Lutein and Zeaxanthin and the Risk of Cataract: The Melbourne Visual Impairment Project

Abstract

purpose. To evaluate the association of cortical, nuclear, or posterior subcapsular (PSC) cataract with dietary intake of lutein-zeaxanthin (LZ) in a population-based sample.

methods. For the study, 3271 (83% of the eligible residents) permanent residents aged ≥40 years were recruited in 1992 to 1994 via a cluster random sampling. In 1997 to 1999, 2594 (79%) attended the follow-up examination including lens photography, a life-style questionnaire, and a food-frequency questionnaire (FFQ). Cases were those with cortical opacity ≥4/16, nuclear opacity grade ≥2.0, or PSC opacity ≥1 mm². Logistic regression was used to calculate the odds ratios for cataract by daily LZ intake, or its quintile indicator with the lowest quintile as the baseline category, controlling for energy-adjusted fat intake and variables previously found to be associated with the cataract outcomes.

results. Of the 2322 participants who attended the follow-up survey and completed the FFQ, 1841 (79%), 1955 (84%), and 1950 (84%) were included in the analyses of cortical, nuclear, and PSC cataract, respectively. There were 182 (9.9%), 387 (19.8%), and 177 (9.1%) cases for cortical, nuclear, and PSC cataract, respectively. Cortical and PSC cataract were not significantly associated with LZ intake. For nuclear cataract the odds ratios were 0.67 (0.46–0.96) and 0.60 (0.40–0.90) for every 1-mg increase in crude and energy-adjusted daily LZ intake, respectively. The odds ratios (95% CI) for those in the top quintile of crude LZ intake was 0.58 (0.37–0.92; P = 0.023 for trend), and it was 0.64 (0.40–1.03) for energy adjusted LZ intake (P = 0.018 for trend).

conclusions. This study found an inverse association between high dietary LZ intake and prevalence of nuclear cataract.

Cataract is one of the five main causes of visual impairment among older adults in two Australian population-based studies: the Melbourne Visual Impairment Project (Melbourne VIP) and the Blue Mountains Eye Study (BMES). ^{1 2} It is also an important cause of disability among older adults in the United States. ³ Cataract surgery is the most commonly performed ophthalmic procedure in Australia, ⁴ and more than 2 million cataract extractions are performed annually in the United States. ³ The prevalence of cataract increases with age ⁵ ; consequently, cataract becomes an important issue in aging populations. Identification of modifiable risk factors that could delay or prevent cataract development would allow development of strategies to reduce cataract incidence. Such strategies have the potential to improve the quality of life for older adults and to reduce medical care costs. 

One potential risk factor for cataract is diet. An inverse relationship between lens optical density and macular pigment optical density has been observed. ⁶ Macular pigment is composed of lutein and zeaxanthin (LZ). This finding suggests that the antioxidant xanthophyll carotenoids LZ, the only carotenoids in human lenses, may protect against the development of cataract. Possible mechanisms include the prevention of the oxidation of lens proteins that plays a central role in the formation of cortical and nuclear cataracts. ^{7 8} Higher LZ intake has been reported to decrease the risk of cataract extraction among women in the Nurses’ Health Study ⁷ and among men in the Health Professionals Follow-up Study. ⁸ Also, the prevalence of nuclear opacity in the highest quintile of LZ intake was significantly lower than that in the lowest quintile in a subsample of 478 nondiabetic women with both lenses intact from the Nurses’ Health Study. ⁹ In the Beaver Dam Eye Study, those in the highest quintile of the prior LZ intake were only half as likely to have an incident nuclear cataract over 5 years as those in the lowest quintile. ¹⁰  

We used data from the urban participants in the follow-up survey of the Melbourne VIP to investigate the hypothesized association between dietary LZ and cataract. The purpose of this study was to evaluate the association of cortical, nuclear, and PSC cataract with dietary intake of the carotenoids LZ in a population-based sample of the Melbourne VIP. 

