Dietary Carotenoids, Vitamins C and E, and Risk of Cataract in Women: A Prospective Study

Objective  To examine in prospective data the relation between dietary intake of carotenoids and vitamins C and E and the risk of cataract in women.

Design  Dietary intake was assessed at baseline in 39 876 female health professionals by using a detailed food frequency questionnaire. A total of 35 551 women provided detailed information on antioxidant nutrient intake from food and supplements and were free of a diagnosis of cataract. The main outcome measure was cataract, defined as an incident, age-related lens opacity responsible for a reduction in best-corrected visual acuity in the worse eye to 20/30 or worse based on self-report confirmed by medical record review.

Results  A total of 2031 cases of incident cataract were confirmed during a mean of 10 years of follow-up. Comparing women in the extreme quintiles, the multivariate relative risk of cataract was 0.82 (95% confidence interval, 0.71-0.95; test for trend, P  = .04) for lutein/zeaxanthin and 0.86 (95% confidence interval, 0.74-1.00; test for trend, P  = .03) for vitamin E from food and supplements.

Conclusion  In these prospective observational data from a large cohort of female health professionals, higher dietary intakes of lutein/zeaxanthin and vitamin E from food and supplements were associated with significantly decreased risks of cataract.

The oxidative hypothesis of cataract formation posits that reactive oxygen species can damage lens proteins and fiber cell membranes and that nutrients with antioxidant capabilities can protect against these changes.^1-3 Results of laboratory studies and studies in animals generally support the antioxidant hypothesis, but results of observational epidemiologic studies in humans have been inconsistent.^4-6Moreover, results of completed randomized trials^7-14 indicate that supplemental use of vitamin E, ascorbic acid (vitamin C), or beta carotene for as long as 6.5 years (12 years for beta carotene⁷) has no marked effect on cataract incidence or progression. Whether a longer duration of treatment with these antioxidants can materially reduce the risk of cataract, which develops slowly across many years, remains to be determined in recently completed and ongoing trials.^15-17

In addition to those nutrients tested in completed and ongoing trials, evidence has also accumulated to suggest a possible role for lutein, a xanthophyll carotenoid, in lowering the risk of cataract. Lutein is concentrated in tissues of the eye, including the lens,^18-20 and may protect against cataract by filtering harmful short-wave blue light and possibly through antioxidant activity.^21-23 Findings from observational epidemiologic studies^24-32 generally support a possible beneficial effect of lutein. However, as encouraging data have accumulated, enthusiasm for its potential benefits has also led to the advocation of lutein-containing supplements to prevent eye disease, although there are no randomized trial data that assess its effectiveness in the eye, and prospective observational data for cataract are limited to 3 studies.^27-29,32 For this reason, the National Eye Institute³³ has cautioned that the possible benefit of lutein on the eye remains uncertain and warrants closer examination before conclusions can be drawn. In this article we examine in prospective data the relation of dietary intake of several carotenoids, including lutein (and its stereoisomer zeaxanthin), and vitamins C and E with the risk of cataract during 10 years of follow-up in a large cohort of female health professionals.

Study participants were women enrolled in the Women's Health Study (WHS), a recently completed randomized, double-masked, placebo-controlled trial of low-dose aspirin, vitamin E, and beta carotene in the primary prevention of cardiovascular disease and cancer in 39 876 apparently healthy female health professionals 45 years and older.^15,34-36 Participants were willing to forego use of individual supplements of beta carotene, vitamin A, and vitamin E but could continue using multivitamins for the duration of the trial. Detailed information on antioxidant intake from food and supplements was provided by 39 310 of the randomized participants (98.58%), who completed a 131-item validated semiquantitative food frequency questionnaire (SFFQ) at baseline in 1993.³⁷ For this analysis we excluded participants who reported total energy intake less than 600 kcal/d or greater than 3500 kcal/d or who had more than 70 blanks on the SFFQ. Of the remaining participants, 35 551 were without a diagnosis of cataract at baseline and are included. This study was conducted according to the ethical guidelines of Brigham and Women's Hospital.

