- Email: [email protected]
Hyperopic shift after refractive keratotomy using the Casebeer System
Hyperopic shift after refractive keratotomy using the Casebeer System Theodore P. Werblin, MD, PhD, G. Michael Stafford, BS ABSTRACT Purpose: To determine the degree of hyperopic shift following refractive keratotomy. Setting: The Werblin Center, Princeton, West Virginia. Methods: The results of 241 consecutive radial/astigmatic keratotomy procedures in 128 patients were studied. All procedures were performed using Casebeer nomograms. Refractive follow-up information was obtained for 78% of patients at 3 months (range 1 to 6 months), 1 year (range 6 to 21 months), and 3 years (range 30 to 44 months). Results: Overall, the amount of hyperopic change decreased significantly (P < .05) during the 3 year period, from + 0.34 diopters (D) per year to + 0.12 D per year. Eyes with more than 6.0 D of preoperative intended correction were more unstable (+0.49 Din the first year and +0.44 Din the second and third years) than eyes with less than 5.0 D (+0.27 D and +0.05, respectively).
Conclusion: Because the average magnitude of the hyperopic shift was +0.6 D in the first 3 years after surgery, a slight undercorrection, -0.5 D to -0.7 D, should be the refractive endpoint for primary and enhanced refractive keratotomy surgery.
J Cataract Refract Surg 1996; 22:1030-1036
I
n a companion paper, 1 we examined the 3 year results of radial and astigmatic keratotomy procedures performed by one surgeon (T.P.W.) using the Casebeer System. In this paper, we focus on the hyperopic shift that occurred during the 3 year follow-up.
Supported in part by a grant from Chiron Vision, Irvine, California. Neither of the authors has a proprietary interest in the development or marketing ofthis or a competing system. Mike Lynn, MS, PERKstatistician, Emory University, provided unpublished PERKdata used in this paper. Dan Krider, PhD, chairman ofthe mathematics department, Concord College, and David Musch, PhD, Kellogg Eye Center, University ofMichigan, provided statistical assistance. Reprint requests to Theodore P. Werblin, MD, PhD, P. 0. Box 5879, Princeton, West Virginia 24740. 1030
Subjects and Methods The study included 241 eyes (128 patients). The patients and procedures and the Casebeer System have been described. 1- 4 To assess whether the degree of preoperative myopia affected postoperative stability, patients were classified into three groups based on the spherical equivalent of the preoperative cycloplegic refraction: low, +0.25 to -3.12 diopters (D); medium, -3.25 to -4.37 D; high, -4.50 to -9.62 D. These groups were similar to those used in the PERK study. Because the PERK groups were somewhat arbitrary, we further analyzed stability by comparing all eyes with more than 5.0 D of myopia preoperatively with those with 5.0 D or less preoperatively (- 5.0 D analysis) and those with more than 6.0 D preoperatively with those with 6.0 D or less ( -6.0 D analysis).
J CATARACT REFRACT SURG-VOL 22, OCTOBER 1996
HYPEROPIC SHIFT AFTER REFRACTIVE KERATOTOMY
For stability analysis, both near- and distance-corrected eyes were analyzed. However, the near-corrected eyes were classified by desired correction (- 2.0 0 of residual myopia), whereas the distance-corrected eyes were classified by total correction (0 0 of residual myopia). Hyperopic shift (stability analysis) was analyzed as follows. The change in cycloplegic spherical equivalent refraction between 3 months and 1 year was divided by the number of days between the two observations and multiplied by 365 to obtain the refractive change per year for the early observation period. The change between 1 and 3 years was divided by the number of days between the two observations and multiplied by 365 to obtain the rate of refractive change for the later period. The rates were compared to determine the progressive refractive change over time. Only eyes with data for all three times prior to a suture or automated lamellar keratoplasty were included in the analysis.
Results In the entire study group, there was a statistically significant decrease in the hyperopic shift between the early and late periods, from 0.33 diopter per year (0/yr) to 0.12 D/yr (P < .05, paired t-test) (Table 1). A more striking finding was seen among the low, medium, and high myopia groups. In the early period, all three had similar hyperopic shifts, 0.25 0, 0.36 0, and 0.40 0, respectively. However, in the late period, the low and medium groups had significantly lower rates of change than the high myopia group, 0.04 D/yr and 0.08 0/yr versus 0.22 0/yr, respectively (P < .05) (Figure 1 and
2A). This differs from the PERK data, which show a more uniform rate of change in all groups (Figure 2B). In the -5.0 and -6.0 D analyses (Table 2), the
shift toward hyperopia was relatively large in the early period. However, in the late period, there was a significantly lower rate of change in eyes with less than 5.0 and 6.0 0 of preoperative myopia (P< .05, paired t-test) but no significant change in eyes with more than 5.0 and 6.0 D (P < .05, paired t-test). There was a clear trend toward a lower rate of change in the later period in all but the more than 6.0 0 group. Thus, the more highly myopic eyes, particularly eyes with 6.12 0 and more, tended to show less stable postoperative refractions. Within the group of 71 highly myopic eyes, we compared the hyperopic shift in eyes that had no enhancements (n = 53) with the shift in those that had enhancements (n = 18). In the early period, the shift was 0.54 0 in the enhanced eyes and 0.32 0 in the nonenhanced eyes, and in the late period, it was 0.27 0 and 0.21 0, respectively; none of the differences was statistically significant. The rate of enhancement procedures was 46% in the highest myopia group and 27% in the lowest. The difference was statistically significant (P < .05).
