Changes in Refractive Error over a 5-Year Interval in the Beaver Dam Eye Study

Abstract

purpose. To examine changes in spherical equivalent over a 5-year period in persons 43 to 84 years of age.

methods. All people 43 to 84 years of age and living in Beaver Dam, Wisconsin, in 1988 were invited for a baseline examination (1988–1990) and a 5-year follow-up examination (1993–1995). Refractions were determined according to the same protocol at both examinations. Aphakic and pseudophakic eyes were excluded as well as eyes with best corrected Snellen visual acuity of 20/40 and worse. After exclusions, refraction was obtained on 3007 right eyes and 3012 left eyes of the 3684 people participating in both examinations.

results. Right and left eyes behaved similarly. Spherical equivalent became more positive in the youngest subjects and more negative in older subjects. After adjusting for other factors, the 5-year change in spherical equivalent of those 45, 55, 65, and 75 years of age was +0.15, +0.18,+ 0.10, and −0.07 D, respectively. Severity of nuclear sclerosis was related to the amount of change. Those with mild nuclear sclerosis at baseline had a change of +0.2 D, whereas those with severe nuclear sclerosis had a change of −0.5 D. The amount of change was also related to gender, diabetes, and age at onset of myopia. It was unrelated to education and baseline spherical equivalent.

conclusions. Changes in spherical equivalent over a 5-year period were small. Before the age of 70, people became more hyperopic. After the age of 70, people became more myopic. Much of the myopic change may be related to increasing nuclear sclerosis.

Many people require the use of corrective lenses or refractive surgery to see clearly because of refractive errors. Myopia is the most common refractive error and is usually related to increased axial length relative to the refractive power of the cornea and lens. ^{1 2} Heredity (primarily among youth-onset myopia) and near-work activities have been investigated as reasons for the excessive axial growth. ^{3 4} Age at onset and degree of myopia, ⁵ lens thickness, ^{6 7} and diabetes ⁸ have been postulated as possible factors leading to changes in refraction. 

Population-based cross-sectional studies in adults show decreasing prevalence rates of myopia with increasing age and less education. ^{9 10 11} The Beaver Dam Eye Study provided prevalence rates of myopia that varied from 43% in those 45 to 54 years of age to 15% in those 65 to 74 years of age. In the Framingham Offspring Eye Study, prevalence of myopia was similar to prevalence rates in the Beaver Dam Eye Study. They also report a prevalence of myopia of 26% in those with 7 to 12 years of education, whereas those with 16 or more years of education had a prevalence of 43%. 

Few studies have investigated changes in refractive error in adults over time. Studies in young adults (20–30 years of age) have found myopic shifts in refraction, regardless of the baseline refraction. ^{2 12 13} Another study involving a range of ages found myopic shifts in refraction before 50 years of age and hyperopic shifts after 50 years of age. ¹⁴ These studies have all been small and often in select populations. They have often studied subjects whose tasks included extensive near work. Methods used were not always consistent over time, such as use of cycloplegia at one examination, but not at another. There remains little information about change in refraction with age. This information is important in anticipating eye care needs and understanding the long-term expectations for patients undergoing refractive surgery. The Beaver Dam Eye Study is a population-based study of adults 43 to 84 years of age observed for 5 years. This article examines changes over time in spherical equivalent as a measure of refraction and its relation to various characteristics. 

Methods

Population

Methods used to identify the population and description of participants and nonparticipants appear in previous reports. ^{15 16} Briefly, a private census of the population of Beaver Dam, Wisconsin, was performed from fall 1987 through spring 1988. All 5925 people identified as living within the township of Beaver Dam and aged 43 to 84 years were eligible for participation and were invited for baseline examinations from March 1988 through September 1990. All people participating in the baseline examination and surviving until March 1993 (n = 4541) were invited to return for follow-up examinations from March 1993 through June 1995. A total of 3684 returned for follow-up examinations. 

Procedures

Similar procedures were used at both the baseline and follow-up examinations. Tenets of the Declaration of Helsinki were followed. Informed consent was obtained from each subject, and institutional human experimentation committee approval was granted. Assessment of the refraction in the participant’s current prescription (if available) was followed by a standardized refraction using an automated refractor. The refraction was refined according to a modification of The Early Treatment Diabetic Retinopathy Study (ETDRS) protocol ¹⁷ to obtain the best corrected visual acuity when the automated refraction yielded visual acuity of 20/40 or worse. Interexaminer and intraexaminer comparisons showed no significant differences over time or among examiners for the Humphrey (San Leandro, CA) refractions obtained. 

