- Email: [email protected]
A comparison of laser photocoagulation with cryotherapy for threshold retinopathy of prematurity at 10 years
A Comparison of Laser Photocoagulation with Cryotherapy for Threshold Retinopathy of Prematurity at 10 Years Part 1. Visual Function and Structural Outcome Eugene Y. J. Ng, BSc,1 Brian P. Connolly, MD,2 J. Arch McNamara, MD, FRCSC,2 Carl D. Regillo, MD,2 James F. Vander, MD,2 William Tasman, MD2 Objective: To assess visual and structural outcomes after laser photocoagulation and transscleral cryotherapy for threshold retinopathy of prematurity (ROP) after 10 years. Design: Extended follow-up of a randomized controlled clinical trial. Participants: One hundred eighteen eyes from 66 patients were randomly assigned to receive either cryotherapy or laser photocoagulation for threshold ROP. Forty-four eyes from 25 patients were examined for 10-year follow-up evaluations. Intervention: Early Treatment of Diabetic Retinopathy Study (ETDRS) visual acuity, slit-lamp, and fundus examination; fundus photography; and B-scans (eyes with retinal detachments) were performed. Patients’ histories were taken to elicit past amblyopia therapy. Based on fundus photographs, independent observers graded the degree of retinal dragging as none, mild, moderate, or severe. Main and Secondary Outcome Measures: Best-corrected ETDRS visual acuity (BCVA). The degree of dragging was determined clinically and photographically. In addition, the presence of strabismus or amblyopia and/or any history of treatment for amblyopia were noted accordingly. Results: Eyes treated with laser had a mean BCVA of 20/66 (Snellen equivalent), whereas cryotherapytreated eyes had a mean BCVA of 20/182 (Snellen equivalent) (P ⫽ 0.015, n ⫽ 42). Compared with eyes treated with cryotherapy, eyes treated with laser photocoagulation were 5.2 times more likely to have a 20/50 or better BCVA (95% confidence interval, 1.37–19.8, n ⫽ 42). Eyes treated with cryotherapy were 7.2 times (95% confidence interval, 1.54 –33.6, n ⫽ 33) more likely to develop retinal dragging compared with laser treatment. By linear regression analysis, ETDRS visual acuity was inversely proportionate to the degree of retinal dragging in both laser (r ⫽ ⫺0.637, P ⫽ 0.006) and cryotherapy (r ⫽ ⫺0.517, P ⫽ 0.040) treated eyes. Among the 21 patients with favorable outcomes in both eyes, 13 had strabismus (62%) and 6 had received amblyopia therapy (29%). Ptosis, loss of cilia, and cortical cataract were among probable treatment-related complications that were noted in this study. Conclusions: Overall, laser-treated eyes had better structural and functional outcome compared with eyes treated with cryotherapy. Ophthalmology 2002;109:928 –935 © 2002 by the American Academy of Ophthalmology. The Multi-center Trial of Cryotherapy for Retinopathy of Prematurity was a landmark study in which ablation of the peripheral avascular retina with cryotherapy was shown to benefit patients with threshold retinopathy of prematurity (ROP).1 More recently, the use of laser photocoagulation to ablate the peripheral retina has gained widespread accep-
tance in the treatment of ROP.2– 6 The patients in our series took part in a previously reported clinical trial from which we concluded that laser was at least as good as cryotherapy.2,3,5 Each of these patients had undergone treatment for threshold ROP in one or both eyes. The treatment modality used for each eye was determined randomly; however, for
Originally received: January 17, 2001. Accepted: July 26, 2001.
None of the authors has any financial or proprietary interest in any of the techniques or equipment discussed in this article.
Manuscript no. 210036.
1
University College of Dublin, National University of Ireland, Dublin, Ireland.
2
Wills Eye Hospital Retina Service, Philadelphia, Pennsylvania. Supported by an unrestricted grant from Research to Prevent Blindness, New York, New York, and by a grant from Iridex Corporation, Mountain View, California.
928
© 2002 by the American Academy of Ophthalmology Published by Elsevier Science Inc.
Reprint requests to William S. Tasman, MD; Ophthalmologist-in-Chief; Wills Eye Hospital; 900 Walnut Street; Philadelphia, PA 19107. Presented in part at the American Academy of Ophthalmology annual meeting, New Orleans, Louisiana, November 2001.
ISSN 0161-6420/02/$–see front matter PII S0161-6420(01)01017-X
Ng et al 䡠 Comparison of Laser and Cryotherapy for ROP patients with bilateral disease, the same modality was never used in both eyes (one eye received laser treatment, and the fellow eye received cryotherapy). At 5-year follow-up, we reported that laser-treated eyes had better visual acuity and less myopia compared with cryotherapy-treated eyes.7 Nonetheless, the previous report had several limitations that were addressed in this study. First, in the earlier study, visual acuities were measured by a variety of methods and observers; whereas in this study, the visual acuities were all measured under standardized Early Treatment of Diabetic Retinopathy Study (ETDRS) conditions by the same observer. Second, in the previous report, the presence of dragging was based on the subjective opinions of multiple examiners; whereas in this study photographic and clinical grading were performed at one center. Finally, this study looked at why the visual acuity was reduced in the cryotherapy-treated eyes, which was a question that our earlier reports did not answer.
unilateral ROP, the randomization center again would instruct the clinician which mode of treatment to administer to the eye selected by a coin toss. The alternate modality was then administered to the fellow eye. Once randomization and treatment were completed, no attempt was made to conceal which treatment had been performed in a given eye, because there was typically marked eyelid swelling in the cryotherapy-treated eyes.
Treatment Administration Cryotherapy and/or laser photocoagulation was administered within 72 hours of the diagnosis of threshold ROP (as described elsewhere).2,3 The first 29 laser treatments performed before October 1990 were performed in an operating room using a liquid cooled argon ion laser (Coherent Medical Group, Palo Alto, CA). In October 1990, the infrared diode laser (IRIS Medical Instruments Inc., Mountain View, CA) became available and was used in the remaining treatments.
Follow-up Examination
Materials and Methods The clinical objective of this study was to evaluate the anatomic and visual results of these patients in the two treatment groups (cryotherapy and laser) 10 years after treatment. The relationship of visual acuity to anatomic outcome was also examined.
Threshold ROP All infants who developed threshold ROP during the study period at one institution were eligible for this study (i.e., five or more contiguous or eight cumulative clock hours of stage 3 ROP in zone I or II in the presence of “plus” disease).1 In accordance with our institutional review board, consent was obtained from either the parents of the children or their legal guardians before randomization and treatment. Patients whose parents or legal guardians did not wish to enroll their children in the study were not included. Sixty-six infants were entered in a prospective randomized clinical trial. This sample size was chosen empirically. Of these, 39 infants had symmetric disease at the onset of disease and underwent bilateral treatment. Another 27 patients had unilateral threshold disease, 15 (56%) of whom were allocated to laser treatment and 12 (44%) of whom were allocated to cryotherapy in the first eye. Of the 15 patients treated initially with laser in one eye, 3 patients (20%) progressed to threshold ROP in their fellow eye and were treated with cryotherapy. In contrast, 10 of 12 (83%) patients whose initial eye was treated with cryotherapy progressed to threshold disease and required laser treatment in the fellow eye.