Discussion In this paper, we addressed the issue of refractive stability. The worst case scenario would be that the 1 to 3 year rate of change of 0.05 0/yr would continue indefinitely. Were this the case, the average patient, about 40 years of age, whose preoperative myopia was less than 5.0 0, would be less than 1.0 0 hyperopic 30 years later
Table 1. Rate of hyperopic change in refraction (cycloplegic) (D/yr). Follow-up Period
Low Myopia Group
Medium Myopia Group
High Myopia Group
All Eyes
Number
Mean ±SO
Number
0.40:!: 0.81 0.22:!: 0.38
71 71
0.33:!: 0.73 0.12:!: 0.32
183 183
41 41
0.43:!: 0.80 0.21 :!: 0.35
68 68
0.35:!: 0.75 0.13:!: 0.31
165 165
2 2
-0.27:!: 0.85 0.51 :!: 0.82
3 3
0.18:!: 0.56 0.07:!: 0.43
18 18
Mean ±SO
Number
Mean ±SO
Number
Early Late
0.25:!: 0.57 0.04:!: 0.22
69 69
0.36:!: 0.84 0.08:!: 0.26 0.05*
42 42
Distance eyes only Early Late
0.24:!: 0.59 0.06:!: 0.21
56 56
0.37:!: 0.85 0.08:!: 0.34
Near eyes only Early Late
0.28:!: 0.51 0.04:!: 0.27
13 13
0.18:!: 0.25 0.12:!: 0.35
Mean ±SO
*Statistical exclusion of "outlier" based on 99.9% tolerance interval ] CATARACT REFRACT SURG-VOL 22, OCTOBER 1996
1031
HYPEROPIC SHIFT AFTER REFRACTIVE KERATOTOMY
Preoperative Myopia +0.25 to -3.12 Rate of Refractive Change 3 Months-1 Year
~
w
0
~
E :l
z
44 42 40 38 36 34 32 30 28 26 24 22 20 18 16
1 Year-3 Years
1/)
44 42 40 38 36 34 32 30 8 6 24 2 0 18 16 14 12 10 8 6 4 2
~ ~
0
~ ~
..-•
0 -2.0
E
:l
z
-1.0
0.0
+1.0
-+2.0
+3.0
'
-2.0
+4.0
---
-1.0
0.0
Diopters/Year
+1.0
+2.0
+3.0
+4.0
+3.0
+4.0
+3.0
+4.0
Diopters/Y ear
Preoperative Myopia -3.25 to -4.37 Rate of Refractive Change
3 Months-1 Year
1 Year-3 Years
38 36 34 32
~
w 0
~
E
:l
z
,., 1/)
Q)
24 22 20 18 16 14 12 10 8 6 4 2 0 -2.0
w 0
~
E
:l
z
•
-~ -1.0
0.0
+1.0
+2.0
+3.0
+4.0
44 42 40 38 36 34 32 30 28 26 24 22 20 18 16 14 12 10 8 6 4 2 0 -2.0
0.0
-1.0
Diopters/Y ear
+1.0
-
+2.0
Diopters/Year
Preoperative Myopia -4.50 to -9.62 Rate of Refractive Change 3 Months-1 Year
,., w 1/)
Q)
0
~
E
:l
z
44 42 40 38 36 34 32 30 28 26 24 22 20 18 16 14 12 10 8 6
4 2 0 -2.0
1 Year-3 Years
,., 1/)
Q)
w 0
-••
=.=1-1.0
0.0
+1.0
Cii .c E :l
z
+2.0
Diopters/Year
--- --+3.0
+4.0
44 42 40 38 36 34 32 30 28 26 24 22 20 18 16 14 12 10 8 6 4 2 0
-2.0
-1.0
+1.0
0.0
+2.0
Diopters/Year
Figure 1. (Werblin) Rate of dioptric change from 3 months to 1 year and from 1 to 3 years in the three myopia groups. 1032
J CATARACT REFRACT SURG-VOL 22, OCTOBER 1996
HYPEROPIC SHIFT AFTER REFRACTIVE KERATOTOMY
Casebeer Stability
[!!
!
0
i5
1.5 1.4 1.3 1.2 1.1 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
PERK Stability 1.5...---------------------, 1.4 1.3 1.2 1.1 1 [!! 0.9 .l!l a. 0.8 0.7 0 i5 0.6 0.5 0.4 0.3 0.2 0.1
-4.50 to -9.62 -3.25 to -4.37 +0.25 to -3.12
0
2
3
4
7
6
5
8
9
0~-.-~"~-,-~~~-r~~,-~~
10
0
2
3
4
5
Years
6
7
8
9
10
11
12
Years
A
B
Figure 2. (Werblin) A: Cumulative change in refraction in the study group by years postoperatively. B: Cumulative change in refraction in the PERK study by years postoperatively.
