- Email: [email protected]
Incidence and Progression of Nuclear Opacities in the Longitudinal Study of Cataract
Incidence and Progression of Nuclear Opacities in the Longitudinal Study of Cataract M. Cristina Leske, MD, MPH/ Leo T. Chylack, Jr., MD, 2 Suh-Yuh Wu, MA/ Elinor Schoenfeld, PhD, 1 Qimei He, PhD, 1 Judith Friend, MA, 2 John Wolfe, PhD, 2 The Longitudinal Study of Cataract Group* Purpose: To estimate incidence and progression rates of nuclear opacities in the Longitudinal Study of Cataract, an epidemiologic study of the natural history of all types of lens opacities. Methods: The Lens Opacities Classification System Ill was used to assess longi tudinal changes between baseline and follow-up lens photographs for the 764 Longi tudinal Study of Cataract participants. Baseline data, collected until December 1988 as part of a case-control study, included color slit, retroillumination, and Scheimpflug pho tographs. The same data were collected by the Longitudinal Study of Cataract at four subsequent visits at yearly intervals. Results: Among patients free of nuclear opacities at baseline, the incidence of new opacities was 6% after 2 years and 8% after 5 years of follow-up. The progression of pre-existing nuclear opacities was much higher. After 2 years, nuclear opacities had progressed in more than one third of the patients with pre-existing opacities; after 5 years, almost half had progressed. Older age was significantly related to higher incidence of new nuclear opacities, but not to progression of pre-existing opacities. Patients with other opacity types had higher nuclear incidence and progression rates. Conclusions: In this clinic-based, older-patient population, new nuclear opacities developed in less than one tenth of the patients after 5 years of follow-up. In contrast, almost one half of the patients with pre-existing opacities had worsened after 5 years. These estimated rates can be used to plan intervention or other studies of nuclear changes in similar populations. Ophthalmology 1996;103:705-712
Age-related cataract is a highly prevalent condition, which is the major cause of visual impairment. 1- 4 In many countries, surgical resources are insufficient to cope with the annual increment, much less the backlog, of unop erated cataracts. 1•2 A preventive or medical approach that Originally received: February 7, 1995. Revision accepted: February 8, 1996. 1 School of Medicine, SUNY, Stony Brook, New York. 2 Harvard Medical School Brigham and Women's Hospital and Center for Ophthalmic Research, Boston. *Members are listed in the Appendix section at the end of this article. Supported by National Eye Institute grant R01EY08291, Bethesda, Maryland. Reprint requests to M. Cristina Leske, MD, MPH, Department of Pre ventive Medicine, SUNY, Z = 8036, Stony Brook, NY 11794-8036.
could delay the need for surgery by 10 years would de crease the annual number of cataract surgeries in the United States by 45%. 3 Such a goal, however, cannot be achieved until more data are available on the causes and natural history of lens opacities. Cataract formation may involve single or multiple re gions of the lens, and each region may opacify at its own rate. However, very little epidemiologic data are available on the development and progression of the various types of lens opacities. 4 The Longitudinal Study of Cataract (LSC) represented a first step toward obtaining such data by providing basic natural history information on longi tudinal changes in the lens. Its major aims were to (1) measure the growth rates of nuclear, cortical, and posterior subcapsular opacities in a clinic-based population, (2) as sess and compare various qualitative and quantitative
705
Ophthalmology
Volume 103, Number 5, May 1996
methods to document changes in opacities and color, and (3) evaluate risk factors. This report presents the estimated incidence rates of new nuclear opacities and progression rates of pre-existing nuclear opacities in the LSC popu lation, using the Lens Opacities Classification System III (LOCS III). 5 The incidence and progression of cortical and posterior subcapsular opacities will be presented in a separate report.
Table 1. Participation and Time Intervals by Study Visit Days Since Last Visit Follow-up Visit
1 2 3
Methods Data Collection The LSC (1989-1993) was an epidemiologic study of the natural history of lens opacities, which followed a subset of participants from the Lens Opacities Case-control Studl ( 1985-1988). Surviving case-control study partic ipants and control subjects were eligible for the LSC unless they had cataract surgery in both eyes, missing or un gradable lens photographs, or were unable to complete yearly follow-up visits. The first follow-up visit for LSC started in September 1989, and subsequent follow-up visits continued at an average of 1-year intervals over a 4-year period. All participants provided informed consent and followed a standardized protocol designed for the study. Each LSC visit included a comprehensive ophthal mologic examination, which determined best-corrected visual acuity after the Early Treatment of Diabetic Reti nopathy Study protocol,7·8 contrast sensitivity using the Pelli-Robson test chart, 9 ocular pathology, and the cause of visual acuity loss, if any. Patient questionnaires were administered to update information on medical history, personal practices, and eye problems, including cataract surgery. Patients who had had unilateral cataract surgery continued to be followed to provide data on fellow eyes. Standardized color slit-lamp (Carl Zeiss, Oberkochen, Germany), retroillumination (Neitz CTR, Neitz Instru ments Co, Torrance, CA), and Scheimpflug (Topcon SL 45, Topcon America Corp, Paramus, NJ) photographs of the lens were taken. Photographs were graded according to the LOCS III method following published protocols. 5 Of the projected 800 eligible case-control participants, 764 (96%) completed the first follow-up visit (Table 1). Subsequently, 621 (81 %) patients remained eligible for at least one additional follow-up visit. Given the advanced age of the cohort (median, 67 years), 143 patients could not be followed due to death or severe illness (n = 60), cataract surgery in both eyes (n = 26), and relocation (n = 57). An additional 16 patients who completed visit 2, as well as 39 who completed visit 3, could not be followed further because the study ended. Among those remaining eligible, the number of patient refusals was 63 for visit 2, 44 for visit 3, and 30 for visit 4, resulting in a total number of patients followed of 618, 524, and 429 for those visits for participation rates of91 %, 92%, and 93%, respectively (Table 1). The average time interval between LSC follow-up visits was approximately 1 year, with standard deviations rang ing from I to 3 months (Table 1). Comparisons were made
706
4 SD
=
Participants/ Eligible
(% eligible)
Mean± SD (median)
764 618/681 524/568 429/459
(91) (92) (93)
331 ± 88 (320) 367 ± 65 (364) 363 ± 34 (364)
standard deviation.
between patients who were lost to follow-up after two visits and those who remained in the study for three or more visits. No differences were found between these groups with respect to their age, sex, and baseline nuclear opalescence status. Classification The LOCS III protocol was followed to grade the type and degree of lens opacification from lens photographs taken with maximal dilatation. 5 A full description and evaluation of the LOCS III method has been published elsewhere. 5 This classification system is based on six slit lamp standard images for grading nuclear opalescence, five retroillumination standard images for grading cortical, and five retroillumination images for grading posterior subcapsular cataract. A decimal scale is used to grade in creasing degrees of lens opacification, which ranges from 0.1 (most clear) to 5.9 (most advanced) for cortical and posterior subcapsular cataract or to 6.9 (most advanced) for nuclear opalescence. For this report, nuclear opacities were considered present if the nuclear opalescence score were 2 or more. All gradings were completed by the same two independent masked graders, and discrepancies were resolved by consensus. Nuclear grading sessions were conducted separately from cortical and posterior subcap sular grading sessions. To avoid drift in grading and to obtain a consistent classification, all photographs for all study visits completed by each patient were ordered ran domly in a carousel and graded at the same session. These gradings were conducted during a 6-month period at the end of the study. Intragrader and intergrader agreement were evaluated by regrading a quality control set of 155 sets of photographs. Intraclass correlation coefficients for agreement within and between observers were above 0.96 for all cataract types. The intergrader and intragrader variation in the LOCS III decimal scale was assessed by 95% tolerance limits, 5 which ranged from ±0.4 to ±0.6 for nuclear opalescence, ±0.6 to ± 1.0 for cortical, and ±0.6 to ±0.9 for posterior subcapsular opacities. The quality control set also was circulated continuously for grading throughout the LOCS III photograph grading ses sions. No signs of drift were detected.
