- Email: [email protected]
A randomized trial of rigid gas permeable contact lenses to reduce progression of children’s myopia
A Randomized Trial of Rigid Gas Permeable Contact Lenses to Reduce Progression of Children’s Myopia JOANNE KATZ, SCD, OLIVER D. SCHEIN, MD, MPH, BRIAN LEVY, OD, MSC, TOM CRUISCULLO, BS, SEANG-MEI SAW, MBBS, MPH, PHD, UMA RAJAN, MBBS, MPH, FAMS, TAT-KEONG CHAN, FRCOPH, FRCS, CHONG YEW KHOO, FRCSOPH, FRACS, FAMS, AND SEK-JIN CHEW, FRCOPH, FAMS, PHD
● PURPOSE: To test whether rigid gas permeable (RGP) contact lens wear can reduced the rate of myopia progression in school age children. ● DESIGN: Randomized clinical trial. ● METHODS:
Setting: Single clinical center. Study Population: Both eyes of 428 Singaporean children. Inclusion Criteria: 6 through 12 years of age with myopia between ⴚ1 and ⴚ4 diopters, astigmatism < 2 diopters, no prior contact lens wear, no other ocular pathologies. Intervention: Spectacle or RGP lens correction for myopia. After a 3-month adaptation period, 383 children were followed, and 298 (78%) remained after 24 months. Outcome measures: Cycloplegic subjective refraction, keratometry, and axial length measured at 12 and 24 months. ● RESULTS: Children who adapted to contact lenses wore them for a median of 7 hours per day, but no more than 40% wore them at least 8 hours per day, 7 days per week. Spectacles were worn for a median of 15 hours per day at the time of the 24-month follow-up. There was an Accepted for publication Jan 9, 2003. InternetAdvance publication at ajo.com Feb 13, 2003. From the Department of International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland (J.K.); Dana Center for Preventive Ophthalmology, Wilmer Eye Institute, Johns Hopkins School of Medicine, Baltimore, Maryland (J.K., O.D.S.); Bausch and Lomb, Rochester, New York (B.L., T.C.); Department of Community, Occupational and Family Medicine, National University of Singapore, Singapore (S.M.S.); School Health Services, Singapore (U.R.), and Singapore Eye Research Institute, Singapore (T.K.C., C.Y.K., S.J.C.). The study was conducted at the Singapore Eye Research Institute, Singapore. Inquiries to Joanne Katz, ScD, the Johns Hopkins Bloomberg School of Public Health, 615 N. Wolfe St., Room W5009, Baltimore, MD 212052103; fax: (410) 955-2029; e-mail: [email protected]
82
©
2003 BY
increase in the spherical equivalent of ⴚ1.33 and ⴚ1.28 diopters (P ⴝ .64), and axial length increased by 0.84 and 0.79 mm (P ⴝ .38) over 2 years among children randomized to contact lenses and spectacles, respectively. Adjustment for baseline differences between the groups and for hours per day of contact lens wear did not alter these findings. ● CONCLUSIONS: Rigid gas permeable lenses did not slow the rate of myopia progression, even among children who used them regularly and consistently. It is unlikely that this intervention holds promise as a method by which to slow the rate of progression of myopia in children. (Am J Ophthalmol 2003;136:82–90. © 2003 by Elsevier Inc. All rights reserved.)
M
YOPIA IS A MAJOR CONCERN FOR ETHNIC CHI-
nese among whom the prevalence is very high and appears to have increased rapidly over a few generations.1–9 The economic burden of myopia is great, with expenditures on spectacles and contact lenses estimated at $ 2 billion annually in the United States alone.10 The use of refractive surgery to correct myopia has also increased considerably ever since lasers have been approved for this purpose.11,12 Severe myopia has also been increasing in ethnic Chinese,2 and this has led to a concern about its ophthalmic consequences such as myopic macular degeneration, retinal holes, tears and detachments, lattice degeneration, cataract, and glaucoma.13–15 It has been postulated that rigid contact lenses could slow the progression of myopia in children.16 –18 Several potential mechanisms have been offered, including improved retinal image quality with contact lens wear; flattening of the cornea, producing a temporary slowing of myopia progression; or an overcorrection for myopia when fitting contact lenses. Several studies have examined this issue, but each has had methodological problems. Stone19 enrolled 124 unmatched, nonrandomized spectacle wearers
ELSEVIER INC. ALL
RIGHTS RESERVED.
0002-9394/03/$30.00 doi:10.1016/S0002-9394(03)00106-5
and 244 contact lens wearers, but only 28 controls and 22 contacts lens patients remained after 5 years of follow-up. Kelly and co-workers20 did not randomize their patients, and they did not provide information about baseline comparability. Grosvenor and colleagues enrolled 100 contact lens wearers and 39 nonrandomized controls. The follow-up was 56% among contact lens wearers and 51% among controls after 3 years.21,22 These studies all showed higher rates of myopia progression among controls than among contact lens wearers, only some of which could be explained by flattening of the cornea due to the contact lens wear. Andreo23 retrospectively analyzed 37 contact lens wearers and 19 spectacle wearers followed for 1 year and found no difference in myopia progression between the two groups. These studies were done in Caucasian populations, where the prevalence and rate of myopia progression was lower than in ethnic Chinese populations. A more recent controlled, but not randomized, study undertaken among school children in Singapore found that there was a slower rate of progression among contact lens wearers compared with spectacle wearers.24 This study also concluded that the difference in progression could not be explained by changes in corneal curvature. However, the study enrolled 105 children in the contact lens group, was not randomized, and the dropout rate was 47% among contact lens wearers over a 3-year period. There were also differences in the baseline severity of myopia between the spectacle and contact lens groups. For this reason, we decided to undertake a larger, randomized trial of the value of rigid gas permeable lenses in reducing progression of myopia in Singaporean school-age children. A prospective, randomized trial with a similar aim but slightly different design is currently underway in the United States.25 The Contact Lens and Myopia Study (CLAMP) randomizes pre-adolescent children to rigid gas permeable or soft contact lenses after an adaptation period and follows them for 3 years. However, the results of this trial are not yet available.
DESIGN THE STUDY WAS A CLINICAL TRIAL, WHERE CHILDREN
were randomized to use either rigid gas permeable contact lenses or spectacle correction for treatment of myopia. Patients were randomized prior to adaptation to contact lens wear.
METHODS ● SETTING:
Children were enrolled and all examinations were conducted at the Myopia Clinic of the Singapore Eye Research Institute. VOL. 136, NO. 1
CONTACT LENSES
● POPULATION: The population for this study was selected because of the high rates of myopia progression observed in Singaporean children.3,9 Participants were recruited through the Refraction Clinics of the School Health Services and through newspaper advertisements, and any child who met the eligibility criteria could participate in the trial. Eligible children were those of Chinese ethnicity, between the ages of 6 and 12, with myopia between ⫺1.0 and ⫺4.0 diopters of sphere, astigmatism of ⱕ2 diopters, no prior history of contact lens wear, who did not have any other ocular or medical pathologies, and who provided informed consent.26 –29 Eligible patients were randomized to one of the two treatment arms using a randomization schedule of block size 6, generated from random number tables in Baltimore and placed in sealed envelopes with sequential patient numbers. Randomization allocation was checked several times during the study against the master list and patient identifiers. Neither patients nor clinical observers were masked to treatment group. The projected sample size for randomization was 150 in each group and was based on a mean change in refractive error of 0.7 diopters per year in the spectacle group, 80% power to detect a 0.2-diopter lower rate of progression per year in the contact lens group, a type I error of 5%, and a dropout rate of approximately 30% in the contact lens group over 2 years of follow-up. Because it was likely that there would be more dropouts among contact lens than among spectacle wearers during the adaptation period, a higher number of patients were recruited to ensure at least 150 children in the contact lens group following adaptation and 100 remaining in the trial at 24 months. The study conformed to the Helsinki guidelines. The trial was approved by the Ethics Committee of the Singapore National Eye Center, and by Bausch and Lomb, which sponsored the trial and provided the contact lenses. After explaining the study to parents and children, written informed consent was obtained from parents and verbal assent from children. ● INTERVENTION: Children were randomized to receive rigid gas permeable contact lenses (Asian Design Lens, Bausch and Lomb, Rochester, New York, USA) or spectacles to correct their refractive error. The contact lens consisted of a front and back junctionless aspheric design with an overall diameter of 9.2 mm. Base curves available for the lenses ranged from 7.3 mm to 8.4 mm in 0.1-mm steps. The lenses were fit to achieve central alignment with light mid peripheral bearing. Apical bearing seen with fluoroscein evaluation was not allowed. The lens was fit to move smoothly 1 to 2 mm after each blink and centered over the cornea with full pupil capture. An automated over-refraction was performed with trial lenses, a full cycloplegic refraction prescription minus 0.5 diopters was provided, and the visual acuity was verified with the trial lenses. Children assigned to spectacle correction were AND
MYOPIA
83
refracted (cyclopleged) by the study personnel and received new spectacles if their current prescription did not conform to the refractive error measured in the study.
axial length were compared using t tests, and the proportion of eyes with changes in these variables above certain cutoff points was compared between treatment groups using chi-square tests. The analysis was done for right and left eyes separately with comparable results. The analysis presented here only uses data from right eyes. A multivariate linear model was used to estimate the difference in the change in refractive errors from baseline through 24 months between the two treatment groups, adjusted for characteristics that were different at baseline.
