Seven-Year Incidence of Age-Related Cataracts Among an Elderly Chinese Population in Shihpai, Taiwan: The Shihpai Eye Study

Abstract

Purpose.: To report the 7-year incidence of age-related cataracts in a metropolitan Chinese elderly population.

Methods.: The Shihpai Eye Study 2006 included 460 (55.8%) of 1361 participants in the 1999 baseline survey for a follow-up eye examination. Cataract was graded by one ophthalmologist using the Lens Opacities Classification System III.

Results.: The 7-year incidence of pure nuclear opacity was 18.5% (95% confidence interval [CI]: 12.7%–24.3%)]; 34.7% (95% CI: 27.6%–41.8%) for cortical opacity; and 0.6% (95% CI: 0%–1.8%) for posterior subcapsular opacity. On the other hand, 11.6% (95% CI: 6.8%–16.4%) participants developed both nuclear and cortical opacity, 3.5% (95% CI: 0.8%–6.2%) nuclear and posterior subcapsular opacity; and 1.7% (95% CI: 0%–3.6%) cortical and posterior subcapsular opacity. A total of 3.5% (95% CI: 0.8%–6.2%) of participants developed all three types of cataract. Nuclear cataract was significantly associated with current smoking (P = 0.04; relative risk [RR]: 2.05; 95% CI: 1.05–3.99). Males were almost half as likely (P = 0.04; RR: 0.57; 95% CI: 0.34–0.97) to develop cortical opacity, whereas participants with a history of diabetes were approximately twice as likely to develop cortical opacity (P = 0.05; RR 2.43; 95% CI: 1.02–5.81). A higher level of education (P = 0.03; RR: 0.40; 95% CI: 0.18–0.91) and a higher body mass index (P = 0.02; RR: 0.28; 95% CI: 0.10–0.79) was protective of posterior subcapsular cataract.

Conclusions.: This study indicated a high incidence of cortical opacities and lower incidence of nuclear opacities compared with Caucasians. Posterior subcapsular opacity incidence and cataract surgery rate were comparable.

Introduction

Age-related cataracts are a major cause of blindness worldwide. ^{1 –4} A mission of the World Health Organization VISION 2020: the Right to Sight is to eliminate avoidable blindness globally to ensure the best possible vision for all people, making cataracts a main target. Although not fatal, age-related cataracts cause visual impairment, ⁵ decrease quality of life, ⁶ and negatively impact daily and social activities. ⁷  

Few longitudinal studies provide information on the incidence of age-related cataracts, ^{8 –16} especially in the Asian population. ¹⁷ Moreover, the longitudinal effects of some predictors of age-related cataracts, such as body mass index (BMI) ^{18 –20} and increased systolic blood pressure, ^21,22 are controversial. The purpose of the present study was to investigate the incidence of different types of age-related cataracts in a metropolitan elderly Chinese population in Shihpai, Taipei, Taiwan, with follow-up at 7 years, and to identify risk factors for various types of age-related cataracts, including demographic variables, lifestyle, and medical history. 

Materials and Methods

The Shihpai Eye Study ²³ was a community-based, cross-sectional survey of vision and eye diseases among noninstitutionalized subjects aged 65 years and older in Shihpai, Taipei, Taiwan. Residents aged 65 years and older were identified using the household registration system. This system officially registers personal information such as date of birth, sex, and home address, as well as family members and relations. According to the official household registration in 1999, the total number of residents aged 65 years and older in Shihpai was 4750; 3746 persons were eligible, and 2045 of these were randomly selected to be invited to participate in the study. Of the 2045 subjects, 1361 (66.6%) participated in both the questionnaire and eye examination. The baseline examination was conducted between July 1, 1999, and December 31, 2000. Follow-up examination of the eye condition of the fixed cohort was conducted from March 25, 2006, and ended December 31, 2007. We planned to invite the 1361 participants in the baseline examination for the follow-up study. A structured questionnaire similar to the baseline survey ²³ was conducted by intensively trained interviewers. The questionnaire obtained information on demographics (age, sex, locality, marital status, and education); personal medical history; and lifestyle (smoking and alcohol intake). Participants' self-reported demographic data were assessed. Personal medical history was assessed by a checklist. Participants were asked whether they had been diagnosed by a physician with a chronic disease such as diabetes (yes/no). Cigarette smoking history was scored as smoker, passive smoker, ex-smoker, or never-smoker. Alcohol consumption was limited to wine and hard alcohol and was scored as no consumption (or frequency of alcohol consumption only once a week) or habit of alcohol consumption (frequency of alcohol consumption more than once a week). Women were also asked whether they ever used estrogen replacement therapy or other hormones (yes/no). Subjects that were interviewed were invited to participate in a comprehensive ophthalmic examination conducted at the Taipei Veterans General Hospital. Ophthalmologists conducted the examinations according to a standardized protocol. Informed consent was obtained from each subject after explaining the purpose and procedure of the study. The survey followed the tenets of the Declaration of Helsinki. 