(Figure 3). Hyperopic shift (Figure 3, PERK5) and myopic shift (Figure 3, excimel or keratomileusis) with "stabilization" over time are such examples. In general, the hyperopic shift seen in RK begins to plateau within a few years of surgery. This tendency is particularly evident with lower myopic corrections. However, truly hyperopic shift, a constant increase in hyperopia over time, is suggested in Figure 2A, high myopia group. This ominous picture however may not be predictive over a longer time since the 3 year data in this study may not be sufficient to demonstrate this phenomenon. The most critical aspect of hyperopic shift is its cumulative effect on the refractive patient. Eyes with less than 6.0 D of myopic correction plateau with time, which means that the 0.05 D/yr shift seen between 1 and 3 years in this study is likely to diminish over time. If it did not diminish for 30 years, a -0.75 D initial result would be under + 1.0 D 30 years later. Most patients whose cycloplegic refraction is +0.50 to +0.75 D have 20/20 to 20/25 acuity without cycloplegia in the "real world," including presbyopia. A result under + 1.0 D at age 70 is still quite satisfactory. Even in this worst-case
if undercorrected the suggested amount -0.75 D; at no time would the uncorrected acuity be less than 20/30. This takes into account the shift of0.268 D noted in the first year and the shift of 0.0486 D/yr noted in the 1 to 3 year data. Patients between -5.0 and -6.0 D have enough hyperopic shift per year to make them significantly hyperopic 30 years after surgery, bur the assumption that the rate of change does not further plateau is likely wrong since the lower degrees of myopia in the PERK 10 year study show a long-term plateauing effect.5 Thus, patients in the -5.0 to -6.0 D range fall into a gray zone, bur clearly, at least for 30 years, patients under -5.0 D should have only minor hyperopic shift. Careful scrutiny of Table 1 raises several critical questions. Refractive stability is critical for all refractive procedures and several outcomes could occur. If, after the recovery from surgical trauma (several weeks), the refractive outcome did not change as is seen in intraocular surgery6 over time, the procedure would be considered stable. However, all corneal refractive procedures, including radial keratotomy (RK), excimer laser, and keratomileusis, show significant degrees of instability Table 2.
Rate of hyperopic change in refraction (cycloplegic) (D/hr) in eyes with 5.0 D and 6.0 D of preoperative myopia.
-5.0 0 Analysis +0.25 to -5.00 0
-6.0 0 Analysis
-5.12 to -9.62 0
+0.25 to -6.00 0
-6.12 to -9.62 0
Follow-up Period
Mean± SO
Number
Mean± SO
Number
Mean± SO
Number
Mean± SO
Number
Early Late
0.28 ± 0.67 0.07 ± 0.28
138 138
0.50 ± 0.88 0.29 ± 0.39
45 45
0.31 ± 0.73 0.09 ± 0.29
165 165
0.49 ± 0.77 0.44 ± 0.41
18 18
J CATARACT REFRACT SURG-VOL 22, OCTOBER 1996
1033
HYPEROPIC SHIFT AFTER REFRACTIVE KERATOTOMY
0.7
1.5..-----------------~
'C'
0.5
~c:
0
'$
c. 0
i5
(tl
0.4
0::
0.3
.E
-0.5
C1)
Ol
c:
0.2
(tl
.t:
-1 -1.5
0.5
0
"13
!!!
~
0.6
()
0
2
3
4
5
6
7
8
9
10
Years
0.1 0
0
2
3
4
5
6
7
8
9
10
Diopters of Intended Correction
Figure 3. (Werblin) Long-term refractive stability of several corneal and intraocular procedures.
Figure 4. (Werblin) Relationship between the rate of refractive change 1 to 3 years postoperatively and the preoperative cycloplegic spherical equivalent refraction.
scenario, I.e., no further plateauing of the hyperopic shift, the long-term effect of the hyperopic shift is negligible. The statistical comparison of the rate of refractive change in the first year, actually between the 3 month and 1 year visit, with the rate between 1 and 3 years is the critical issue. The high myopia group (Figure 2A), in contrast to the low and medium myopia groups, showed a more persistent rate of hyperopic change between 3 months and 1 year. In addition, the rates of hyperopic change for low, medium, and high myopia groups did not differ statistically for the 3 month to 1 year period but were significantly different later. Therefore, a better indicator oflong-term instability is the rate of hyperopic change 2 to 3 years after the final surgical procedure. This is shown graphically in Figure, 4 in which the low, medium, and high myopia groups are divided into six subgroups based on preoperative intended cycloplegic refraction. A nonlinear relationship is demonstrated, indicating increasing instability as a function of intended correction particularly above -6.0 D. Eyes with less than 4.5 D of myopia showed little evidence of progression. In eyes with 4.5 to 6.0 D of myopia, a greater hyperopic shift was seen but it appeared to plateau ( +0.16 D/yr in the 1 to 3 year period). Eyes with more than 6.0 D of intended correction showed a greater progressive shift ( +0.44 D/yr for the 1 to 3 year period) (Figure 4). Even in this group, however, many eyes did not show consistent changes in both periods but had a large shift in either the 3 month to 1 year or the 1 to 3 year period. In the lower myopia 1034
groups, patients occasionally demonstrated major hyperopic shifts, overcorrections, or both. Therefore, throughout this population there was a general trend toward a small hyperopic shift. Exceptions were seen occasionally in lower degrees of myopia and more commonly in higher degrees of myopia (Figure 2). Eleven eyes (9%) showed evidence of a constant increase in hyperopia of 0.5 D or more in each year of the study rather than plateauing of the hyperopic change or an isolated hyperopic change. These progressive changes represent a real challenge to the refractive keratotomy surgeon. Six of the eyes have had suture enhancements, which can reverse the problem. However, the additional cost and effort involved in these subsequent procedures are significant. In only one patient was this process bilateral. It is interesting to speculate why these data appear to stabilize more rapidly than data presented in the PERK studies (Figure 2A, B). The PERK data showed that shorter incisions, i.e., larger optical zones, were more stable. 5 Newer concepts of limited amounts of incisional surgery have also demonstrated more corneal stability with shorter incisions. 8 Since our execution of the Casebeer System starts the incisions well inside the limbus, unlike the PERK system, the shorter incisions may correlate with the greater degree of corneal stability. To date there has been no clear explanation of why the cornea is destabilized by refractive keratotomy in some patients. Generally, but not always, larger surgical corrections are correlated with higher rates of instability. The greatest surgical effect usually occurs the moment
J CATARACT REFRACT SURG-VOL 22, OCTOBER 1996
HYPEROPIC SHIFT AFTER REFRACTIVE KERATOTOMY
the surgery is completed. Subsequent healing correlates with a variable amount of regression. Clinically, the wounds become slightly opalescent and have a feathery appearance along the wound edges. The opalescent appearance resolves over time. Are we sure that corneal wound healing or lack thereof is totally responsible for refractive instability following RK? A recent report on long-term RK results emphasized the "natural" hyperopic refractive progression of "normal" eyes as a function of age. 9 Previous reports 10 - 12 indicate that between ages 31 and 64 years, an average 0.035 D/yr shift toward hyperopia occurs. To look at this problem analytically, we took the data from Table 1 and separated patients below age 31 from the rest of the group and recalculated the change in refraction as a function of time postoperatively (Table 3). For the 1 to 3 year period, patients below 31 years had a much lower rate of refractive change than patients over 31 years. Even though the 31 and older patients had a relatively higher rate of change, 0.048 D/yr, when the "natural" hyperopic drift of the eye is subtracted (0.035 D/yr), both age groups seem to behave the same, 0.014 D/yr and 0.013 D/yr, respectively. Despite this finding, whether the shift is "natural" or the result of surgical wound healing, small surgical undercorrections are still appropriate. Hyperopic shift should be a serious concern after routine refractive keratotomy procedures for certain, rei-
Table 3.