Leske et al · Incidence and Progression of Nuclear Opacities
Analysis Eye-based Analyses The initial analyses evaluated the presence of changes in nuclear opalescence for each eye. To allow for "noise" due to grading or photograph variation, a change in grad ing was defined as present ifthe difference between nuclear opalescence scores at any two visits exceeded a given threshold. This threshold was defined as a difference in nuclear opalescence scores of 0. 7 or more in the LOCS III decimal scale. The threshold was based on the highest variation observed in the 95% tolerance limits for nuclear gradings, as derived both from the quality control results from this study (limits of ±0.4 to ±0.6) and from previous LOCS III reproducibility evaluations. 5 Therefore, a change in opalescence was considered present ifthe nuclear opal escence score at one visit was at least 0. 7 higher (or lower) than a previous nuclear opalescence score for that eye. For eyes of patients with only two visits, increases (or decreases) in opalescence at follow-up were simply defined according to these criteria, that is, whenever the difference between the two available nuclear opalescence scores was at least 0.7. For eyes of the majority of patients, who had three or more visits, all available nuclear scores were used to define confirmed increases (or decreases) throughout the follow-up period. The definition of confirmed increase was (1) a difference in scores of at least 0. 7 from baseline to the last visit, plus (2) a similar increase in nuclear scores between any two LSC visits, which was either followed by another increase in scores or by no change in scores at subsequent visits. The definition of confirmed decrease was the same, if the scores retrogressed over the follow up period. If these criteria for confirmed increase or de crease were not met, eyes were classified as having no change in opalescence. Time of the change, if any, was defined as the midpoint between the follow-up visit where the change was first observed and its preceding visit. For eyes that had cataract surgery, an increase in opal escence also was considered present if nuclear opacities were the most severe cataract type (i.e., the nuclear score was proportionally higher than the score for cortical and posterior subcapsular opacities) at the last visit before sur gery. Otherwise, no change in nuclear opalescence was assumed. The time of increase was defined as the mid year of the cataract surgery; if the year was unknown, the time of increase was defined as the midpoint between the last visit before surgery and the date when cataract surgery was reported.
Patient-based Analyses After determining increases (or decreases) for each eye, the subsequent analyses were patient-based. As a first step, we determined the percent change in nuclear scores over the entire period of the study, regardless of the length of time of follow-up. Separate analyses were conducted for persons with both lenses and a single lens, as well as for patients with two visits only, three or more visits, and with cataract surgery. As a second step, we estimated in
cidence and progression rates, using Kaplan-Meier prod uct-limit estimates 10 to account for the variation in in dividual follow-up periods. Incidence was defined as the development of new nuclear opacities in patients without such opacities at baseline (i.e., scores increased from nu clear opalescence < 2 at baseline to nuclear opalescence 2. 2 at follow-up in either eye). Progression was defined as the worsening of pre-existing nuclear opacities in pa tients who had these opacities at baseline (i.e., scores in creased from nuclear opalescence 2. 2 at baseline to higher follow-up scores in either eye). The rates were based on persons, rather than eyes. The analyses excluded two pa tients with nuclear opalescence scores of more than 6.2 in at least one eye at baseline and who did not have cat aract surgery during LSC, because these patients could not have progressed by our criteria. The rates were esti mated using a "worst eye" approach, that is, a patient was considered to have a nuclear change whenever an increase in nuclear opalescence occurred in either eye. The time when that change occurred was defined as the time when the opalescence change occurred in the first eye. For patients who had no change in scores or who had a retrogression in scores, the follow-up time interval was defined as the difference between the date of the baseline visit and the date of the last visit. (An alternative to the "worst eye" approach is to estimate the person-based rates by adjusting for the correlation between the status of the eyes from the same person. Such an approach will be the subject of a separate methodologic report.) A log-rank test was used to test for difference in rates. 10 Because most patients had their first LSC follow-up visit more than 2 years after baseline (median, 2. 7 years; mean ± standard deviation, 2.4 ± 1.1 years), the estimates of incidence or progression rates for the first year would be grossly underestimated and invalid. For that reason, only rates for years 2 to 5 are presented. Because rates may vary according to the presence of co-existing opacities, separate analyses were conducted for patients with and without other types of opacities at baseline.
Results Of the 764 patients who completed the first LSC follow up visit, 20 had no gradable nuclear photographs. Table 2 presents the baseline characteristics of the 744 patients with nuclear opalescence gradings. Median age of these patients was 65 years, and 45% were males. At baseline, approximately 53% had no nuclear opacities (score, < 2.0 in both eyes), and approximately 47% had nuclear opac ities (score, 2. 2 in either eye). Table 3 shows the frequency and type of nuclear opal escence change over the course of the study, without ac counting for patient follow-up times. It is based on the 698 patients with both lenses gradable and the 46 patients with only one lens gradable. Of the 698 patients with both lenses gradable, 75.5% had no nuclear change in either eye, whereas 22.6% ([64 + 94]1698) had increases in nu clear opacities in at least one eye and l. 7% ([2 + l 0]/698) had decreases in at least one eye. Only one patient had
707
Ophthalmology
Volume 103, Number 5, May 1996 (P = 0.00 I) higher for those 65 years of age and older
Table 2. Baseline Characteristics (n = 744) Characteristics Age (yrs) Mean± SD Median Males(%) Nuclear opacities at baseline No opacities (%) (LOCS III nuclear score < 2.0 in both eyes) Opacities (%) (LOCS III nuclear score ~ 2.0 in at least 1 eye) SO
=
63.8 ± 8.0 65.0 44.9 53.0 47.0
standard deviation.
an increase in scores in one eye, whereas the fellow eye had a decrease (Table 3). Of the 46 patients with only one lens gradable, 58.7% had no change; 39.1% had increases; and 2.2% had decreases in nuclear gradings (Table 3). These data included a total of88 patients who had cataract surgery in at least one eye after baseline. Seventy of these patients, or approximately four fifths, had increases in nuclear opalescence. Table 4 presents incidence rates for development of new nuclear opacities after 2 to 5 years of follow-up. As indicated in the Methods section, !-year rates were not presented because most patients began follow-up visits at least I year after baseline. The results show that among the 394 patients without nuclear opacities at baseline, the risk of new nuclear opacities developing during follow up was 6% to 8%. The incidence rates were significantly
than for younger patients. Similar incidence rates were found in males and females, although rates for males younger than 65 years of age tended to be lower than for females. Table 5 presents progression rates of pre-existing opac ities after 2 to 5 years of follow-up. Of the patients with pre-existing nuclear opacities at baseline, 35.8% had worsened after 2 years and 45.8% had worsened after 5 years of follow-up. Therefore, more than one half of pre existing opacities (54.2%) had not progressed 5 years later. The progression rate was slightly, but not significantly, higher in older ages. There was no difference in progression by sex. Figure I presents the variation in rates according to the presence of coexisting cortical and posterior subcap sular opacities. In the group of patients younger than 65 years of age, persons with these coexisting opacities had higher incidence (P = 0.00 I) and progression (P = 0.0 I) than others. In the group of patients 65 years of age and older, a higher progression (P = 0.03) also was suggested in persons with coexisting opacities.
