BIFOCAL CONTROL OF MYOPIA
Kenneth H. Oakley, MD.
Francis A. Young, Ph.D.
Primate Research Center
Washington State University
Forty-three Native American bifocal wearers grouped by yearly age levels from 9 to 15 with a mixed group of 6 to 8 year olds are matched on beginning age, sex, beginning refractive error and ending age with 104 Native American control subjects. Similarly, 226 Caucasian bifocal wearers are matched on the same criteria against 382 control subjects. Although the comparisons are made on each age group, the average annual rate of progression for the bifocal Native American subjects is -0.12 and -0.10 diopters in the right and left eyes respectively against a comparable rate of progression of -0.38 and -0.36 diopters for the control subjects These differences are significant but not as significant as those found on the Caucasian subjects of -0.02 and -0.03 diopters right and left eyes against -0.53 and -0.52 diopters for the controls. The meaning of these differences is discussed.
Recently a number of investigators have reported successful control of the progression of myopia in children through the use of 1% atropine sulfate drops on a daily or alternate day basis. (Gostin, 1962; Bedrossian, 1966; Boyd, 1969; Dyer and Thel, Jr., 1970.) The success of the use of atropine for the control of myopia is believed to be related to the reduction of the accommodative response under the cycloplegic action of atropine. Young (1965) reported similar results on monkeys.
If the reduction of the accommodative response is related to the progression of myopia, it seems reasonable that the reduction of the accommodative response through the use of plus reading glasses or through the use of bifocals on already myopic children should also have the effect of reducing the rate of progression of myopia in children. Such has been reported by a number of investigators (Betz, 1949; Gamble, 1949; Miles, 1957, 1962; Parker, 1958; and Warren, 1955.)
Mandell (1959) found no evidence to support the concept of control of myopia progression through the use of bifocals. Mandell indicated that what is needed to establish the merits of bifocal control is a study in which bifocals are fitted to one of two groups of myopic patients comparable in age, degree of myopia, rate of progression before correction and environment. Under these conditions the rate of progression of the group given bifocals could then be easily checked against the rate of the control group. He further points out that a study of this type presents difficult operational procedures due to the problem of obtaining suitable subjects for the necessary length of time.
Since Mandell was not able to carry out such a study he substituted an evaluation of patients in the clinical records of a practicing optometrist and proceeded to violate his own suggested criteria. Thus the patients who were fitted with bifocals had an average initial refractive error of 2.75 diopters with an average initial age of 14.3 years, while his control patients had an average initial refractive error of 1.48 diopters with an average initial age of 17.1 years. Clearly, the subjects who were fitted with bifocals were progressing at a higher rate since they had developed almost twice as much myopia by age 14 than the control subjects had by age 17. Also, since myopia is supposed to more or less stop progressing in the late teens (at the end of high school for individuals who do not go beyond high school) one could expect the 17 year old subjects to show less progression with or without bifocals than the younger subjects who were wearing bifocals. This lack of matching between the bifocal wearers and the controls make it difficult to draw conclusions, although Mandell does conclude that the wearing of bifocals had little or no effect on the progression of myopia.
The present study represents an attempt on the part of the authors to achieve the suggestions made by Mandell in designing a study to determine the effect of bifocals on the progression of myopia and to compare these subjects with control subjects who demonstrate a similar age and initial refraction and who have been followed for a number of years.
One of the difficulties of carrying on a longitudinal study which requires cooperation over time on the part of the subjects involved in the study is to achieve such cooperation consistently. This is particularly important in attempting to evaluate the effect of such drugs as atropine or devices as bifocals on the progression of myopia in children. If the drug is not used properly or at all, or if the bifocal is not fit properly or used, the investigator usually assumes that his instructions have been followed. However, without adequate checkups, he may be mislead by his subjects.
In an attempt to evaluate this possibility, the present study utilized two groups for the investigation: one group of Caucasian children and the other of Native American children. Continued contact with the two groups of children clearly indicates that the Caucasian children were much more compulsive in wearing and using their bifocals than were the Native American children. Since there was a consistent difference in the two groups, one might expect a greater effect among the Caucasian children than among the Native American children.
There were two groups of children available as subjects. The Native American population consisted of 156 children ranging in age from 6 to 21 with 54 children in the bifocal population and 102 in the control population. The Caucasian population consisted of 441 subjects who are divided into 226 bifocal subjects and 215 control subjects with the same age range as the Native American subjects.
In the Native American population 16 (29.6%) of the bifocal children were males and 38 (70.4%) were females while 36 (35.3%) of the control children were males and 66 (64.7%) were females. A Chi-square test of the sex distribution between the two groups indicates that there is no significant difference (Chi2 = 0.29 with 1 degree of freedom). Correspondingly, in the Caucasian population 118 (50.2%) of the bifocal children were males and 117 (49.8%) were females while 99 (41.6%) of the controls were males and 139 (58.4%) were females. A Chi-square test of the sex distribution between the bifocal and control groups for the Caucasian population yields a value of 3.20 with 1 degree of freedom, which is not significant.