● MAIN OUTCOME MEASURES:
The principal outcome was the change in refractive error measured by subjective cycloplegic refraction from postadaptation through 2 years of follow-up. Secondary outcomes were the change in keratometry (autokeratometry) and total axial length (AScan ultrasonography) over the same time period. Refractive error was obtained using both automated refraction (Nidek ARK 900, Nidek Co., Tokyo, Japan) and subjective refraction (using autorefraction as the starting point for refining the refraction) 30 minutes after the third drop of 1% cyclogyl ophthalmic solution was instilled. Total axial length was measured by placing the probe in the center of the patient’s cornea without applanation. Additional readings were taken until the SD was 0.12 mm or below. Children were dispensed contact lenses or spectacles based on the refraction obtained during the preadaptation examination. A 3-month period was provided to allow children to adapt to wearing the contact lenses or spectacles before the start of the trial. Children returned for an examination every 3 months to undergo a slit-lamp examination for detection of any adverse events, and the contact lens fit was evaluated at these visits. Children were asked to wear their contact lenses for 4 hours prior to the ocular examination. Cycloplegic refractions were obtained at 6-month intervals following the end of the adaptation period. Cycloplegic autorefraction and keratometry were performed within 2 hours of lens removal. Children could also come to the clinic at any time if they believed there were any adverse events or concerns. Spectacles and contact lenses were replaced annually for all participants. If the visual acuity at any of the three monthly visits fell below 20/40, then the study team replaced the contact lenses or spectacles. The total follow-up period was 24 months or five 6-month visits in which cycloplegic refractive errors were measured. Two measures of adherence were taken at each follow-up visit, one from children and one from parents. Each was asked how many hours per day the contact lenses or spectacles were worn and the number of days per week. The agreement between children and parents was almost 100%, so the parent data for compliance measures were selected for analysis. Adherence was defined as use of contact lenses (if assigned to this group) or spectacle use (if assigned to spectacle wear) for at least 8 hours per day, 7 days per week as reported at each scheduled visit. Statistical analysis included comparability of baseline characteristics in the two treatment groups, a comparison of those who completed the study and those who dropped out for various reasons, and a comparison of the main outcomes by treatment group. The mean changes in spherical equivalent refraction, keratometry, and total
84
AMERICAN JOURNAL
RESULTS AS SHOWN IN FIGURE 1, A TOTAL OF 564 ELIGIBLE PATIENTS
were randomized to contact lenses (281) or spectacles (283). Because of initial problems with the supply of contact lenses for the trial, a lag time existed between randomization and the initial preadaptation visit. Hence, 195 patients completed the initial visit in the contact lens group and 233 in the spectacle group (Figure 1). Of these participants, 158 in the contact lens group and 225 in the spectacle group completed the adaptation phase. The proportion of patients remaining in the trial declined with each successive visit. After 24 months, 105 patients remained in the contact lens group and 192 in the spectacle group. At the initial preadaptation visit, the age distribution, mean age (8.3 years), axial length (24.3 mm), astigmatism (⫺0.4 diopters), and prior spectacle wear (85%) were comparable, but mean keratometry was 44.1 mm and 43.6 mm in the contact and spectacle groups, respectively (P ⫽ .001). Mean subjective spherical equivalent was ⫺2.56 and ⫺2.39 diopters in the contact lens and spectacle groups (P ⫽ .02). The proportion of girls in the contact lens group was higher than in the spectacle group (56% vs 41%, P ⫽ .002). At the end of the adaptation phase (the baseline for change in refractive error over the subsequent 24 months), the mean age and age distribution, history of prior spectacle wear, axial length, and astigmatism were comparable in the two groups and similar to the initial visit values (Table 1). The keratometric readings remained slightly higher in the contact lens than in the spectacle group (P ⫽ .01), and the children assigned to contact lenses remained more myopic by 0.2 diopters than did those in the spectacle group (P ⫽ .02). The gender imbalance remained the same, with 55% and 41% girls in the contact lens and spectacle groups, respectively (P ⫽ .006). Among the 158 assigned to contact lenses at the start of the postadaptation phase, 53 patients did not complete the 24-month examination. Forty-eight patients withdrew from the study and 5 remained in the study but did not complete the 24-month examination and were considered “censored” from the study. Of the 48 patients who withdrew, 31 dropped out prior to 24 months owing to OF
OPHTHALMOLOGY
JULY 2003
FIGURE 1. Flow chart giving the number of eligible patients who were randomized and how many remained in the study at each successive visit.
contact-lens–related problems, accounting for 65% of the withdrawals in the contact lens group. Only one of these 31 patients dropped out because of positive slit-lamp findings. The remainder refused to wear the contact lenses
for a variety of reasons related to lifestyle and comfort. Girls assigned to contact lenses were more likely to complete the 24-month follow-up than were boys (P ⫽ .16), but the ages of those who did and did not complete
VOL. 136, NO. 1
AND
CONTACT LENSES
MYOPIA
85
TABLE 1. Demographic and Ocular Characteristics of Eligible Patients at the End of the Adaptation Period by Assignment Group Contact Lenses (n ⫽ 158) n
Age (years) 6–7 8–9 10–12 Mean (SD) Female Prior spectacle wear Ocular Measures
%
51 70 37 8.4 87 136 Mean
Subjective spherical equivalent (diopters) Keratometry (mm) Axial length (mm) Astigmatism (diopters)
Spectacles (n ⫽ 225) n
32.3 44.3 23.4 1.5 55.1 86.1 SD
%
90 81 54 8.3 92 193 Mean
40.0 36.0 24.0 1.6 40.9 85.8 SD
P P P P
⫽ ⫽ ⫽ ⫽
.21 .26 .006 .93
⫺2.84
0.83
⫺2.63
0.85
P ⫽ .02
43.91 24.39 ⫺0.41
1.39 0.79 0.40
43.56 24.46 ⫺0.43
1.35 0.79 0.38
P ⫽ .01 P ⫽ .40 P ⫽ .69
SD ⫽ standard deviation.
TABLE 2. Difference Between Those Followed From the Start of the Postadaptation Period Through 24 Months and Those Who Dropped Out Prior to 24 Months of Follow-up Contact Lenses Dropped out (n ⫽ 53)
Age (years) 6–7 8–9 10–12 Mean (SD) Female Prior spectacle wear Ocular Measures Spherical equivalent (diopters) Keratometry (mm) Axial length (mm) Astigmatism (diopters)
Spectacles
Followed (n ⫽ 105)
Dropped out (n ⫽ 33)
Followed (n ⫽ 192)
n
%
n
%
n
%
n
%
15 28 10 8.5 25 42 Mean
28.3 52.3 18.9 1.5 47.2 79.3 (SD)
36 42 27 8.4 62 94 Mean
34.3 40.0 25.7 1.6 59.1 89.5 (SD)
13 12 8 8.4 12 25 Mean
39.4 36.4 24.2 1.7 36.4 75.8 (SD)
77 69 37 8.3 80 168 Mean
40.0 35.9 19.3 1.6 41.7 87.5 (SD)
⫺2.63 43.8 24.3 ⫺0.36
0.83 1.4 0.8 0.37
⫺2.95 44.0 24.4 ⫺0.43
0.81 1.4 0.8 0.42
⫺2.58 43.2 24.7 ⫺0.35
0.81 1.3 0.8 0.36
⫺2.64 43.6 24.4 ⫺0.44
0.87 1.4 0.8 0.39
SD ⫽ standard deviation.