Discussion This study provides longitudinal data on the growth of nuclear opacities in an older patient population. Among patients free of nuclear opacities at baseline, new opacities developed in 6% after 2 years of follow-up, an incidence rate that rose to 8% at 5 years. Incidence rates were higher in older patients and were similar for men and women.
Table 3. Percent Change of Nuclear Opacities 2 Visits Only No Surgery
3+ Visits
With Surgery
No Surgery
With Surgery
Total (%)
No. of Patients with Both Lenses Gradable at Baseline Increase, OU Increase and no change No change, OU Decrease and increase Decrease, OU Decrease and no change Subtotal
1 0
7 7 2 0 0 0
21 46 436 0
107
16
11 17
77
10
25 24 12 0 0 0
64 (9.2) 94 (13.5) 527 (75.5) 1 (0.1) 2 (0.3) 10 (1.4)
514
61
698 (100)
No. of Patients with One Lens Gradable at Baseline Increase No 1=hange Decrease
3 11 1
5 3 0
8 12 0
2 1 0
18 (39.1) 27 (58.7) 1 (2.2)
Subtotal
15
8
20
3
46 (100)
122
24
534
64
Total OU
708
=
both eyes.
744
Leske et al · Incidence and Progression of Nuclear Opacities Table 4. Age-Sex Specific and Overall Incidence Rates (%) for Nuclear Opacities Using the Product-Limit Method Year 2
Year 3
Year 4
Year 5
No.
Estimate ± SE
Estimate ± SE
Estimate ± SE
Estimate ± SE
<65 Vt"s Males Females Both sexes
121 127 248
1.7±1.2 5.0 ± 2.0 3.4 ± 1.2
2.7 ± 1.5 5.0 ± 2.0 3.8 ± 1.3
2.7 ± 1.5 6.0 ± 2.2 4.3 ± 1.4
4.6 ± 2.5 6.0 ± 2.2 5.3 ± 1.6
~ 65 Yrs Males Females Both sexes
67 79 146
10.8 ± 3.9 9.7 ± 3.5 10.3 ± 2.6
11.0 ± 3.9 11.2±3.7 11.0 ± 2.7
11.0±3.9 13.1 ± 4.1 12.0 ± 2.8
11.0±3.9 13.1 ± 4.1 12.0 ± 2.8
Total
394
5.9 ± 1.2
6.5 ± 1.3
7.1 ± 1.4
7.7 ± 1.5
SE
=
standard error.
In patients 65 years of age and older, rates ranged from I 0% to 12% and were approximately twice as high than for younger patients (Table 4). The progression of pre-existing opacities was much higher than the incidence of new opacities (Table 5). Of the patients who already had nuclear opacities at baseline, more than one third had worsened after 2 years and almost half had worsened after 5 years. These rates were not sig nificantly higher at older ages, suggesting that age is a more important factor for developing new nuclear opac ities than for worsening of pre-existing opacities. Although increases in nuclear opacification were common, more than one half of patients with opacities at baseline did not progress after 5 years of follow-up. However, incidence and progression rates were generally higher in patients who had other types of coexisting opacities (i.e., cortical, posterior subcapsular) than in other patients (Fig 1). All the incidence and progression analyses were conducted separately for patients with three or more visits and for those with only two visits. Similar patterns were found in both patient groups. Few epidemiologic data on the development and pro gression oflens opacities are available for comparison with our results. Podgor et al 11 used age-specific prevalence data from the Framingham Eye Study to estimate 5-year incidence rates for lens opacities and cataract. For lens opacities, the 5-year incidence estimates for patients 55, 60, 65, 70, and 75 years of age were 10%, 16%,23%,31 %, and 37%, respectively. Although the rates in this report are lower, they refer to nuclear cataract only and not to all cataract types. Taylor and Munoz 12 evaluated selected patients from a clinic who were 15 to 88 years of age and entered into a prospective study to assess the incidence and rate of progression of lens opacities. This study used a different cataract classification system, had a smaller sample size (n = 97), and a shorter duration of follow-up (median, 16 months). Hence, only first-year rates were estimated from their study, where similar rates of inci dence (range, 11 %-20%) and progression (range, 14%
15%) were reported. Differences in methodology between studies and the absence of a longer follow-up period pre clude any firm conclusions. Other studies have assessed cataract natural history us ing LOCS II, our earlier classification system. Because LOCS II is based on ordinal scores and has a less-refined grading scale than LOCS III, its use is likely to result in higher frequencies of increases and decreases in scores. In a follow-up study of 50 patients with age-related cat aract and 17 control subjects, Magno et al 13 reported a 6 month incidence of 22% and progression of 50% for nu clear cataract using the LOCS II system. The Italian American Study of the Natural History ofCataract, a large study that paralleled our design in following a clinic-based case-control study, reported incidence and progression for all opacity types. 14 Although they used the LOCS II classification system and had a different definition of change, their estimates on the incidence of nuclear cataract also indicated higher rates with older age. That study also found that progression was much higher than incidence for each type of opacity. For example, the 3-year cumu lative incidence and progression rates of nuclear cataract for persons 65 to 74 years of age were 6.5% and 67.0%, respectively. However, their rates of progression, as well as of retrogression, were considerably higher than our es timates. In addition to the different classification system used, these differences could be partially due to our more stringent definition for change (i.e., in addition to 2 or more consecutive changes in scores, our confirmed change method required a change in scores from baseline to the last visit). In addition, because of the small number of persons with moderate or advanced opacities, the Italian American study considered progression to exist in patients who started with early opacities, whereas our study based progression rates on persons who had higher nuclear opalescence scores at baseline (i.e., between 2.0 and 6.2). In addition to providing data on cataract growth in a patient population, our results contribute information on
709
Ophthalmology
Volume 103, Number 5, May 1996
Table 5. Age-Sex Specific and Overall Progression Rates(%) for Nuclear Opacities Using the Product Limit Method Year 2
Year3
Year 4
Year 5 Estimate ± SE
No.
Estimate ± SE
Estimate ± SE
Estimate ± SE
54 55 109
36.6 ± 6.7 28.5 ± 6.2 32.6 ± 4.6
38.9 ± 6.9 30.4 ± 6.4 34.7 ± 4.7
45.6 ± 7.6 32.6 ± 6.5 38.9 ± 5.0
45.6 ± 7.6 38.5 ± 7.2 42.1 ± 5.2
Males Females Both sexes
92 149 241
38.8 ± 5.1 36.2 ± 4.1 37.2 ± 3.2
41.1 ± 5.2 40.5 ± 4.3 40.7 ± 3.3
42.5 ± 5.3 44.3 ± 4.4 43.6 ± 3.4
47.3 ± 5.9 47.8 ± 4.6 47.5 ± 3.6
Total
350
35.8 ± 2.6
38.8 ± 2.7
42.1 ± 2.8
45.8 ± 3.0
<65 Yrs Males Females Both sexes ~ 65
SE
=
Yrs
standard error.
the methods to measure cataract growth in follow-up studies. To assess cataract growth, it is essential to use a grading system that is sensitive enough to detect lens changes over time, but not overly influenced by grading scales, grader variation, or difficulty to apply. Bailey et al 15 addressed the various issues involved in using grading scales that are too coarse. Based on our experience with developing and testing various classification methods, we decided to use LOCS III as the method of choice for as sessing natural history in LSC. An improved and ex panded LOCS system, LOCS III uses a decimal, rather than an ordinal scale, thus allowing better quantification than other methods. It also extends the scale for nuclear opalescence, so that early stages of nuclear change are better represented. In addition, LOCS III uses scaling in tervals, which are equal on a computerized image analysis system, for measuring nuclear opacification. The results presented here confirm the ability of LOCS III to detect longitudinal changes in nuclear opacities. In addition to grading scales, methodologic issues con cerning the measurement ofcataract growth are complex. Changes in lens gradings may reflect (1) "noise" inherent to the grading methods, which are dependent on factors such as grader variation, photographic techniques or pho tographic artifacts; (2) true biologic retrogression or pro gression; or (3) a combination of these components. We explored several methods to enhance the validity of LSC results. To evaluate the possible effects of grading "noise," the LSC design included ongoing quality control evalu ations that measured intragrader and intergrader repro ducibility. These evaluations showed that a highly satis factory reproducibility was achieved. The issue of tem poral drift in gradings also was considered. Such drift is difficult to avoid if photographs are graded over several years of follow-up. This problem was addressed by grading all the photographs for each patient during the same LOCS III grading session. In addition, all gradings for all patients were completed over a 6-month period, which was sched uled at the end of the study.