Randomization At the outset of the study, a stack of opaque envelopes was prepared. Half of the envelopes contained a piece of paper with the word “cryotherapy” printed on it, and the remaining half stated “laser.” These envelopes were shuffled and stored at the randomization center. When a patient was diagnosed with unilateral threshold ROP, the randomization center was instructed to open the envelope at the top of the stack, and the treatment modality printed on the piece of paper was administered. If the fellow eye subsequently developed threshold ROP, the alternative modality was then administered. In cases with bilateral threshold disease, a coin was tossed to choose one eye for randomization. As with
In June 2000, we invited all patients we could contact to return for a follow-up eye examination. Twenty-five patients agreed to attend. Two patients were deceased, two patients had relocated to a distant location, 11 patients refused examination, and 26 were lost to follow-up. Figure 1 shows a detailed flow chart of all the patients from this study. Of the 25 patients who returned for follow-up, 19 had undergone bilateral treatment, and six had undergone unilateral treatment. Detailed examinations were performed on a total of 23 laser-treated eyes and 21 cryotherapy-treated eyes. During the visit, a history was taken with attention to prior amblyopia therapy. Hirschberg tests and/or cross-cover tests were performed to detect ocular misalignment. Best-corrected visual acuity (BCVA) and uncorrected visual acuity were measured with a backlit ETDRS acuity chart (Lighthouse Low Vision, Long Island, NY). After administration of tropicamide 1.0%, phenylephrine hydrochloride 2.5%, and cyclopentolate hydrochloride 1.0%, fundus photography (Kodachrome), slit-lamp examination, and indirect ophthalmoscopy were carried out. The structural outcome of the fundus was assessed clinically by indirect ophthalmoscopy. In accordance with the Multi-center Trial of Cryotherapy for ROP, macular fold, retinal detachment, retrolental mass, and phthisis bulbi were recorded as unfavorable outcomes. The amount of dragging present was assessed clinically as none, mild, moderate, or severe. In situations in which media clarity was unsatisfactory, through the lid ultrasonic B-scan (Innovative Imaging, Sacramento, CA) was performed. Finally, the degree of retinal dragging was assessed on the basis of fundus photographs. Three independent observers (BPC, JAM, WT) graded photographic slide(s) from each eye of each patient. These observers were masked to the modality of treatment in a given eye. Only photographs showing both optic disc and macula were used. Areas showing treatment lesions on four photographs (from three patients) were covered to prevent the reader from recognizing the treatment modality. The slides were shuffled before being graded. The degree of dragging present in each photograph was categorized as follows: none, mild, moderate, severe. Figures 2 through 4 are representative of the mild, moderate, and severe categories. There was 100% agreement among the graders for the photographs seen in Figures 2 through 4. Figure 2 shows the mildest dragging among our fundus photographs and was classified as mild dragging by all observers. Figure 3 shows the minimum amount of dragging necessary to be classified as moderate.
929
Ophthalmology Volume 109, Number 5, May 2002
Figure 1. Detailed chart summarizing patient flow and completeness of follow-up. Pt. ⫽ patient, pts. ⫽ patients.
Figure 4 shows severe dragging. Eyes with retinal detachments were not included in this analysis of photographic grading. Multiple views of photographs of the same eye served as built-in controls. A total of 45 slides were graded in a single session by each observer. Although the degree of dragging was a subjective measure, there was good interobserver agreement. In four slides, one of the three observers graded the slides as normal, whereas the other two graded them as mild dragging. Two other slides were graded as moderate dragging by two observers, but the other graded them as severe dragging. In these cases, the slides were reported according to the observation of the two agreeing observers. Hence there were interobserver differences in 6 of the 45 slides (13%). Within each observer, only 1 out of 10 built-in controls differed in grading.
Data Analysis and Recording Patient data and results obtained were entered into a Microsoft Access 2000 database, and statistical analysis was performed using Microsoft Excel 2000. Mean age, birth weight, and gestational age were calculated. Gender and racial distribution of these patients were treated as binomials and are represented as percentages. To ensure uniform baseline characteristics of disease at treatment, the mean number of clock hours of stage 3 disease and the percentage of eyes with zone 1 disease at the time of treatment were calculated for both the laser-treated eyes and the cryotherapy-treated eyes of those patients who returned for follow-up examination. Unpaired twotailed Student’s t tests were performed to determine whether the number of clock-hours of stage 3 disease differed between the treatment groups. Chi-square analysis was used to ensure uniformity of the zone of disease between the two groups at the time of treatment. Both paired and unpaired two-tailed Student’s t tests were administered to compare the two treatment groups’ ETDRS BCVA scores. Eyes with visual acuities of finger counting (less than one ETDRS visual acuity score), hand motion vision, and no light perception were assigned an ETDRS visual acuity score of zero.
930
An unpaired test was administered to analyze all treated eyes. In addition, a paired two-tailed Student’s t test was administered to analyze only patients who had received bilateral treatment. In keeping with our previous report,7 a two-by-two contingency table was constructed to determine the odds of laser treatment resulting in visual acuity of 20/50 or better (ETDRS score ⱖ65) compared with cryotherapy. Because of cognitive limitations, patients 17 and 20 were not cooperative with ETDRS acuity testing and fundus photography. Data for patient 20 were not analyzed. Because one eye of patient 17 was unmistakably blind, it was accorded a visual acuity score of zero and used in the analysis (unpaired) of data. She was not cooperative with testing of the seeing eye. Patient 16 had moderate cognitive impairment; however, she was able to perform a Tumbling-E Test for visual acuity testing. The results were translated into approximate ETDRS units and then to ETDRS visual acuity scores for our calculations. Fundus photographs of patients 5 and 13 were unavailable, because the negatives were inadvertently destroyed while being processed. For statistical analysis, an ordinal value of 0 to 3 was assigned to the photographically assessed degree of dragging. Those eyes with no dragging were assigned a value of 0 points, mild dragging was assigned 1 point, moderate dragging 2 points, and severe dragging 3 points. Macular fold of retina was also accorded 3 points. Eyes with retinal detachments were not included in this analysis. The difference in the degree of dragging between eyes treated with the two different modalities was compared in the same manner as visual acuity. In addition, linear regression analysis and analysis of variance were performed to determine whether the BCVA of cryotherapytreated eyes was correlated to that of laser-treated eyes among bilaterally treated patients and whether visual outcome achieved was associated with the anatomic outcome of treatment. This requires inclusion of only eyes for which both BCVA and the degree of dragging were known. Both the correlation coefficient (r) and the P value were calculated. Where appropriate, axis intercepts of the linear trend line together with their P values were
Ng et al 䡠 Comparison of Laser and Cryotherapy for ROP
Figure Figure Figure Figure Figure Figure
2. 3. 4. 5. 6. 7.
Fundus photograph showing mild dragging. Fundus photograph showing moderate dragging. Fundus photograph showing severe dragging. Loss of cilia in a cryotherapy-treated eye. Fundus photograph, patient 12, shortly after argon laser photocoagulation, Fundus photograph, patient 12, at 10 year follow-up.
calculated and included for discussion. Statistical significance was defined as a P value ⱕ 0.05 for this report.
Results The mean age at follow-up was 9.9 years (range, 8.6 –11.1 years). Of the 25 patients, 11 (44%) were male and four (16%) were
African American. This cohort of patients had an average gestational age of 26 weeks (range, 23–33 weeks). Their mean birth weight was 807 g (range, 440 –1477 g). Cryotherapy-treated and laser-treated eyes had similar preoperative characteristics. The mean number of clock hours of stage 3 disease and percentage of eyes with zone 1 disease for each of the treatment groups are shown in Table 1.