the study were similar (Table 2). Those in the contact lens group who completed the study had 0.3 diopters more myopia at baseline than those who did not (P ⫽ .02), but astigmatism, keratometry, and axial length were similar. In the spectacle group, 28 patients withdrew prior to the 24-month visit, and 5 were censored at 24 months. Lack of motivation, refusal to accept the spectacle assignment, and patients who moved were the majority of reasons for dropout (79%). Among patients assigned to spectacles, there were no statistically significant demographic or ocular differences between those who did and those who did not complete the study. 86
AMERICAN JOURNAL
Among those who completed the study, the ages and prior spectacle wear of those in the contact lens and spectacle groups were similar (Table 2), but a higher proportion of girls in the contact lens (59%) than spectacle group (42%) completed the study (P ⫽ .004). Axial length and astigmatism were similar between the treatment groups that completed the study, but the contact lens group that completed the study had 0.3 diopters more myopia at baseline than did those in the spectacle group that completed the study (P ⫽ .003). The mean baseline keratometry was 44.0 mm and 43.6 mm in the contact lens and spectacle groups that completed the study (P ⫽ .04). OF
OPHTHALMOLOGY
JULY 2003
TABLE 3. Changes in Refraction and Biometry at 1 and 2 Years of Follow-up by Assignment Group (Change Equals Follow-up Minus Immediate Postadaptation Measurements) Contact Lenses
Change in subjective spherical equivalent (diopters) 0 to 12 months 0 to 24 months Change in keratometry (mm) 0 to 12 months 0 to 24 months Change in axial length (mm) 0 to 12 months 0 to 24 months
Spectacles
n
Mean
SD
n
Mean
SD
120 97
⫺0.65 ⫺1.33
0.55 0.84
186 188
⫺0.63 ⫺1.28
0.49 0.78
P ⫽ .74 P ⫽ .64
120 97
⫺0.08 ⫺0.13
0.33 0.33
183 185
⫺0.002 ⫺0.07
0.20 0.33
P ⫽ .006 P ⫽ .20
118 97
0.35 0.84
0.41 0.47
183 184
0.33 0.79
0.40 0.45
P ⫽ .65 P ⫽ .38
SD ⫽ standard deviation.
The median hours per day that contact lenses were worn if they were being used daily varied from 5.7 hours at the 18-month visit to 7.1 hours at the 6- and 12-month visits. In contrast, spectacles were worn for a median of 12 hours per day at the 6-month visit and 15 hours per day by the 24-month visit. Using the definition of “adherence” as wearing either contact lenses or spectacles for at least 8 hours per day, 7 days per week, 39.8% and 91.1% of children were adherent at the 6-month visit in the contact lens and spectacle groups, respectively. The adherence with contact lens wear declined over time, reaching 31.5% at the 24-month visit, while adherence with spectacles increased to 98.4%. The increase in spherical equivalent for subjective refraction was approximately ⫺0.6 diopters per year for subjective cycloplegic refraction (Table 3). No difference existed between the two treatment groups with respect to the change in refraction from the start of the trial through 1 and 2 years of follow-up. Similarly, no difference was seen in the distribution of change in refractive error over 24 months (P ⫽ .35). Twenty-four percent of eyes in the contact lens group had an increase of 2 or more diopters, whereas 16% of eyes in the spectacle group increased by that amount or more over 2 years (P ⫽ .11). A similar increase occurred in total axial length in both the spectacle and contact lens groups over the course of the study, as would be expected with a similar increase in severity of myopia over the same time period (Table 3). A flattening of the cornea occurred, as measured by keratometry, in the contact-lens–wearing group, but the difference between the spectacle and contact lens groups was statistically significant only at 12 months of follow-up. Among children who wore their contact lenses for 12 or more hours per day at the 24-month visit, no difference existed in the change in refractive error, compared with all children in the spectacle group or among children who wore their spectacles for a similar amount of time per day
as the contact lens wearers (Table 4). No evidence was seen of decreasing myopia progression with increasing contact lens wear. Among children who wore contact lenses for 12 or more hours per day, there was a difference in the keratometry between the contact lens and spectacle groups, but no difference in total axial length. A multivariate model to predict the change in spherical equivalent from baseline through 24 months was fit with treatment group, age, gender, subjective refractive error, and keratometry at baseline as explanatory variables. These variables were selected because they were different at baseline between those in the spectacle and contact lens groups (subjective spherical equivalent and keratometry), or because they were predictive of myopia progression over 2 years of follow-up in bivariate models (progression of myopia decreased with increasing age and more progression among girls than boys). The adjusted difference in the change in refractive error from baseline through 24 months between the two groups (⫺0.033 diopters, 95% confidence interval [CI] from ⫺0.204, 0.138) was slightly less than the unadjusted difference of ⫺0.047 diopters. None of these differences was statistically significantly different from zero. Predictors that remained statistically significant in the multivariate model were age and gender.
VOL. 136, NO. 1
AND
CONTACT LENSES
DISCUSSION BASED ON THIS RANDOMIZED 2-YEAR TRIAL, NO EVIDENCE
was found that rigid gas permeable (RGP) lenses reduced the progression of myopia in preadolescent school children in Singapore. Among the children with 12 or more hours per day of contact lens wear, there was a slight, but not statistically significant difference of 0.2 diopters between the rate of progression in the contact lenses compared to the spectacle group, but a substantial proportion of chilMYOPIA
87
TABLE 4. Change in Subjective Refraction and Biometry From Postadaptation Through 24 Months by Assignment Group and Degree of Adherence (hours per week) at the 24-month Visit Contact Lens (n ⫽ 89)
Subjective spherical equivalent (diopters) ⬍ 4 hr/day 4 to 7 hr/days 8 to 11 hr/day 12 or more hr Keratometry (mm) ⬍ 4 hr/day 4 to 7 hr/days 8 to 11 hr/day 12 or more hr Axial length (mm) ⬍ 4 hr/day 4 to 7 hr/days 8 to 11 hr/day 12 or more hr
Spectacles (n ⫽ 186)
n
Mean
SD
n
Mean
SD
31* 20 16 21
⫺1.55 ⫺1.16 ⫺1.28 ⫺1.09
0.75 0.78 1.01 0.86
0 1 9 176
⫺0.25 ⫺1.21 ⫺1.30
0.46 0.78
P ⫽ .25
31* 20 16 21
⫺0.02 ⫺0.01 ⫺0.09 0.41
0.28 0.23 0.27 0.37
0 1 9 173
0.00 ⫺0.05 ⫺0.09
0.11 0.33
P ⫽ .0001
31* 20 16 21
0.90 0.68 0.79 0.91
0.44 0.55 0.61 0.41
0 1 9 172
⫺1.01 0.80 0.80
0.32 0.43
P ⫽ .28
*One patient missing refractive error, keratometry, and axial length data. Sixteen contact lens and six spectacle patients were missing adherence data. SD ⫽ standard deviation.