710
Our experience with the repeated gradings ofphotographs using LOCS III permitted us to determine the expected vari ability in gradings. This information allowed the use of a pre-set threshold, which had to be exceeded to define true change. Additionally, because most patients had three or more visits, it was possible to define confirmed change, using all the longitudinal data, to assess natural history for most of the cohort. In addition to being based on several data points, this corifirmed change method has the advantage of measuring consistent change over time, thus further reducing "noise." As a consequence, results based on confirmed change would be more conservative than those based on two data points alone. This trend is suggested by data in Table 3, where patients with three or more visits have some what lower percent change than patients with two visits only. This trend is not apparent in patients with one lens only, but the numbers in the table are small. In addition to the confirmed change method, we used an additional method to assess cataract changes, which was based on regression slopes of LOCS III scores over time. The results obtained (data not shown) were very similar. Percent increase was higher in patients with one grad able lens than in those with both gradable lenses (Table 3). One possible explanation is that patients with one gradable lens had a cataract extraction in the other eye and, therefore, had more progression because of the cor relation between fellow eyes. As also shown in Table 3, only one patient had discrepant results in both eyes (i.e., 1 eye had an increase in scores, whereas the fellow eye showed a decrease). Of the remaining patients, 23.7% ([94 + 64 + 18]/744) had increases in nuclear opalescence in at least one eye, whereas only 1.7% ([2 + 10 + 1]/744) had decreases in at least one eye over the entire follow-up period. Retrogression in gradings also has been re ported in several studies.' 2- 14•16 Our results suggest that true biologic retrogression of nuclear opalescence is possible, al though infrequent. The finding of higher retrogression rates than those reported here could be due to the coarseness of the grading scales used by other studies.
Leske et al · Incidence and Progression of Nuclear Opacities 70 00
p=.001 *
50
1:40 !! Q. 30
•
•
I
20
N=36
10
I
0
A
..a.e9S
'2.
;~ '\
..a.e9S
.
I N=212
'\e9S
~
70~--------------------------------~
eo
p=.32
50
20
10
.•.
*
r-:.
?.
N=55 I.-------~
0~----~--------~------~------~
c
10
N=91
~'2.
..a.e
0~--------------~------~------~
D
~'2.
'\e
Figure 1. Incidence and progression rates for nuclear opalescence among persons with and without other opacities. A, cumulative incidence rate: 64 or younger. B, cumulative progression rate: 64 or younger. C, cumulative incidence rate: 64 or older. D, cumulative progression rate: 65 or older. Straight lines = with other opacities at baseline; broken lines = without other opacities at baseline; asterisks = P of chi-square for log-rank test.
Another factor affecting the estimates of change is the occurrence ofcataract surgery. From Table 3, 88 patients, or approximately 12% of the LSC population, had cataract surgery in at least one eye during follow-up. Because these surgeries were performed by many different ophthalmol ogists and standardization of cataract classification could not be ensured, we assumed that nuclear opalescence in creased when nuclear cataract was the most severe type present at the last visit before surgery. Although such as sumption could introduce a classification bias, the exclu sion of all the patients who had cataract surgery would lead to underestimation of progression rates. Thus, if cat aract surgery were not considered and results were based only on photographic gradings, only 44% (rather than 80%) of the patients who underwent surgery would have been classified as having incidence or progression. The LSC is based on older patients originally identified in a hospital-based study several years earlier. As such,
population representativeness is not claimed or intended by the LSC and is outside the scope of the study. Although patients eligible for follow-up maintained high partici pation in LSC, losses between visits were inevitable, given the age and composition of the cohort. These losses would affect our results if the characteristics of persons lost to follow-up were very different from others (e.g., rates would be underestimated if some nonparticipants were much older or had more cataract surgery than those who re mained). These features of our study population must be considered in interpreting the results. When planning studies on similar clinic-based populations, however, the estimated incidence and progression rates provide much needed data. For example, our results could be used to plan intervention studies to evaluate the effect of anti cataract agents on nuclear progression. Such studies might lead to medical and/or preventive means to decelerate current rates of cataract growth.
711
Ophthalmology
Volume 103, Number 5, May 1996
Appendix Members ofthe Longitudinal Study of Cataract Group: M. Cris tina Leske, MD, MPH, Principal Investigator; University Medical Center at Stony Brook, Stony Brook, New York: M. Cristina Leske, MD, MPH, Ho Cheung, MS, Judith M. Greene, MPH, Qimei He, PhD, Phyllis Neuschwender, Elinor R. Schoenfeld, PhD, Suh-Yuh Wu, MA; The Center for Ophthalmic Research, Boston, Massachusetts: Leo T. Chylack, Jr., MD, Director, Margarett Baker, Laura Bury, BS, COA, Judith Friend, MA, Grazyna Jakubicz, MD, Mohammad Karabassi, MD, Patricia Khu, MD, Judith Libman, BA, Betty McDonald, David Singer, BS, John Wolfe, PhD.
References I. International Agency for the Prevention of Blindness. In:
2. 3.
4. 5.
712
Wilson J, ed. World Blindness and Its Prevention. Oxford; New York: Oxford University Press, Oxford, England, 1984; vol. 2. To restore sight: the global conquest of cataract blindness. New York: Helen Keller International, 1986;5. Report of the Cataract Panel. Vision research: a national plan, 1983-1987: 1987 evaluation and update. National Advisory Eye Council. Bethesda, MD: U.S. Dept. of Health and Human Services, Public Health Service, National In stitute of Health, 1987; vol. 2, pt (NIH publication no. 87 2755). Leske MC, Sperduto RD. The epidemiology of senile cat aracts: a review. Am J Epidemio11983;118:152-65. Chylack LT Jr, Wolfe JK, Singer DM, et a!. for the Lon gitudinal Study of Cataract Study Group. The Lens Opacities Classification System III. Arch Ophthalmol 1993; 111:831 6.
6. Leske MC, Chylack LT Jr, Wu S-Y, the Lens Opacities Case-Control Study Group. The Lens Opacities Case-control Study: risk factors for cataract. Arch Ophthalmol 1991; 109: 244-51. 7. Ferris FL III, Kassoff A, Bresnick GH, Bailey I. New visual acuity charts for clinical research. Am J Ophthalmol 1982;94:91-6. 8. Ferris FL III, Sperduto RD. Standardized illumination for visual acuity testing in clinical research. Am J Ophthalmol 1982;94:97-8. 9. Pelli DG, Robson JG, Wilkins AJ. The design of a new letter chart for measuring contrast sensitivity. Clin Vis Sci 1988;2:187-99. 10. Kalbfleisch JD, Prentice RL. The statistical analysis offailure time data. New York: John Wiley & Sons, Inc, 1980;143 62. 11. Podgor MJ, Leske MC, Ederer F. Incidence estimates for lens changes, macular changes, open-angle glaucoma and diabetic retinopathy. Am J Epidemiol1983;118:206 12. 12. Taylor HR, Munoz B. The incidence and progression of lens opacities. Aus N Z J Ophthalmol 1991;19:353
6.