This study was approved by the Institutional Review Board of the Taipei Veterans General Hospital. 

Definitions

Three major types of age-related cataracts (nuclear, cortical, and posterior subcapsular) were assessed based on the Lens Opacity Classification system III (LOCS III) by one ophthalmologist at the slit-lamp under maximum dilatation with tropicamide. Subjects were categorized as having an age-related cataract if any type of opacity with an LOCS III grade >2 was present in one or both eyes. We chose grade >2 as the cutoff, because this grade was used in our prevalence analysis ²³ and it is also the definition of cataracts adopted by previous major population-based studies. ^12,13,15 The intraobserver kappa statistics for reproducibility of grading cataracts (grade >2 versus grade ≤2 ) were 0.86 for cortical opacity, 0.84 for posterior subcapsular opacity, and 0.87 for nuclear opacity. 

Incidence was estimated by participants without any type of opacity (grade ≤2) in both eyes at baseline that developed lens opacity (grade >2) in either eye at the 7-year follow-up, and were categorized into pure form of lens opacity, mixture of two types of opacity, and presence of all three types of opacities. In cases of discordance between eyes, the eye with more types of lens opacity was taken as the cataract type for that participant. 

For comparison with other population-based studies, the development of any cortical, nuclear, or posterior subcapsular opacity (grade >2) in at least one eye among persons without that opacity type (grade ≤2) at baseline was also calculated. An eye that had one opacity type at baseline was considered to be at risk for the other incident opacity types. Participants who were pseudophakic or aphakic at baseline were excluded. Participants that had cataract surgery during this period were excluded in the incidence analysis, but were assessed separately for the cataract surgery rate. 

Height and weight were measured and BMI was calculated by the formula weight in kilograms divided by the square of the height in meters (kg/m²) and was dichotomized into overweight and not overweight according to World Health Organization criteria (overweight, ≥25 kg/m² versus not overweight, <25 kg/m²). Three consecutive seated blood pressure readings, at least 5 minutes apart, were taken from the right arm and averaged. Waist girth was measured at the minimum circumference, whereas hip girth was measured at the maximum circumference. The definition of regional or central obesity by waist-to-hip ratio was ≥0.92 for men or ≥0.88 for women. ²⁴  

Statistical Analysis

Dependent variables in the analysis were incidence of nuclear, cortical, or posterior subcapsular opacities. Independent variables tested were age (≥80 years vs. 72–79 years); sex (female versus male); education (high school and above versus secondary school and below); marital status (married and living with spouse versus single/separated/divorced/widowed), BMI (≥25 vs. <25); waist-to-hip ratio (high: waist-to-hip ratio ≥0.92 for males, ≥0.88 for females versus normal: <0.92 for males, <0.88 for females); systolic blood pressure (≥160 mm Hg versus <160 mm Hg); history of hypertension (yes/no); diabetes (yes/no); cardiovascular disease (yes/no); stroke (yes/no); smoking history (current smoking versus nonsmoking; ex-smoking versus nonsmoking); alcohol drinking (yes/no); and history of hormone use (yes/no) among women. Univariate analysis was performed to test for an association between each independent variable and dependent variable by χ² analysis. Generalized estimating equations were used to fit the best model for independent variables. Models were developed separately for each type of opacity—that is, those with grade ≤2 at baseline for one type of opacity that developed to grade >2 at follow-up, regardless of the status of the other two types of opacity. Another analysis was performed with the dependent variable as pure incident nuclear opacity and pure incident cortical opacity, controls were participants without any type of lens opacity (grade ≤2 for all three types of opacity) in both eyes at baseline and follow-up. For posterior subcapsular opacity, because incident case number was small, mixture of posterior subcapsular opacity with nuclear and/or cortical opacity were also included as cases. Sex and age as well as independent variables with a P value of <0.20 under univariate analysis were analyzed in the multivariate models. A P value of <0.05 was considered to be statistically significant. 