The subjects in all populations were grouped by ages using the age at which the bifocal subjects began to wear bifocals as the "beginning age" with the control subjects matched to these ages. There were sufficient subjects at all age levels between 9 and 15 inclusive in the Native American and Caucasian populations to form yearly groups as well as a Caucasian group with beginning age 16. The subjects in both groups between the ages of 6 and 8 inclusive were grouped into a mixed group as were the 17 and 18 year old Caucasian subjects to form 8 age groups of Native American subjects and 10 age groups of Caucasian subjects with a bifocal group and a control group at each age level.
The bifocal and control subjects were matched on beginning refractive error as a group as well as on age and sex since other studies have shown that the earlier the child or animal becomes myopic the higher the yearly rate of progression (Young et al, 1954 a,b; Young, 1961, 1963). Unfortunately, when the two groups at each age level were matched for beginning refractive error, a number of subjects were lost as indicated in Table 1 for the Native American subjects and in Table 2 for the Caucasian subjects.
TABLES 1 AND 2
None of the bifocal subjects were used more than once and were used only at the age level at which they began to wear bifocals. However, since the control subjects were fitted with full distance corrections and wore these for near work, it was possible to use them at more than one beginning age level if there were a sufficient number of refractions to permit this. Thus such a subject might be used in the control group for age 9 untill he ended at age 15. If he had an intermediate refraction at age 10 he might also be used as a control subject beginning at age 10 and ending at age 15.
The relationship between the number of males and females in the bifocal and control groups was tested by the Chi-square test for each age level in both tables and no significant differences were found between the bifocal and control subjects.
Eleven of the Native American bifocal subjects were not used along with 19 of the control subjects. A total of 83 control subjects were used with 63 being used once, 19 used twice and 1 used thrice to provide the total 104 control subjects. The dropped subjects in the bifocal and control groups were compared with the subjects which were used in each group to determine what the differences were and whether any differences were significant. These comparisons were made in terms of beginning age level, ending age level, beginning refraction in each eye and the annual rate of progression for each eye (the main dependent measure). Since all refractions were carried out at approximately one, two, three, etc. year intervals, the total change over the period was divided by the number of years in the period to determine the annual rate of progression or change.
When the 11 Native American bifocal subjects who were not used are compared with those who were used, there was no difference in age (11.67 years for used and 11.18 years for not-used) or in the criterion measure, the annual rate of progression (for used the means were OD -0.12 D and OS -0.09 D while for the not-used the means were OD -0.12 D and OS -0.08 D). However, when the beginning refractions were compared there were differences significant at the 0.01 level between the groups for each eye. Thus the mean beginning refraction for the right eye was -1.28 D and -1.36 for the left eye for the used subjects and -2.83 for the right eye and -2.92 for the left eye for the not-used subjects with respective t-test values of 2.73 and 2.65 with 52 degrees of freedom. The subjects were discarded since they could not be matched with control subjects of equal beginning refractions. The similarity of the annual progression rates indicates that no bias was introduced by dropping these subjects.
A comparison of the used and not-used control subjects yields a similar picture. If the 19 not-used control subjects are compared in age with the total group of used subjects, the difference between the means (11.01 years for used against 8.84 years for not-used) is significant at the 0.01 level while none of the other differences approach significance (the used subjects beginning refractive errors are -1.23 D for right and -1.28 D for left eyes against -1.35 D right and -1.45 D left eyes of the not-used subjects; the annual rates for progression are identical for the right eyes of both groups at -0.83 D and almost identical for the left eyes at -0.36 D for the used and -0.37 D for the not-used).
Since the age difference was significant, we decided to compare the not-used subjects with a group of used subjects with approximately the same age. To accomplish this, all control subjects in the first four groups of Table 1 were used. The mean ages for the two groups became 8.53 years for the used subjects and 8.84 for the not-used subjects. While the differences between the mean beginning refractive errors for the two groups are not significant they are larger with -0.68 D for the right eye and -0.60 D for the left eye of the used subjects and -1.35 D and -1.45 D for right and left eyes respectively for the not-used subjects. A similar situation holds for the annual rate of progression with means of -0.45 D and -0.43 D, right and left eyes of used subjects, and -0.38 D and -0.37 D respectively for the not-used subjects. These comparisons still indicate that no significant bias is introduced by eliminating the 19 control subjects who could not be matched with the used bifocal subjects in terms of beginning refractive errors. The ending age means ranged from 12.49 to 14.63 and none of the comparisons were significant.