Methods

Study Design

The Melbourne VIP is a population-based epidemiologic study of eye diseases. ¹¹ A household census was conducted to identify those who were 40 years of age or older and had been residents for at least 6 months in one of nine randomly selected pairs of census collector districts within the Melbourne statistical division. A group of 3271 (83% of the eligible residents) residents were recruited from 1992 through 1994. ¹¹ All urban participants from the baseline survey were invited to participate in the 5-year follow-up survey conducted from 1997 through 1999. ¹² At both time points, participants signed an informed consent form, completed a standardized questionnaire, and underwent a detailed ophthalmic examination. The standardized questionnaire was used to obtain information on socioeconomic and demographic characteristics, historic, and current symptoms of eye diseases, medical history, and use of medication and vitamin supplements. The follow-up survey also included a food-frequency questionnaire (FFQ) covering dietary intake in the previous 12 months. The study protocol was approved by the Human Research and Ethics Committee of the Royal Victorian Eye and Ear Hospital and complied with the Declaration of Helsinki for research involving human subjects. 

Cataract Grading

Cataract grading in the Melbourne VIP has been described in detail. ⁵ Briefly, lens opacities were graded clinically and from photographs by the Wilmer cataract photograph grading system. ¹³ Two research assistants graded photographs separately, and an independent reviewer adjudicated discrepancies. The cortical and PSC opacities were photographed with a retroillumination camera, and a photograph slit lamp was used to photograph nuclear opacities. Clinical grades were used only if corresponding photographic grades were missing. Of 4461 eyes with final nuclear cataract grading, 3293 (74%) eyes had photographic grades. Of 4451 eyes with final cortical or PSC cataract grading, 2348 (53%) had photographic grading. Cortical opacity was the proportion measured in sixteenths of pupil circumference occupied by opacity, and cortical cataract was defined as cortical opacity of four sixteenths or greater. Posterior subcapsular cataract was defined as opacity of 1 mm² or greater. Nuclear cataract was defined as opacity greater than or equal to Wilmer standard 2.0. We used data from the treating ophthalmologist to classify cataract type at the time of cataract surgery for those lenses removed before the follow-up survey. 

Information on Covariates

We collected data on medication use including current and previous medication supplements and total duration of use. In the analysis, β-blocker use refers to total duration of use for cortical cataract, whereas it refers to ever having used (currently and/or previously) them for nuclear cataract. Also we refer to having ever used ACE inhibitors, loop diuretics, acetaminophen, and thiazide diuretics. 

The logarithm of average annual ultraviolet B exposure is used in the multivariate logistic regressions, where average annual ultraviolet B exposure was the total ultraviolet B exposure divided by age at examination. For nuclear cataract, we used the variable of interaction between daily total vitamin E intake (in grams) and the logarithm of average annual ocular exposure. This interaction between UVB exposure that causes oxidative lens damage and vitamin E intake that provides an antioxidative effect has been explored in detail for the baseline Melbourne VIP, where ocular ultraviolet B exposure was not independently related to nuclear cataract. ⁵  

Age-related macular degeneration (AMD) was diagnosed by either clinical or photographic examination. ¹⁴ We used two different definitions of early AMD, the international classification (IC) and Wisconsin early AMD, but only one for late AMD. ¹⁴ In this article, AMD refers to both early and late AMD, and we present only the results with IC AMD, as both definitions produced essentially the same results. 

Nutritional Analyses

Of the 2594 participants who attended the follow-up survey, 2322 (90%) completed an FFQ developed and validated by the Cancer Council Victoria ^{15 16} to collect dietary intake. The mean age of those who completed the FFQ was 62.1 ± 10.6 years, whereas the mean age of those who did not complete the FFQ was 65.8 ± 12.8 years. The probability derived from the t-test for the difference in age was <0.001, but there was no difference in gender between these two groups of participants (P = 0.61). 

The FFQ included 13 different fruit items and 25 vegetable items, and each item had 10 available frequency options from never to three or more times per day. Software developed by the Cancer Council Victoria was used to compute nutrient intakes from the FFQ, by using carotenoid composition data from the U.S. Department of Agriculture (USDA) ¹⁷ to estimate carotenoid intakes including LZ. 

We also calculated energy-adjusted LZ intake according to Willett’s method. ¹⁸ The logarithm of the energy-adjusted nutrient intake was the difference between the observed and expected values of the logarithm of the nutrient intake plus the predicted logarithm of nutrient intake at the population mean of the logarithm of energy intake, where the expected and predicted values were calculated from the simple linear regression of the logarithm of the nutrient intake on the logarithm of total energy intake. 