Statistical analysis was performed using commercial software (SAS 6.12; SAS Institute, Cary, NC). 

Results

Of the 1361 participants who attended the baseline examination, 205 (15.1%) residents died before they were interviewed, 301 (22.1%) residents moved away, and 31 (2.3%) were institutionalized, thus 824 (60.5%) subjects were eligible for the study, 725 (87.4%) of whom agreed to be interviewed for the questionnaire. Among those interviewed, 460 (55.8% of those eligible, or 39.8% of survivors) participated in the ophthalmic examination. A comparison of the demographics and some of the variables in subjects with and without the eye examination are shown in Table 1. Participants were younger (78.1 ± 4.1 years versus 80.4 ± 5.4 years, P < 0.001), more likely to be male (P < 0.01), married and living with spouse (P < 0.001), and were more highly educated (P < 0.001). Participants were less likely to have a history of stroke (P = 0.03) and more likely to be current smokers (P = 0.03). 

Table 1

 

Descriptive Characteristics of Participants in Shihpai, Taipei, Taiwan, 2006 to 2007

Characteristic	Participants, n = 460	Non participants, n = 265	P Value
Age, y
72–79	333	150	<0.001*
≥80	127	115
Sex
Male	304	128	<0.001*
Female	156	137
Education
Secondary education and below	251	191	<0.001*
High school and above	209	74
Marital status†
With spouse	365	191	0.03*
Without spouse	95	74
BMI, kg/m2
<25	292	158	0.30
≥25	168	107
History of hypertension
Yes	213	138	0.53
No	208	122
History of diabetes
Yes	85	53	0.99
No	331	206
History of cardiovascular disease
Yes	167	85	0.06
No	249	173
History of stroke
Yes	14	18	0.03*
No	407	241
Smoking (current vs. never)
Yes	43	14	0.03*
No	335	218
Ex-smoking (quit vs. never)
Yes	46	29	0.90
No	335	218
Alcohol
Yes	20	7	0.10
No	322	233

*   P < 0.05.

†  “With spouse” included participants who were married and living with spouse. “Without spouse” included participants who were single, separated, divorced, or widowed.

Among the participants, males were significantly more highly educated (P < 0.001) and married, living with their spouse (P < 0.001) than female participants. There was no significant difference among other demographic variables for participants. 

The number of participants without any type of opacity in both eyes at baseline that participated in the follow-up was 184. After excluding subjects who were unable to cooperate, declined slit-lamp biomicroscopy, or had undergone cataract surgery, information on 173 participants was assessed. A total of 32 (18.5%; 95% confidence interval [CI]: 12.7%–24.3%) participants had incident pure nuclear opacity; 60 participants (34.7%; 95% CI: 27.6%–41.8%) had incident pure cortical opacity; and 1 participant (0.6%; 95% CI: 0%–1.8%) developed incident pure posterior subcapsular opacity. On the other hand, 20 (11.6%; 95% CI: 6.8%–16.4%) participants had both nuclear and cortical opacity; 6 (3.5%; 95% CI: 0.8%–6.2%) had a mixture of nuclear and posterior subcapsular opacity; and 3 (1.7%; 95% CI: 0%–3.6%) participants developed both cortical and posterior subcapsular opacity; 6 (3.5%; 95% CI: 0.8%–6.2%) participants developed all three types of cataract after a 7-year follow-up. 