For each food item, a standard unit or portion size was specified and participants were asked how often, on average, during the previous year they had consumed that amount. Nine responses were possible, ranging from “never” to “6 or more times per day.” Responses to the individual food items were converted to average daily intake of each nutrient based on food tables maintained by the Harvard School of Public Health. The carotenoid content of food items was determined by using the US Department of Agriculture–National Cancer Institute carotenoid food composition databases.^38,39 Participants also provided information about their current use of multivitamins and supplements of ascorbic acid, vitamin E, and beta carotene. The total intake of antioxidants was calculated by adding the contributions from vitamin supplements and foods. For beta carotene, ascorbic acid, and vitamin E, we also calculated intake from food sources alone. In the databases, the data for lutein and zeaxanthin have been combined.

The reproducibility and validity of SFFQ estimates of vitamin and carotenoid intake have been examined in a similar population of female nurses. The Pearson correlation coefficient between estimates from the SFFQ and the average of two 1-week diet records was 0.76 for energy-adjusted total ascorbic acid.⁴⁰ For vitamin E, the correlation between estimates of intake from the SFFQ and plasma concentrations of alpha tocopherol was 0.41.⁴¹ For carotenoids, correlations between plasma concentrations and the SFFQ assessments in nonsmoking women were 0.27 for beta carotene, 0.27 for lutein, 0.32 for beta cryptoxanthin, 0.48 for alpha carotene, and 0.21 for lycopene.⁴² These estimates were similar to those found between plasma levels of these nutrients and estimates from food records.⁴³

Information on possible risk factors for cataract was collected on the WHS baseline questionnaire and included age, height and weight, smoking status, alcohol use, frequency of exercise, parental history of myocardial infarction at younger than 60 years, history of hypertension, history of diabetes mellitus, history of hypercholesterolemia, postmenopausal hormone use, and history of an eye examination in the past 2 years.

Following the report of a cataract diagnosis, written consent identifying the diagnosing ophthalmologist or optometrist was obtained. The ophthalmologist or optometrist was asked to complete a cataract questionnaire supplying information about the presence of lens opacities, date of diagnosis, visual acuity loss, cataract extraction, other ocular abnormalities that could explain visual acuity loss, cataract type, and etiology. Medical record information was obtained for 95% of participants reporting cataract. The study end point was incident cataract, defined as a self-report confirmed by medical record review to be initially diagnosed after randomization, age-related in origin, with best-corrected visual acuity of 20/30 or worse and with no alternative ocular pathologic abnormalities to explain the visual acuity loss.

The unit of analysis was individuals, rather than eyes, because eyes were not examined independently, and participants were classified according to the status of the worse eye based on disease severity. Participants were followed up until the time of diagnosis of cataract or until death or February 1, 2004, whichever came first. Nutrient intake was considered as a categorical variable (in quintiles), with adjustment for total energy using the residual method.⁴⁰ Estimates of relative risks (RRs) were computed as the rate of cataract in a specific quintile of intake divided by the corresponding rate in the lowest quintile (reference). Age- and treatment-adjusted RR estimates were obtained by adjusting for age (in years) and randomized treatment assignment in Cox proportional hazards regression models.⁴⁴ Multivariate RRs were obtained by further adjusting for smoking, alcohol use, history of diabetes, history of hypertension, history of hypercholesterolemia, body mass index, physical activity, parental history of myocardial infarction, postmenopausal hormone use, and history of an eye examination in the past 2 years. For each RR, 2-sided P values and 95% confidence intervals (CIs) were calculated.⁴⁵ Tests of linear trend across increasing quintiles of nutrient intake used the medians of intakes within quintiles as scores. We used an interaction term between ordinal scores for each nutrient and length of follow-up to evaluate the adequacy of the proportional hazards assumption across time. For each nutrient, P > .05, indicating that the proportionality assumption was not violated. We also performed tests of interaction to evaluate the statistical significance of any modifying effect of age and baseline smoking status (current vs not current) on the association of nutrient intake with cataract.