subjective refraction between the two groups remained 0.2 diopters following adaptation. The third source of differential dropout was during the 2-year follow-up period, with a 33% and 14% loss to follow-up in the contact lens and spectacle groups, respectively. Again, those in the contact lens group who dropped out after adaptation were less myopic by 0.1 diopters than those who remained in the study, and more boys dropped out in the contact lens than in the spectacle group. In the spectacle group, no difference existed in refractive error between those who dropped out during the trial and those who remained and there was no differential dropout among boys in this group. It is likely that the spectacle group at the end of the study was quite similar to those who were initially randomized, but that the children who continued to wear contact lenses through 2 years were different from those initially randomized to this group. The likely differences were that more girls remained in the study and that these children were more myopic at randomization than those who dropped out, regardless of the whether the dropout was due to logistics, adaptation, or contact-lens– related events. In this study, girls had higher rates of progression, but initial refractive error did not predict the rate of progression of myopia. No evidence was seen that the rate of progression was different in the two treatment groups at 1 or 2 years of follow-up. Although the follow-up rates were relatively lower at 2 years than at 1 year (66% and 85% in the contact lens and spectacle groups, respectively, at 2 years
dren were unable to adapt to or to maintain sustained RGP contact lens wear over the course of the trial. One strength of this trial lay in the relatively large sample size and the randomization of children to contact lens or spectacle wear. Given the problems of adaptation to contact lens wear and self-selection of patients for contact lenses or spectacle wear, randomization is essential to eliminate investigator and patient selection bias and improve internal validity of the study. Despite initial randomization, there was differential dropout in the contact lens group from randomization through the end of the study. Three key sources of bias were potentially associated with this dropout. It is assumed that the probability is high that randomization produces comparable groups of patients in each of the treatment groups. Following randomization, a differential loss of children occurred in the contact lens group (31%) due to initial problems with the supply of contact lenses. This resulted in slightly more girls presenting for the initial visit in the contact lens than spectacle group. At this visit, children assigned to contact lenses were 0.2 diopters more myopic than were those in the spectacle group. The second source of differential dropout was during the adaptation period, with 19% dropping out in the contact lens group compared with 3% in the spectacle group. No gender difference occurred in adaptation. Those children who were slightly less myopic were more likely not to adapt to contact lenses, but they were also more likely to drop out of the spectacle group. Hence, the differential in 88
AMERICAN JOURNAL
OF
OPHTHALMOLOGY
JULY 2003
vs 85% and 94% at 1 year), the rate of progression was ⫺0.65 diopters and ⫺0.63 diopters in the contact lens and spectacle groups at 1 year. Hence, even with a relatively high follow-up rate and less differential loss at 1 year of follow-up, there was no evidence of an impact of contact lens wear on myopia progression. This consistency from the first to the second year of follow-up adds strength to the evidence supporting the no-treatment effect. There was also no pattern of decreasing progression with increasing daily hours of contact lens wear during the study, further evidence of a lack of a dose response to treatment. Finally, despite the slight difference in initial refractive error and gender distribution associated with differential dropout, after adjusting for all factors that were different between the two treatment groups at baseline, or that might predict myopia progression, there was still no difference in progression between the contact lens and spectacle wearers. In this study, children were randomized prior to adaptation to contact lens wear. A design where all eligible children were provided contact lenses and only those who could adapt to lens wear would be eligible for randomization would not have been practical or logistically feasible. Such children would not have agreed to be randomized to spectacles following successful adaptation to contact lenses. In this sense, the current design allowed for an “effectiveness” analysis in which children who could adapt to contact lenses were compared with children who could wear spectacles, resulting in less internal, but greater external, validity. The CLAMP study now underway in the United States, has resolved this problem by randomizing after adaptation, but the “control” group subjects are assigned to soft contact lenses, thus removing the problem of refusal to enter the control group after successful adaptation to contact lenses.25 Investigators found that 78.4% were able to adapt to RGP contact lens wear following adaptation, comparable to the 81% in our population. Dropout rates from that study25 have not yet been reported, so they cannot be compared to our findings. Because of the control group, the CLAMP trial will answer a slightly different question; that of whether RGP lenses rather than soft contact lenses can reduce the rate of myopia progression. Despite the adaptation period in our trial, a higher dropout rate occurred in the contact lens than the spectacle group during the 2 years following the adaptation period, with the excess dropout being associated with contact lens wear problems. Among those contact lens wearers who remained in the trial, at least 35% wore the contact lenses for fewer than 4 hours per day. Hence, contact lens wear for this age group of school children was not easy to maintain over long periods of time, except in a relatively small group of patients. In summary, the rate of myopia progression was about 0.65 diopters per year among this group of Singaporean children. The use of RGP lenses did not slow the rate of
myopia progression, and there was only a slightly lower progression rate among children who tolerated the contact lenses well and used them regularly and consistently. Adaptation to contact lenses and continued contact lens wear in this age group of children was relatively high. It is unlikely that this intervention holds promise as a method by which to slow the rate of progression of myopia in such children.
VOL. 136, NO. 1
AND
CONTACT LENSES
REFERENCES 1. Lin LL, Chen CJ, Hung PT, et al. Nation-wide survey of myopia among school children in Taiwan, 1986. Acta Ophthalmol Suppl 1988;185:29 –33. 2. Lin LL, Shih YF, Tsai CB, et al. Epidemiologic study of ocular refraction among schoolchildren in Taiwan in 1995. Optom Vis Sci 1999;76:275–281. 3. Zhan MZ, Saw SM, Hong RZ, et al. Refractive errors in Singapore and Ziamen, China—a comparative study in school children aged 6 to 7 years. Optom Vis Sci 2000;77: 302–308. 4. Au Eong KG, Tay TH, Lim MK. Race, culture and myopia in 110,236 young Singaporean males. Singapore Med J 1993; 34:29 –32. 5. Edwards MH. The development of myopia in Hong Kong children between the ages of 7 and 12 years: a five-year longitudinal study. Ophthalmic Physiol Opt 1999;19:286 – 294. 6. Wu MM, Edwards MH. The effect of having myopic parents: an analysis of myopia in three generations. Optom Vis Sci 1999;76:387–392. 7. Chew SJ, Chia SC, Lee LK. The pattern of myopia in young Singaporean men. Singapore Med J 1988;29:201–211. 8. Chow YC, Dhillon B, Chew PT, et al. Refractive errors in Singapore medical students. Singapore Med J 1990;31:472– 473. 9. Rajan U, Tan FT, Chan TK, et al. Increasing prevalence of myopia in Singapore school children: In: Chew SJ, Weintraub J, editors. Proceedings of the Fifth International Conference on Myopia, Toronto, Canada, June 22–24, 1994. New York: Myopia International Research Foundation, 1995, 41–46. 10. National Advisory Eye Council (US). Vision Research: a national plan, 1983–1987. US Department of Health and Human Services, Public Health Service, National Institutes of Health, 1983 (NIH Publication No. 83-2469). 11. Mandelbaum S, de Luise VP, Driebe W, Groden LR, McLeod SD, Sher NA, Talamo JH, Katz J. Refractive Errors Preferred Practice Pattern. San Francisco: American Academy of Ophthalmology, 1997. 12. American Academy of Ophthalmology. Excimer lazer photorefractive keratectomy (PRK) for myopia and astigmatism. Ophthalmology 1999;106:422–437. 13. Celorio JM, Pruett RC. Prevalence of lattice degeneration and its relation to axial length in severe myopia. Am J Ophthalmol 1991;111:20 –23. 14. Sorensen KE, Baggesen K. Retinal detachment following intracapsular cataract extraction. Acta Ophthalmol 1990;68: 549 –553. 15. Saw SM, Katz J, Schein OD, Chew SJ, Chan TK. Epidemiology of myopia. Epidemiol Rev 1996;18:175–187. 16. Jessen GN. Contact lenses as a therapeutic device. Am J Optom Arch Am Acad Optom 1964;41:429 –435.
MYOPIA
89
17. Morrison RJ. The use of contact lenses in adolescent myopic patients. Am J Optom 1960;37:165–168. 18. Stone J. Contact lens wear in the young myope. Br J Physiol Opt 1973;28:90 –134. 19. Stone J. The possible influence of contact lenses on myopia. Br J Physiol Opt 1976;31:89 –114. 20. Kelly T, Chatfield C, Tustin G. Clinical assessment of the arrest of myopia. Br J Ophthalmol 1975;59:529 –538. 21. Grosvenor T, Perrigin J, Perrigin D, Quintero S. Use of silicone-acrylate contact lenses for the control of myopia. Results after two years of lens wear. Optom Vis Sci 1989;66: 41–47. 22. Perrigin J, Perrigin D, Quintero S, Grosvenor T. Siliconeacrylate contact lenses for myopia control: 3-year results. Optom Vis Sci 1990;67:764 –769. 23. Andreo LK. Long-term effects of hydrophilic contact lenses on myopia. Ann Ophthalmol 1990;22:224 –229. 24. Khoo CY, Chong J, Rajan U. A 3-year study on the effect of
90
AMERICAN JOURNAL
25. 26.
27. 28. 29.
OF
RGP contact lenses on myopic children. Singapore Med J 1999;40:230 –237. The Contact Lens And Myopia Progression (CLAMP) study: design and baseline data. Optom Vis Sci 2001;78:223– 233. Saw SM, Nieto JF, Katz J, Chew SJ. Estimating the magnitude of close-up work in school-age children: a comparison of questionnaire and diary instruments. Ophthalmic Epidemiol 1999;6:291–301. Saw SM, Nieto JF, Katz J, Chew SJ. Distance, lighting, and parental beliefs: understanding near work in epidemiologic studies of myopia. Optom Vis Sci 1999;76:355–362. Saw SM, Nieto JF, Katz J, Schein OD, Levy B, Chew SJ. Factors related to the progression of myopia in Singaporean children. Optom Vis Sci 2000;77:549 –554. Saw SM, Nieto JF, Katz J, Schein OD, Levy B, Chew SK. Familial clustering and myopia progression in Singapore school children. Ophthalmic Epidemiol 2001;8:227–236.