Statistical Analysis

All statistical analyses were conducted (SAS, ver. 9.1 for Windows; SAS Institute, Cary, NC). We fitted the logistic regressions for the by-person analyses and the generalized linear models for the by-eye analyses. The generalized linear models with an unstructured working correlation matrix were used to account for dependency between two eyes from the same participant. We compared eyes with pure cortical cataract versus eyes with no cataract in the by-eye analyses. We did similar analyses for nuclear and PSC cataract. 

Multivariate models included either crude or energy-adjusted LZ intakes, and covariates found important in the baseline analysis of the VIP cohort. ⁵ For nuclear cataract, we also considered the interaction between smoking and daily LZ intake, as smoking has been known to decrease some plasma antioxidants including LZ ¹⁹ by further including an indicator for current smokers and a product of this indicator and LZ variables. However, this interaction was not significant and was not included in the multivariate analyses. The probability for trend in a multivariate analysis was derived from the test of whether the coefficient for the quintile medians was equal to zero. A test result with P < 0.05 was considered to be significant. 

Results

Of 3040 participants who were still alive at the follow-up survey, 2594 (85%) participated, 51 (1.7%) moved from Melbourne, 312 (10%) refused to participate, and 83 (2.7%) were not traceable. Of the 2322 participants who attended the follow-up survey and completed the FFQ, 1841 (79%) participants had nonmissing values from the follow-up survey for cortical cataract and covariates (Table 1) . Similarly, the corresponding number of participants was 1955 (84%) in the nuclear cataract analysis, 1950 (84%) in the PSC cataract analysis (Table 1) . Of the total respective samples, 182 (9.9%) had cortical cataract, 387 (19.8%) had nuclear cataract, and 177 (9.1%) had PSC cataract. The daily intakes from the first to the fifth quintiles were (in micrograms) ≤454, 454 to 640, 640 to 812, 812 to 1104, and >1104 for crude LZ intake, and they were ≤472, 472 to 639, 639 to 808, 808 to 1037, and >1037 for adjusted LZ intake. We present the by-person analyses (Table 2) . There were 2998, 3438, and 3114 eyes included in the by-eye analyses for cortical, nuclear, and PSC cataract, respectively. In these respective samples, 151 (5%) eyes had pure cortical cataract, 437 (12.7%) had pure nuclear cataract, and 113 (3.6%) had pure PSC cataract. 

From the by-person analyses, older people had a significantly higher risk of having any kind of cataract. Female gender, higher accumulated sunlight exposure, and myopia ≥1 D were significantly associated with higher risk of cortical cataract and therefore were included in multivariate analysis. The factors found to be significantly associated with increased risk of nuclear cataract and included in the multivariate analyses for nuclear cataract were female gender, brown to dark brown colored irides, use of acetaminophen, myopia ≥1 D, smoking duration >30 years, and increased product of accumulated sunlight exposure and total vitamin E (including intake and supplement). Use of thiazide diuretic and myopia ≥1 D significantly increased the risk of PSC cataract. 

There was no association between daily LZ intake and either cortical or PSC cataract (Table 2) . For nuclear cataract, the multivariate odds ratios (95% CI) were 0.67 (0.46–0.96) and 0.60 (0.40–0.90) for every 1-mg increase in crude and energy-adjusted daily LZ intake, respectively (Table 2) . The odds ratios were all approximately equal to 0.60 for those in the third, fourth, and fifth crude LZ intake quintiles (P ≤ 0.032) and for those in the fourth and fifth energy-adjusted LZ intake quintiles (P = 0.017 and 0.066, respectively). The decreasing trend for quintile medians was also significant for crude and energy-adjusted LZ intake (P = 0.023 and 0.018, respectively). 

The odds ratios and probabilities for LZ intake from the by-person analyses were similar to those from the by-eye analyses, when we compared eyes with pure cataract of specific type versus eyes having had no cataract of this type (data not shown). We also repeated the by-person analyses for those not taking supplements, and the associations between nuclear cataract and LZ intake became more significant, although the sample size was reduced approximately 21%. For cortical and PSC cataract, the results for excluding those who took supplements were not much different from those for all participants. 