When analyzed according to eyes, 27.1% (95% CI: 22.2%–32.6%) of the right eye, 21.8% of the left eye (95% CI: 17.4%–27.0%), and 29.4% (95% CI: 24.6%–34.8%) for either eye developed nuclear opacity. For cortical cataract, the percentage was 44.5% (95% CI: 38.9%–50.3%) for the right eye, 41.0% (95% CI: 35.5%–46.7%) for the left eye, and 49.1% (95% CI: 44.3%–55.1%) for either eye. For posterior subcapsular cataract, 5.9% (95% CI: 3.8%–8.9%) of the right eye, 5.8% (95% CI: 3.7%–8.8%) for the left eye, and 8.1% (95% CI: 5.7%–11.3%) for either eye developed posterior subcapsular cataract. The distribution of cataracts among participants is shown in Table 2. During this period of time, 41 (10.6%; 95% CI: 7.9%–14.1%) right eyes and 38 (9.8%; 95% CI: 7.2%–13.2%) left eyes received cataract surgery. 

Table 2

 

Seven-Year Incidence of Each Type of Age-Related Cataract in Either Eye in Shihpai, Taipei, Taiwan, 2006 to 2007

Type of Cataract	Male		Female		Total
	n	%	n	%	n	%
Nuclear cataract	65/220	29.6	26/89	29.2	91/309	29.4
Cortical cataract	102/228	39.8	60/98	49.7	162/326	49.1
Posterior subcapsular cataract	20/252	7.9	10/120	8.3	30/372	8.1
Pseudophakia	27/255	10.6	30/132	22.7	57/387	14.7

In univariate analysis, nuclear cataract was significantly associated with smoking (P = 0.04). Females (P = 0.01) were more likely to have cortical cataracts than males. Moreover, alcohol drinking (P = 0.01) and a history of diabetes (P = 0.04) were also significantly associated with cortical opacity. For posterior subcapsular cataract, BMI (P = 0.01) was a significant predisposing factor (Table 3). 

Table 3

 

Univariate Analysis on Age-Related Cataracts Among Residents in Shihpai, Taipei, Taiwan, 2006 to 2007

Variable	Nuclear		Cortical		Posterior Subcapsular
	n	P Value	n	P Value	n	P Value
Age, y
72–79	71/236	0.66	116/243	0.23	20/280	0.25
80–93	20/73		46/83		10/92
Sex
Male	65/220	0.95	102/228	0.01*	20/252	0.90
Female	26/89		60/98		10/120
Education
≤Secondary	45/154	0.93	81/160	0.74	20/195	0.10
≥High school	46/155		81/166		10/177
BMI, kg/m2
<25	56/193	0.83	102/199	0.48	25/230	0.01*
≥25	35/116		60/127		5/142
Waist-to-hip ratio†
High	29/97	0.91	40/96	0.06	10/112	0.69
Not high	62/212		122/230		20/260
Systolic blood pressure, mm Hg
≥160	9/25	0.45	10/25	0.31	3/33	0.82
<160	82/284		152/301		27/339
Smoking
Current	23/56	0.04*	33/62	0.58	7/68	0.44
Never	52/195		99/201		17/230
Smoking
Past	15/54	0.87	28/59	0.81	6/67	0.67
Never	52/195		99/201		17/230
Alcohol
Yes	12/43	0.73	14/44	0.01*	5/49	0.58
No	76/249		140/266		24/306
Hypertension history
Yes	36/118	0.75	62/119	0.51	14/140	0.29
No	55/191		100/207		16/232
Diabetes history
Yes	10/38	0.69	21/31	0.04*	5/47	0.46
No	78/265		138/288		24/318
Cardiovascular disease history
Yes	25/80	0.62	35/83	0.11	10/92	0.17
No	64/226		126/241		18/276
Stroke history
Yes	2/8	0.78	3/8	0.48	1/11	0.90
No	89/301		159/318		29/361
History of hormone use in women
Yes	6/24	0.57	12/23	0.48	3/28	0.63
No	20/64		46/73		7/90

*   P < 0.05.