13. Magno BV, Datiles MB III, Lasa SM. Senile cataract pro gression studies using the Lens Opacities Classification Sys tem II. Invest Ophthalmol Vis Sci 1993;34:2138-41. 14. The Italian-American Cataract Study Group. Incidence and progression of cortical, nuclear, and posterior subcapsular cataracts. Am J Ophthalmol 1994;118:623-31. 15. Bailey IL, Bullimore MA, Raasch TW, Taylor HR. Clinical grading and the effects of scaling. Invest Ophthalmol Vis Sci 1991 ;32:422-32. 16. West SK, Munoz B, Wang F, Taylor H. Measuring pro gression of lens opacities for longitudinal studies. Curr Eye Res 1993;12: 123-32.
Age-related cataract is a highly prevalent condition, which is the major cause of visual impairment. 1- 4 In many countries, surgical resources are insufficient to cope with the annual increment, much less the backlog, of unop erated cataracts. 1•2 A preventive or medical approach that Originally received: February 7, 1995. Revision accepted: February 8, 1996. 1 School of Medicine, SUNY, Stony Brook, New York. 2 Harvard Medical School Brigham and Women's Hospital and Center for Ophthalmic Research, Boston. *Members are listed in the Appendix section at the end of this article. Supported by National Eye Institute grant R01EY08291, Bethesda, Maryland. Reprint requests to M. Cristina Leske, MD, MPH, Department of Pre ventive Medicine, SUNY, Z = 8036, Stony Brook, NY 11794-8036.
could delay the need for surgery by 10 years would de crease the annual number of cataract surgeries in the United States by 45%. 3 Such a goal, however, cannot be achieved until more data are available on the causes and natural history of lens opacities. Cataract formation may involve single or multiple re gions of the lens, and each region may opacify at its own rate. However, very little epidemiologic data are available on the development and progression of the various types of lens opacities. 4 The Longitudinal Study of Cataract (LSC) represented a first step toward obtaining such data by providing basic natural history information on longi tudinal changes in the lens. Its major aims were to (1) measure the growth rates of nuclear, cortical, and posterior subcapsular opacities in a clinic-based population, (2) as sess and compare various qualitative and quantitative
705
Ophthalmology
Volume 103, Number 5, May 1996
methods to document changes in opacities and color, and (3) evaluate risk factors. This report presents the estimated incidence rates of new nuclear opacities and progression rates of pre-existing nuclear opacities in the LSC popu lation, using the Lens Opacities Classification System III (LOCS III). 5 The incidence and progression of cortical and posterior subcapsular opacities will be presented in a separate report.
Table 1. Participation and Time Intervals by Study Visit Days Since Last Visit Follow-up Visit
1 2 3
Methods Data Collection The LSC (1989-1993) was an epidemiologic study of the natural history of lens opacities, which followed a subset of participants from the Lens Opacities Case-control Studl ( 1985-1988). Surviving case-control study partic ipants and control subjects were eligible for the LSC unless they had cataract surgery in both eyes, missing or un gradable lens photographs, or were unable to complete yearly follow-up visits. The first follow-up visit for LSC started in September 1989, and subsequent follow-up visits continued at an average of 1-year intervals over a 4-year period. All participants provided informed consent and followed a standardized protocol designed for the study. Each LSC visit included a comprehensive ophthal mologic examination, which determined best-corrected visual acuity after the Early Treatment of Diabetic Reti nopathy Study protocol,7·8 contrast sensitivity using the Pelli-Robson test chart, 9 ocular pathology, and the cause of visual acuity loss, if any. Patient questionnaires were administered to update information on medical history, personal practices, and eye problems, including cataract surgery. Patients who had had unilateral cataract surgery continued to be followed to provide data on fellow eyes. Standardized color slit-lamp (Carl Zeiss, Oberkochen, Germany), retroillumination (Neitz CTR, Neitz Instru ments Co, Torrance, CA), and Scheimpflug (Topcon SL 45, Topcon America Corp, Paramus, NJ) photographs of the lens were taken. Photographs were graded according to the LOCS III method following published protocols. 5 Of the projected 800 eligible case-control participants, 764 (96%) completed the first follow-up visit (Table 1). Subsequently, 621 (81 %) patients remained eligible for at least one additional follow-up visit. Given the advanced age of the cohort (median, 67 years), 143 patients could not be followed due to death or severe illness (n = 60), cataract surgery in both eyes (n = 26), and relocation (n = 57). An additional 16 patients who completed visit 2, as well as 39 who completed visit 3, could not be followed further because the study ended. Among those remaining eligible, the number of patient refusals was 63 for visit 2, 44 for visit 3, and 30 for visit 4, resulting in a total number of patients followed of 618, 524, and 429 for those visits for participation rates of91 %, 92%, and 93%, respectively (Table 1). The average time interval between LSC follow-up visits was approximately 1 year, with standard deviations rang ing from I to 3 months (Table 1). Comparisons were made
706
4 SD
=
Participants/ Eligible
(% eligible)
Mean± SD (median)
764 618/681 524/568 429/459
(91) (92) (93)
331 ± 88 (320) 367 ± 65 (364) 363 ± 34 (364)
standard deviation.
between patients who were lost to follow-up after two visits and those who remained in the study for three or more visits. No differences were found between these groups with respect to their age, sex, and baseline nuclear opalescence status. Classification The LOCS III protocol was followed to grade the type and degree of lens opacification from lens photographs taken with maximal dilatation. 5 A full description and evaluation of the LOCS III method has been published elsewhere. 5 This classification system is based on six slit lamp standard images for grading nuclear opalescence, five retroillumination standard images for grading cortical, and five retroillumination images for grading posterior subcapsular cataract. A decimal scale is used to grade in creasing degrees of lens opacification, which ranges from 0.1 (most clear) to 5.9 (most advanced) for cortical and posterior subcapsular cataract or to 6.9 (most advanced) for nuclear opalescence. For this report, nuclear opacities were considered present if the nuclear opalescence score were 2 or more. All gradings were completed by the same two independent masked graders, and discrepancies were resolved by consensus. Nuclear grading sessions were conducted separately from cortical and posterior subcap sular grading sessions. To avoid drift in grading and to obtain a consistent classification, all photographs for all study visits completed by each patient were ordered ran domly in a carousel and graded at the same session. These gradings were conducted during a 6-month period at the end of the study. Intragrader and intergrader agreement were evaluated by regrading a quality control set of 155 sets of photographs. Intraclass correlation coefficients for agreement within and between observers were above 0.96 for all cataract types. The intergrader and intragrader variation in the LOCS III decimal scale was assessed by 95% tolerance limits, 5 which ranged from ±0.4 to ±0.6 for nuclear opalescence, ±0.6 to ± 1.0 for cortical, and ±0.6 to ±0.9 for posterior subcapsular opacities. The quality control set also was circulated continuously for grading throughout the LOCS III photograph grading ses sions. No signs of drift were detected.