We examined the independent contribution of each nutrient to cataract risk by simultaneously entering all nutrients into a Cox regression model using the median scores from quintiles and adjusting for other cataract risk factors. Using a backward selection procedure, nutrients were removed according to level of significance until only those nutrients that were significantly associated with cataract at the P < .05 level remained. We also fit a model that retained only those nutrients that were significantly associated with cataract at the P < .20 level.

Baseline characteristics of the study population are given in Table 1. Women who were newly diagnosed as having cataract during follow-up were older and, after adjusting for age, were more likely to smoke than women without a diagnosis of cataract. Women with diagnosed cataract also had a higher body mass index, exercised less, and were more likely to report a history of hypertension, diabetes, and high cholesterol.

Mean intakes of lutein/zeaxanthin and several other nutrients were highly correlated with one another (P < .001 for all correlation coefficients; data not shown). The correlation coefficients between lutein/zeaxanthin and the other carotenoids ranged from 0.22 for lutein/zeaxanthin and beta cryptoxanthin to 0.72 for lutein/zeaxanthin and beta carotene from food sources (0.67 for beta carotene from food and supplements). The correlation coefficients between lutein/zeaxanthin and vitamins C and E ranged from 0.17 for total vitamin E to 0.45 for vitamin C from food sources only.

During an average of 10 years of follow-up, 2031 cases of incident cataract were confirmed.Significant inverse trends with cataract were observed for dietary intakes of lutein/zeaxanthin and vitamin E.Comparing women in the highest quintile of intake of lutein/zeaxanthin with those in the lowest quintile, the multivariate RR of cataract was 0.82 (95% CI, 0.71-0.95; test for trend, P = .04) (Table 2). The RR for vitamin E from food and supplements, comparing women in the extreme quintiles, was 0.86 (95% CI, 0.74-1.00; test for trend, P = .03) (Table 2). Vitamin E from food sources alone was not significantly associated with risk of cataract. For beta cryptoxanthin, there was a significant inverse trend with cataract in age- and treatment-adjusted analyses (test for trend, P = .04) but not in analyses that also adjusted for other cataract risk factors (test for trend, P = .19). Multivariate RRs for the other nutrients (alpha carotene, beta carotene, lycopene, and ascorbic acid) were generally less than 1.0, but none of the tests for trend across quintiles attained statistical significance.

We examined the independent contribution of each nutrient in Cox regression models using backward selection procedures. Values for beta carotene, vitamin E, and ascorbic acid in this analysis included intake from both food and supplements. After stepwise removal of nonsignificant nutrients, only vitamin E remained significantly associated with the risk of cataract (test for trend, P = .03). When the significance level for retention was set at 0.20, vitamin E (test for trend, P = .04) and lutein/zeaxanthin (test for trend, P = .06) were retained in the final model.

We also examined whether the associations between nutrient intake and cataract differed by age and smoking status at baseline. We found no evidence that associations between nutrient status and cataract differed by age. For smoking, inverse associations tended to be stronger in women who were nonsmokers at baseline, but formal tests for interaction were not statistically significant for any nutrient.

Because of the inverse association between cataract and lutein/zeaxanthin, we also examined the association of cataract with specific food sources of lutein/zeaxanthin and other carotenoids (Table 3). Women with high overall intake of fruits and vegetables had an approximately 10% lower risk of cataract that was not statistically significant. There was, however, a borderline significant inverse trend between higher intake of green leafy vegetables and risk of cataract (test for trend, P = .06). When we considered specific foods that are important contributors to lutein/zeaxanthin intake, raw spinach showed a borderline significant inverse relation with risk of cataract in age- and treatment-adjusted analyses (test for trend, P = .06), but not after adjustment for other cataract risk factors.

In this large population of female health professionals, significant inverse trends with risk of cataract were observed for dietary intake of lutein/zeaxanthin and vitamin E. Comparing women in the extreme quintiles, women with high intake of lutein/zeaxanthin had an 18% lower risk of cataract in multivariate analysis. High intake of vitamin E from food and supplements was associated with a 14% lower risk of cataract. The inverse associations for lutein/zeaxanthin and vitamin E from food and supplements persisted in models that mutually adjusted for intake of several other carotenoids and ascorbic acid.