atively "high risk" patients. It is clear from the experience of most refractive keratotomy surgeons that myopic patients who become symptomatically hyperopic after surgery are extremely unhappy, often more unhappy than when they were myopic. Therefore, most surgeons use nomograms that attempt to achieve a slight undercorrection. Most patients who are accustomed to 20/20 or better corrected acuity prior to surgery, however, are not ecstatic about less than 20/25 or 20/30 uncorrected acuity postoperatively. Some are even dissatisfied with 20/20 acuity, especially if the fellow eye is20/15. A delicate balance between the amount of undercorrection that is surgically advisable and that which is tolerated and functional for the patient has to be achieved. Because the average hyperopic shift appears to be about 0.60 D in the first 3 years following surgery, the ideal patient will be undercorrected by 0.50 to 0. 75 D immediately after surgery. This would allow the expected hyperopic shift to place the patient close to emmetropia, while leaving him or her at worst with 20/25 to 20/30 uncorrected acuity immediately after surgery. Although this may not be the patient's ideal result, the physician should make every attempt to explain the likely improvement in time with healing, rather than relying too heavily on enhancement procedures for small amounts of residual myopia. A better solution may move our thinking away from the cornea to avoid
Rate of hyperopic change in refraction (cycloplegic) (D/yr) as a function of preoperative age.
Preoperative Range in Myopia
-4.50 to -9.62 0
+0.25 to -4.37 0 Mean± SO
Number
Mean± SO
Number
All eyes First year 1 to 3 years
0.290 ± 0.685 0.043 ± 0.236
111 111
0.398 ± 0.808 0.224 ± 0.375
71 71
Below 31 years First year 1 to 3 years
0.252 ± 0.677 0.014 ± 0.232
17 17
0.454 ± 1.009 0.214 ± 0.246
20 20
31 years and above First year 1 to 3 years
0.297 ± 0.690 0.048 ± 0.238
94 94
0.376 ± 0. 724 0.229 ± 0.447
51 51
Age/Follow-up
31 years and above (minus normal hyperopic "drift") First year 1 to 3 years
0.262 0.013
J CATARACT REFRACT SURG-VOL 22, OCTOBER 1996
0.341 0.194
I035
HYPEROPIC SHIFT AFTER REFRACTIVE KERATOTOMY
wound healing and optical distortion difficulties inherent in all forms of corneal surgery. 13 7.
References 1. Werblin TP, Stafford GM. Three year results of refractive keratotomy using the Casebeer System. J Cataract Refract Surg 1996; 22: 1023-1029 2. Werblin TP, Stafford GM. The Casebeer System for predictable keratorefractive surgery; one-year evaluation of 205 consecutive eyes. Ophthalmology 1993; 100:10951102 3. Casebeer JC. Casebeer lncisional Keratotomy. Thorofare, NJ, Slack Inc, 1995 4. Friedlander MH, Nordan LT, Maxwell WA, eta!. Radial Keratotomy predictability (letters to the editor). Ophthalmology 1994; 101:411-416 5. Waring GO Ill, Lynn MJ, McDonnell PJ, and the PERK Study Group. Results of the Prospective Evaluation of Radial Keratotomy (PERK) study 10 years after surgery. Arch Ophthalmol 1994; 112: 1298 -1308 6. Werblin TP. Refractive stability after cataract extraction
1036
8. 9.
10.
11.