Blood pressures were measured according to the Hypertension Detection and Follow-up Program protocol. ¹⁸ Pupils were pharmacologically dilated and an interview schedule was administered. During the interview, participants were asked about history of glasses use for distance, years of education, income, history of diabetes, cardiovascular diseases (myocardial infarction, angina, and stroke), smoking, and alcohol consumption. After pupil dilation, slit lamp examination of the lens was performed, and lens status was recorded. Photographs were then taken of the lens of each eye using modified cameras. Slit lamp photographs were subsequently graded for lens status, severity of nuclear sclerosis, and lens thickness. ¹⁹ Nuclear sclerosis was graded on a five-level scale by comparing lens density to a set of standards. Serum glucose and glycosylated hemoglobin from a casual blood specimen were measured for each subject. ^{20 21}  

Definitions

The spherical equivalent was calculated from one of three possible methods of refraction. The formula for calculating spherical equivalent was spherical power (in diopters) + one half cylinder power (in diopters). The results of the Humphrey refraction were used in the analyses for 96% of eyes at baseline and for 93% of eyes at follow-up. When ETDRS refraction (as modified for this study and described) was performed, that refraction was used in the analyses (4% of eyes at baseline and 5% of eyes at follow-up). In the remaining people, refraction from the current prescription was used (<1% of eyes at baseline, 2% of eyes at follow-up). Eyes without a lens, with an intraocular lens, or with best corrected visual acuity 20/40 and worse were excluded from analyses reported here because of diminished reliability and increased variability of refractions in those with impaired vision. 

A person was considered to have diabetes if there was a self-report of diabetes accompanied by treatment (insulin or diet) or elevated glucose or glycosylated hemoglobin. Age was defined by the baseline value. Education level was categorized as fewer than 12 years, 12 years, 13 to 15 years, and 16 or more years. Myopia was defined as a spherical equivalent less than −0.5 D. Hyperopia was defined as a spherical equivalent greater than +0.5 D. When discussing the direction of change in spherical equivalent, a change in the positive direction was considered a hyperopic shift, whereas a change in the negative direction was considered a myopic shift. Youth-onset myopia was defined as self-reported use of glasses for distance before age 20 years, whereas adult-onset myopia was defined as a history of wearing glasses for distance after age 20 years. 

Analyses

The Statistical Analysis System (SAS; Cary, NC) was used for analyses. ²² Relationships to amount of change were examined through analysis of variance, the Pearsonχ ² test, and the Cochran–Mantel–Haenszel test of general correlation. 

Results

Changes in spherical equivalent were analyzed for the 3007 right eyes with refraction at both examinations after exclusion (92 refractions not measured, 262 cataract extractions, 323 with visual acuity of 20/40 or worse). After similar exclusions, 3012 left eyes were available for analyses. Inclusion of those with visual acuity of 20/40 or worse did not consistently affect the results. Because results from right and left eyes were similar, only the results from the right eyes are presented. Comparisons of various characteristics among those included for analyses and those excluded are shown in Table 1 . Those excluded tended to be older, female, and less educated. They had lower income, drank less alcohol, and had higher systolic and lower diastolic blood pressures. They were also more likely to have more severe nuclear sclerosis, cardiovascular disease, and diabetes. 

The distribution of changes in spherical equivalent by age and gender is shown in Table 2 . Over the 5-year period, there was a small hyperopic shift of +0.12 D across the entire population. Hyperopic (positive) change in spherical equivalent occurred in those younger than 65 years. Thirty-nine percent had a positive shift of at least 0.5 D with an average change of approximately +0.20 D. After age 65 years, there was a myopic change in spherical equivalent of −0.12 D over 5 years, with only 24% having a positive shift. Those aged more than 75 years had larger myopic changes than those 65 to 74 years of age. Although statistically significant, men and women had similar changes (14.4% versus 13.4% with changes less than −0.5 D). 