The prospective design of this study precluded the possibility that participant reports of nutrient intake at baseline were associated with subsequent cataract status. However, random or nondifferential misclassification of dietary intake was likely and would tend to underestimate any association of diet with risk of cataract. Changes in dietary behavior during follow-up seem unlikely to be differential with respect to the cataract end point and, thus, would also attenuate the true associations. Random misclassification of the cataract end point was reduced by the use of medical records to confirm the participant reports and by the use of strict diagnostic criteria. To control for possible surveillance bias, we included a term for the baseline report of an eye examination in the past 2 years in multivariate analyses. Finally, we controlled for a variety of measured confounders (Table 1), but other potential confounders that were either unmeasured or unknown may have contributed to the findings.

There have been 3 other prospective studies that have examined the relationship between dietary intake of lutein and risk of cataract. In the Nurses' Health Study of 77 466 female nurses, women in the top 10% of lutein/zeaxanthin intake, compared with those in the bottom quintile, had a 22% lower risk of cataract extraction (RR, 0.78; 95% CI, 0.63-0.95; test for trend, P = .04) during 12 years of follow-up.²⁸ In another study³² from that cohort, based on a small subsample of 408 participants, there was no association between intake of lutein/zeaxanthin and 5-year change in nuclear density as measured by analysis of digital images. In a second study,²⁷ data from the Health Professionals Study of 36 644 male health professionals showed that men in the highest quintile of lutein/zeaxanthin intake, compared with those in the lowest quintile, had a 19% lower risk of cataract extraction (RR, 0.81; 95% CI, 0.65-1.01; test for trend, P = .03) during 8 years of follow-up. A third study,²⁹ based on data from 1354 men and women participating in a nutrition substudy of the Beaver Dam Eye Study, showed that those in the highest quintile of intake of lutein/zeaxanthin in the distant past (10 years before baseline), compared with those in the lowest quintile, had a 50% decreased risk of incident nuclear opacity (odds ratio, 0.5; 95% CI, 0.3-0.8; test for trend, P = .002) at 5 years of follow-up. The present data from a large cohort of female health professionals indicate an approximately 20% decreased risk of cataract for those with high dietary intake of lutein/zeaxanthin and, thus, seem most consistent with the findings for cataract extraction reported in the Nurses' Health Study and the Health Professionals Study.^27,28 Of note, lutein/zeaxanthin intake in the reference group in the WHS and the 2 other cohorts of health professionals^27,28 seem markedly higher than the reference intake for lutein/zeaxanthin in the population-based sample from Beaver Dam, Wisconsin,²⁹ which may at least partially explain the smaller risk reductions observed in the former.

Of the other carotenoids examined in this study, only beta carotene from food and supplements showed a possible inverse relation with risk of cataract. Women in the highest, compared with the lowest, quintile of intake had a borderline significant 13% reduced risk of cataract in multivariate analysis. However, the test for trend across quintiles was not significant in the multivariate model or after adjustment for intake of other nutrients.These findings seem to be consistent with most earlier prospective studies^{27-29,32,46,47} that report a weak and statistically nonsignificant inverse trend between beta carotene level in the diet or blood and risk of cataract. More important, the results of 5 randomized trials^7,8,10,11,14 indicate that supplemental use of beta carotene (with or without other antioxidant supplements) for as long as 12 years has little impact on risk of cataract.

We observed a significant inverse trend between vitamin E intake from food and supplements and risk of cataract in this population of women. This inverse trend persisted after adjustment for other nutrients and seemed to be due largely to a 14% reduced risk of cataract for women in the highest quintile of intake. Median intake of vitamin E for this group of women was 262.4 mg/d, a level of intake difficult to attain from food sources alone. The reduction in risk seemed to reflect supplemental use of vitamin E rather than multivitamins. Seventy-one percent of women in the highest quintile of vitamin E intake reported using supplements of vitamin E at baseline, and adjustment for use of multivitamins had little impact on the RR estimate (RR comparing the extreme quintiles of vitamin E intake, 0.86; 95% CI, 0.73-1.00; test for trend, P = .048). Results of other prospective studies have been mixed, with some supporting an inverse association between dietary or serum vitamin E and cataract^32,46-50 and others reporting no association.^29,51,52 Data from 5 randomized trials^8,10-13 completed to date provide little evidence that use of vitamin E supplements, alone or in combination with other vitamin supplements, for as long as 6.5 years has any material impact on cataract development and progression. The final results for cataract during the 10-year treatment period for vitamin E in the WHS will be reported elsewhere.