931
Ophthalmology Volume 109, Number 5, May 2002 Table 1. Comparison of Baseline Characteristics (Clock Hours of Stage 3 and Zone) Between the Treatment Groups Laser-treated Eyes Mean number of clock hours stage 3 Standard error of the mean P value Percentage of zone 1 eyes P value for 2
Cryotherapy-treated Eyes
7.7
7.0
0.61
0.54 0.44
9% (6 of 64)
11% (6 of 54) 0.95
P values based on unpaired, two-tailed Student’s t test results. There were no significant differences between the two treatment groups in terms of these baseline characteristics.
Functional and Visual Outcome Overall, those treated with laser had a mean ETDRS visual acuity score of 59.3, corresponding to a Snellen acuity-equivalent of 20/66 (standard error of the mean, 5.8). The mean ETDRS visual acuity score of the cryotherapy-treated eyes was 37.5, corresponding to a Snellen acuity of 20/182 (standard error of the mean, 6.3). An unpaired two-tailed Student’s t test used to compare the two treatment groups showed a statistically significant difference (P ⫽ 0.0152, n ⫽ 42 eyes; see Table 2). A two-by-two contingency table was constructed to compare the visual acuity of the two treatment groups (see Table 3). An odds ratio was calculated, and a 95% confidence interval was
determined by Woolfe’s method.8 The laser-treated eyes were 5.20 times more likely to have 20/50 or better BCVA (95% confidence interval, 1.37–19.77, n ⫽ 42 eyes). When 18 patients who had undergone bilateral treatment (and who were capable of ETDRS vision testing) were analyzed separately by paired Student’s t test analysis, a significant difference between the treatment groups (P ⫽ 0.033, n ⫽ 18 pairs of eyes) was again evident. The laser-treated eyes had a mean ETDRS visual acuity score of 59.1 (standard error of the mean, 6.7) comparable to 20/66 Snellen acuity, whereas cryotherapy-treated eyes had a mean ETDRS visual acuity score of 41.4 (standard error of the mean, 6.8) corresponding to a 20/152 Snellen equivalent. Although it would have been interesting to stratify patients in each treatment group by initial ROP severity (i.e., zone, clock hours of disease), the sample size was not large enough to do so. Amblyopia therapy had been administered to 6 of the 21 binocular patients (29%). Of these six amblyopic eyes, four were treated with cryotherapy compared with two who were treated with laser. Five had mild dragging, whereas one had moderate dragging. In all six cases, the fundi of the fellow eyes were anatomically normal. Four of these six also had strabismus in the same eye.
Anatomic Outcome Of the 21 eyes treated with cryotherapy, four (19%) eyes had unfavorable outcomes (two eyes developed stage 5 ROP, one eye developed stage 4B ROP, and another eye developed a macular fold). There were 2 (10%) unfavorable outcomes among the 23 eyes treated with laser; both developed stage 5 ROP. Of 21 patients with favorable outcomes in both eyes, 13 (62%)
Table 2. Summary Table of Visual Acuities and Dragging For All Patients Laser-treated Eyes
Cryotherapy-treated Eyes
Patient No.
Best-corrected Visual Acuity
Snellen Visual Acuity
Dragging
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
88 83 78 75 73 73 73 70 69 67 67 62 57 51 1 76 ? 0 0 ? 76 61 45
20/16 20/20 20/25 20/25 20/32 20/32 20/32 20/32 20/40 20/40 20/40 20/50 20/63 20/80 20/800 20/25 CSM NLP NLP CSM 20/25 20/50 20/100
Normal Normal Normal Normal * Normal Normal Normal Normal Normal Normal Moderate dragging * Normal Moderate dragging Mild dragging * Stage 5 Stage 5 * Normal Moderate dragging Moderate dragging
Best-corrected Visual Acuity
Snellen Visual Acuity
Dragging
68 40 74 0 77 69 7 54 57 64 63 10 8 4 46 70 0 35 0
20/40 20/125 20/32 NLP 20/25 20/40 20/640 20/80 20/63 20/50 20/50 20/640 20/640 20/800 20/100 20/40 NLP 20/200 NLP
Mild dragging Mild dragging Normal Stage 4B * Normal Mild dragging Moderate dragging Mild dragging Mild dragging Mild dragging Moderate dragging * Retinal fold of macula Moderate dragging Normal Stage 5 Severe dragging Stage 5
36 5
20/160 20/800
Normal Mild dragging
Synopsis of patient’s visual acuity and fundus anatomy. Patients 20 thru 25 received only unilateral treatment. BCVA ⫽ best-corrected visual acuity; CSM ⫽ central steady and maintained; NLP ⫽ no light perception. * Favorable outcome but fundus photograph not available for systematic assessment of dragging.
932
Ng et al 䡠 Comparison of Laser and Cryotherapy for ROP Table 3. A Comparison of Visual Acuities between Lasertreated Eyes and Cryotherapy-treated Eyes
Table 4. A Comparison of Dragging between Laser-treated Eyes and Cryotherapy-treated Eyes
Treatment
20/50 or Better
20/60 or Worse
Total
Treatment
No Dragging
Any Dragging
Total
Laser Cryotherapy Total
13 5 18
8 16 24
21 21 42
Laser Cryotherapy Total
12 4 16
5 12 17
17 16 33
Patients 17 and 20 were cognitively impaired and unable to perform acuity testing. Patient 17 had a retinal detachment in the cryotherapy-treated eye, so that eye was clearly blind and was included in this table. Although the patient’s laser-treated eye was able to fixate steadily, it was excluded from this analysis because visual acuity could not be determined accurately. In addition, patient 20, who had unilateral treatment with laser, was unable to perform acuity testing because of cognitive impairment and was excluded from this analysis.
patients had strabismus. Ten of these 13 had nonaccommodative esotropia. One patient had an alternating esotropia, and two patients demonstrated intermittent exotropia. In addition, among eyes with a favorable outcome, one patient had a small (one clock hour in width) cortical cataract in the laser-treated eye. Furthermore, loss of cilia was observed in one eye treated with cryotherapy (Fig 5). Two other patients developed blepharoptosis in their cryotherapy-treated eyes. One of these eyes had a macular fold and an axial length of 20.1 mm, whereas the other eye had normal fundus anatomy and an axial length of 22.2 mm.
Fundus Photography Because of lack of cooperation or somewhat opaque ocular media, fundus photography was not possible in some patients, but we were able to obtain photographs in 16 cryotherapy-treated eyes and 17 laser-treated eyes. Of the 16 cryotherapy-treated eyes photographed, 7 had mild dragging, 3 had moderate dragging, two had severe dragging, and the remaining 4 were normal. In comparison, 12 of the 17 (71%) photographed laser-treated eyes were normal, 1 (6%) had mild dragging, 4 (24%) had moderate dragging, and none had severe dragging (see Table 2). Eyes with retinal detachments were not photographed. Of the 13 patients who received bilateral treatment (and for whom fundus photographs were taken), laser-treated eyes had significantly less dragging. The laser-treated eyes had a mean scaled score of 0.39, whereas cryotherapy-treated eyes were more likely to have dragging and had a mean score of 1.15 (P ⫽ 0.018, n ⫽ 13 pairs of eyes, paired Student’s t test). When all eyes for which photographs were available were included in this analysis (unpaired Student’s t test), the mean score in laser-treated eyes was 0.53 (n ⫽ 17 eyes) compared with 1.19 (n ⫽ 16 eyes) in cryotherapy-treated eyes (P ⫽ 0.05, n ⫽ 33 eyes). Because the unpaired Student’s t test is a less powerful statistical analysis, more patients would be needed to detect a true difference of similar magnitude compared with the more powerful paired test. Nonetheless, eyes treated with cryotherapy were 7.2 times (95% confidence interval, 1.54 –33.6, n ⫽ 33) more likely to develop dragging compared with fellow laser-treated eyes (see Table 4). This odds ratio was calculated by Woolfe’s method.8 Linear regression analysis revealed the following findings. First, when comparing the BCVA of one eye with the fellow eye, a poor correlation coefficient was obtained (r ⫽ 0.36, P ⫽ 0.14, n ⫽ 18). Because this implies that the BCVA of one eye is poorly
Eyes with mild, moderate, and severe dragging were included in the “any dragging” category for this table.