All of the bifocal subjects were used in the Caucasian study and only the control subjects are involved in an evaluation of the used and not-used subjects. The 23 not-used subjects had a mean age of 8.78 years and are compared with the 142 used control subjects in the first three groups of Table 2 who had a mean age of 8.74 years. The ending ages were 13.30 years for the discarded subjects and 12.76 years for the used subjects. The differences between the used and not-used subjects are not significant for either age comparisons. The average refractive error for the not-used subjects was -3.11 D for the right eye and -2.78 D for the left eye compared with the values of -0.99 D and -0.96 D for the same eyes on the used subjects. The t-test values are 2.93 significant at the 0.01 level for the right eyes and 2.44 for the left eyes which is significant at the 0.02 level. The annual rates of progression into myopia are -0.61 D and -0.68 D for the used and not-used subjects' right eyes and -0.59 D and -0.70 D for the left eyes of the respective groups. These differences are not significant. As indicated, earlier subjects showing an early development of myopia usually show a faster rate of progression and develop a higher level of myopia. Such subjects could not be matched with bifocal subjects and were discarded even though their annual rate of progression was higher than that of the subjects actually used.
All refractions and prescriptions were performed by the senior author. When a young child was referred for an initial refraction which indicated that the child was in or close to myopia, the author discussed with the parents the possibility of fitting a reading lens or bifocal which would provide 3/4 to 1 diopter (D) of plus lens magnification over the minus distance prescription which was usually under-corrected by 0.50 diopter. For example, if the child's refraction indicated -1.00 diopters, the prescription would be written for -0.50 D with a plus 1.50 diopter add. If the child responded positively to this lens combination, the minus distance correction would be dropped but a reading lens of +1.00 diopter would be retained. In the case of older children who were already myopic more than -2.00 diopters the distance correction would be cut 0.50 diopter and a plus 1.50 to 2.00 diopter add would be prescribed. In all cases a flat top add would be prescribed and positioned so that the top of the add would reach the middle of the pupil when the eyes are in the primary position. This location of the add would require the child to tilt his head slightly forward and to look slightly up to avoid the add at distance but would make it virtually impossible to read without using the add even if the glasses slide downward on the nose as they so frequently do in children.
Virtually all children fitted with bifocals in this study demonstrated a near point esophoria as did most of the control children who were progressing. If the parents did not wish to try the bifocal approach, the child was fitted with a slight under-correction for distance (0.50 diopters) and told to wear his glasses at all times.
Since the author carried out all refractions, there is no inter-investigator variability although there is intra-investigator variability as well as the possibility of bias in favor of the bifocal subjects. Unfortunately, this possibility was ignored until it was too late to meaningfully institute a control for this possibility. Ideally, a second refractionist should have performed initial and final refractions on all subjects without knowledge as to whether the subjects were in the bifocal or control groups. Even though this blind control was not carried out, we did attempt to determine whether bias was present by comparing rates of change, variability and patterns of change with age within and between the bifocal and control groups and between the Native American and Caucasian subjects since the use of the bifocals differed in the two groups.
The minimum number of refractions carried out on a subject was two but the average number of refractions on the Native American subjects was 3.21 for the bifocal subjects with a range from 2 to 6 and 3.17 for the control subjects with a range from 2 to 8. The corresponding medians are 3.53 and 3.35 respectively. With standard deviations of 1.17 and 1.32 respectively the difference between the means was not significant. For the Caucasian subjects the average number of refractions was 4.12 with an SD of 1.68 for the bifocal subjects and 3.20 with an SD of 1.59 for the control subjects. The range of refractions was 2 to 11 for the bifocal and 2 to 8 for the control subjects. The difference between the means was significant at the 0.001 level (t=5.91). The median number of refractions for the Caucasian bifocal group was 4.40 and for the control group 3.43.
The results obtained on the Native American bifocal and control subjects are presented in terms of means, standard deviations (SD) and t-tests in Table 3. Comparable information for the Caucasian subjects is presented in Table 4 and for the combined groups in Table 5.
TABLE 3 TABLE 4 TABLE 5
The matching criteria employed were ranked in order of importance from beginning age as the most important through beginning refractive error, sex and ending age as the least important. As might be expected with these criteria, the corresponding matches between the bifocal and control groups reflect this order as may be seen from an inspection of Tables 1 through 5. While there are no significant differences between beginning ages, sex or beginning refractive errors in either eye, there are significant differences in ending ages for the Caucasian subjects in age levels 12 and 13 and for the combined subjects in age levels 12, 13 and 14.
