Cataract is a clouding or opacification of the lens inside the eye that obstructs the passage of light [1
]. Age-related cataract (ARC) is the leading causes of blindness and vision impairment worldwide [2
]. It was estimated that 20 million people older than 40 years were visually impaired due to ARC in the United States [3
]. Although new therapeutic methods emerged in recent years and most ARC cases can be cured, the high treatment costs and increasing demands for therapy will challenge the long-term economic stability of health care systems [4
]. With the rapidly aging population, ARC has brought a massive burden on health care and become an important public health issue. Thus, identifying modifiable factors available to prevent or delay the development of ARC is a crucial strategy.
Light-initiated oxidative damages are hypothesized to be the mechanism involved in ARC [5
]. The xanthophyll carotenoids lutein and zeaxanthin are uniquely concentrated in the lens, where they can attenuate photochemical damage by filtering high-energy short-wavelength light [6
]. In addition, they serve to protect the lens from oxidative damage by scavenging reactive oxygen species (ROS), indicating that these carotenoids may play a potentially important role in the prevention of ARC [8
]. Numerous epidemiological studies have investigated the relationship between dietary intake and blood levels of lutein and zeaxanthin and the risk of ARC [9
]. Because the accuracy of dietary intake measurements is greatly influenced by the different dietary assessment methods across the studies and the individual differences in utilization and absorption, blood concentrations appears to be a stronger predictor of nutritional status [11
]. Many studies reported that high serum concentrations in lutein and zeaxanthin were inversely associated with the prevalence of ARC; however, others failed to find such association or the results regarding certain subtypes of ARC were inconsistent [13
]. Therefore, we conducted this meta-analysis to evaluate the relationship between serum concentration of lutein and zeaxanthin and risk of difference ARC subtypes.
2. Experimental Section
This meta-analysis was performed in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [15
2.1. Search Strategy
We searched MEDLINE, EMBASE, ISI Web of Science and Cochrane Library for relevant articles up to April 2013. Following MeSH words and text words were used: “lutein”, “zeaxanthin”, “xanthophyll”, or “carotenoid”, together with each of the following words “cataract”, “age-related cataract”, “ARC”, or “lens opacities”. Language restrictions were not imposed. Besides, manual searches of references cited by the retrieved articles were performed for additional literature. Authors and experts in the field were also contacted for relevant unpublished studies.
2.2. Study Selection
All retrieved articles were examined by scanning their titles and abstracts and any clearly irrelevant studies were excluded. Full texts of the remaining articles were further reviewed. Studies were included in this meta-analysis if they met the following criteria: (1) the primary outcome was clearly defined as ARC; (2) the exposure of interest was blood (plasma or serum) levels of lutein and zeaxanthin; (3) odds ratio (OR) or relative risk (RR) estimates and their 95% confidence intervals (CIs) were provided; and (4) potential confounders were controlled for by matching or multivariable analysis in the studies. If a series of articles were published from the same study, the report with the most updated data was selected for analysis. Literature search and screening were conducted independently by three investigators (RBY, RL, and ZXH) in a standardized manner, and any disagreements were resolved by discussion.
2.3. Data Abstraction and Quality Assessment
Information from eligible studies was abstracted independently by three investigators (X.H.L., R.B.Y., and R.L.), and discrepancies were resolved by consensus. We extracted the name of the first author, year of publication, country of origin, study design, characteristics of the study population (sample size, distribution of age, sex and ethnicity), diagnostic method of ARC, classification and grading systems for ARC, blood concentration assessment methods, fully adjusted odds ratio or relative risk, and the adjustment factors. If a study provided several risk estimates, the most completely adjusted estimate was extracted.
2.4. Statistical Analyses
For all studies, RRs or ORs and the corresponding 95% CIs were extracted from each article. Inverse variance-weighted risk estimates for ARC comparing the highest-versus-lowest categories of lutein and zeaxanthin were pooled with the fixed effects models or DerSimonian-Laird random effects models. The DerSimonian and Laird’s Q statistic and I
² statistic were performed to evaluate statistical heterogeneity. Forest plots were used to visually assess the pooled ORs estimates across studies. Sensitivity analyses were performed to examine the influence of each individual study by iteratively excluding each study from the meta-analysis and comparing the point estimates including and excluding the study to determine their magnitude of influence on the overall estimate. Stratification and meta-regression analyses were performed to explore the source of heterogeneity among variables, such as country of origin, mean age of participants, method of ARC diagnosis and classification criteria of ARC. We assessed the potential of publication bias by using a funnel plot, Egger regression asymmetry test and Begg adjusted rank correlation test [16
]. All statistical analyses were conducted by using Stata version 10.0 (Stata Corporation, College Station, TX, USA).
The findings from this meta-analysis indicated that the reduction in the occurrence of nuclear cataract was significantly associated with high concentrations of lutein and zeaxanthin in serum, especially for zeaxanthin. Moreover, a stronger association between nuclear cataract and blood zeaxanthin might be noted for the studies conducted in the European Nations. However, no significant protective effects were found for each of these carotenoids against either cortical cataract or posterior subcapsular cataract, except a borderline significant association between blood lutein and subcapsular cataract.