OPHTHALMOLOGY
JULY 2003
● PURPOSE: To test whether rigid gas permeable (RGP) contact lens wear can reduced the rate of myopia progression in school age children. ● DESIGN: Randomized clinical trial. ● METHODS:
Setting: Single clinical center. Study Population: Both eyes of 428 Singaporean children. Inclusion Criteria: 6 through 12 years of age with myopia between ⴚ1 and ⴚ4 diopters, astigmatism < 2 diopters, no prior contact lens wear, no other ocular pathologies. Intervention: Spectacle or RGP lens correction for myopia. After a 3-month adaptation period, 383 children were followed, and 298 (78%) remained after 24 months. Outcome measures: Cycloplegic subjective refraction, keratometry, and axial length measured at 12 and 24 months. ● RESULTS: Children who adapted to contact lenses wore them for a median of 7 hours per day, but no more than 40% wore them at least 8 hours per day, 7 days per week. Spectacles were worn for a median of 15 hours per day at the time of the 24-month follow-up. There was an Accepted for publication Jan 9, 2003. InternetAdvance publication at ajo.com Feb 13, 2003. From the Department of International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland (J.K.); Dana Center for Preventive Ophthalmology, Wilmer Eye Institute, Johns Hopkins School of Medicine, Baltimore, Maryland (J.K., O.D.S.); Bausch and Lomb, Rochester, New York (B.L., T.C.); Department of Community, Occupational and Family Medicine, National University of Singapore, Singapore (S.M.S.); School Health Services, Singapore (U.R.), and Singapore Eye Research Institute, Singapore (T.K.C., C.Y.K., S.J.C.). The study was conducted at the Singapore Eye Research Institute, Singapore. Inquiries to Joanne Katz, ScD, the Johns Hopkins Bloomberg School of Public Health, 615 N. Wolfe St., Room W5009, Baltimore, MD 212052103; fax: (410) 955-2029; e-mail: [email protected]
82
©
2003 BY
increase in the spherical equivalent of ⴚ1.33 and ⴚ1.28 diopters (P ⴝ .64), and axial length increased by 0.84 and 0.79 mm (P ⴝ .38) over 2 years among children randomized to contact lenses and spectacles, respectively. Adjustment for baseline differences between the groups and for hours per day of contact lens wear did not alter these findings. ● CONCLUSIONS: Rigid gas permeable lenses did not slow the rate of myopia progression, even among children who used them regularly and consistently. It is unlikely that this intervention holds promise as a method by which to slow the rate of progression of myopia in children. (Am J Ophthalmol 2003;136:82–90. © 2003 by Elsevier Inc. All rights reserved.)
M
YOPIA IS A MAJOR CONCERN FOR ETHNIC CHI-
nese among whom the prevalence is very high and appears to have increased rapidly over a few generations.1–9 The economic burden of myopia is great, with expenditures on spectacles and contact lenses estimated at $ 2 billion annually in the United States alone.10 The use of refractive surgery to correct myopia has also increased considerably ever since lasers have been approved for this purpose.11,12 Severe myopia has also been increasing in ethnic Chinese,2 and this has led to a concern about its ophthalmic consequences such as myopic macular degeneration, retinal holes, tears and detachments, lattice degeneration, cataract, and glaucoma.13–15 It has been postulated that rigid contact lenses could slow the progression of myopia in children.16 –18 Several potential mechanisms have been offered, including improved retinal image quality with contact lens wear; flattening of the cornea, producing a temporary slowing of myopia progression; or an overcorrection for myopia when fitting contact lenses. Several studies have examined this issue, but each has had methodological problems. Stone19 enrolled 124 unmatched, nonrandomized spectacle wearers
ELSEVIER INC. ALL
RIGHTS RESERVED.
0002-9394/03/$30.00 doi:10.1016/S0002-9394(03)00106-5
and 244 contact lens wearers, but only 28 controls and 22 contacts lens patients remained after 5 years of follow-up. Kelly and co-workers20 did not randomize their patients, and they did not provide information about baseline comparability. Grosvenor and colleagues enrolled 100 contact lens wearers and 39 nonrandomized controls. The follow-up was 56% among contact lens wearers and 51% among controls after 3 years.21,22 These studies all showed higher rates of myopia progression among controls than among contact lens wearers, only some of which could be explained by flattening of the cornea due to the contact lens wear. Andreo23 retrospectively analyzed 37 contact lens wearers and 19 spectacle wearers followed for 1 year and found no difference in myopia progression between the two groups. These studies were done in Caucasian populations, where the prevalence and rate of myopia progression was lower than in ethnic Chinese populations. A more recent controlled, but not randomized, study undertaken among school children in Singapore found that there was a slower rate of progression among contact lens wearers compared with spectacle wearers.24 This study also concluded that the difference in progression could not be explained by changes in corneal curvature. However, the study enrolled 105 children in the contact lens group, was not randomized, and the dropout rate was 47% among contact lens wearers over a 3-year period. There were also differences in the baseline severity of myopia between the spectacle and contact lens groups. For this reason, we decided to undertake a larger, randomized trial of the value of rigid gas permeable lenses in reducing progression of myopia in Singaporean school-age children. A prospective, randomized trial with a similar aim but slightly different design is currently underway in the United States.25 The Contact Lens and Myopia Study (CLAMP) randomizes pre-adolescent children to rigid gas permeable or soft contact lenses after an adaptation period and follows them for 3 years. However, the results of this trial are not yet available.
DESIGN THE STUDY WAS A CLINICAL TRIAL, WHERE CHILDREN
were randomized to use either rigid gas permeable contact lenses or spectacle correction for treatment of myopia. Patients were randomized prior to adaptation to contact lens wear.
METHODS ● SETTING:
Children were enrolled and all examinations were conducted at the Myopia Clinic of the Singapore Eye Research Institute. VOL. 136, NO. 1
CONTACT LENSES
● POPULATION: The population for this study was selected because of the high rates of myopia progression observed in Singaporean children.3,9 Participants were recruited through the Refraction Clinics of the School Health Services and through newspaper advertisements, and any child who met the eligibility criteria could participate in the trial. Eligible children were those of Chinese ethnicity, between the ages of 6 and 12, with myopia between ⫺1.0 and ⫺4.0 diopters of sphere, astigmatism of ⱕ2 diopters, no prior history of contact lens wear, who did not have any other ocular or medical pathologies, and who provided informed consent.26 –29 Eligible patients were randomized to one of the two treatment arms using a randomization schedule of block size 6, generated from random number tables in Baltimore and placed in sealed envelopes with sequential patient numbers. Randomization allocation was checked several times during the study against the master list and patient identifiers. Neither patients nor clinical observers were masked to treatment group. The projected sample size for randomization was 150 in each group and was based on a mean change in refractive error of 0.7 diopters per year in the spectacle group, 80% power to detect a 0.2-diopter lower rate of progression per year in the contact lens group, a type I error of 5%, and a dropout rate of approximately 30% in the contact lens group over 2 years of follow-up. Because it was likely that there would be more dropouts among contact lens than among spectacle wearers during the adaptation period, a higher number of patients were recruited to ensure at least 150 children in the contact lens group following adaptation and 100 remaining in the trial at 24 months. The study conformed to the Helsinki guidelines. The trial was approved by the Ethics Committee of the Singapore National Eye Center, and by Bausch and Lomb, which sponsored the trial and provided the contact lenses. After explaining the study to parents and children, written informed consent was obtained from parents and verbal assent from children. ● INTERVENTION: Children were randomized to receive rigid gas permeable contact lenses (Asian Design Lens, Bausch and Lomb, Rochester, New York, USA) or spectacles to correct their refractive error. The contact lens consisted of a front and back junctionless aspheric design with an overall diameter of 9.2 mm. Base curves available for the lenses ranged from 7.3 mm to 8.4 mm in 0.1-mm steps. The lenses were fit to achieve central alignment with light mid peripheral bearing. Apical bearing seen with fluoroscein evaluation was not allowed. The lens was fit to move smoothly 1 to 2 mm after each blink and centered over the cornea with full pupil capture. An automated over-refraction was performed with trial lenses, a full cycloplegic refraction prescription minus 0.5 diopters was provided, and the visual acuity was verified with the trial lenses. Children assigned to spectacle correction were AND
MYOPIA
83
refracted (cyclopleged) by the study personnel and received new spectacles if their current prescription did not conform to the refractive error measured in the study.
axial length were compared using t tests, and the proportion of eyes with changes in these variables above certain cutoff points was compared between treatment groups using chi-square tests. The analysis was done for right and left eyes separately with comparable results. The analysis presented here only uses data from right eyes. A multivariate linear model was used to estimate the difference in the change in refractive errors from baseline through 24 months between the two treatment groups, adjusted for characteristics that were different at baseline.