†  Waist-to-hip ratio: ≥0.92 for men or ≥0.88 for women. ²⁴

In multivariate analysis, nuclear cataract was significantly associated with smoking (P = 0.04; relative risk [RR]: 2.05; 95% CI: 1.05–3.99). On the other hand, males were almost half as likely to develop cortical opacity compared with females (P = 0.04; RR: 0.57; 95% CI: 0.34–0.97) and those with a history of diabetes were 2.43 times (P = 0.05; RR 2.43; 95% CI: 1.02–5.81) more likely to have cortical opacity. For posterior subcapsular cataract, a higher level of education was protective against posterior subcapsular cataract (P = 0.03; RR: 0.40; 95% CI: 0.18–0.91) and a higher BMI was associated with a lower risk of developing posterior subcapsular opacity (P = 0.02; RR 0.28; 95% CI: 0.10–0.79) (Table 4). 

Table 4

 

Multivariate Analysis on Age-Related Cataracts Among Residents in Shihpai, Taipei, Taiwan, 2006 to 2007

Variable	Nuclear		Cortical		Posterior Subcapsular
	RR (95% CI)	P Value	RR (95% CI)	P Value	RR (95% CI)	P Value
Age, y
80–93 vs. 72–79	0.84 (0.43–1.64)	0.60	1.23 (0.71–2.12)	0.47	1.24 (0.52–3.00)	0.63
Sex
Male vs. female	0.85 (0.46–1.57)	0.60	0.57 (0.34–0.97)	0.04*	1.24 (0.54–2.81)	0.61
Education
≥High school vs. ≤secondary	–	–	–	–	0.40 (0.18–0.91)	0.03*
BMI, kg/m2
≥25 vs. <25	–	–	–	–	0.28 (0.10–0.79)	0.02*
Waist-to-hip ratio†
High vs. not high	–	–	0.63 (0.37–1.07)	0.09	–	–
Smoking
Current vs. never smoking	2.05 (1.05–3.99)	0.04*	–	–	–	–
Alcohol
Yes vs. no	–	–	0.53 (0.26–1.07)	0.08	–	–
History of diabetes
Yes vs. no	–	–	2.43 (1.02–5.81)	0.05*	–	–
History of cardiovascular disease
Yes vs. no	–	–	0.68 (0.39–1.18)	0.17	1.88 (0.80–4.43)	0.15

*   P < 0.05.

†  Waist-to-hip ratio: ≥0.92 for men or ≥0.88 for women. ²⁴

Separate multivariate analysis performed for the development of a pure type of lens opacity showed that results were generally consistent with the previous analysis. Smoking (P = 0.01; RR 5.87; 95% CI: 1.53–22.48) was a significant predictor for nuclear opacity. For cortical opacity, age (P = 0.03; RR 4.01; 95% CI: 1.16–13.84) was a significant predictor, whereas sex (P = 0.87) became statistically insignificant. Diabetes mellitus (P = 0.05; RR: 6.63; 95% CI: 1.01–45.88) was borderline significant. For posterior subcapsular opacity, BMI (P = 0.03; RR 0.15; 95% CI: 0.03–0.80) became the only significant predictor. 

Discussion

The Shihpai Eye Study provides the longest follow-up data on the natural history of cataract development in a Chinese elderly population. A particularly high incidence of cortical opacity was noted in our study. This observation was very similar to the Barbados Eye Study. ^12,13 When compared with the Beaver Dam Eye Study ^{8 –10} and the Blue Mountains Eye Study ¹¹ (Table 5), whites had a much higher incidence of nuclear opacity and a lower incidence of cortical opacity. Interestingly, even among different races, posterior subcapsular opacity had the lowest incidence of the three types of cataract. 

