983;23(3):93-118. 4. Waring GO III, Lynn MJ, Gelender H, et al. Results of the Prospective Evaluation of Radial Keratotomy (PERK) study one year after surgery. Ophthalmology 1985;92:177-98. 5. Waring GO III, Lynn MJ, Nizam A, et al. Results of the Prospective Evaluation of Radial Keratotomy (PERK) study five years after surgery. Ophthalmology 1991;98:1164-76. 6. Melles GRJ, Binder PS. Effect of radial keratotomy incision direction on wound depth. Refract Corneal Surg 1990;6: 394-403. 7. Melles GRJ, Wijdh RHJ, Cost B, et al. Effect of blade configuration, knife action, and intraocular pressure on keratotomy incision depth and shape. Cornea 1993;12:299309. 8. Berkeley RG, Sanders DR, Piccolo MG. Effect of incision direction on radial keratotomy outcome. J Cataract Refract Surg 1991; 17:819-23. 9. Assil KK, Kassoff J, Schanzlin DJ, Quantock AJ. A combined incision technique ofradial keratotomy. A comparison to centripetal and centrifugal incision techniques in human donor eyes. Ophthalmology 1994;101:746-54. 10. Assil KK, Schanzlin DJ. Radial and Astigmatic Keratotomy. A Complete Handbook for the Successful Practice of Incisional Keratotomy Using the Combined Technique. Thorofare, NJ: Slack Press, 1994. 11. Werblin TP, Stafford GM. The Casebeer system for predictable keratorefractive surgery: one-year evaluation of 205 consecutive eyes. Ophthalmology 1993;100:1095-102. 12. Schachar RA, Black TD, Huang T. A physicist view ofradial keratotomy with practical surgical implications. In: Schachar RA, Levy NS, Schachar L, eds. Keratorefraction: Proceedings of the Keratorefractive Society meeting. Denison, TX: LAL Publishing, 1980; 195-220.

13. Villasenor RA, Salz J, Steel D, Krasnow MA. Changes in corneal thickness during radial keratotomy. Ophthalmic Surg 1981;12:341-42. 14. Dietz MR, Sanders DR, Raanan MG. Progressive hyperopia in radial keratotomy: long-term follow-up of diamond-knife and metal-blade series. Ophthalmology 1986;93: 1284-9. 15. Assil KK, Quantock AJ. Wound healing in response to keratorefractive surgery. Surv Ophthalmol 1993;38:289-302. 16. Kwito S, Sinbawy A, Lee M, McDonnell PJ. Pharmacologic alteration of corneal topography after radial keratotomy. Ophthalmic Surg 1992;23:738-41. 17. Holladay JT, Prager TC. Mean visual acuity [letter]. Am J OphthalmoI1991;lll:372-4. 18. Moseley MJ, Jones HS. Visual acuity. Calculating appropriate averages. Acta Ophthalmol 1993;71 :296-300. 19. Waring GO III, Lynn MJ, McDonnell PJ, et al. Results of the Prospective Evaluation of Radial Keratotomy (PERK) study 10 years after surgery. Arch Ophthalmol 1994;112: 1298-308. 20. Rowsey JJ, Balyeat HD, Rabinovitch B, et al. Predicting the results of radial keratotomy. Ophthalmology 1983;90: 642-54. 21. Sanders DR, Dietz MR, Gallagher D . Factors affecting predictability of radial keratotomy. Ophthalmology 1985;92: 1237-43. 22. Deitz MR, Sanders DR, Raanan MG. A consecutive series (1982-1985) of radial keratotomies performed with the diamond blade. Am J OphthalmoI1987;103:417-22. 23. Salz JJ, Salz JM, Salz M, Jones D. Ten years experience with a conservative approach to radial keratotomy. Refract Corneal Surg 1991;7:12-22. 24. Spigelman AV, Williams PA, Lindstrom RL. Further studies of four incision radial keratotomy. Refract Corneal Surg 1989;5:292-5.

Discussion by George O. Waring III, MD, FACS, FRCOphth Three large multicenter trials of refractive keratotomy have been reported within the last decade: the Prospective Evaluation of Radial Keratotomy (PERK) Study of 793 eyes funded by the National Eye Institute, I the Casebeer Study of 615 eyes funded by Chiron Vision," and the current study of 375 eyes carried out privately by The Refractive Keratoplasty Study Group. The PERK Study serves as a benchmark for measuring progress and the study by The Refractive Keratoplasty Study Group represents contemporary refractive keratotomy techniques as done in the early 1990s. Three important contemporary techniques of changes that have taken place with radial keratotomy are used in the current study'; (1) intentionally staged procedures, (2) bidirectional incisions with a partially dull diamond blade, and (3) shorter incisions.

Staged Procedures Because a single refractive keratotomy procedure performed on an individual eye cannot yield an exactly predictable result, surgeons have altered their approach by constructing nomograms that produce an intentional undercorrection in a large percent of eyes, preventing overcorrections in the vast majority." The Dr. Waring is a consultant for Chiron Vision. From Emory University, Atlanta.