12. 13.
using a 6.5-millimeter scleral pocket incision with horizontal or radial sutures. J Refract Corneal Surg 1994; 10:339-342 Maguen E, Salz JJ, N esburn AB, et a!. Results of excimer laser photorefractive keratectomy for the correction of myopia. Ophthalmology 1994; 101: 1548-15 56 Lindstrom RL. Minimally invasive radial keratotomy: mini-RK. J Cataract Refract Surg 1995; 21:27-34 Sawelson H, Marks RG. Ten-year refractive and visual results of radial keratotomy. Ophthalmology 1995; 102: 1892-1901 Hirsch MJ. Changes in refractive state after the age of forty-five. Am J Optom Arch Am Acad Optom 1958; 35:229-237 Walton WG Jr. Refractive changes in the eye over aperiod of years. Am J Optom Arch Am Acad Optom 1950; 27:267-286 Slataper FJ. Age norms of refraction and vision. Arch Ophthalmol1950; 43:466-481 Werblin TP. Should we consider clear lens extraction for routine refractive surgery? Refract Corneal Surg 1992; 8:480-481
J CATARACT REFRACT SURG-VOL 22, OCTOBER 1996
Conclusion: Because the average magnitude of the hyperopic shift was +0.6 D in the first 3 years after surgery, a slight undercorrection, -0.5 D to -0.7 D, should be the refractive endpoint for primary and enhanced refractive keratotomy surgery.
J Cataract Refract Surg 1996; 22:1030-1036
I
n a companion paper, 1 we examined the 3 year results of radial and astigmatic keratotomy procedures performed by one surgeon (T.P.W.) using the Casebeer System. In this paper, we focus on the hyperopic shift that occurred during the 3 year follow-up.
Supported in part by a grant from Chiron Vision, Irvine, California. Neither of the authors has a proprietary interest in the development or marketing ofthis or a competing system. Mike Lynn, MS, PERKstatistician, Emory University, provided unpublished PERKdata used in this paper. Dan Krider, PhD, chairman ofthe mathematics department, Concord College, and David Musch, PhD, Kellogg Eye Center, University ofMichigan, provided statistical assistance. Reprint requests to Theodore P. Werblin, MD, PhD, P. 0. Box 5879, Princeton, West Virginia 24740. 1030
Subjects and Methods The study included 241 eyes (128 patients). The patients and procedures and the Casebeer System have been described. 1- 4 To assess whether the degree of preoperative myopia affected postoperative stability, patients were classified into three groups based on the spherical equivalent of the preoperative cycloplegic refraction: low, +0.25 to -3.12 diopters (D); medium, -3.25 to -4.37 D; high, -4.50 to -9.62 D. These groups were similar to those used in the PERK study. Because the PERK groups were somewhat arbitrary, we further analyzed stability by comparing all eyes with more than 5.0 D of myopia preoperatively with those with 5.0 D or less preoperatively (- 5.0 D analysis) and those with more than 6.0 D preoperatively with those with 6.0 D or less ( -6.0 D analysis).
J CATARACT REFRACT SURG-VOL 22, OCTOBER 1996
HYPEROPIC SHIFT AFTER REFRACTIVE KERATOTOMY
For stability analysis, both near- and distance-corrected eyes were analyzed. However, the near-corrected eyes were classified by desired correction (- 2.0 0 of residual myopia), whereas the distance-corrected eyes were classified by total correction (0 0 of residual myopia). Hyperopic shift (stability analysis) was analyzed as follows. The change in cycloplegic spherical equivalent refraction between 3 months and 1 year was divided by the number of days between the two observations and multiplied by 365 to obtain the refractive change per year for the early observation period. The change between 1 and 3 years was divided by the number of days between the two observations and multiplied by 365 to obtain the rate of refractive change for the later period. The rates were compared to determine the progressive refractive change over time. Only eyes with data for all three times prior to a suture or automated lamellar keratoplasty were included in the analysis.
Results In the entire study group, there was a statistically significant decrease in the hyperopic shift between the early and late periods, from 0.33 diopter per year (0/yr) to 0.12 D/yr (P < .05, paired t-test) (Table 1). A more striking finding was seen among the low, medium, and high myopia groups. In the early period, all three had similar hyperopic shifts, 0.25 0, 0.36 0, and 0.40 0, respectively. However, in the late period, the low and medium groups had significantly lower rates of change than the high myopia group, 0.04 D/yr and 0.08 0/yr versus 0.22 0/yr, respectively (P < .05) (Figure 1 and
2A). This differs from the PERK data, which show a more uniform rate of change in all groups (Figure 2B). In the -5.0 and -6.0 D analyses (Table 2), the
shift toward hyperopia was relatively large in the early period. However, in the late period, there was a significantly lower rate of change in eyes with less than 5.0 and 6.0 0 of preoperative myopia (P< .05, paired t-test) but no significant change in eyes with more than 5.0 and 6.0 D (P < .05, paired t-test). There was a clear trend toward a lower rate of change in the later period in all but the more than 6.0 0 group. Thus, the more highly myopic eyes, particularly eyes with 6.12 0 and more, tended to show less stable postoperative refractions. Within the group of 71 highly myopic eyes, we compared the hyperopic shift in eyes that had no enhancements (n = 53) with the shift in those that had enhancements (n = 18). In the early period, the shift was 0.54 0 in the enhanced eyes and 0.32 0 in the nonenhanced eyes, and in the late period, it was 0.27 0 and 0.21 0, respectively; none of the differences was statistically significant. The rate of enhancement procedures was 46% in the highest myopia group and 27% in the lowest. The difference was statistically significant (P < .05).