Leske et al · Incidence and Progression of Nuclear Opacities
Analysis Eye-based Analyses The initial analyses evaluated the presence of changes in nuclear opalescence for each eye. To allow for "noise" due to grading or photograph variation, a change in grad ing was defined as present ifthe difference between nuclear opalescence scores at any two visits exceeded a given threshold. This threshold was defined as a difference in nuclear opalescence scores of 0. 7 or more in the LOCS III decimal scale. The threshold was based on the highest variation observed in the 95% tolerance limits for nuclear gradings, as derived both from the quality control results from this study (limits of ±0.4 to ±0.6) and from previous LOCS III reproducibility evaluations. 5 Therefore, a change in opalescence was considered present ifthe nuclear opal escence score at one visit was at least 0. 7 higher (or lower) than a previous nuclear opalescence score for that eye. For eyes of patients with only two visits, increases (or decreases) in opalescence at follow-up were simply defined according to these criteria, that is, whenever the difference between the two available nuclear opalescence scores was at least 0.7. For eyes of the majority of patients, who had three or more visits, all available nuclear scores were used to define confirmed increases (or decreases) throughout the follow-up period. The definition of confirmed increase was (1) a difference in scores of at least 0. 7 from baseline to the last visit, plus (2) a similar increase in nuclear scores between any two LSC visits, which was either followed by another increase in scores or by no change in scores at subsequent visits. The definition of confirmed decrease was the same, if the scores retrogressed over the follow up period. If these criteria for confirmed increase or de crease were not met, eyes were classified as having no change in opalescence. Time of the change, if any, was defined as the midpoint between the follow-up visit where the change was first observed and its preceding visit. For eyes that had cataract surgery, an increase in opal escence also was considered present if nuclear opacities were the most severe cataract type (i.e., the nuclear score was proportionally higher than the score for cortical and posterior subcapsular opacities) at the last visit before sur gery. Otherwise, no change in nuclear opalescence was assumed. The time of increase was defined as the mid year of the cataract surgery; if the year was unknown, the time of increase was defined as the midpoint between the last visit before surgery and the date when cataract surgery was reported.
Patient-based Analyses After determining increases (or decreases) for each eye, the subsequent analyses were patient-based. As a first step, we determined the percent change in nuclear scores over the entire period of the study, regardless of the length of time of follow-up. Separate analyses were conducted for persons with both lenses and a single lens, as well as for patients with two visits only, three or more visits, and with cataract surgery. As a second step, we estimated in
cidence and progression rates, using Kaplan-Meier prod uct-limit estimates 10 to account for the variation in in dividual follow-up periods. Incidence was defined as the development of new nuclear opacities in patients without such opacities at baseline (i.e., scores increased from nu clear opalescence < 2 at baseline to nuclear opalescence 2. 2 at follow-up in either eye). Progression was defined as the worsening of pre-existing nuclear opacities in pa tients who had these opacities at baseline (i.e., scores in creased from nuclear opalescence 2. 2 at baseline to higher follow-up scores in either eye). The rates were based on persons, rather than eyes. The analyses excluded two pa tients with nuclear opalescence scores of more than 6.2 in at least one eye at baseline and who did not have cat aract surgery during LSC, because these patients could not have progressed by our criteria. The rates were esti mated using a "worst eye" approach, that is, a patient was considered to have a nuclear change whenever an increase in nuclear opalescence occurred in either eye. The time when that change occurred was defined as the time when the opalescence change occurred in the first eye. For patients who had no change in scores or who had a retrogression in scores, the follow-up time interval was defined as the difference between the date of the baseline visit and the date of the last visit. (An alternative to the "worst eye" approach is to estimate the person-based rates by adjusting for the correlation between the status of the eyes from the same person. Such an approach will be the subject of a separate methodologic report.) A log-rank test was used to test for difference in rates. 10 Because most patients had their first LSC follow-up visit more than 2 years after baseline (median, 2. 7 years; mean ± standard deviation, 2.4 ± 1.1 years), the estimates of incidence or progression rates for the first year would be grossly underestimated and invalid. For that reason, only rates for years 2 to 5 are presented. Because rates may vary according to the presence of co-existing opacities, separate analyses were conducted for patients with and without other types of opacities at baseline.
Results Of the 764 patients who completed the first LSC follow up visit, 20 had no gradable nuclear photographs. Table 2 presents the baseline characteristics of the 744 patients with nuclear opalescence gradings. Median age of these patients was 65 years, and 45% were males. At baseline, approximately 53% had no nuclear opacities (score, < 2.0 in both eyes), and approximately 47% had nuclear opac ities (score, 2. 2 in either eye). Table 3 shows the frequency and type of nuclear opal escence change over the course of the study, without ac counting for patient follow-up times. It is based on the 698 patients with both lenses gradable and the 46 patients with only one lens gradable. Of the 698 patients with both lenses gradable, 75.5% had no nuclear change in either eye, whereas 22.6% ([64 + 94]1698) had increases in nu clear opacities in at least one eye and l. 7% ([2 + l 0]/698) had decreases in at least one eye. Only one patient had
707
Ophthalmology
Volume 103, Number 5, May 1996 (P = 0.00 I) higher for those 65 years of age and older
Table 2. Baseline Characteristics (n = 744) Characteristics Age (yrs) Mean± SD Median Males(%) Nuclear opacities at baseline No opacities (%) (LOCS III nuclear score < 2.0 in both eyes) Opacities (%) (LOCS III nuclear score ~ 2.0 in at least 1 eye) SO
=
63.8 ± 8.0 65.0 44.9 53.0 47.0
standard deviation.
an increase in scores in one eye, whereas the fellow eye had a decrease (Table 3). Of the 46 patients with only one lens gradable, 58.7% had no change; 39.1% had increases; and 2.2% had decreases in nuclear gradings (Table 3). These data included a total of88 patients who had cataract surgery in at least one eye after baseline. Seventy of these patients, or approximately four fifths, had increases in nuclear opalescence. Table 4 presents incidence rates for development of new nuclear opacities after 2 to 5 years of follow-up. As indicated in the Methods section, !-year rates were not presented because most patients began follow-up visits at least I year after baseline. The results show that among the 394 patients without nuclear opacities at baseline, the risk of new nuclear opacities developing during follow up was 6% to 8%. The incidence rates were significantly
than for younger patients. Similar incidence rates were found in males and females, although rates for males younger than 65 years of age tended to be lower than for females. Table 5 presents progression rates of pre-existing opac ities after 2 to 5 years of follow-up. Of the patients with pre-existing nuclear opacities at baseline, 35.8% had worsened after 2 years and 45.8% had worsened after 5 years of follow-up. Therefore, more than one half of pre existing opacities (54.2%) had not progressed 5 years later. The progression rate was slightly, but not significantly, higher in older ages. There was no difference in progression by sex. Figure I presents the variation in rates according to the presence of coexisting cortical and posterior subcap sular opacities. In the group of patients younger than 65 years of age, persons with these coexisting opacities had higher incidence (P = 0.00 I) and progression (P = 0.0 I) than others. In the group of patients 65 years of age and older, a higher progression (P = 0.03) also was suggested in persons with coexisting opacities.
Discussion This study provides longitudinal data on the growth of nuclear opacities in an older patient population. Among patients free of nuclear opacities at baseline, new opacities developed in 6% after 2 years of follow-up, an incidence rate that rose to 8% at 5 years. Incidence rates were higher in older patients and were similar for men and women.
Table 3. Percent Change of Nuclear Opacities 2 Visits Only No Surgery
3+ Visits
With Surgery
No Surgery
With Surgery
Total (%)
No. of Patients with Both Lenses Gradable at Baseline Increase, OU Increase and no change No change, OU Decrease and increase Decrease, OU Decrease and no change Subtotal
1 0
7 7 2 0 0 0
21 46 436 0
107
16
11 17
77
10
25 24 12 0 0 0
64 (9.2) 94 (13.5) 527 (75.5) 1 (0.1) 2 (0.3) 10 (1.4)
514
61
698 (100)
No. of Patients with One Lens Gradable at Baseline Increase No 1=hange Decrease
3 11 1
5 3 0
8 12 0
2 1 0
18 (39.1) 27 (58.7) 1 (2.2)
Subtotal
15
8
20
3
46 (100)
122
24
534
64
Total OU
708
=
both eyes.