1916

undercorrected eyes can receive repeat surgeries that lengthen, deepen, or increase the number of incisions to tailor the outcome to the individual eye. 4•5 This approach was used in the current study; 27% of eyes required a single additional procedure at 6 weeks or more after surgery. Only 1.3% of eyes had more than one repeat surgery. This methodology achieved the goal ofa low rate of overcorrection, 1.3% of eyes with a refraction of more than + 1.00 diopter. Eyes that were fixated mechanically during the incision had approximately half the number of repeat surgeries compared with eyes that were not fixated mechanically, suggesting that surgeons made more reliable incisions-presumably deeper and of a more correct length-when the globe was mechanically fixated. . The drawbacks of repeat surgeries are increased time, effort, cost, and morbidity. In addition, repeated surgeries are likely to make more irregular incisions, both in the depth of the incision and at their extensions. Whether more glare or refractive instability occurs in eyes that have had repeat surgeries is unknown.

Bidirectional Incisions The bidirectional incision has evolved gradually over the last decade." In the early 1980s, surgeons chose between the Russian centripetal incision made with a vertical blade and the American centrifugal incision made with an oblique blade.' Deeper incisions were achieved with a double centrifugal pass with the blade

Verity et al . Combined-technique (Genesis) RK of the same extension or peripheral deepening incisions with increased blade extension. Clinical experiences and histopathologic studies?demonstrated that the Russian centripetal incision with the vertical blade made more uniformly deep incisions, and in the late 1980s this technique was taught actively in the United States using new, thinner (100-lLm), double-edged diamond blades. However, new problems emerged: it was difficult to guide the blade in a straight line, to stop the blade exactly at the edge of the clear zone mark, and to make a small lengthening cut without extending into the clear zone. These problems were solved by designing knife blades that were sharp on the oblique side and sharp only in the distal 200 to 250 ILm on the vertical side. This allowed the surgeon to make an initial centrifugal incision with the oblique edge-a safe move because the incision was shallow and perforation was unlikely and because the incision moved away from the clear zone-followed by a centripetal incision back toward the clear zone using the initial incision as a guide and the dull edge of the vertical blade as a stop so the incision could not extend into the clear zone. The authors have demonstrated the use and effectiveness of this bidirectional technique and have pointed out that it may result in a double cut in the deep apex of the incision. The authors did not emphasize a practical clinical problem-the dull part of the vertical blade often hangs up in the incision during the centripetal pass because the blade gets out of the track and bumps into the uncut tissue or because the initial centrifugal incision is too shallow, and the dull part of the blade bumps into the uncut tissue. This problem can be solved by making two centrifugal passes, creating a broad, deeper track followed by the centripetal one, which then can be done more smoothly.

Shorter Incisions The authors emphasize stopping the incisions I mm inside the limbus, a technique that recognizes the fact that a shorter radial

incision can be almost as effective as a long incision!" and that has led to the current use of "mini-radial keratotomy." References 1. Waring GO, Lynn MJ, Nizam A, et al. Results of the Prospective Evaluation of Radial Keratotomy (PERK) Study five years after surgery. Ophthalmology 1991;98:1164-76. 2. Assil KK, Schanzlin DJ. Radial and Astigmatic Keratotomy: A Complete Handbook for the Successful Practice of Incisional Keratotomy Using the Combined Technique. St. Louis, MO; Poole Press/Slack, 1994. 3. Waring GO, Gordon JF, Lee PA, Dru RM, Keratorefractive Study Group. A comparison of the Duotrak'" blade technique and the Russian technique in prospective multicenter clinical studies of refractive keratotomy for myopia and astigmatism. Invest Ophthalmol Vis Sci 1994:35(Suppl): 1637. 4. Buzard KA. Deepening of incisions after radial keratotomy using the "tickle" technique. Refract Corneal Surg 1991;7: 348-55. 5. Franks S. Radial keratotomy undercorrections: a new approach. J Refract Surg 1986;2:171-3. 6. Merlin U, Bordin P, Rimondi AP, et al. Factors that affect keratotomy depth. Refract Corneal Surg 1991;7:356-9. 7. Waring GO. Atlas of surgical techniques of radial keratotomy. In: Waring GO, ed. Refractive Keratotomy for Myopia and Astigmatism. St. Louis, MO; Mosby-Year Book, Inc, 1992;507-639. 8. Berkeley RG, Sanders DR, Piccolo MG. Effect of incision direction on radial keratotomy outcome. J Cataract Refract Surg 1991;17:819-23. 9. Melles GRJ, Wijdh RHJ, Cost B, et al. Effect of blade configuration, knife action, and intraocular pressure on keratotomy incision depth and shape. Cornea 1993;12:299-309. 10. O'Donnell FE. Short incision radial keratotomy: a comparative study in rabbits. J Refract Surg 1987;3:102-3.

1917