Discussion In this paper, we addressed the issue of refractive stability. The worst case scenario would be that the 1 to 3 year rate of change of 0.05 0/yr would continue indefinitely. Were this the case, the average patient, about 40 years of age, whose preoperative myopia was less than 5.0 0, would be less than 1.0 0 hyperopic 30 years later
Table 1. Rate of hyperopic change in refraction (cycloplegic) (D/yr). Follow-up Period
Low Myopia Group
Medium Myopia Group
High Myopia Group
All Eyes
Number
Mean ±SO
Number
0.40:!: 0.81 0.22:!: 0.38
71 71
0.33:!: 0.73 0.12:!: 0.32
183 183
41 41
0.43:!: 0.80 0.21 :!: 0.35
68 68
0.35:!: 0.75 0.13:!: 0.31
165 165
2 2
-0.27:!: 0.85 0.51 :!: 0.82
3 3
0.18:!: 0.56 0.07:!: 0.43
18 18
Mean ±SO
Number
Mean ±SO
Number
Early Late
0.25:!: 0.57 0.04:!: 0.22
69 69
0.36:!: 0.84 0.08:!: 0.26 0.05*
42 42
Distance eyes only Early Late
0.24:!: 0.59 0.06:!: 0.21
56 56
0.37:!: 0.85 0.08:!: 0.34
Near eyes only Early Late
0.28:!: 0.51 0.04:!: 0.27
13 13
0.18:!: 0.25 0.12:!: 0.35
Mean ±SO
*Statistical exclusion of "outlier" based on 99.9% tolerance interval ] CATARACT REFRACT SURG-VOL 22, OCTOBER 1996
1031
HYPEROPIC SHIFT AFTER REFRACTIVE KERATOTOMY
Preoperative Myopia +0.25 to -3.12 Rate of Refractive Change 3 Months-1 Year
~
w
0
~
E :l
z
44 42 40 38 36 34 32 30 28 26 24 22 20 18 16
1 Year-3 Years
1/)
44 42 40 38 36 34 32 30 8 6 24 2 0 18 16 14 12 10 8 6 4 2
~ ~
0
~ ~
..-•
0 -2.0
E
:l
z
-1.0
0.0
+1.0
-+2.0
+3.0
'
-2.0
+4.0
---
-1.0
0.0
Diopters/Year
+1.0
+2.0
+3.0
+4.0
+3.0
+4.0
+3.0
+4.0
Diopters/Y ear
Preoperative Myopia -3.25 to -4.37 Rate of Refractive Change
3 Months-1 Year
1 Year-3 Years
38 36 34 32
~
w 0
~
E
:l
z
,., 1/)
Q)
24 22 20 18 16 14 12 10 8 6 4 2 0 -2.0
w 0
~
E
:l
z
•
-~ -1.0
0.0
+1.0
+2.0
+3.0
+4.0
44 42 40 38 36 34 32 30 28 26 24 22 20 18 16 14 12 10 8 6 4 2 0 -2.0
0.0
-1.0
Diopters/Y ear
+1.0
-
+2.0
Diopters/Year
Preoperative Myopia -4.50 to -9.62 Rate of Refractive Change 3 Months-1 Year
,., w 1/)
Q)
0
~
E
:l
z
44 42 40 38 36 34 32 30 28 26 24 22 20 18 16 14 12 10 8 6
4 2 0 -2.0
1 Year-3 Years
,., 1/)
Q)
w 0
-••
=.=1-1.0
0.0
+1.0
Cii .c E :l
z
+2.0
Diopters/Year
--- --+3.0
+4.0
44 42 40 38 36 34 32 30 28 26 24 22 20 18 16 14 12 10 8 6 4 2 0
-2.0
-1.0
+1.0
0.0
+2.0
Diopters/Year
Figure 1. (Werblin) Rate of dioptric change from 3 months to 1 year and from 1 to 3 years in the three myopia groups. 1032
J CATARACT REFRACT SURG-VOL 22, OCTOBER 1996
HYPEROPIC SHIFT AFTER REFRACTIVE KERATOTOMY
Casebeer Stability
[!!
!
0
i5
1.5 1.4 1.3 1.2 1.1 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
PERK Stability 1.5...---------------------, 1.4 1.3 1.2 1.1 1 [!! 0.9 .l!l a. 0.8 0.7 0 i5 0.6 0.5 0.4 0.3 0.2 0.1
-4.50 to -9.62 -3.25 to -4.37 +0.25 to -3.12
0
2
3
4
7
6
5
8
9
0~-.-~"~-,-~~~-r~~,-~~
10
0
2
3
4
5
Years
6
7
8
9
10
11
12
Years
A
B
Figure 2. (Werblin) A: Cumulative change in refraction in the study group by years postoperatively. B: Cumulative change in refraction in the PERK study by years postoperatively.
(Figure 3). Hyperopic shift (Figure 3, PERK5) and myopic shift (Figure 3, excimel or keratomileusis) with "stabilization" over time are such examples. In general, the hyperopic shift seen in RK begins to plateau within a few years of surgery. This tendency is particularly evident with lower myopic corrections. However, truly hyperopic shift, a constant increase in hyperopia over time, is suggested in Figure 2A, high myopia group. This ominous picture however may not be predictive over a longer time since the 3 year data in this study may not be sufficient to demonstrate this phenomenon. The most critical aspect of hyperopic shift is its cumulative effect on the refractive patient. Eyes with less than 6.0 D of myopic correction plateau with time, which means that the 0.05 D/yr shift seen between 1 and 3 years in this study is likely to diminish over time. If it did not diminish for 30 years, a -0.75 D initial result would be under + 1.0 D 30 years later. Most patients whose cycloplegic refraction is +0.50 to +0.75 D have 20/20 to 20/25 acuity without cycloplegia in the "real world," including presbyopia. A result under + 1.0 D at age 70 is still quite satisfactory. Even in this worst-case
if undercorrected the suggested amount -0.75 D; at no time would the uncorrected acuity be less than 20/30. This takes into account the shift of0.268 D noted in the first year and the shift of 0.0486 D/yr noted in the 1 to 3 year data. Patients between -5.0 and -6.0 D have enough hyperopic shift per year to make them significantly hyperopic 30 years after surgery, bur the assumption that the rate of change does not further plateau is likely wrong since the lower degrees of myopia in the PERK 10 year study show a long-term plateauing effect.5 Thus, patients in the -5.0 to -6.0 D range fall into a gray zone, bur clearly, at least for 30 years, patients under -5.0 D should have only minor hyperopic shift. Careful scrutiny of Table 1 raises several critical questions. Refractive stability is critical for all refractive procedures and several outcomes could occur. If, after the recovery from surgical trauma (several weeks), the refractive outcome did not change as is seen in intraocular surgery6 over time, the procedure would be considered stable. However, all corneal refractive procedures, including radial keratotomy (RK), excimer laser, and keratomileusis, show significant degrees of instability Table 2.