Table 3 describes the distribution of change in spherical equivalent by age for potentially related characteristics. Data are presented for those less than 65 years of age and those 65 years of age and older because of small sample sizes in the older age groups and similar behavior in the younger age groups. Significance testing was performed by these age strata and for all ages combined, adjusting for the age strata. Although those aged more than 65 years with more education had positive changes, there was no significant relationship between education and refractive changes when age and sex were accounted for. Baseline refractive state was not related to amount of refractive shifts. However, persons with youth-onset myopia were more likely to have myopic shifts than persons with adult-onset myopia or those never needing glasses. People with diabetes were more likely to have hyperopic shifts, although this was only significant after adjusting for age and gender. Those with higher levels of nuclear sclerosis at baseline were more likely to have myopic changes. 

A linear model was fit to assess the relationship of all related factors to change in spherical equivalent (Table 4) . Factors were included based on a stepwise method of including the most important factors one at a time until no other factors contributed further information. A quadratic term for age was included because of the apparent nonlinear relationship, as seen in Table 2 . Partitioning the sums of squares shows the relative importance of each factor by describing the amount of the variability explained by the addition of that factor after all others are in the model. Severe nuclear sclerosis (levels 4, 5) contributed the most information to the model, with an average decrease in spherical equivalent change of −0.7 D compared with those with mild nuclear sclerosis (levels 1, 2). Age, gender, moderate nuclear sclerosis and youth-onset myopia were similar to each other in their contribution to explaining change in refraction. Diabetes status at baseline explained a smaller amount. The model may be used to calculate the 5-year change in spherical equivalent for those 45, 55, 65, and 75 years of age as +0.15, +0.18, +0.10, and− 0.07 D, respectively, after controlling for other factors. Similarly, those with mild, moderate, and severe nuclear sclerosis had changes of+ 0.20, +0.07, and −0.52, respectively. 

Discussion

The presence and severity of nuclear sclerosis at baseline was an important determinant of the amount of change in spherical equivalent in this population. From the multivariate model, after controlling for other factors, we found that those with mild nuclear sclerosis (levels 1, 2) had hyperopic shifts of +0.20 D, whereas those with severe nuclear sclerosis (levels 4, 5) had large myopic shifts of −0.52 D (a decrease of −0.72). Other studies have not reported this finding. Most studies that examined change in spherical equivalent have been in young people, presumably free of cataract. One study evaluating changes in older people ¹⁴ excluded those with cataract. They found 10-year changes of +0.3 D in those 48 to 52 years of age at baseline and +0.4 in those 58 to 62 years of age at baseline. Excluding those with severe nuclear sclerosis (levels 4, 5), we found 5-year changes of+ 0.3, +0.3, +0.2, and +0.1 D, respectively, in those 48 to 52, 53 to 57, 58 to 62, and 65 to 68 years of age at baseline. Although our rates of hyperopic shifts decreased more for older ages, they were comparable to the rates found by Ellingsen et al. ¹⁴  

After accounting for nuclear sclerosis, there was still a strong relationship to age and gender. Spherical equivalent increased in younger age groups and decreased in older age groups. After adjusting for nuclear sclerosis and other factors, those 45, 55, 65, and 75 years of age had 5-year changes in spherical equivalent of +0.15, +0.18,+ 0.10, and −0.07 D, respectively. Women had changes of +0.16 D and men had changes of +0.06 D. This was consistent with the increased prevalence of hyperopia with age that tapers around the age of 60 years, seen in most prevalence studies. ^{9 10 11}  

We found no relationship among the amount of change for various baseline spherical equivalents, although there was a trend for those with myopia at baseline to have smaller hyperopic changes. We found a relationship, however, to the age at onset of myopia. After adjusting for other factors, those who had worn glasses since childhood had on average a +0.02-D change in spherical equivalent, whereas those never needing glasses for distance vision had +0.16-D change. The difference in the amount of change (−0.14 D) between youth-onset myopes and others was consistent with studies involving younger subjects. In a study of 53 university students (aged 18–26 years), adult-onset myopes (after the age of 16 years) and emmetropes had a 3-year decrease in spherical equivalent of −0.18 and −0.15 D, respectively, whereas youth-onset myopes had a decrease of −0.26 D. ² On further investigation, we found that those wearing glasses since childhood were also more likely to be myopic at our baseline examination (data not shown). It is possible that the trend observed in change in spherical equivalent for myopes reflects the experience of those people wearing glasses for distance since childhood. 

Prevalence studies have found relationships of education and diabetes to refraction. ^{8 9 10 11} Level of education may represent a propensity for near-work activities throughout life, which is thought to affect change in spherical equivalent. We found no relationship between education level and amount of change. After adjusting for other factors, those with diabetes had a +0.22-D change in spherical equivalent, whereas those without diabetes had a +0.10-D change. 