The criterion measure, annual rate of progression, is based upon the difference between each subject's beginning and ending ages which ranged between one and sixteen years. The average difference between the beginning and ending age levels across all ages for the Native American bifocal group is 2.79 years with a range of 2.33 to 3.40 years and 3.39 years for the control group with a range of 2.33 to 4.50 years. Correspondingly, the Caucasian bifocal subjects have a mean difference of 3.53 years with a range of 2.33 to 4.50 years while the control subjects have a mean of 3.29 years with a range of 2.50 to 4.00 years. The mean rate of progression for the Native American bifocal subjects for all age levels is -0.12 diopters for the right eye and -0.10 diopters for the left eye compared with the corresponding values for the Native American control subjects of -0.38 and -0.36 diopters for the right and left eyes respectively. Thus, the annual rate of progression in the bifocal group is one-third or less of the rate of progression in the control subjects. For the Caucasian subjects the annual rate across all age groups is -0.02 diopters for the right eye and -0.03 diopters for the left eye of the bifocal subjects. The control subjects have a mean annual rate of -0.53 for the right eye and -0.52 for the left eye; under these conditions the annual rate of progression for the bifocal subjects is approximately 5 per cent (4 percent for right eyes and 6 per cent for left eyes) of that shown by the control subjects. The corresponding values for the combined group are -0.04 diopters for each eye of the bifocal group and -0.51 for the right eyes and -0.49 for the left eyes of the control group. The overall annual rate of progression for the bifocal subjects is 8 per cent of that demonstrated by the control subjects who were matched against the bifocal subjects on beginning age, sex, beginning refractive error and to some extent on total time.
For ages 6 through 12, except for age 11, the Caucasian control subjects have annual rates of progression ranging between -0.56 D and -0.67 D in either eye while the rates at age 11 are -0.51 D for the right eye and -0.53 D for the left eye. At age 13 the rate drops to -0.49 D and -0.48 D for right and left eyes and continues to drop reaching -0.42 and -0.41 D at age 14, -0.43 and -0.41 D at age 15, -0.38 and -0.39 D at age 16 and, finally, -0.24 and -0.28 D for the right and left eyes respectively at age 17. If a child is referred for examination at age 8 with 1 diopter of myopia and continues to progress at the average rate for each year until he is 18 he will develop approximately 5 diopters of myopia by the time he reaches 18 years of age. The bifocal subject would fall between 0 and 1.40 D of myopia.
The Native American subjects in the control group vary between -0.37 D and -0.46 D for the ages 6 through 12 except for age 11 which jumps to -0.60 D in each eye. At age 13 the annual rate drops to -0.32 D for the right eye and -0.25 D for the left eye. The rate moves up to -0.42 D for each eye at age 14 and then drops to -0.20 D for each eye at age 15. While there is some consistency in the pattern that the rate of progression is higher at the younger ages and drops at the older ages in both the Caucasian and the Native American groups, the rate of change in the Native American groups is only 2/3 the rate in the Caucasian group. Conversely, when the two groups are compared in terms of the annual rate of change for the bifocal subjects, the Native American subjects change at an average rate which is 4 to 6 times as high as that found in the Caucasian subjects. The Native American bifocal group changes more rapidly than the Caucasian bifocal subjects while the control subjects change less rapidly than the Caucasian control subjects to reduce the differences between the bifocal and control Native American subjects.
While this difference in pattern may be due to several variables, one variable which operates differentially in this situation is the reading variable. The Native American subjects generally read less and less intensively than the Caucasian subjects. They also tend to drop out of school earlier and do not wear their bifocals as compulsively as do the Caucasian subjects. Since the bifocal wearers do not use their bifocals as consistently as the Caucasian subjects, the bifocals should be less effective, which they apparently are. On the other hand, since the Native American subjects do less reading, the control subjects should show lower rates of progression, which they do. Thus the reading variable alone could account for the differences found between the two groups. Since even under these conditions the bifocals seem to have a significant effect on the annual rates of progression and clearly have a significant effect on the Caucasian subjects who follow the instructions better and also do more reading, the bifocals seem to be a relatively effective means of controlling the progression of myopia although probably not as effective as atropine.
The rates of progression of about -0.50 diopters per year among the control subjects at the younger age levels are commonly found among myopes at these age levels and suggest that the control subjects in this study do not differ from myopic children who are fitted with a virtually full correction which is worn constantly. The annual rate of progression of -0.04 diopters per year found among the bifocal subjects is uncommonly found among myopes of these age levels and suggests that the bifocals are having a controlling and reducing effect upon the rate of progression. The effectiveness of the bifocal in this study may well depend upon the very high position of the add fitted to the child. The bifocal can only have an effect, if it has an effect at all, if it is used. The bifocal fitted too low or too small a bifocal or a bifocal which is not used does not provide a proper test of the effectiveness of the bifocal. Under reasonably well controlled conditions the bifocal appears to be effective in controlling the progression of myopia.