● MAIN OUTCOME MEASURES:
The principal outcome was the change in refractive error measured by subjective cycloplegic refraction from postadaptation through 2 years of follow-up. Secondary outcomes were the change in keratometry (autokeratometry) and total axial length (AScan ultrasonography) over the same time period. Refractive error was obtained using both automated refraction (Nidek ARK 900, Nidek Co., Tokyo, Japan) and subjective refraction (using autorefraction as the starting point for refining the refraction) 30 minutes after the third drop of 1% cyclogyl ophthalmic solution was instilled. Total axial length was measured by placing the probe in the center of the patient’s cornea without applanation. Additional readings were taken until the SD was 0.12 mm or below. Children were dispensed contact lenses or spectacles based on the refraction obtained during the preadaptation examination. A 3-month period was provided to allow children to adapt to wearing the contact lenses or spectacles before the start of the trial. Children returned for an examination every 3 months to undergo a slit-lamp examination for detection of any adverse events, and the contact lens fit was evaluated at these visits. Children were asked to wear their contact lenses for 4 hours prior to the ocular examination. Cycloplegic refractions were obtained at 6-month intervals following the end of the adaptation period. Cycloplegic autorefraction and keratometry were performed within 2 hours of lens removal. Children could also come to the clinic at any time if they believed there were any adverse events or concerns. Spectacles and contact lenses were replaced annually for all participants. If the visual acuity at any of the three monthly visits fell below 20/40, then the study team replaced the contact lenses or spectacles. The total follow-up period was 24 months or five 6-month visits in which cycloplegic refractive errors were measured. Two measures of adherence were taken at each follow-up visit, one from children and one from parents. Each was asked how many hours per day the contact lenses or spectacles were worn and the number of days per week. The agreement between children and parents was almost 100%, so the parent data for compliance measures were selected for analysis. Adherence was defined as use of contact lenses (if assigned to this group) or spectacle use (if assigned to spectacle wear) for at least 8 hours per day, 7 days per week as reported at each scheduled visit. Statistical analysis included comparability of baseline characteristics in the two treatment groups, a comparison of those who completed the study and those who dropped out for various reasons, and a comparison of the main outcomes by treatment group. The mean changes in spherical equivalent refraction, keratometry, and total
84
AMERICAN JOURNAL
RESULTS AS SHOWN IN FIGURE 1, A TOTAL OF 564 ELIGIBLE PATIENTS
were randomized to contact lenses (281) or spectacles (283). Because of initial problems with the supply of contact lenses for the trial, a lag time existed between randomization and the initial preadaptation visit. Hence, 195 patients completed the initial visit in the contact lens group and 233 in the spectacle group (Figure 1). Of these participants, 158 in the contact lens group and 225 in the spectacle group completed the adaptation phase. The proportion of patients remaining in the trial declined with each successive visit. After 24 months, 105 patients remained in the contact lens group and 192 in the spectacle group. At the initial preadaptation visit, the age distribution, mean age (8.3 years), axial length (24.3 mm), astigmatism (⫺0.4 diopters), and prior spectacle wear (85%) were comparable, but mean keratometry was 44.1 mm and 43.6 mm in the contact and spectacle groups, respectively (P ⫽ .001). Mean subjective spherical equivalent was ⫺2.56 and ⫺2.39 diopters in the contact lens and spectacle groups (P ⫽ .02). The proportion of girls in the contact lens group was higher than in the spectacle group (56% vs 41%, P ⫽ .002). At the end of the adaptation phase (the baseline for change in refractive error over the subsequent 24 months), the mean age and age distribution, history of prior spectacle wear, axial length, and astigmatism were comparable in the two groups and similar to the initial visit values (Table 1). The keratometric readings remained slightly higher in the contact lens than in the spectacle group (P ⫽ .01), and the children assigned to contact lenses remained more myopic by 0.2 diopters than did those in the spectacle group (P ⫽ .02). The gender imbalance remained the same, with 55% and 41% girls in the contact lens and spectacle groups, respectively (P ⫽ .006). Among the 158 assigned to contact lenses at the start of the postadaptation phase, 53 patients did not complete the 24-month examination. Forty-eight patients withdrew from the study and 5 remained in the study but did not complete the 24-month examination and were considered “censored” from the study. Of the 48 patients who withdrew, 31 dropped out prior to 24 months owing to OF
OPHTHALMOLOGY
JULY 2003
FIGURE 1. Flow chart giving the number of eligible patients who were randomized and how many remained in the study at each successive visit.
contact-lens–related problems, accounting for 65% of the withdrawals in the contact lens group. Only one of these 31 patients dropped out because of positive slit-lamp findings. The remainder refused to wear the contact lenses
for a variety of reasons related to lifestyle and comfort. Girls assigned to contact lenses were more likely to complete the 24-month follow-up than were boys (P ⫽ .16), but the ages of those who did and did not complete
VOL. 136, NO. 1
AND
CONTACT LENSES
MYOPIA
85
TABLE 1. Demographic and Ocular Characteristics of Eligible Patients at the End of the Adaptation Period by Assignment Group Contact Lenses (n ⫽ 158) n
Age (years) 6–7 8–9 10–12 Mean (SD) Female Prior spectacle wear Ocular Measures
%
51 70 37 8.4 87 136 Mean
Subjective spherical equivalent (diopters) Keratometry (mm) Axial length (mm) Astigmatism (diopters)
Spectacles (n ⫽ 225) n
32.3 44.3 23.4 1.5 55.1 86.1 SD
%
90 81 54 8.3 92 193 Mean
40.0 36.0 24.0 1.6 40.9 85.8 SD
P P P P
⫽ ⫽ ⫽ ⫽
.21 .26 .006 .93
⫺2.84
0.83
⫺2.63
0.85
P ⫽ .02
43.91 24.39 ⫺0.41
1.39 0.79 0.40
43.56 24.46 ⫺0.43
1.35 0.79 0.38
P ⫽ .01 P ⫽ .40 P ⫽ .69
SD ⫽ standard deviation.
TABLE 2. Difference Between Those Followed From the Start of the Postadaptation Period Through 24 Months and Those Who Dropped Out Prior to 24 Months of Follow-up Contact Lenses Dropped out (n ⫽ 53)
Age (years) 6–7 8–9 10–12 Mean (SD) Female Prior spectacle wear Ocular Measures Spherical equivalent (diopters) Keratometry (mm) Axial length (mm) Astigmatism (diopters)
Spectacles
Followed (n ⫽ 105)
Dropped out (n ⫽ 33)
Followed (n ⫽ 192)
n
%
n
%
n
%
n
%
15 28 10 8.5 25 42 Mean
28.3 52.3 18.9 1.5 47.2 79.3 (SD)
36 42 27 8.4 62 94 Mean
34.3 40.0 25.7 1.6 59.1 89.5 (SD)
13 12 8 8.4 12 25 Mean
39.4 36.4 24.2 1.7 36.4 75.8 (SD)
77 69 37 8.3 80 168 Mean
40.0 35.9 19.3 1.6 41.7 87.5 (SD)
⫺2.63 43.8 24.3 ⫺0.36
0.83 1.4 0.8 0.37
⫺2.95 44.0 24.4 ⫺0.43
0.81 1.4 0.8 0.42
⫺2.58 43.2 24.7 ⫺0.35
0.81 1.3 0.8 0.36
⫺2.64 43.6 24.4 ⫺0.44
0.87 1.4 0.8 0.39
SD ⫽ standard deviation.