The potential mechanisms for the positive effects of these carotenoids on ARC are not yet completely understood, but are thought to be primarily through their antioxidant properties and blue-light filtering activities [23
]. The lens of the eye is particularly susceptible to oxidative damage; increased generation of ROS and high-energy short-wavelength light exposure leads to cross-linking or aggregation of the crystalline proteins in lens epithelial cells and thereby results in the generation of cataract [24
]. Lutein and its isomer zeaxanthin are uniquely concentrated in the lens, indicating that each may have a possible particular function in this vital ocular tissue. They also have the potent antioxidant properties based on their abilities to quench singlet oxygen, scavenge superoxide and hydroxyl radicals, protect membrane phospholipids against UV-induced peroxidation, and reduce lipofuscin formation. In addition, absorption spectrum of these carotenoids peaks at 450 nm, consistent with the action spectrum for light-induced damage [25
]. Moreover, lutein and zeaxanthin were incorporated in higher amounts into cell membranes in a single orientation, making them ideal optical filters. Therefore these carotenoids can absorb and attenuate the photic damage in the human lens [26
Results from the present study indicated that blood concentration of lutein and zeaxanthin was significantly associated with the decreased risk of nuclear cataract. The consistency between our findings for each of these xanthophyll carotenoids indicated that such associations were robust. It should be noted that, compared to blood lutein, blood zeaxanthin showed a stronger inverse association with the occurrence of nuclear cataract. The seeming discrepancy was presumably explained by the difference in distribution pattern and biological properties of these two carotenoids. The ratio of zeaxanthin to lutein is much higher in human lens (0.8:1) than that in the plasma (0.2:1), suggesting that a selective uptake of zeaxanthin may occur in the lens tissue [27
]. Moreover, zeaxanthin adopts a roughly perpendicular orientation with respect to the plane of the membrane, which makes zeaxanthin more effective in protecting the lens against the UV-induced membrane damage [29
]. In addition, zeaxanthin appears to afford more protection to the liposomal lipids from light-induced oxidative stress, perhaps because zeaxanthin is a more effective singlet oxygen scavenger than lutein [31
The stratified analyses showed that the association between nuclear cataract and blood zeaxanthin was significantly stronger for Europeans than it was for the Indians. This might be partly explained by the difference in zeaxanthin status for different populations. Compared with the Indians, blood zeaxanthin in Europeans was relatively higher [12
]. Meanwhile, epidemiological data indicated that the average daily intake of lutein plus zeaxanthin was only about 2 mg per day and tended to less in Europe [32
]. The consumption of these carotenoids was far below the level (above 6 mg per day) that had been associated with reduced risks of ARC [9
]. Therefore, it is possible that Europeans with suboptimal zeaxanthin status might benefit more from improving blood zeaxanthin level via increasing consumption of foods rich in lutein and zeaxanthin. On the other hand, as the prevalence of the subtypes of ARC differs in different regions and populations, ethnic and racial differences might also have the potential to influence the association between nuclear cataract and zeaxanthin status [33
With respect to cortical cataract and subcapsular cataract, we found no significant associations of these types of cataract with blood zeaxanthin or lutein. The difference in associations between cataract subtypes and serum zeaxanthin and lutein was probably due to the distinct pathogenesis for each type of ARC. With increasing age, a lower percentage of reduced glutathione can reach the lens nucleus, which makes the nucleus become less able to repair oxidative damage [34
]. In contrast, glutathione levels in the outer cortex of the lens remain at high levels, even in nuclear cataracts [36
]. Therefore, nuclear cataract may be more prone to a significant association with serum zeaxanthin and lutein. In addition, a marginally inverse relationship had been found between blood lutein and subcapsular cataract in the present study; however, subcapsular cataract is least common among these three main types of ARC and further studies are needed to confirm such finding [37
The present study also has several potential limitations. First, our study is a meta-analysis of observational studies because of no randomized controlled trials (RCTs) found. It is commonly considered that observational studies are more likely to be subject to confounding and bias than RCTs, which can confound the results of our analysis. Second, the relatively small number of trials available evaluated the relationship between lutein and zeaxanthin status and the risk of cortical cataract or posterior subcapsular cataract; thus, our analysis was unlikely to detect substantially significant effect of xanthophyll carotenoids on opacities in these regions. Third, it may be that persons who have high blood levels of lutein and zeaxanthin are more likely to adhere to a healthier diet or lifestyle, which is associated with lower risks of ARC. Most studies adjusted for some major potential confounders; however, it was possible that residual or unknown confounding might bias the overall results in either direction. Fourth, although most subjects in studies from the European Nations were Europeans, it is possible that these study groups included participants of Asian and other non-European participants, which might lead to a mis-estimation of the association. Finally, although there was little evidence of publication bias from our funnel plot and rank correlation analysis, this possibility could be not completely excluded.