The present data for vitamin C indicate a weak, and statistically nonsignificant, inverse association with risk of cataract.This finding seems to conflict with cross-sectional data presented in 2 recent substudies of the Nurses' Health Study^30,31 but is consistent with the results of several other prospective studies,^29,49-52 including 5-year follow-up data in the Nurses' Health Study subsample.³² Furthermore, findings from 3 randomized trials^8,11,12 indicate no major benefit of combined treatment with ascorbic acid and other antioxidants for as long as 6.5 years.

The hypothesis that antioxidant nutrients may protect against age-related damage to the human lens was derived from laboratory and animal studies and has been generally supported by findings of observational epidemiologic studies in humans. However, the results of completed randomized trials testing vitamin E, ascorbic acid, or beta carotene have been disappointing, and ongoing trials will determine whether observable benefits on cataract can emerge with longer-term treatment with these antioxidant vitamins. In the meantime, the results of the present study add to the growing body of observational evidence that suggests a possible beneficial effect of lutein/zeaxanthin in delaying cataract formation. Lutein and zeaxanthin are the only carotenoids detected in the human lens,^18-20 and the presence of oxidation products of lutein and zeaxanthin in the lens⁵³ further supports a functional role for xanthophylls in maintaining lens clarity.

In conclusion, these prospective data from a large cohort of female health professionals indicate that higher intakes of lutein/zeaxanthin and vitamin E are associated with decreased risk of cataract. Although reliable data from randomized trials are accumulating for vitamin E and other antioxidant vitamins, randomized trial data for lutein/zeaxanthin are lacking. Such information will help to clarify the benefits of supplemental use of lutein/zeaxanthin and provide the most reliable evidence on which to base public health recommendations for cataract prevention by vitamin supplementation.

Correspondence: William G. Christen, ScD, 900 Commonwealth Ave E, Boston, MA 02215-1204 (wchristen@rics.bwh.harvard.edu).

Submitted for Publication: April 10, 2006; final revision received December 6, 2006; accepted January 15, 2007.

Author Contributions: Dr Christen had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Financial Disclosure: Dr Christen has received research funding support from the National Institutes of Health, Harvard University (Clinical Nutrition Research Center), and DSM Nutritional Products, Inc (Roche). Dr Glynn has been funded by grants to the Brigham and Women's Hospital from AstraZeneca, Bristol-Myers Squibb, Merck, and Novartis. Dr Gaziano has received investigator-initiated federal funding from the National Institutes of Health (National Cancer Institute, National Heart, Lung, and Blood Institute, National Institute on Aging, and National Eye Institute) and the Department of Veterans Affairs (Cooperative Studies Program) and nonfederal investigator-initiated funding from McNeil Consumer Products and PLIVA; has received research support in the form of pills and/or packaging from BASF, DMS Pharmaceuticals, and Wyeth Pharmaceuticals; has received honoraria from Bayer and Pfizer for speaking engagements; has served as a consultant for McNeil Consumer Products and Wyeth Pharmaceuticals; and has served as an expert witness for Merck, Nutraquest, and GlaxoSmithKline. Dr Buring has received investigator-initiated research funding and support as Principal Investigator from the National Institutes of Health (National Heart, Lung, and Blood Institute, National Cancer Institute, and National Institute on Aging) and Dow Corning Corporation; has received research support for pills and/or packaging from Bayer Health Care and the Natural Source Vitamin E Association; has received honoraria from Bayer for speaking engagements; and serves on an external scientific advisory committee for a study by Procter & Gamble.

Funding/Support: This study was supported by research grants CA 47988, HL 43851, and EY 06633 from the National Institutes of Health and by DSM Nutritional Products, Inc.

Role of the Sponsor: DSM Nutritional Products, Inc, had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; or preparation, review, and approval of this manuscript.