744
Leske et al · Incidence and Progression of Nuclear Opacities Table 4. Age-Sex Specific and Overall Incidence Rates (%) for Nuclear Opacities Using the Product-Limit Method Year 2
Year 3
Year 4
Year 5
No.
Estimate ± SE
Estimate ± SE
Estimate ± SE
Estimate ± SE
<65 Vt"s Males Females Both sexes
121 127 248
1.7±1.2 5.0 ± 2.0 3.4 ± 1.2
2.7 ± 1.5 5.0 ± 2.0 3.8 ± 1.3
2.7 ± 1.5 6.0 ± 2.2 4.3 ± 1.4
4.6 ± 2.5 6.0 ± 2.2 5.3 ± 1.6
~ 65 Yrs Males Females Both sexes
67 79 146
10.8 ± 3.9 9.7 ± 3.5 10.3 ± 2.6
11.0 ± 3.9 11.2±3.7 11.0 ± 2.7
11.0±3.9 13.1 ± 4.1 12.0 ± 2.8
11.0±3.9 13.1 ± 4.1 12.0 ± 2.8
Total
394
5.9 ± 1.2
6.5 ± 1.3
7.1 ± 1.4
7.7 ± 1.5
SE
=
standard error.
In patients 65 years of age and older, rates ranged from I 0% to 12% and were approximately twice as high than for younger patients (Table 4). The progression of pre-existing opacities was much higher than the incidence of new opacities (Table 5). Of the patients who already had nuclear opacities at baseline, more than one third had worsened after 2 years and almost half had worsened after 5 years. These rates were not sig nificantly higher at older ages, suggesting that age is a more important factor for developing new nuclear opac ities than for worsening of pre-existing opacities. Although increases in nuclear opacification were common, more than one half of patients with opacities at baseline did not progress after 5 years of follow-up. However, incidence and progression rates were generally higher in patients who had other types of coexisting opacities (i.e., cortical, posterior subcapsular) than in other patients (Fig 1). All the incidence and progression analyses were conducted separately for patients with three or more visits and for those with only two visits. Similar patterns were found in both patient groups. Few epidemiologic data on the development and pro gression oflens opacities are available for comparison with our results. Podgor et al 11 used age-specific prevalence data from the Framingham Eye Study to estimate 5-year incidence rates for lens opacities and cataract. For lens opacities, the 5-year incidence estimates for patients 55, 60, 65, 70, and 75 years of age were 10%, 16%,23%,31 %, and 37%, respectively. Although the rates in this report are lower, they refer to nuclear cataract only and not to all cataract types. Taylor and Munoz 12 evaluated selected patients from a clinic who were 15 to 88 years of age and entered into a prospective study to assess the incidence and rate of progression of lens opacities. This study used a different cataract classification system, had a smaller sample size (n = 97), and a shorter duration of follow-up (median, 16 months). Hence, only first-year rates were estimated from their study, where similar rates of inci dence (range, 11 %-20%) and progression (range, 14%
15%) were reported. Differences in methodology between studies and the absence of a longer follow-up period pre clude any firm conclusions. Other studies have assessed cataract natural history us ing LOCS II, our earlier classification system. Because LOCS II is based on ordinal scores and has a less-refined grading scale than LOCS III, its use is likely to result in higher frequencies of increases and decreases in scores. In a follow-up study of 50 patients with age-related cat aract and 17 control subjects, Magno et al 13 reported a 6 month incidence of 22% and progression of 50% for nu clear cataract using the LOCS II system. The Italian American Study of the Natural History ofCataract, a large study that paralleled our design in following a clinic-based case-control study, reported incidence and progression for all opacity types. 14 Although they used the LOCS II classification system and had a different definition of change, their estimates on the incidence of nuclear cataract also indicated higher rates with older age. That study also found that progression was much higher than incidence for each type of opacity. For example, the 3-year cumu lative incidence and progression rates of nuclear cataract for persons 65 to 74 years of age were 6.5% and 67.0%, respectively. However, their rates of progression, as well as of retrogression, were considerably higher than our es timates. In addition to the different classification system used, these differences could be partially due to our more stringent definition for change (i.e., in addition to 2 or more consecutive changes in scores, our confirmed change method required a change in scores from baseline to the last visit). In addition, because of the small number of persons with moderate or advanced opacities, the Italian American study considered progression to exist in patients who started with early opacities, whereas our study based progression rates on persons who had higher nuclear opalescence scores at baseline (i.e., between 2.0 and 6.2). In addition to providing data on cataract growth in a patient population, our results contribute information on
709
Ophthalmology
Volume 103, Number 5, May 1996
Table 5. Age-Sex Specific and Overall Progression Rates(%) for Nuclear Opacities Using the Product Limit Method Year 2
Year3
Year 4
Year 5 Estimate ± SE
No.
Estimate ± SE
Estimate ± SE
Estimate ± SE
54 55 109
36.6 ± 6.7 28.5 ± 6.2 32.6 ± 4.6
38.9 ± 6.9 30.4 ± 6.4 34.7 ± 4.7
45.6 ± 7.6 32.6 ± 6.5 38.9 ± 5.0
45.6 ± 7.6 38.5 ± 7.2 42.1 ± 5.2
Males Females Both sexes
92 149 241
38.8 ± 5.1 36.2 ± 4.1 37.2 ± 3.2
41.1 ± 5.2 40.5 ± 4.3 40.7 ± 3.3
42.5 ± 5.3 44.3 ± 4.4 43.6 ± 3.4
47.3 ± 5.9 47.8 ± 4.6 47.5 ± 3.6
Total
350
35.8 ± 2.6
38.8 ± 2.7
42.1 ± 2.8
45.8 ± 3.0
<65 Yrs Males Females Both sexes ~ 65
SE
=
Yrs
standard error.
the methods to measure cataract growth in follow-up studies. To assess cataract growth, it is essential to use a grading system that is sensitive enough to detect lens changes over time, but not overly influenced by grading scales, grader variation, or difficulty to apply. Bailey et al 15 addressed the various issues involved in using grading scales that are too coarse. Based on our experience with developing and testing various classification methods, we decided to use LOCS III as the method of choice for as sessing natural history in LSC. An improved and ex panded LOCS system, LOCS III uses a decimal, rather than an ordinal scale, thus allowing better quantification than other methods. It also extends the scale for nuclear opalescence, so that early stages of nuclear change are better represented. In addition, LOCS III uses scaling in tervals, which are equal on a computerized image analysis system, for measuring nuclear opacification. The results presented here confirm the ability of LOCS III to detect longitudinal changes in nuclear opacities. In addition to grading scales, methodologic issues con cerning the measurement ofcataract growth are complex. Changes in lens gradings may reflect (1) "noise" inherent to the grading methods, which are dependent on factors such as grader variation, photographic techniques or pho tographic artifacts; (2) true biologic retrogression or pro gression; or (3) a combination of these components. We explored several methods to enhance the validity of LSC results. To evaluate the possible effects of grading "noise," the LSC design included ongoing quality control evalu ations that measured intragrader and intergrader repro ducibility. These evaluations showed that a highly satis factory reproducibility was achieved. The issue of tem poral drift in gradings also was considered. Such drift is difficult to avoid if photographs are graded over several years of follow-up. This problem was addressed by grading all the photographs for each patient during the same LOCS III grading session. In addition, all gradings for all patients were completed over a 6-month period, which was sched uled at the end of the study.