Table 5

 

Incidence of Various Types of Age-Related Cataract and Cataract surgery in Population-Based Studies

Study	Race	Follow-up Years	Age at Baseline	Baseline Examination	Grading System	Nuclear		Cortical		Posterior Subcapsular		Surgery
						At Risk, n	%	At Risk, n	%	At Risk, n	%	At Risk, n	%
Beaver Dam Eye Study8	Caucasians	4.8	65–74	1988–1990	Wisconsin	565	40.2	582	18.9	647	7.6	909	11.2
	(either eye)		≥75			107	55.1	121	26.5	158	11.4	314	15.0
Beaver Dam Eye Study9	Caucasians	10	65–74	1988–1990	Wisconsin	566	53.3	584	36.9	712	12.9	909	26.4
	(either eye)		≥75			107	72.7	123	48.2	194	14.0	314	29.8
Blue Mountains Eye Study11	Caucasians	5	65–74	1992–1994	Wisconsin	407	35.1	563	14.6	679	4.9	847	7.3
	(persons)		>74			99	61.6	146	21.2	214	6.1	296	16.6
Blue Mountains Eye Study11	Caucasians	10	65–74	1992–1994	Wisconsin	407	54.2	563	33.9	679	12.9	847	28.0
	(persons)		>74			99	78.7	146	50.8	214	15.1	296	45.4
Barbados Eye Study12	African descent	4	60–69	1987–1992	LOCS II	567	18.9	343	40.5	681	3.8	–	–
	(persons)		70+			251	41.8	135	56.3	473	10.4	–	–
Barbados Eye Study13	African descent	9	60–69	1987–1992	LOCS II	610	81.9	365	55.7	732	9.1	779	5.7
	(persons)		70+			262	94.6	141	72.9	494	21.2	602	14.4
Los Angeles Latino Eye Study16	Latinos	4	60–69	2000–2003	LOCS II	701	26.5	553	17.2	711	4.9	3668	1.48
	(either eye)		70+			194	53.6	170	25.9	271	10.3	(all age groups)
Beijing Eye Study17	Chinese	5	65−74	2001	AREDS	362	16.9	485	29.1	536	15.5	5.4
	(either eye)		75+			24	33.3	65	24.6	75	25.3	9.5
This Study	Chinese	7	65−74	1999–2000	LOCS III	236	29.6	243	48.5	280	6.5	12.6
	(either eye)		>74			73	28.9	83	55.4	92	15.4	24.6

The incidence of different types of cataract differed between our study and the Beijing Eye Study, ¹⁷ although the two studies included subjects of the same race. There are several possible explanations. First, different grading systems were used (AREDS in the Beijing Eye study ¹⁷ and LOCS III in our study). Second, the definition of cataract development differed. Third, the target population in our study was aged 65 and older at baseline, whereas the recruitment age for the Beijing Eye Study was 40 years and older. The elderly group, especially those aged 75 years and older, represented only a small proportion in the age structure of their participants. Hence, one subject might contribute to a substantial percentage in the incidence rate. Moreover, the follow-up period was 5 years in their study as opposed to 7 years in our survey. Environmental and lifestyle factors may play a role as well. A higher incidence of nuclear and cortical opacity was observed in our study compared with the Beijing Eye Study. ¹⁷ On the other hand, the incidence of posterior subcapsular cataract was much lower in our population. The cataract operation rate was approximately double in our participants compared with the Beijing Eye Study (Table 5). The cataract surgery rate in our population was comparable with the Beaver Dam Eye Study ^8,9 and higher than the Barbados Eye Study. ^12,13 Posterior subcapsular opacity is more visually disturbing than the other two types of cataract and thus participants might be more inclined to undergo cataract operation. 

Among the potential risk factors, our study revealed that current smokers had a higher chance of developing nuclear opacity compared with nonsmokers. On the other hand, there was no significant difference in the risk of developing all three types of cataract between previous smokers and nonsmokers. This finding was largely confirmed by earlier studies. ^{23,25 –31} Our results based on longitudinal follow-up of a fixed cohort further suggested a possible effect that smoking imposed on the development of nuclear cataract. However, a higher proportion of current smokers participated in the eye examination, which may have biased the strength of the relationship between current smoking and nuclear cataracts in our study. 