the study were similar (Table 2). Those in the contact lens group who completed the study had 0.3 diopters more myopia at baseline than those who did not (P ⫽ .02), but astigmatism, keratometry, and axial length were similar. In the spectacle group, 28 patients withdrew prior to the 24-month visit, and 5 were censored at 24 months. Lack of motivation, refusal to accept the spectacle assignment, and patients who moved were the majority of reasons for dropout (79%). Among patients assigned to spectacles, there were no statistically significant demographic or ocular differences between those who did and those who did not complete the study. 86
AMERICAN JOURNAL
Among those who completed the study, the ages and prior spectacle wear of those in the contact lens and spectacle groups were similar (Table 2), but a higher proportion of girls in the contact lens (59%) than spectacle group (42%) completed the study (P ⫽ .004). Axial length and astigmatism were similar between the treatment groups that completed the study, but the contact lens group that completed the study had 0.3 diopters more myopia at baseline than did those in the spectacle group that completed the study (P ⫽ .003). The mean baseline keratometry was 44.0 mm and 43.6 mm in the contact lens and spectacle groups that completed the study (P ⫽ .04). OF
OPHTHALMOLOGY
JULY 2003
TABLE 3. Changes in Refraction and Biometry at 1 and 2 Years of Follow-up by Assignment Group (Change Equals Follow-up Minus Immediate Postadaptation Measurements) Contact Lenses
Change in subjective spherical equivalent (diopters) 0 to 12 months 0 to 24 months Change in keratometry (mm) 0 to 12 months 0 to 24 months Change in axial length (mm) 0 to 12 months 0 to 24 months
Spectacles
n
Mean
SD
n
Mean
SD
120 97
⫺0.65 ⫺1.33
0.55 0.84
186 188
⫺0.63 ⫺1.28
0.49 0.78
P ⫽ .74 P ⫽ .64
120 97
⫺0.08 ⫺0.13
0.33 0.33
183 185
⫺0.002 ⫺0.07
0.20 0.33
P ⫽ .006 P ⫽ .20
118 97
0.35 0.84
0.41 0.47
183 184
0.33 0.79
0.40 0.45
P ⫽ .65 P ⫽ .38
SD ⫽ standard deviation.
The median hours per day that contact lenses were worn if they were being used daily varied from 5.7 hours at the 18-month visit to 7.1 hours at the 6- and 12-month visits. In contrast, spectacles were worn for a median of 12 hours per day at the 6-month visit and 15 hours per day by the 24-month visit. Using the definition of “adherence” as wearing either contact lenses or spectacles for at least 8 hours per day, 7 days per week, 39.8% and 91.1% of children were adherent at the 6-month visit in the contact lens and spectacle groups, respectively. The adherence with contact lens wear declined over time, reaching 31.5% at the 24-month visit, while adherence with spectacles increased to 98.4%. The increase in spherical equivalent for subjective refraction was approximately ⫺0.6 diopters per year for subjective cycloplegic refraction (Table 3). No difference existed between the two treatment groups with respect to the change in refraction from the start of the trial through 1 and 2 years of follow-up. Similarly, no difference was seen in the distribution of change in refractive error over 24 months (P ⫽ .35). Twenty-four percent of eyes in the contact lens group had an increase of 2 or more diopters, whereas 16% of eyes in the spectacle group increased by that amount or more over 2 years (P ⫽ .11). A similar increase occurred in total axial length in both the spectacle and contact lens groups over the course of the study, as would be expected with a similar increase in severity of myopia over the same time period (Table 3). A flattening of the cornea occurred, as measured by keratometry, in the contact-lens–wearing group, but the difference between the spectacle and contact lens groups was statistically significant only at 12 months of follow-up. Among children who wore their contact lenses for 12 or more hours per day at the 24-month visit, no difference existed in the change in refractive error, compared with all children in the spectacle group or among children who wore their spectacles for a similar amount of time per day
as the contact lens wearers (Table 4). No evidence was seen of decreasing myopia progression with increasing contact lens wear. Among children who wore contact lenses for 12 or more hours per day, there was a difference in the keratometry between the contact lens and spectacle groups, but no difference in total axial length. A multivariate model to predict the change in spherical equivalent from baseline through 24 months was fit with treatment group, age, gender, subjective refractive error, and keratometry at baseline as explanatory variables. These variables were selected because they were different at baseline between those in the spectacle and contact lens groups (subjective spherical equivalent and keratometry), or because they were predictive of myopia progression over 2 years of follow-up in bivariate models (progression of myopia decreased with increasing age and more progression among girls than boys). The adjusted difference in the change in refractive error from baseline through 24 months between the two groups (⫺0.033 diopters, 95% confidence interval [CI] from ⫺0.204, 0.138) was slightly less than the unadjusted difference of ⫺0.047 diopters. None of these differences was statistically significantly different from zero. Predictors that remained statistically significant in the multivariate model were age and gender.
VOL. 136, NO. 1
AND
CONTACT LENSES
DISCUSSION BASED ON THIS RANDOMIZED 2-YEAR TRIAL, NO EVIDENCE
was found that rigid gas permeable (RGP) lenses reduced the progression of myopia in preadolescent school children in Singapore. Among the children with 12 or more hours per day of contact lens wear, there was a slight, but not statistically significant difference of 0.2 diopters between the rate of progression in the contact lenses compared to the spectacle group, but a substantial proportion of chilMYOPIA
87
TABLE 4. Change in Subjective Refraction and Biometry From Postadaptation Through 24 Months by Assignment Group and Degree of Adherence (hours per week) at the 24-month Visit Contact Lens (n ⫽ 89)
Subjective spherical equivalent (diopters) ⬍ 4 hr/day 4 to 7 hr/days 8 to 11 hr/day 12 or more hr Keratometry (mm) ⬍ 4 hr/day 4 to 7 hr/days 8 to 11 hr/day 12 or more hr Axial length (mm) ⬍ 4 hr/day 4 to 7 hr/days 8 to 11 hr/day 12 or more hr
Spectacles (n ⫽ 186)
n
Mean
SD
n
Mean
SD
31* 20 16 21
⫺1.55 ⫺1.16 ⫺1.28 ⫺1.09
0.75 0.78 1.01 0.86
0 1 9 176
⫺0.25 ⫺1.21 ⫺1.30
0.46 0.78
P ⫽ .25
31* 20 16 21
⫺0.02 ⫺0.01 ⫺0.09 0.41
0.28 0.23 0.27 0.37
0 1 9 173
0.00 ⫺0.05 ⫺0.09
0.11 0.33
P ⫽ .0001
31* 20 16 21
0.90 0.68 0.79 0.91
0.44 0.55 0.61 0.41
0 1 9 172
⫺1.01 0.80 0.80
0.32 0.43
P ⫽ .28
*One patient missing refractive error, keratometry, and axial length data. Sixteen contact lens and six spectacle patients were missing adherence data. SD ⫽ standard deviation.