Discussion

Our study showed an inverse association between higher dietary LZ intake and the presence of nuclear cataract. This finding was consistent with other epidemiologic studies. ^{9 10} The main strengths of the study are the use of a comprehensive ophthalmic examination including lens photography, the large, population-based sample and large number of cases, especially for nuclear cataract. Also, we collected information for many confounders. One limitation of this study is that the data are cross-sectional. Another limitation includes a possible selection bias due to missing values and/or loss to follow-up. The mean age of those included in the multivariate analyses was significantly less than the mean age of those lost to follow-up or excluded from the multivariate analyses for each kind of cataract; the differences in mean ages were 3.1, 3.3, and 3.6 years for cortical, nuclear, and PSC cataract, respectively. However, most of the odds ratios were similar to those in the previously published baseline analyses, ⁵ suggesting they were not affected by selection bias. 

Most previous studies of cataract have collected the cataract outcomes and dietary intake data at the same time, ^{7 8} whereas an association between past LZ intake and cataract incidence would be more valuable in studying cause–effect associations. Thus, in the Beaver Dam Eye Study, LZ intake 10 years before baseline was associated with incidence of nuclear cataract over the after 5 years, whereas current dietary intake showed little association. ¹⁰ Although it is not clear whether the observed lower risks in those with high LZ intake resulted from the protective effect of high LZ intake or from the lifestyle of those consuming foods rich in LZ. ²⁰  

The possible reason for detecting significant association for only nuclear cataract in this study is that, there were approximately twice as many cases of nuclear cataract as there were of either cortical or PSC cataract. However, the odds ratios for every 1-mg increase in both crude or energy-adjusted daily LZ intake for either cortical or PSC cataract were similar to those for nuclear cataract. Therefore, a larger sample size would be needed to detect a possible association between LZ intake and risk of either cortical or PSC cataract. In the analyses with LZ intake as a continuous variable, the powers of detecting a significantly lower risk associated with high crude daily LZ intake were 53%, 70%, and 20% for cortical, nuclear, and PSC cataract, respectively, and they were 34%, 80%, and 23% for energy-adjusted LZ intake. We used the estimates derived from the multivariate logistic regressions to calculate these powers. 

For any type of cataract and any type of analysis, the means of LZ intake were greater for noncases than for cases, and the means of LZ intake were greater for those who took supplements than for those who did not take supplements. For nuclear cataract, the difference in crude daily LZ intake means of noncases and cases was 109 μg (P < 0.0001) for those not taking supplements and 47μg (P = 0.40) for those taking supplements. The difference in adjusted daily LZ intake means was 76 μg (P = 0.0009) for those not taking supplements and 52 μg (P = 0.30) for those taking supplements. 

The disadvantage of this study is that we did not have an opportunity to examine LZ concentration in the serum or macula. Although we found an inverse association between LZ intake and nuclear cataract prevalence, bioavailability of the nutrients varies significantly in individuals and dietary intake is not necessarily reflected in either serum level of these carotenoids or macular tissue. Given the variability and possibility of the significant bias in cross-sectional studies, there seems to be good reason to proceed with a randomized controlled trial to confirm whether increase one’s dietary intake of LZ would significantly reduce the risk of development of nuclear cataract. 

 Supported, in part, by grants from the National Health and Medical Research Council, Canberra, Australia; the Jack Brockhoff Foundation, Melbourne, Australia; the Ansell Ophthalmology Foundation, Melbourne, Australia; the Eye, Ear, Nose and Throat Research Institute; Edols Bequest; Appel Family Bequest; and the Hugh Williamson Foundation and Ian Potter Foundation.

 Submitted for publication May 12, 2005; revised August 31, 2005, and February 7, and May 16, 2006; accepted July 17, 2006.

 Disclosure: H.T.V. Vu, none; L. Robman, None; A. Hodge, None; C.A. McCarty, None; H.R. Taylor, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Characteristics Included in the Multivariate Analyses for Three Types of Cataract among the Melbourne VIP Study Participants