Because we did not measure many parameters of the eye (e.g., axial length), it is not possible to say whether the changes in refraction by age and gender that we found were caused by specific components of the refractive system of the eye other than the relationship to nuclear sclerosis. Residual accommodative ability may have influenced the changes observed in the younger subjects who may have had greater accommodative ability at the baseline examination than 5 years later. Further research is needed to understand fully the natural changes in refraction in adults. Population-based studies with cycloplegic refraction performed on subjects of a wide range of ages are needed to explore the magnitude and direction of change by age. 

In conclusion, refraction continues to change throughout adulthood. We cannot assess which anatomic and physiologic components contribute to the changes. We note that changes observed over 5 years are small but may have a cumulative effect over many more years. As these shifts in refraction occur, the use of glasses may be required. This affects all people, regardless of initial refractive status, including emmetropes, naturally, or as a result of surgery. 

 Reprint requests: Kristine E. Lee, Dept. of Ophthalmology and Visual Sciences, University of Wisconsin-Madison, 610 North Walnut Street, 460 WARF, Madison, WI 53705-2397.

 Supported by Grant EY06594 (RK, BEKK) from the National Institutes of Health.

 Submitted for publication July 29, 1998; revised November 30, 1998; accepted February 9, 1999.

 Proprietary interest category: N.

 

Table 1.

 

Comparison of Participant Characteristics by Inclusion for Analyses, Right Eyes

Characteristic	Included			Excluded			P
		n	Mean ± SD	n	Mean ± SD
Age (y)	3007	58.5 ± 9.5	677	68.9 ± 10.3	<0.001
Systolic blood pressure (mm Hg)	3006	129.6 ± 18.6	677	135.9 ± 22.0	<0.001
Diastolic blood pressure (mm Hg)	3006	78.4 ± 10.2	677	75.1 ± 11.1	<0.001
Education (y)	3005	12.4 ± 2.8	677	11.6 ± 2.9	<0.001
Pack-years smoked	2998	16.3 ± 24.2	674	17.4 ± 29.6	0.36
Alcohol (g ethanol)	2998	58.6 ± 121.0	676	40.4 ± 85.5	<0.001
	n	%	n	%
Gender (female)	3007	54.9	677	65.0	<0.001
Income ($)	2916		632		<0.001
10–19,000		24.6		33.2
≥60,000		18.9		9.2
Diabetes (yes)	2994	6.5	673	12.6	<0.001
Cardiovascular disease history	2969	13.7	662	26.0	<0.001
Severe nuclear sclerosis (grade 4, 5)	2963	6.0	513	35.1	<0.001

Table 2.

 

Five-Year Changes in Spherical Equivalent by Age and Gender, Right Eyes

Gender and Age (y)	n	Amount of Change (D)					Percent with Changes (D)
				Mean	SD	P	Less than −0.5	−0.5–+0.5	Greater than +0.5	P
Women
43–54	636	0.27	0.58	<0.001	6.6	52.4	41.0	≤0.001
55–64	508	0.23	0.62		10.2	45.7	44.1
65–74	411	−0.01	0.69		21.9	51.8	26.3
75+	97	−0.37	1.08		39.2	44.3	16.5
Men
43–54	557	0.19	0.48	<0.001	9.0	55.5	35.6	≤0.001
55–64	438	0.18	0.62		8.2	57.8	34.0
65–74	295	−0.16	0.79		28.5	47.1	24.4
75+	65	−0.21	0.74		38.5	40.0	21.5
Women (all)	1652	0.15	0.68	<0.003*	13.4	49.7	36.9	<0.002*
Men (all)	1355	0.09	0.63		14.4	53.7	32.0
Women and men
43–54	1193	0.23	0.54	<0.001, †	7.7	53.8	38.5	≤0.001, †
55–64	946	0.21	0.62		9.3	51.3	39.4
65–74	706	−0.07	0.73		24.7	49.9	25.5
75+	162	−0.30	0.96		38.9	42.6	18.5
All	3007	0.12	0.66		13.9	51.5	34.7

*  Age-adjusted.

†  Gender-adjusted.

Table 3.