correlated and not predictive of the BCVA of the fellow eye, it further strengthens our observation of a real difference in BCVA between the two different treatment modalities. Second, the absence of dragging was a predictor of good visual acuity in both laser-treated and cryotherapy-treated eyes. There was a strong negative correlation between the BCVA and the degree of dragging in laser (r ⫽ ⫺0.64, P ⫽ 0.0059, n ⫽ 17) and cryotherapy eyes (r ⫽ ⫺0.52, P ⫽ 0.04, n ⫽ 16). Third, the intercept values obtained from line fit plot statistics suggest a predicted BCVA score of 73 (Snellen equivalent 20/32) in the absence of retinal dragging for laser-treated eyes and a predicted BCVA score of only 60 (Snellen equivalent 20/50) in the absence of retinal dragging for cryotherapy-treated eyes. This raises the possibility that laser treatment may allow for better BCVA potential than cryotherapy even when minimal dragging is present. Because fundus photographs of patient 12 were taken in the operating room at the time of argon laser treatment, we were able to monitor the progression of retinal dragging. Figure 6 is a fundus photograph taken soon after treatment, and Figure 7 was obtained during this follow-up 10 years later. It was interesting to note that the dragging had progressed in both eyes. Unfortunately, we only have postoperative photographs of one of these eyes for comparison.
Discussion After roughly a decade, many of the children who returned for follow-up evaluations were cognitively capable of cooperating with our study. Our results show that visual acuity is significantly better in those eyes treated with laser photocoagulation into late childhood. This significant difference persisted despite administration of amblyopia therapy (where appropriate) in these patients. The probability that chance alone accounted for the difference in visual acuity between the two treatment groups was approximately 3%. The observation that visual acuity was better in cryotherapy-treated eyes was consistent with our earlier report.7 Another randomized prospective clinical trial comparing cryotherapy to laser photocoagulation for threshold ROP was published by White and Repka9 They also observed that visual acuity was somewhat better in the laser-treated eyes; however, this trend was not statistically significant because of their relatively small sample size. In addition, two nonconcurrent retrospective comparisons of cryotherapy and laser photocoagulation by Paysse et al10 and Pearce et al11 demonstrated that the visual outcomes were superior in the laser-treated eyes.
933
Ophthalmology Volume 109, Number 5, May 2002 The underlying reason for the difference in the long-term visual acuities after these two different treatments is not clear. In this study, cryotherapy-treated eyes were shown to be more likely to develop dragging, and linear regression analysis indicated that dragging was associated with poorer visual acuities (not surprisingly). This would suggest that inferior macular structural outcomes were an important factor in the cryotherapy-treated eyes. Nonetheless, it is difficult to explain why dragging was more prevalent in the cryotherapy-treated eyes. Although both laser treatment and cryotherapy are ablative treatments, the type of tissue damage differs. Transscleral cryotherapy to the developing sclera causes retinal, choroidal, and scleral disruption and may perpetuate disorganized scar tissue formation.12 The photocoagulation burns produced by a transpupillary diode laser (810 nm) produce retinal and choroidal changes. Compared with comparable argon laser spots, diode laser burns show more of an effect in the choroid.13 In a rabbit model, Smiddy and Hernandez14 reported that the effect of laser burns on the sclera tends to be fairly minimal, except when high energy levels were used to treat atrophic areas. When present in the area of a photocoagulation burn, ciliary nerves in the choroid or sclera consistently showed scarring in a cynomolgus monkey model.13 Perhaps the more widespread nature of the thermal injury that occurs with cryotherapy accounts for part of the additional dragging seen in these patients. When linear regression analysis was used to predict the mean BCVA in the absence of any dragging, cryotherapytreated eyes seemed to have reduced potential visual acuity even in the absence of dragging compared with laser-treated eyes. Thus, dragging alone did not seem to account for all of the reduction in visual acuity in the cryotherapy-treated eyes. The cryotherapy-treated eye of patient 24 was fairly amblyopic with an ETDRS score of 36 (Snellen equivalent 20/160), and this may have accounted for some of this discrepancy. Nonetheless, when this patient was excluded from the linear regression analysis, the intercept was 66 (Snellen equivalent 20/40)—still somewhat less than that for laser-treated eyes. Although this study detected a difference in the amount of dragging and a possible difference in the average visual potential between the two treatment modalities, even in the absence of dragging, it was not designed to determine the underlying reasons for these discrepancies. Late-appearing or persistent treatment complications were relatively rare. Those that were apparently related to cryotherapy treatment included a juxtamacular chorioretinal scar, loss of cilia, and blepharoptosis. In addition, a cortical cataract was seen in a laser-treated eye. Some limitations of this study included a relatively small number of patients and a recall rate of only 38%. Furthermore, a number of patients examined were cognitively limited, which led to variable participation in the different tests performed. Nonetheless, these patients are a crosssection of the premature patients who are affected by this condition. Additional amblyopia therapy would be unlikely to benefit these children; thus, it would be unlikely to find
934
any significant new trends in visual acuity after a 10-year follow-up. Despite some limitations, the randomized design of this original study controls for many potential confounding variables. The 10-year follow-up results of this study suggested that laser photocoagulation used in the treatment of ROP resulted in better visual and anatomic outcome and was associated with less long-term morbidity than cryotherapy treatment. Although this trial was smaller than the Multicenter Trial of Cryotherapy for ROP, it is unlikely that another trial of similar scale will be undertaken to compare cryotherapy with laser photocoagulation. Acknowledgment. The authors thank Andrew Smith, PhD, for his extensive statistical assistance with this article.