710
Our experience with the repeated gradings ofphotographs using LOCS III permitted us to determine the expected vari ability in gradings. This information allowed the use of a pre-set threshold, which had to be exceeded to define true change. Additionally, because most patients had three or more visits, it was possible to define confirmed change, using all the longitudinal data, to assess natural history for most of the cohort. In addition to being based on several data points, this corifirmed change method has the advantage of measuring consistent change over time, thus further reducing "noise." As a consequence, results based on confirmed change would be more conservative than those based on two data points alone. This trend is suggested by data in Table 3, where patients with three or more visits have some what lower percent change than patients with two visits only. This trend is not apparent in patients with one lens only, but the numbers in the table are small. In addition to the confirmed change method, we used an additional method to assess cataract changes, which was based on regression slopes of LOCS III scores over time. The results obtained (data not shown) were very similar. Percent increase was higher in patients with one grad able lens than in those with both gradable lenses (Table 3). One possible explanation is that patients with one gradable lens had a cataract extraction in the other eye and, therefore, had more progression because of the cor relation between fellow eyes. As also shown in Table 3, only one patient had discrepant results in both eyes (i.e., 1 eye had an increase in scores, whereas the fellow eye showed a decrease). Of the remaining patients, 23.7% ([94 + 64 + 18]/744) had increases in nuclear opalescence in at least one eye, whereas only 1.7% ([2 + 10 + 1]/744) had decreases in at least one eye over the entire follow-up period. Retrogression in gradings also has been re ported in several studies.' 2- 14•16 Our results suggest that true biologic retrogression of nuclear opalescence is possible, al though infrequent. The finding of higher retrogression rates than those reported here could be due to the coarseness of the grading scales used by other studies.
Leske et al · Incidence and Progression of Nuclear Opacities 70 00
p=.001 *
50
1:40 !! Q. 30
•
•
I
20
N=36
10
I
0
A
..a.e9S
'2.
;~ '\
..a.e9S
.
I N=212
'\e9S
~
70~--------------------------------~
eo
p=.32
50
20
10
.•.
*
r-:.
?.
N=55 I.-------~
0~----~--------~------~------~
c
10
N=91
~'2.
..a.e
0~--------------~------~------~
D
~'2.
'\e
Figure 1. Incidence and progression rates for nuclear opalescence among persons with and without other opacities. A, cumulative incidence rate: 64 or younger. B, cumulative progression rate: 64 or younger. C, cumulative incidence rate: 64 or older. D, cumulative progression rate: 65 or older. Straight lines = with other opacities at baseline; broken lines = without other opacities at baseline; asterisks = P of chi-square for log-rank test.
Another factor affecting the estimates of change is the occurrence ofcataract surgery. From Table 3, 88 patients, or approximately 12% of the LSC population, had cataract surgery in at least one eye during follow-up. Because these surgeries were performed by many different ophthalmol ogists and standardization of cataract classification could not be ensured, we assumed that nuclear opalescence in creased when nuclear cataract was the most severe type present at the last visit before surgery. Although such as sumption could introduce a classification bias, the exclu sion of all the patients who had cataract surgery would lead to underestimation of progression rates. Thus, if cat aract surgery were not considered and results were based only on photographic gradings, only 44% (rather than 80%) of the patients who underwent surgery would have been classified as having incidence or progression. The LSC is based on older patients originally identified in a hospital-based study several years earlier. As such,
population representativeness is not claimed or intended by the LSC and is outside the scope of the study. Although patients eligible for follow-up maintained high partici pation in LSC, losses between visits were inevitable, given the age and composition of the cohort. These losses would affect our results if the characteristics of persons lost to follow-up were very different from others (e.g., rates would be underestimated if some nonparticipants were much older or had more cataract surgery than those who re mained). These features of our study population must be considered in interpreting the results. When planning studies on similar clinic-based populations, however, the estimated incidence and progression rates provide much needed data. For example, our results could be used to plan intervention studies to evaluate the effect of anti cataract agents on nuclear progression. Such studies might lead to medical and/or preventive means to decelerate current rates of cataract growth.
711
Ophthalmology
Volume 103, Number 5, May 1996
Appendix Members ofthe Longitudinal Study of Cataract Group: M. Cris tina Leske, MD, MPH, Principal Investigator; University Medical Center at Stony Brook, Stony Brook, New York: M. Cristina Leske, MD, MPH, Ho Cheung, MS, Judith M. Greene, MPH, Qimei He, PhD, Phyllis Neuschwender, Elinor R. Schoenfeld, PhD, Suh-Yuh Wu, MA; The Center for Ophthalmic Research, Boston, Massachusetts: Leo T. Chylack, Jr., MD, Director, Margarett Baker, Laura Bury, BS, COA, Judith Friend, MA, Grazyna Jakubicz, MD, Mohammad Karabassi, MD, Patricia Khu, MD, Judith Libman, BA, Betty McDonald, David Singer, BS, John Wolfe, PhD.
References I. International Agency for the Prevention of Blindness. In:
2. 3.
4. 5.
712
Wilson J, ed. World Blindness and Its Prevention. Oxford; New York: Oxford University Press, Oxford, England, 1984; vol. 2. To restore sight: the global conquest of cataract blindness. New York: Helen Keller International, 1986;5. Report of the Cataract Panel. Vision research: a national plan, 1983-1987: 1987 evaluation and update. National Advisory Eye Council. Bethesda, MD: U.S. Dept. of Health and Human Services, Public Health Service, National In stitute of Health, 1987; vol. 2, pt (NIH publication no. 87 2755). Leske MC, Sperduto RD. The epidemiology of senile cat aracts: a review. Am J Epidemio11983;118:152-65. Chylack LT Jr, Wolfe JK, Singer DM, et a!. for the Lon gitudinal Study of Cataract Study Group. The Lens Opacities Classification System III. Arch Ophthalmol 1993; 111:831 6.
6. Leske MC, Chylack LT Jr, Wu S-Y, the Lens Opacities Case-Control Study Group. The Lens Opacities Case-control Study: risk factors for cataract. Arch Ophthalmol 1991; 109: 244-51. 7. Ferris FL III, Kassoff A, Bresnick GH, Bailey I. New visual acuity charts for clinical research. Am J Ophthalmol 1982;94:91-6. 8. Ferris FL III, Sperduto RD. Standardized illumination for visual acuity testing in clinical research. Am J Ophthalmol 1982;94:97-8. 9. Pelli DG, Robson JG, Wilkins AJ. The design of a new letter chart for measuring contrast sensitivity. Clin Vis Sci 1988;2:187-99. 10. Kalbfleisch JD, Prentice RL. The statistical analysis offailure time data. New York: John Wiley & Sons, Inc, 1980;143 62. 11. Podgor MJ, Leske MC, Ederer F. Incidence estimates for lens changes, macular changes, open-angle glaucoma and diabetic retinopathy. Am J Epidemiol1983;118:206 12. 12. Taylor HR, Munoz B. The incidence and progression of lens opacities. Aus N Z J Ophthalmol 1991;19:353
6.
13. Magno BV, Datiles MB III, Lasa SM. Senile cataract pro gression studies using the Lens Opacities Classification Sys tem II. Invest Ophthalmol Vis Sci 1993;34:2138-41. 14. The Italian-American Cataract Study Group. Incidence and progression of cortical, nuclear, and posterior subcapsular cataracts. Am J Ophthalmol 1994;118:623-31. 15. Bailey IL, Bullimore MA, Raasch TW, Taylor HR. Clinical grading and the effects of scaling. Invest Ophthalmol Vis Sci 1991 ;32:422-32. 16. West SK, Munoz B, Wang F, Taylor H. Measuring pro gression of lens opacities for longitudinal studies. Curr Eye Res 1993;12: 123-32.