subjective refraction between the two groups remained 0.2 diopters following adaptation. The third source of differential dropout was during the 2-year follow-up period, with a 33% and 14% loss to follow-up in the contact lens and spectacle groups, respectively. Again, those in the contact lens group who dropped out after adaptation were less myopic by 0.1 diopters than those who remained in the study, and more boys dropped out in the contact lens than in the spectacle group. In the spectacle group, no difference existed in refractive error between those who dropped out during the trial and those who remained and there was no differential dropout among boys in this group. It is likely that the spectacle group at the end of the study was quite similar to those who were initially randomized, but that the children who continued to wear contact lenses through 2 years were different from those initially randomized to this group. The likely differences were that more girls remained in the study and that these children were more myopic at randomization than those who dropped out, regardless of the whether the dropout was due to logistics, adaptation, or contact-lens– related events. In this study, girls had higher rates of progression, but initial refractive error did not predict the rate of progression of myopia. No evidence was seen that the rate of progression was different in the two treatment groups at 1 or 2 years of follow-up. Although the follow-up rates were relatively lower at 2 years than at 1 year (66% and 85% in the contact lens and spectacle groups, respectively, at 2 years
dren were unable to adapt to or to maintain sustained RGP contact lens wear over the course of the trial. One strength of this trial lay in the relatively large sample size and the randomization of children to contact lens or spectacle wear. Given the problems of adaptation to contact lens wear and self-selection of patients for contact lenses or spectacle wear, randomization is essential to eliminate investigator and patient selection bias and improve internal validity of the study. Despite initial randomization, there was differential dropout in the contact lens group from randomization through the end of the study. Three key sources of bias were potentially associated with this dropout. It is assumed that the probability is high that randomization produces comparable groups of patients in each of the treatment groups. Following randomization, a differential loss of children occurred in the contact lens group (31%) due to initial problems with the supply of contact lenses. This resulted in slightly more girls presenting for the initial visit in the contact lens than spectacle group. At this visit, children assigned to contact lenses were 0.2 diopters more myopic than were those in the spectacle group. The second source of differential dropout was during the adaptation period, with 19% dropping out in the contact lens group compared with 3% in the spectacle group. No gender difference occurred in adaptation. Those children who were slightly less myopic were more likely not to adapt to contact lenses, but they were also more likely to drop out of the spectacle group. Hence, the differential in 88
AMERICAN JOURNAL
OF
OPHTHALMOLOGY
JULY 2003
vs 85% and 94% at 1 year), the rate of progression was ⫺0.65 diopters and ⫺0.63 diopters in the contact lens and spectacle groups at 1 year. Hence, even with a relatively high follow-up rate and less differential loss at 1 year of follow-up, there was no evidence of an impact of contact lens wear on myopia progression. This consistency from the first to the second year of follow-up adds strength to the evidence supporting the no-treatment effect. There was also no pattern of decreasing progression with increasing daily hours of contact lens wear during the study, further evidence of a lack of a dose response to treatment. Finally, despite the slight difference in initial refractive error and gender distribution associated with differential dropout, after adjusting for all factors that were different between the two treatment groups at baseline, or that might predict myopia progression, there was still no difference in progression between the contact lens and spectacle wearers. In this study, children were randomized prior to adaptation to contact lens wear. A design where all eligible children were provided contact lenses and only those who could adapt to lens wear would be eligible for randomization would not have been practical or logistically feasible. Such children would not have agreed to be randomized to spectacles following successful adaptation to contact lenses. In this sense, the current design allowed for an “effectiveness” analysis in which children who could adapt to contact lenses were compared with children who could wear spectacles, resulting in less internal, but greater external, validity. The CLAMP study now underway in the United States, has resolved this problem by randomizing after adaptation, but the “control” group subjects are assigned to soft contact lenses, thus removing the problem of refusal to enter the control group after successful adaptation to contact lenses.25 Investigators found that 78.4% were able to adapt to RGP contact lens wear following adaptation, comparable to the 81% in our population. Dropout rates from that study25 have not yet been reported, so they cannot be compared to our findings. Because of the control group, the CLAMP trial will answer a slightly different question; that of whether RGP lenses rather than soft contact lenses can reduce the rate of myopia progression. Despite the adaptation period in our trial, a higher dropout rate occurred in the contact lens than the spectacle group during the 2 years following the adaptation period, with the excess dropout being associated with contact lens wear problems. Among those contact lens wearers who remained in the trial, at least 35% wore the contact lenses for fewer than 4 hours per day. Hence, contact lens wear for this age group of school children was not easy to maintain over long periods of time, except in a relatively small group of patients. In summary, the rate of myopia progression was about 0.65 diopters per year among this group of Singaporean children. The use of RGP lenses did not slow the rate of
myopia progression, and there was only a slightly lower progression rate among children who tolerated the contact lenses well and used them regularly and consistently. Adaptation to contact lenses and continued contact lens wear in this age group of children was relatively high. It is unlikely that this intervention holds promise as a method by which to slow the rate of progression of myopia in such children.
VOL. 136, NO. 1
AND
CONTACT LENSES
REFERENCES 1. Lin LL, Chen CJ, Hung PT, et al. Nation-wide survey of myopia among school children in Taiwan, 1986. Acta Ophthalmol Suppl 1988;185:29 –33. 2. Lin LL, Shih YF, Tsai CB, et al. Epidemiologic study of ocular refraction among schoolchildren in Taiwan in 1995. Optom Vis Sci 1999;76:275–281. 3. Zhan MZ, Saw SM, Hong RZ, et al. Refractive errors in Singapore and Ziamen, China—a comparative study in school children aged 6 to 7 years. Optom Vis Sci 2000;77: 302–308. 4. Au Eong KG, Tay TH, Lim MK. Race, culture and myopia in 110,236 young Singaporean males. Singapore Med J 1993; 34:29 –32. 5. Edwards MH. The development of myopia in Hong Kong children between the ages of 7 and 12 years: a five-year longitudinal study. Ophthalmic Physiol Opt 1999;19:286 – 294. 6. Wu MM, Edwards MH. The effect of having myopic parents: an analysis of myopia in three generations. Optom Vis Sci 1999;76:387–392. 7. Chew SJ, Chia SC, Lee LK. The pattern of myopia in young Singaporean men. Singapore Med J 1988;29:201–211. 8. Chow YC, Dhillon B, Chew PT, et al. Refractive errors in Singapore medical students. Singapore Med J 1990;31:472– 473. 9. Rajan U, Tan FT, Chan TK, et al. Increasing prevalence of myopia in Singapore school children: In: Chew SJ, Weintraub J, editors. Proceedings of the Fifth International Conference on Myopia, Toronto, Canada, June 22–24, 1994. New York: Myopia International Research Foundation, 1995, 41–46. 10. National Advisory Eye Council (US). Vision Research: a national plan, 1983–1987. US Department of Health and Human Services, Public Health Service, National Institutes of Health, 1983 (NIH Publication No. 83-2469). 11. Mandelbaum S, de Luise VP, Driebe W, Groden LR, McLeod SD, Sher NA, Talamo JH, Katz J. Refractive Errors Preferred Practice Pattern. San Francisco: American Academy of Ophthalmology, 1997. 12. American Academy of Ophthalmology. Excimer lazer photorefractive keratectomy (PRK) for myopia and astigmatism. Ophthalmology 1999;106:422–437. 13. Celorio JM, Pruett RC. Prevalence of lattice degeneration and its relation to axial length in severe myopia. Am J Ophthalmol 1991;111:20 –23. 14. Sorensen KE, Baggesen K. Retinal detachment following intracapsular cataract extraction. Acta Ophthalmol 1990;68: 549 –553. 15. Saw SM, Katz J, Schein OD, Chew SJ, Chan TK. Epidemiology of myopia. Epidemiol Rev 1996;18:175–187. 16. Jessen GN. Contact lenses as a therapeutic device. Am J Optom Arch Am Acad Optom 1964;41:429 –435.
MYOPIA
89
17. Morrison RJ. The use of contact lenses in adolescent myopic patients. Am J Optom 1960;37:165–168. 18. Stone J. Contact lens wear in the young myope. Br J Physiol Opt 1973;28:90 –134. 19. Stone J. The possible influence of contact lenses on myopia. Br J Physiol Opt 1976;31:89 –114. 20. Kelly T, Chatfield C, Tustin G. Clinical assessment of the arrest of myopia. Br J Ophthalmol 1975;59:529 –538. 21. Grosvenor T, Perrigin J, Perrigin D, Quintero S. Use of silicone-acrylate contact lenses for the control of myopia. Results after two years of lens wear. Optom Vis Sci 1989;66: 41–47. 22. Perrigin J, Perrigin D, Quintero S, Grosvenor T. Siliconeacrylate contact lenses for myopia control: 3-year results. Optom Vis Sci 1990;67:764 –769. 23. Andreo LK. Long-term effects of hydrophilic contact lenses on myopia. Ann Ophthalmol 1990;22:224 –229. 24. Khoo CY, Chong J, Rajan U. A 3-year study on the effect of
90
AMERICAN JOURNAL
25. 26.
27. 28. 29.
OF
RGP contact lenses on myopic children. Singapore Med J 1999;40:230 –237. The Contact Lens And Myopia Progression (CLAMP) study: design and baseline data. Optom Vis Sci 2001;78:223– 233. Saw SM, Nieto JF, Katz J, Chew SJ. Estimating the magnitude of close-up work in school-age children: a comparison of questionnaire and diary instruments. Ophthalmic Epidemiol 1999;6:291–301. Saw SM, Nieto JF, Katz J, Chew SJ. Distance, lighting, and parental beliefs: understanding near work in epidemiologic studies of myopia. Optom Vis Sci 1999;76:355–362. Saw SM, Nieto JF, Katz J, Schein OD, Levy B, Chew SJ. Factors related to the progression of myopia in Singaporean children. Optom Vis Sci 2000;77:549 –554. Saw SM, Nieto JF, Katz J, Schein OD, Levy B, Chew SK. Familial clustering and myopia progression in Singapore school children. Ophthalmic Epidemiol 2001;8:227–236.
OPHTHALMOLOGY
JULY 2003