References 1. Multicenter trial of cryotherapy for retinopathy of prematurity. Preliminary results. Cryotherapy for Retinopathy of Prematurity Cooperative Group. Arch Ophthalmol 1988;106:471–9. 2. McNamara JA, Tasman W, Vander JF, et al. Diode laser photocoagulation for retinopathy of prematurity. Preliminary results [see comments]. Arch Ophthalmol 1992;110:1714 – 6. 3. McNamara JA, Tasman W, Brown GC, et al. Laser photocoagulation for stage 3⫹ retinopathy of prematurity. Ophthalmology 1991;98:576 – 80. 4. Landers MBD, Semple HC, Ruben JB, et al. Argon laser photocoagulation for advanced retinopathy of prematurity. Am J Ophthalmol 1990;110:429 –31. 5. Laser therapy for retinopathy of prematurity. Laser ROP Study Group [letter]. Arch Ophthalmol 1994;112:154 – 6. 6. Knight-Nanan DM, O’Keefe M. Refractive outcome in eyes with retinopathy of prematurity treated with cryotherapy or diode laser: 3 year follow up. Br J Ophthalmol 1996;80:998 – 1001. 7. Connolly BP, McNamara JA, Sharma S, et al. A comparison of laser photocoagulation with trans-scleral cryotherapy in the treatment of threshold retinopathy of prematurity. Ophthalmology 1998;105:1628 –31. 8. Knapp RG, Miller MC. Clinical epidemiology and biostatistics. Baltimore: Williams & Wilkins, 1992. 9. White JE, Repka MX. Randomized comparison of diode laser photocoagulation versus cryotherapy for threshold retinopathy of prematurity: 3-year outcome [see comments]. J Pediatr Ophthalmol Strabismus 1997;34:83–7; quiz 121–2. 10. Paysse EA, Lindsey JL, Coats DK, et al. Therapeutic outcomes of cryotherapy versus transpupillary diode laser photocoagulation for threshold retinopathy of prematurity. J AAPOS 1999;3:234 – 40. 11. Pearce IA, Pennie FC, Gannon LM, et al. Three year visual outcome for treated stage 3 retinopathy of prematurity: cryotherapy versus laser. Br J Ophthalmol 1998;82:1254 –9. 12. Vrabec TR, McNamara JA, Eagle RC Jr, et al. Cryotherapy for retinopathy of prematurity: A histopathologic comparison of a treated and untreated eye. Ophthalmic Surg 1994;25:38 – 41. 13. Wallow IH, Sponsel WE, Stevens TS. Clinicopathologic correlation of diode laser burns in monkeys [see comments]. Arch Ophthalmol 1991;109:648 –53. 14. Smiddy WE, Hernandez E. Histopathologic results of retinal diode laser photocoagulation in rabbit eyes. Arch Ophthalmol 1992;110:693– 8.
tance in the treatment of ROP.2– 6 The patients in our series took part in a previously reported clinical trial from which we concluded that laser was at least as good as cryotherapy.2,3,5 Each of these patients had undergone treatment for threshold ROP in one or both eyes. The treatment modality used for each eye was determined randomly; however, for
Originally received: January 17, 2001. Accepted: July 26, 2001.
None of the authors has any financial or proprietary interest in any of the techniques or equipment discussed in this article.
Manuscript no. 210036.
1
University College of Dublin, National University of Ireland, Dublin, Ireland.
2
Wills Eye Hospital Retina Service, Philadelphia, Pennsylvania. Supported by an unrestricted grant from Research to Prevent Blindness, New York, New York, and by a grant from Iridex Corporation, Mountain View, California.
928
© 2002 by the American Academy of Ophthalmology Published by Elsevier Science Inc.
Reprint requests to William S. Tasman, MD; Ophthalmologist-in-Chief; Wills Eye Hospital; 900 Walnut Street; Philadelphia, PA 19107. Presented in part at the American Academy of Ophthalmology annual meeting, New Orleans, Louisiana, November 2001.
ISSN 0161-6420/02/$–see front matter PII S0161-6420(01)01017-X
Ng et al 䡠 Comparison of Laser and Cryotherapy for ROP patients with bilateral disease, the same modality was never used in both eyes (one eye received laser treatment, and the fellow eye received cryotherapy). At 5-year follow-up, we reported that laser-treated eyes had better visual acuity and less myopia compared with cryotherapy-treated eyes.7 Nonetheless, the previous report had several limitations that were addressed in this study. First, in the earlier study, visual acuities were measured by a variety of methods and observers; whereas in this study, the visual acuities were all measured under standardized Early Treatment of Diabetic Retinopathy Study (ETDRS) conditions by the same observer. Second, in the previous report, the presence of dragging was based on the subjective opinions of multiple examiners; whereas in this study photographic and clinical grading were performed at one center. Finally, this study looked at why the visual acuity was reduced in the cryotherapy-treated eyes, which was a question that our earlier reports did not answer.
unilateral ROP, the randomization center again would instruct the clinician which mode of treatment to administer to the eye selected by a coin toss. The alternate modality was then administered to the fellow eye. Once randomization and treatment were completed, no attempt was made to conceal which treatment had been performed in a given eye, because there was typically marked eyelid swelling in the cryotherapy-treated eyes.
Treatment Administration Cryotherapy and/or laser photocoagulation was administered within 72 hours of the diagnosis of threshold ROP (as described elsewhere).2,3 The first 29 laser treatments performed before October 1990 were performed in an operating room using a liquid cooled argon ion laser (Coherent Medical Group, Palo Alto, CA). In October 1990, the infrared diode laser (IRIS Medical Instruments Inc., Mountain View, CA) became available and was used in the remaining treatments.
Follow-up Examination
Materials and Methods The clinical objective of this study was to evaluate the anatomic and visual results of these patients in the two treatment groups (cryotherapy and laser) 10 years after treatment. The relationship of visual acuity to anatomic outcome was also examined.
Threshold ROP All infants who developed threshold ROP during the study period at one institution were eligible for this study (i.e., five or more contiguous or eight cumulative clock hours of stage 3 ROP in zone I or II in the presence of “plus” disease).1 In accordance with our institutional review board, consent was obtained from either the parents of the children or their legal guardians before randomization and treatment. Patients whose parents or legal guardians did not wish to enroll their children in the study were not included. Sixty-six infants were entered in a prospective randomized clinical trial. This sample size was chosen empirically. Of these, 39 infants had symmetric disease at the onset of disease and underwent bilateral treatment. Another 27 patients had unilateral threshold disease, 15 (56%) of whom were allocated to laser treatment and 12 (44%) of whom were allocated to cryotherapy in the first eye. Of the 15 patients treated initially with laser in one eye, 3 patients (20%) progressed to threshold ROP in their fellow eye and were treated with cryotherapy. In contrast, 10 of 12 (83%) patients whose initial eye was treated with cryotherapy progressed to threshold disease and required laser treatment in the fellow eye.
Randomization At the outset of the study, a stack of opaque envelopes was prepared. Half of the envelopes contained a piece of paper with the word “cryotherapy” printed on it, and the remaining half stated “laser.” These envelopes were shuffled and stored at the randomization center. When a patient was diagnosed with unilateral threshold ROP, the randomization center was instructed to open the envelope at the top of the stack, and the treatment modality printed on the piece of paper was administered. If the fellow eye subsequently developed threshold ROP, the alternative modality was then administered. In cases with bilateral threshold disease, a coin was tossed to choose one eye for randomization. As with
In June 2000, we invited all patients we could contact to return for a follow-up eye examination. Twenty-five patients agreed to attend. Two patients were deceased, two patients had relocated to a distant location, 11 patients refused examination, and 26 were lost to follow-up. Figure 1 shows a detailed flow chart of all the patients from this study. Of the 25 patients who returned for follow-up, 19 had undergone bilateral treatment, and six had undergone unilateral treatment. Detailed examinations were performed on a total of 23 laser-treated eyes and 21 cryotherapy-treated eyes. During the visit, a history was taken with attention to prior amblyopia therapy. Hirschberg tests and/or cross-cover tests were performed to detect ocular misalignment. Best-corrected visual acuity (BCVA) and uncorrected visual acuity were measured with a backlit ETDRS acuity chart (Lighthouse Low Vision, Long Island, NY). After administration of tropicamide 1.0%, phenylephrine hydrochloride 2.5%, and cyclopentolate hydrochloride 1.0%, fundus photography (Kodachrome), slit-lamp examination, and indirect ophthalmoscopy were carried out. The structural outcome of the fundus was assessed clinically by indirect ophthalmoscopy. In accordance with the Multi-center Trial of Cryotherapy for ROP, macular fold, retinal detachment, retrolental mass, and phthisis bulbi were recorded as unfavorable outcomes. The amount of dragging present was assessed clinically as none, mild, moderate, or severe. In situations in which media clarity was unsatisfactory, through the lid ultrasonic B-scan (Innovative Imaging, Sacramento, CA) was performed. Finally, the degree of retinal dragging was assessed on the basis of fundus photographs. Three independent observers (BPC, JAM, WT) graded photographic slide(s) from each eye of each patient. These observers were masked to the modality of treatment in a given eye. Only photographs showing both optic disc and macula were used. Areas showing treatment lesions on four photographs (from three patients) were covered to prevent the reader from recognizing the treatment modality. The slides were shuffled before being graded. The degree of dragging present in each photograph was categorized as follows: none, mild, moderate, severe. Figures 2 through 4 are representative of the mild, moderate, and severe categories. There was 100% agreement among the graders for the photographs seen in Figures 2 through 4. Figure 2 shows the mildest dragging among our fundus photographs and was classified as mild dragging by all observers. Figure 3 shows the minimum amount of dragging necessary to be classified as moderate.