On the other hand, female sex and a history of diabetes were associated with a higher risk of cortical opacity. This finding was concordant with the Los Angeles Latino Eye Study ²⁵ and most other population-based studies. ^18,21,23  

Our findings indicated that having a higher level of education and higher BMI were protective of posterior subcapsular opacity. One possible reason is that lower levels of education might be related to lower income and a higher probability of outdoor jobs ³⁰ and hence more UV-light exposure. Moreover, a lower income might also lead to a low diet quality with insufficient antioxidants and vitamins. ^30,32  

In our study, higher BMI was noted to be protective of posterior subcapsular opacity. Higher education provides a better chance of indoor jobs and a more sedentary lifestyle. Hence, this may be related to a lower chance of UV-light exposure as well as a higher chance of a greater BMI. Interestingly, there are abundant studies in the literature suggesting that BMI is related to age-related cataract, although the pathway is unclear. Moreover, the relationship is inconsistent with regard to the type of cataract involved as well as the direction of the relationship. ^{18 –20,33} Because BMI is a modifiable risk factor, however, this aspect deserves further evaluation and analysis. 

There were some limitations to our study. The response rate in our study was relatively low (55.8% of those eligible). Obtaining population-based prevalence estimates of eye disease among elderly persons is challenging because this group of individuals is less likely to participate in research studies. ³⁴ The inclusion rate in the Rotterdam Study ³⁵ ranged from 59% in the group aged 75 to 84 years to 28% in the group aged 85 years and older. Similarly, in the Baltimore Study, ³⁶ inclusion rates were 48% in the group aged 70 to 79 years and 21% in the group aged 80 years and older. Another potential reason for the low participation rate is that lack of utilization of ophthalmologic care for prevention and treatment has created the impression that loss of vision is expected in senior life and the idea that nothing can be done to improve the situation among elderly people, particularly among the less-educated. ²³  

Nonparticipants were older, more likely to be female or illiterate, and more likely to have a history of stroke and to be current smokers. These unexamined subjects remain a potential source of bias. Our study populations were noninstitutionalized survivors; excluding those who are inpatients or who have paralysis or disability in the survey probably removes a disproportionate number of potential participants with functional or physical impairment and/or declining health-related quality of life and might bias the results of the study. Furthermore, mortality of nonparticipants might lead to an underestimation of the incidence rate. Third, the assessment of comorbidities by adopting a dichotomized classification was simplistic. Moreover, the possibility of a chance finding cannot be completely excluded. 

On the other hand, interobserver agreement of graders between baseline and follow-up was unavailable. The interobserver reproducibility of lens grading between two study ophthalmologists at baseline ²³ was 0.86 for cortical opacity, 0.82 for posterior subcapsular opacity, and 0.85 for nuclear opacity. The intraobserver agreement for reproducibility of grading cataracts was 0.83 for cortical opacity, 0.80 for posterior subcapsular opacity, and 0.86 for nuclear opacity for one ophthalmologist and 0.87, 0.85, and 0.84, respectively, for the other ophthalmologist. The possibility of a contemporaneous variability and temporal drift in lens grading cannot be eliminated. A previous study ³⁷ showed that when temporal drift was present, baseline scores tended to be higher than subsequent scores for eyes with higher severity grades at baseline. Hence, the effect on our study should be minimal. 

In conclusion, the incidence of cortical opacity is highest among the three types of cataract after a follow-up of 7 years in our urban elderly Chinese population, followed by nuclear opacity, then the incidence of posterior subcapsular opacity. Understanding the incidence of various types of cataracts and their risk factors can facilitate the development of educational programs and campaigns to prevent and delay the onset of cataracts and the impact it may bring to the quality of life of the elderly. 

Acknowledgments

Supported by a grant from Taipei Veterans General Hospital, Taipei, Taiwan (V95S3-001). The authors alone are responsible for the content and writing of the paper. 

Disclosure: T.-M. Kuang, None; S.-Y. Tsai, None; C.J.-L. Liu, None; Y.-C. Ko, None; S.-M. Lee, None; P. Chou, None