	Total Sample Analyzed for Cortical Cataract Risk Factors (n = 1841)	Total Sample Analyzed for Nuclear Cataract Risk Factors (n = 1955)	Total Sample Analyzed for PSC Risk Factors (n = 1950)
Continuous variable, mean (SD)
Age	61.3 (10.3)	61.7 (10.4)	61.6 (10.4)
Daily energy intake (kJ)	7377 (2754)	7386 (2769)	7394 (2769)
Daily LZ intake (μg)	792 (409)	796 (411)	798 (413)
Adjusted daily LZ intake (μg)	772 (367)	775 (369)	775 (370)
Daily fat intake (g)	65.7 (30.1)	65.9 (30.3)	65.9 (30.3)
Energy-adjusted daily fat intake (g)	59.8 (11.2)	59.8 (11.2)	59.8 (11.2)
Categorical variable, frequency (%)
Female gender	1027 (56)	1084 (55)	VNI
Brown or dark brown iris	441 (24)	461 (24)	VNI
Arthritis diagnosis	616 (33)	VNI	VNI
Diabetes duration >5 years	69 (3.7)	78 (4.0)	VNI
Gout duration >10 years	88 (4.8)	VNI	VNI
Ever used ACE inhibitor (yes/no)	VNI	292 (15)	VNI
Ever used loop diuretic (yes/no)	VNI	32 (1.6)	VNI
Ever used acetaminophen (yes/no)	VNI	207 (11)	VNI
Ever used β-blocker (yes/no)	224 (12.2)	238 (12)	VNI
Ever use of thiazide diuretic (yes/no)	VNI	VNI	140 (7.2)
Myopia ≥1 D	304 (17)	323 (17)	315 (16)
AMD (international classification)	VNI	320 (16)	VNI
AMD (Wisconsin definition)	VNI	135 (6.9)	VNI
Presence of glaucoma	80 (4.3)	VNI	VNI
Smoking duration >30 years	VNI	356 (18)	VNI
Taking vitamin A supplement (yes/no)	VNI	VNI	170 (8.7)
Daily vitamin E (mg)
≤5	VNI	VNI	649 (33)
5–10	VNI	VNI	941 (48)
>10	VNI	VNI	360 (18)
Taking supplement(s) (yes/no)	379/1800* (21.1)	397/1912* (20.8)	403/1950* (20.7)

 VNI, variable not included in the cataract-specific multivariate models.

*  Number of complete records for the corresponding variable.

Table 2.

 

Multivariate Analyses of the Relationships between Three Types of Cataract and Dietary Intake of LZ

Variable	Cortical Cataract*							Nuclear Cataract, †							PSC, ‡
		Crude LZ Intake			Energy-Adjusted LZ Intake			Crude LZ Intake			Energy-Adjusted LZ Intake			Crude LZ Intake			Energy-Adjusted LZ Intake
		OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI
Daily LZ intake (mg)ξ	0.68	0.44–1.05	0.74	0.46–1.19	0.67	0.46–0.96	0.60	0.40–0.90	0.83	0.53–1.31	0.81	0.51–1.28
Daily LZ intake quintiles
First	1.00	—	1.00	—	1.00	—	1.00	—	1.00	—	1.00	—
Second	1.17	0.73–1.86	0.85	0.52–1.40	0.68	0.44–1.04	0.90	0.59–1.39	1.02	0.62–1.68	1.13	0.69–1.85
Third	0.63	0.37–1.08	0.92	0.56–1.50	0.62	0.40–0.96	0.85	0.55–1.31	0.91	0.54–1.54	0.85	0.51–1.43
Fourth	0.93	0.56–1.53	0.61	0.35–1.03	0.59	0.38–0.93	0.58	0.37–0.91	1.09	0.64–1.85	0.90	0.54–1.51
Fifth	0.74	0.43–1.28	0.84	0.50–1.41	0.58	0.37–0.92	0.64	0.40–1.03	0.75	0.42–1.35	0.94	0.55–1.59
P for trend, ¶		0.190		0.328		0.023		0.018		0.406		0.601

 Statistically significant data are shown in bold.

*  Odds ratios adjusted for fat intake, age, gender, iris color, arthritis, diabetes >5 years, gout >10 years, use of β-blockers >10 years, UV-B exposure, and myopia ≥1 D.

†  Odds ratios adjusted for fat intake, age, gender, iris color, diabetes >5 years, use of ACE inhibitor, use of β-blocker, use of loop diuretic, use of acetaminophen, myopia ≥1 D, AMD (international classification), and smoking for >30 years.

‡  Odds ratios adjusted for fat intake, age, use of thiazide diuretics, myopia ≥1 D, vitamin E 5–10 mg/d vs. ≤ 5 mg/day, vitamin E >10 mg/day vs. ≤5 mg/day, use of vitamin A supplements, and UV-B exposure.

 ξ LZ as a continuous variable.

¶   P for daily LZ quintile medians.