929
Ophthalmology Volume 109, Number 5, May 2002
Figure 1. Detailed chart summarizing patient flow and completeness of follow-up. Pt. ⫽ patient, pts. ⫽ patients.
Figure 4 shows severe dragging. Eyes with retinal detachments were not included in this analysis of photographic grading. Multiple views of photographs of the same eye served as built-in controls. A total of 45 slides were graded in a single session by each observer. Although the degree of dragging was a subjective measure, there was good interobserver agreement. In four slides, one of the three observers graded the slides as normal, whereas the other two graded them as mild dragging. Two other slides were graded as moderate dragging by two observers, but the other graded them as severe dragging. In these cases, the slides were reported according to the observation of the two agreeing observers. Hence there were interobserver differences in 6 of the 45 slides (13%). Within each observer, only 1 out of 10 built-in controls differed in grading.
Data Analysis and Recording Patient data and results obtained were entered into a Microsoft Access 2000 database, and statistical analysis was performed using Microsoft Excel 2000. Mean age, birth weight, and gestational age were calculated. Gender and racial distribution of these patients were treated as binomials and are represented as percentages. To ensure uniform baseline characteristics of disease at treatment, the mean number of clock hours of stage 3 disease and the percentage of eyes with zone 1 disease at the time of treatment were calculated for both the laser-treated eyes and the cryotherapy-treated eyes of those patients who returned for follow-up examination. Unpaired twotailed Student’s t tests were performed to determine whether the number of clock-hours of stage 3 disease differed between the treatment groups. Chi-square analysis was used to ensure uniformity of the zone of disease between the two groups at the time of treatment. Both paired and unpaired two-tailed Student’s t tests were administered to compare the two treatment groups’ ETDRS BCVA scores. Eyes with visual acuities of finger counting (less than one ETDRS visual acuity score), hand motion vision, and no light perception were assigned an ETDRS visual acuity score of zero.
930
An unpaired test was administered to analyze all treated eyes. In addition, a paired two-tailed Student’s t test was administered to analyze only patients who had received bilateral treatment. In keeping with our previous report,7 a two-by-two contingency table was constructed to determine the odds of laser treatment resulting in visual acuity of 20/50 or better (ETDRS score ⱖ65) compared with cryotherapy. Because of cognitive limitations, patients 17 and 20 were not cooperative with ETDRS acuity testing and fundus photography. Data for patient 20 were not analyzed. Because one eye of patient 17 was unmistakably blind, it was accorded a visual acuity score of zero and used in the analysis (unpaired) of data. She was not cooperative with testing of the seeing eye. Patient 16 had moderate cognitive impairment; however, she was able to perform a Tumbling-E Test for visual acuity testing. The results were translated into approximate ETDRS units and then to ETDRS visual acuity scores for our calculations. Fundus photographs of patients 5 and 13 were unavailable, because the negatives were inadvertently destroyed while being processed. For statistical analysis, an ordinal value of 0 to 3 was assigned to the photographically assessed degree of dragging. Those eyes with no dragging were assigned a value of 0 points, mild dragging was assigned 1 point, moderate dragging 2 points, and severe dragging 3 points. Macular fold of retina was also accorded 3 points. Eyes with retinal detachments were not included in this analysis. The difference in the degree of dragging between eyes treated with the two different modalities was compared in the same manner as visual acuity. In addition, linear regression analysis and analysis of variance were performed to determine whether the BCVA of cryotherapytreated eyes was correlated to that of laser-treated eyes among bilaterally treated patients and whether visual outcome achieved was associated with the anatomic outcome of treatment. This requires inclusion of only eyes for which both BCVA and the degree of dragging were known. Both the correlation coefficient (r) and the P value were calculated. Where appropriate, axis intercepts of the linear trend line together with their P values were
Ng et al 䡠 Comparison of Laser and Cryotherapy for ROP
Figure Figure Figure Figure Figure Figure
2. 3. 4. 5. 6. 7.
Fundus photograph showing mild dragging. Fundus photograph showing moderate dragging. Fundus photograph showing severe dragging. Loss of cilia in a cryotherapy-treated eye. Fundus photograph, patient 12, shortly after argon laser photocoagulation, Fundus photograph, patient 12, at 10 year follow-up.
calculated and included for discussion. Statistical significance was defined as a P value ⱕ 0.05 for this report.
Results The mean age at follow-up was 9.9 years (range, 8.6 –11.1 years). Of the 25 patients, 11 (44%) were male and four (16%) were
African American. This cohort of patients had an average gestational age of 26 weeks (range, 23–33 weeks). Their mean birth weight was 807 g (range, 440 –1477 g). Cryotherapy-treated and laser-treated eyes had similar preoperative characteristics. The mean number of clock hours of stage 3 disease and percentage of eyes with zone 1 disease for each of the treatment groups are shown in Table 1.
931
Ophthalmology Volume 109, Number 5, May 2002 Table 1. Comparison of Baseline Characteristics (Clock Hours of Stage 3 and Zone) Between the Treatment Groups Laser-treated Eyes Mean number of clock hours stage 3 Standard error of the mean P value Percentage of zone 1 eyes P value for 2
Cryotherapy-treated Eyes
7.7
7.0
0.61
0.54 0.44
9% (6 of 64)
11% (6 of 54) 0.95
P values based on unpaired, two-tailed Student’s t test results. There were no significant differences between the two treatment groups in terms of these baseline characteristics.
Functional and Visual Outcome Overall, those treated with laser had a mean ETDRS visual acuity score of 59.3, corresponding to a Snellen acuity-equivalent of 20/66 (standard error of the mean, 5.8). The mean ETDRS visual acuity score of the cryotherapy-treated eyes was 37.5, corresponding to a Snellen acuity of 20/182 (standard error of the mean, 6.3). An unpaired two-tailed Student’s t test used to compare the two treatment groups showed a statistically significant difference (P ⫽ 0.0152, n ⫽ 42 eyes; see Table 2). A two-by-two contingency table was constructed to compare the visual acuity of the two treatment groups (see Table 3). An odds ratio was calculated, and a 95% confidence interval was
determined by Woolfe’s method.8 The laser-treated eyes were 5.20 times more likely to have 20/50 or better BCVA (95% confidence interval, 1.37–19.77, n ⫽ 42 eyes). When 18 patients who had undergone bilateral treatment (and who were capable of ETDRS vision testing) were analyzed separately by paired Student’s t test analysis, a significant difference between the treatment groups (P ⫽ 0.033, n ⫽ 18 pairs of eyes) was again evident. The laser-treated eyes had a mean ETDRS visual acuity score of 59.1 (standard error of the mean, 6.7) comparable to 20/66 Snellen acuity, whereas cryotherapy-treated eyes had a mean ETDRS visual acuity score of 41.4 (standard error of the mean, 6.8) corresponding to a 20/152 Snellen equivalent. Although it would have been interesting to stratify patients in each treatment group by initial ROP severity (i.e., zone, clock hours of disease), the sample size was not large enough to do so. Amblyopia therapy had been administered to 6 of the 21 binocular patients (29%). Of these six amblyopic eyes, four were treated with cryotherapy compared with two who were treated with laser. Five had mild dragging, whereas one had moderate dragging. In all six cases, the fundi of the fellow eyes were anatomically normal. Four of these six also had strabismus in the same eye.