 

Relationship of Baseline Characteristics to Changes in Spherical Equivalent, Right Eyes

	<65 Years							65+ Years							All Ages
		n	Change	Percent			n	Change	Percent			n	Change	Percent
				(Mean ± SD)	Neg*	Pos, †			(Mean ± SD)	Neg*	Pos, †			(Mean ± SD)	Neg*	Pos, †
Education
< High school	298	0.17 ± 0.59	9.7	36.2	346	−0.17 ± 0.75	30.9	22.0	644	−0.01 ± 0.70	21.1	28.6
High school	1074	0.23 ± 0.56	8.6	40.4	349	−0.09 ± 0.73	26.7	24.9	1423	0.16 ± 0.63	13.0	36.6
College	365	0.24 ± 0.51	8.0	40.8	105	−0.09 ± 0.76	22.9	25.7	470	0.17 ± 0.59	11.3	37.5
Post-college	400	0.20 ± 0.64	7.3	35.3	68	−0.05 ± 1.17	19.1	29.4	468	0.16 ± 0.75	9.0	34.4
P		0.29	0.99				0.49	0.02				≤0.001	≤0.001
P , ‡										0.24	0.15
Refraction status (baseline)
Emmetrope	369	0.27 ± 0.54	8.9	47.4	64	0.02 ± 0.78	20.3	29.7	433	0.23 ± 0.58	10.6	44.8
Myope	889	0.21 ± 0.54	8.4	34.4	157	−0.09 ± 0.82	28.0	27.4	1046	0.16 ± 0.60	11.4	33.4
Hyperope	881	0.21 ± 0.62	8.2	39.8	647	−0.14 ± 0.78	27.8	22.9	1528	0.07 ± 0.71	16.5	32.7
P		0.21	0.36				0.27	0.12				≤0.001	≤0.001
P , ‡										0.11	0.12
Glasses use
Never	864	0.28 ± 0.48	6.0	43.9	205	−0.10 ± 0.78	27.8	28.3	1069	0.21 ± 0.57	10.2	40.9
Youth-onset	588	0.14 ± 0.70	11.2	33.0	128	−0.20 ± 1.02	28.9	20.3	716	0.08 ± 0.78	14.4	30.7
Adult-onset	678	0.21 ± 0.57	9.1	37.8	513	−0.10 ± 0.72	26.7	23.4	1191	0.08 ± 0.65	16.7	31.6
P		≤0.001	≤0.001				0.48	0.43				≤0.001	≤0.001
P , ‡										≤0.001	0.003
Diabetes status (baseline)
No	1980	0.22 ± 0.58	8.3	38.9	779	−0.14 ± 0.80	27.6	23.0	2759	0.12 ± 0.67	13.7	34.4
Yes	124	0.28 ± 0.53	10.5	38.7	72	0.05 ± 0.68	25.0	36.1	196	0.20 ± 0.60	15.8	37.8
P		0.21	0.68				0.06	0.08				0.11	0.80
P , ‡										0.02	0.36
Nuclear sclerosis severity (baseline)
Mild (1, 2)	1713	0.25 ± 0.53	7.0	39.1	250	0.08 ± 0.73	14.0	28.8	1963	0.22 ± 0.56	7.9	37.8
Moderate (3)	376	0.17 ± 0.66	12.0	40.2	445	−0.06 ± 0.68	26.1	26.1	821	0.04 ± 0.68	19.6	32.5
Severe (4, 5)	25	−0.83 ± 1.22	56.0	12.0	154	−0.58 ± 0.97	51.3	11.7	179	−0.62 ± 1.01	52.0	11.7
P		≤0.001	≤0.001				≤0.001	≤0.001				≤0.001	≤0.001
P , ‡										≤0.001	≤0.001

*  Negative change: −0.5 D or less.

†  Positive change: +0.5 D or more.

‡  Age- and sex-adjusted.

§  Exclude those with diabetes.

Table 4.

 

Multivariable Model of the Amount of Change in Refraction

	Amount of Change	P	Contribution to Sum of Squares
Age (/year)
Linear term	+0.05	<0.001	5.09
Quadratic term	−0.001	<0.001	6.57
Sex (M)	−0.10	<0.001	7.16
Nuclear sclerosis
Moderate vs mild	−0.13	<0.001	6.73
Severe vs mild	−0.72	<0.001	62.79
Glasses use
Before age 20 vs never	−0.14	<0.001	8.24
After age 20 vs never	−0.05	0.10	1.07
Diabetes (present)	0.12	0.01	2.42