Rate of hyperopic change in refraction (cycloplegic) (D/hr) in eyes with 5.0 D and 6.0 D of preoperative myopia.
-5.0 0 Analysis +0.25 to -5.00 0
-6.0 0 Analysis
-5.12 to -9.62 0
+0.25 to -6.00 0
-6.12 to -9.62 0
Follow-up Period
Mean± SO
Number
Mean± SO
Number
Mean± SO
Number
Mean± SO
Number
Early Late
0.28 ± 0.67 0.07 ± 0.28
138 138
0.50 ± 0.88 0.29 ± 0.39
45 45
0.31 ± 0.73 0.09 ± 0.29
165 165
0.49 ± 0.77 0.44 ± 0.41
18 18
J CATARACT REFRACT SURG-VOL 22, OCTOBER 1996
1033
HYPEROPIC SHIFT AFTER REFRACTIVE KERATOTOMY
0.7
1.5..-----------------~
'C'
0.5
~c:
0
'$
c. 0
i5
(tl
0.4
0::
0.3
.E
-0.5
C1)
Ol
c:
0.2
(tl
.t:
-1 -1.5
0.5
0
"13
!!!
~
0.6
()
0
2
3
4
5
6
7
8
9
10
Years
0.1 0
0
2
3
4
5
6
7
8
9
10
Diopters of Intended Correction
Figure 3. (Werblin) Long-term refractive stability of several corneal and intraocular procedures.
Figure 4. (Werblin) Relationship between the rate of refractive change 1 to 3 years postoperatively and the preoperative cycloplegic spherical equivalent refraction.
scenario, I.e., no further plateauing of the hyperopic shift, the long-term effect of the hyperopic shift is negligible. The statistical comparison of the rate of refractive change in the first year, actually between the 3 month and 1 year visit, with the rate between 1 and 3 years is the critical issue. The high myopia group (Figure 2A), in contrast to the low and medium myopia groups, showed a more persistent rate of hyperopic change between 3 months and 1 year. In addition, the rates of hyperopic change for low, medium, and high myopia groups did not differ statistically for the 3 month to 1 year period but were significantly different later. Therefore, a better indicator oflong-term instability is the rate of hyperopic change 2 to 3 years after the final surgical procedure. This is shown graphically in Figure, 4 in which the low, medium, and high myopia groups are divided into six subgroups based on preoperative intended cycloplegic refraction. A nonlinear relationship is demonstrated, indicating increasing instability as a function of intended correction particularly above -6.0 D. Eyes with less than 4.5 D of myopia showed little evidence of progression. In eyes with 4.5 to 6.0 D of myopia, a greater hyperopic shift was seen but it appeared to plateau ( +0.16 D/yr in the 1 to 3 year period). Eyes with more than 6.0 D of intended correction showed a greater progressive shift ( +0.44 D/yr for the 1 to 3 year period) (Figure 4). Even in this group, however, many eyes did not show consistent changes in both periods but had a large shift in either the 3 month to 1 year or the 1 to 3 year period. In the lower myopia 1034
groups, patients occasionally demonstrated major hyperopic shifts, overcorrections, or both. Therefore, throughout this population there was a general trend toward a small hyperopic shift. Exceptions were seen occasionally in lower degrees of myopia and more commonly in higher degrees of myopia (Figure 2). Eleven eyes (9%) showed evidence of a constant increase in hyperopia of 0.5 D or more in each year of the study rather than plateauing of the hyperopic change or an isolated hyperopic change. These progressive changes represent a real challenge to the refractive keratotomy surgeon. Six of the eyes have had suture enhancements, which can reverse the problem. However, the additional cost and effort involved in these subsequent procedures are significant. In only one patient was this process bilateral. It is interesting to speculate why these data appear to stabilize more rapidly than data presented in the PERK studies (Figure 2A, B). The PERK data showed that shorter incisions, i.e., larger optical zones, were more stable. 5 Newer concepts of limited amounts of incisional surgery have also demonstrated more corneal stability with shorter incisions. 8 Since our execution of the Casebeer System starts the incisions well inside the limbus, unlike the PERK system, the shorter incisions may correlate with the greater degree of corneal stability. To date there has been no clear explanation of why the cornea is destabilized by refractive keratotomy in some patients. Generally, but not always, larger surgical corrections are correlated with higher rates of instability. The greatest surgical effect usually occurs the moment
J CATARACT REFRACT SURG-VOL 22, OCTOBER 1996
HYPEROPIC SHIFT AFTER REFRACTIVE KERATOTOMY
the surgery is completed. Subsequent healing correlates with a variable amount of regression. Clinically, the wounds become slightly opalescent and have a feathery appearance along the wound edges. The opalescent appearance resolves over time. Are we sure that corneal wound healing or lack thereof is totally responsible for refractive instability following RK? A recent report on long-term RK results emphasized the "natural" hyperopic refractive progression of "normal" eyes as a function of age. 9 Previous reports 10 - 12 indicate that between ages 31 and 64 years, an average 0.035 D/yr shift toward hyperopia occurs. To look at this problem analytically, we took the data from Table 1 and separated patients below age 31 from the rest of the group and recalculated the change in refraction as a function of time postoperatively (Table 3). For the 1 to 3 year period, patients below 31 years had a much lower rate of refractive change than patients over 31 years. Even though the 31 and older patients had a relatively higher rate of change, 0.048 D/yr, when the "natural" hyperopic drift of the eye is subtracted (0.035 D/yr), both age groups seem to behave the same, 0.014 D/yr and 0.013 D/yr, respectively. Despite this finding, whether the shift is "natural" or the result of surgical wound healing, small surgical undercorrections are still appropriate. Hyperopic shift should be a serious concern after routine refractive keratotomy procedures for certain, rei-
Table 3.