Anatomic Outcome Of the 21 eyes treated with cryotherapy, four (19%) eyes had unfavorable outcomes (two eyes developed stage 5 ROP, one eye developed stage 4B ROP, and another eye developed a macular fold). There were 2 (10%) unfavorable outcomes among the 23 eyes treated with laser; both developed stage 5 ROP. Of 21 patients with favorable outcomes in both eyes, 13 (62%)
Table 2. Summary Table of Visual Acuities and Dragging For All Patients Laser-treated Eyes
Cryotherapy-treated Eyes
Patient No.
Best-corrected Visual Acuity
Snellen Visual Acuity
Dragging
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
88 83 78 75 73 73 73 70 69 67 67 62 57 51 1 76 ? 0 0 ? 76 61 45
20/16 20/20 20/25 20/25 20/32 20/32 20/32 20/32 20/40 20/40 20/40 20/50 20/63 20/80 20/800 20/25 CSM NLP NLP CSM 20/25 20/50 20/100
Normal Normal Normal Normal * Normal Normal Normal Normal Normal Normal Moderate dragging * Normal Moderate dragging Mild dragging * Stage 5 Stage 5 * Normal Moderate dragging Moderate dragging
Best-corrected Visual Acuity
Snellen Visual Acuity
Dragging
68 40 74 0 77 69 7 54 57 64 63 10 8 4 46 70 0 35 0
20/40 20/125 20/32 NLP 20/25 20/40 20/640 20/80 20/63 20/50 20/50 20/640 20/640 20/800 20/100 20/40 NLP 20/200 NLP
Mild dragging Mild dragging Normal Stage 4B * Normal Mild dragging Moderate dragging Mild dragging Mild dragging Mild dragging Moderate dragging * Retinal fold of macula Moderate dragging Normal Stage 5 Severe dragging Stage 5
36 5
20/160 20/800
Normal Mild dragging
Synopsis of patient’s visual acuity and fundus anatomy. Patients 20 thru 25 received only unilateral treatment. BCVA ⫽ best-corrected visual acuity; CSM ⫽ central steady and maintained; NLP ⫽ no light perception. * Favorable outcome but fundus photograph not available for systematic assessment of dragging.
932
Ng et al 䡠 Comparison of Laser and Cryotherapy for ROP Table 3. A Comparison of Visual Acuities between Lasertreated Eyes and Cryotherapy-treated Eyes
Table 4. A Comparison of Dragging between Laser-treated Eyes and Cryotherapy-treated Eyes
Treatment
20/50 or Better
20/60 or Worse
Total
Treatment
No Dragging
Any Dragging
Total
Laser Cryotherapy Total
13 5 18
8 16 24
21 21 42
Laser Cryotherapy Total
12 4 16
5 12 17
17 16 33
Patients 17 and 20 were cognitively impaired and unable to perform acuity testing. Patient 17 had a retinal detachment in the cryotherapy-treated eye, so that eye was clearly blind and was included in this table. Although the patient’s laser-treated eye was able to fixate steadily, it was excluded from this analysis because visual acuity could not be determined accurately. In addition, patient 20, who had unilateral treatment with laser, was unable to perform acuity testing because of cognitive impairment and was excluded from this analysis.
patients had strabismus. Ten of these 13 had nonaccommodative esotropia. One patient had an alternating esotropia, and two patients demonstrated intermittent exotropia. In addition, among eyes with a favorable outcome, one patient had a small (one clock hour in width) cortical cataract in the laser-treated eye. Furthermore, loss of cilia was observed in one eye treated with cryotherapy (Fig 5). Two other patients developed blepharoptosis in their cryotherapy-treated eyes. One of these eyes had a macular fold and an axial length of 20.1 mm, whereas the other eye had normal fundus anatomy and an axial length of 22.2 mm.
Fundus Photography Because of lack of cooperation or somewhat opaque ocular media, fundus photography was not possible in some patients, but we were able to obtain photographs in 16 cryotherapy-treated eyes and 17 laser-treated eyes. Of the 16 cryotherapy-treated eyes photographed, 7 had mild dragging, 3 had moderate dragging, two had severe dragging, and the remaining 4 were normal. In comparison, 12 of the 17 (71%) photographed laser-treated eyes were normal, 1 (6%) had mild dragging, 4 (24%) had moderate dragging, and none had severe dragging (see Table 2). Eyes with retinal detachments were not photographed. Of the 13 patients who received bilateral treatment (and for whom fundus photographs were taken), laser-treated eyes had significantly less dragging. The laser-treated eyes had a mean scaled score of 0.39, whereas cryotherapy-treated eyes were more likely to have dragging and had a mean score of 1.15 (P ⫽ 0.018, n ⫽ 13 pairs of eyes, paired Student’s t test). When all eyes for which photographs were available were included in this analysis (unpaired Student’s t test), the mean score in laser-treated eyes was 0.53 (n ⫽ 17 eyes) compared with 1.19 (n ⫽ 16 eyes) in cryotherapy-treated eyes (P ⫽ 0.05, n ⫽ 33 eyes). Because the unpaired Student’s t test is a less powerful statistical analysis, more patients would be needed to detect a true difference of similar magnitude compared with the more powerful paired test. Nonetheless, eyes treated with cryotherapy were 7.2 times (95% confidence interval, 1.54 –33.6, n ⫽ 33) more likely to develop dragging compared with fellow laser-treated eyes (see Table 4). This odds ratio was calculated by Woolfe’s method.8 Linear regression analysis revealed the following findings. First, when comparing the BCVA of one eye with the fellow eye, a poor correlation coefficient was obtained (r ⫽ 0.36, P ⫽ 0.14, n ⫽ 18). Because this implies that the BCVA of one eye is poorly
Eyes with mild, moderate, and severe dragging were included in the “any dragging” category for this table.
correlated and not predictive of the BCVA of the fellow eye, it further strengthens our observation of a real difference in BCVA between the two different treatment modalities. Second, the absence of dragging was a predictor of good visual acuity in both laser-treated and cryotherapy-treated eyes. There was a strong negative correlation between the BCVA and the degree of dragging in laser (r ⫽ ⫺0.64, P ⫽ 0.0059, n ⫽ 17) and cryotherapy eyes (r ⫽ ⫺0.52, P ⫽ 0.04, n ⫽ 16). Third, the intercept values obtained from line fit plot statistics suggest a predicted BCVA score of 73 (Snellen equivalent 20/32) in the absence of retinal dragging for laser-treated eyes and a predicted BCVA score of only 60 (Snellen equivalent 20/50) in the absence of retinal dragging for cryotherapy-treated eyes. This raises the possibility that laser treatment may allow for better BCVA potential than cryotherapy even when minimal dragging is present. Because fundus photographs of patient 12 were taken in the operating room at the time of argon laser treatment, we were able to monitor the progression of retinal dragging. Figure 6 is a fundus photograph taken soon after treatment, and Figure 7 was obtained during this follow-up 10 years later. It was interesting to note that the dragging had progressed in both eyes. Unfortunately, we only have postoperative photographs of one of these eyes for comparison.