atively "high risk" patients. It is clear from the experience of most refractive keratotomy surgeons that myopic patients who become symptomatically hyperopic after surgery are extremely unhappy, often more unhappy than when they were myopic. Therefore, most surgeons use nomograms that attempt to achieve a slight undercorrection. Most patients who are accustomed to 20/20 or better corrected acuity prior to surgery, however, are not ecstatic about less than 20/25 or 20/30 uncorrected acuity postoperatively. Some are even dissatisfied with 20/20 acuity, especially if the fellow eye is20/15. A delicate balance between the amount of undercorrection that is surgically advisable and that which is tolerated and functional for the patient has to be achieved. Because the average hyperopic shift appears to be about 0.60 D in the first 3 years following surgery, the ideal patient will be undercorrected by 0.50 to 0. 75 D immediately after surgery. This would allow the expected hyperopic shift to place the patient close to emmetropia, while leaving him or her at worst with 20/25 to 20/30 uncorrected acuity immediately after surgery. Although this may not be the patient's ideal result, the physician should make every attempt to explain the likely improvement in time with healing, rather than relying too heavily on enhancement procedures for small amounts of residual myopia. A better solution may move our thinking away from the cornea to avoid
Rate of hyperopic change in refraction (cycloplegic) (D/yr) as a function of preoperative age.
Preoperative Range in Myopia
-4.50 to -9.62 0
+0.25 to -4.37 0 Mean± SO
Number
Mean± SO
Number
All eyes First year 1 to 3 years
0.290 ± 0.685 0.043 ± 0.236
111 111
0.398 ± 0.808 0.224 ± 0.375
71 71
Below 31 years First year 1 to 3 years
0.252 ± 0.677 0.014 ± 0.232
17 17
0.454 ± 1.009 0.214 ± 0.246
20 20
31 years and above First year 1 to 3 years
0.297 ± 0.690 0.048 ± 0.238
94 94
0.376 ± 0. 724 0.229 ± 0.447
51 51
Age/Follow-up
31 years and above (minus normal hyperopic "drift") First year 1 to 3 years
0.262 0.013
J CATARACT REFRACT SURG-VOL 22, OCTOBER 1996
0.341 0.194
I035
HYPEROPIC SHIFT AFTER REFRACTIVE KERATOTOMY
wound healing and optical distortion difficulties inherent in all forms of corneal surgery. 13 7.
References 1. Werblin TP, Stafford GM. Three year results of refractive keratotomy using the Casebeer System. J Cataract Refract Surg 1996; 22: 1023-1029 2. Werblin TP, Stafford GM. The Casebeer System for predictable keratorefractive surgery; one-year evaluation of 205 consecutive eyes. Ophthalmology 1993; 100:10951102 3. Casebeer JC. Casebeer lncisional Keratotomy. Thorofare, NJ, Slack Inc, 1995 4. Friedlander MH, Nordan LT, Maxwell WA, eta!. Radial Keratotomy predictability (letters to the editor). Ophthalmology 1994; 101:411-416 5. Waring GO Ill, Lynn MJ, McDonnell PJ, and the PERK Study Group. Results of the Prospective Evaluation of Radial Keratotomy (PERK) study 10 years after surgery. Arch Ophthalmol 1994; 112: 1298 -1308 6. Werblin TP. Refractive stability after cataract extraction
1036
8. 9.
10.
11.
12. 13.
using a 6.5-millimeter scleral pocket incision with horizontal or radial sutures. J Refract Corneal Surg 1994; 10:339-342 Maguen E, Salz JJ, N esburn AB, et a!. Results of excimer laser photorefractive keratectomy for the correction of myopia. Ophthalmology 1994; 101: 1548-15 56 Lindstrom RL. Minimally invasive radial keratotomy: mini-RK. J Cataract Refract Surg 1995; 21:27-34 Sawelson H, Marks RG. Ten-year refractive and visual results of radial keratotomy. Ophthalmology 1995; 102: 1892-1901 Hirsch MJ. Changes in refractive state after the age of forty-five. Am J Optom Arch Am Acad Optom 1958; 35:229-237 Walton WG Jr. Refractive changes in the eye over aperiod of years. Am J Optom Arch Am Acad Optom 1950; 27:267-286 Slataper FJ. Age norms of refraction and vision. Arch Ophthalmol1950; 43:466-481 Werblin TP. Should we consider clear lens extraction for routine refractive surgery? Refract Corneal Surg 1992; 8:480-481
J CATARACT REFRACT SURG-VOL 22, OCTOBER 1996