Discussion After roughly a decade, many of the children who returned for follow-up evaluations were cognitively capable of cooperating with our study. Our results show that visual acuity is significantly better in those eyes treated with laser photocoagulation into late childhood. This significant difference persisted despite administration of amblyopia therapy (where appropriate) in these patients. The probability that chance alone accounted for the difference in visual acuity between the two treatment groups was approximately 3%. The observation that visual acuity was better in cryotherapy-treated eyes was consistent with our earlier report.7 Another randomized prospective clinical trial comparing cryotherapy to laser photocoagulation for threshold ROP was published by White and Repka9 They also observed that visual acuity was somewhat better in the laser-treated eyes; however, this trend was not statistically significant because of their relatively small sample size. In addition, two nonconcurrent retrospective comparisons of cryotherapy and laser photocoagulation by Paysse et al10 and Pearce et al11 demonstrated that the visual outcomes were superior in the laser-treated eyes.
933
Ophthalmology Volume 109, Number 5, May 2002 The underlying reason for the difference in the long-term visual acuities after these two different treatments is not clear. In this study, cryotherapy-treated eyes were shown to be more likely to develop dragging, and linear regression analysis indicated that dragging was associated with poorer visual acuities (not surprisingly). This would suggest that inferior macular structural outcomes were an important factor in the cryotherapy-treated eyes. Nonetheless, it is difficult to explain why dragging was more prevalent in the cryotherapy-treated eyes. Although both laser treatment and cryotherapy are ablative treatments, the type of tissue damage differs. Transscleral cryotherapy to the developing sclera causes retinal, choroidal, and scleral disruption and may perpetuate disorganized scar tissue formation.12 The photocoagulation burns produced by a transpupillary diode laser (810 nm) produce retinal and choroidal changes. Compared with comparable argon laser spots, diode laser burns show more of an effect in the choroid.13 In a rabbit model, Smiddy and Hernandez14 reported that the effect of laser burns on the sclera tends to be fairly minimal, except when high energy levels were used to treat atrophic areas. When present in the area of a photocoagulation burn, ciliary nerves in the choroid or sclera consistently showed scarring in a cynomolgus monkey model.13 Perhaps the more widespread nature of the thermal injury that occurs with cryotherapy accounts for part of the additional dragging seen in these patients. When linear regression analysis was used to predict the mean BCVA in the absence of any dragging, cryotherapytreated eyes seemed to have reduced potential visual acuity even in the absence of dragging compared with laser-treated eyes. Thus, dragging alone did not seem to account for all of the reduction in visual acuity in the cryotherapy-treated eyes. The cryotherapy-treated eye of patient 24 was fairly amblyopic with an ETDRS score of 36 (Snellen equivalent 20/160), and this may have accounted for some of this discrepancy. Nonetheless, when this patient was excluded from the linear regression analysis, the intercept was 66 (Snellen equivalent 20/40)—still somewhat less than that for laser-treated eyes. Although this study detected a difference in the amount of dragging and a possible difference in the average visual potential between the two treatment modalities, even in the absence of dragging, it was not designed to determine the underlying reasons for these discrepancies. Late-appearing or persistent treatment complications were relatively rare. Those that were apparently related to cryotherapy treatment included a juxtamacular chorioretinal scar, loss of cilia, and blepharoptosis. In addition, a cortical cataract was seen in a laser-treated eye. Some limitations of this study included a relatively small number of patients and a recall rate of only 38%. Furthermore, a number of patients examined were cognitively limited, which led to variable participation in the different tests performed. Nonetheless, these patients are a crosssection of the premature patients who are affected by this condition. Additional amblyopia therapy would be unlikely to benefit these children; thus, it would be unlikely to find
934
any significant new trends in visual acuity after a 10-year follow-up. Despite some limitations, the randomized design of this original study controls for many potential confounding variables. The 10-year follow-up results of this study suggested that laser photocoagulation used in the treatment of ROP resulted in better visual and anatomic outcome and was associated with less long-term morbidity than cryotherapy treatment. Although this trial was smaller than the Multicenter Trial of Cryotherapy for ROP, it is unlikely that another trial of similar scale will be undertaken to compare cryotherapy with laser photocoagulation. Acknowledgment. The authors thank Andrew Smith, PhD, for his extensive statistical assistance with this article.
References 1. Multicenter trial of cryotherapy for retinopathy of prematurity. Preliminary results. Cryotherapy for Retinopathy of Prematurity Cooperative Group. Arch Ophthalmol 1988;106:471–9. 2. McNamara JA, Tasman W, Vander JF, et al. Diode laser photocoagulation for retinopathy of prematurity. Preliminary results [see comments]. Arch Ophthalmol 1992;110:1714 – 6. 3. McNamara JA, Tasman W, Brown GC, et al. Laser photocoagulation for stage 3⫹ retinopathy of prematurity. Ophthalmology 1991;98:576 – 80. 4. Landers MBD, Semple HC, Ruben JB, et al. Argon laser photocoagulation for advanced retinopathy of prematurity. Am J Ophthalmol 1990;110:429 –31. 5. Laser therapy for retinopathy of prematurity. Laser ROP Study Group [letter]. Arch Ophthalmol 1994;112:154 – 6. 6. Knight-Nanan DM, O’Keefe M. Refractive outcome in eyes with retinopathy of prematurity treated with cryotherapy or diode laser: 3 year follow up. Br J Ophthalmol 1996;80:998 – 1001. 7. Connolly BP, McNamara JA, Sharma S, et al. A comparison of laser photocoagulation with trans-scleral cryotherapy in the treatment of threshold retinopathy of prematurity. Ophthalmology 1998;105:1628 –31. 8. Knapp RG, Miller MC. Clinical epidemiology and biostatistics. Baltimore: Williams & Wilkins, 1992. 9. White JE, Repka MX. Randomized comparison of diode laser photocoagulation versus cryotherapy for threshold retinopathy of prematurity: 3-year outcome [see comments]. J Pediatr Ophthalmol Strabismus 1997;34:83–7; quiz 121–2. 10. Paysse EA, Lindsey JL, Coats DK, et al. Therapeutic outcomes of cryotherapy versus transpupillary diode laser photocoagulation for threshold retinopathy of prematurity. J AAPOS 1999;3:234 – 40. 11. Pearce IA, Pennie FC, Gannon LM, et al. Three year visual outcome for treated stage 3 retinopathy of prematurity: cryotherapy versus laser. Br J Ophthalmol 1998;82:1254 –9. 12. Vrabec TR, McNamara JA, Eagle RC Jr, et al. Cryotherapy for retinopathy of prematurity: A histopathologic comparison of a treated and untreated eye. Ophthalmic Surg 1994;25:38 – 41. 13. Wallow IH, Sponsel WE, Stevens TS. Clinicopathologic correlation of diode laser burns in monkeys [see comments]. Arch Ophthalmol 1991;109:648 –53. 14. Smiddy WE, Hernandez E. Histopathologic results of retinal diode laser photocoagulation in rabbit eyes. Arch Ophthalmol 1992;110:693– 8.