The prevalence of cataract is high among elderly individuals, and it is the primary cause of blindness worldwide [1
]. The number of people with age-related cataract has been predicted to increase dramatically in the next 20 years worldwide, especially in Western countries, because of the increasing life expectancy [2
]. The only available treatment to restore vision after cataract formation is via intraocular surgery, and this imposes a major medical cost and heavy workload burden on the health care systems [3
Oxidative stress is believed to be involved in the formation and maturation of cataract, causing damage to the proteins and lipids of the lens epithelium [4
]. Age, ultraviolet light exposure, smoking, and corticosteroid use are recognized risk factors associated with cataract formation [7
]. Besides other risk factors such as abdominal obesity [9
] and hormonal therapy [10
], the production of proinflammatory components increases the levels of free oxygen radicals [11
], which have been associated with an increasing risk of cataract.
A significant and growing body of observational research suggests that an appropriate nutritional intervention may offer a way to decrease the risk of cataract [13
]. Accordingly, the association between cataract progression and nutritional exposure is a matter of great scientific interest. Recently, an inverse association between the total antioxidant capacity (TAC) and the risk of age-related cataract has been reported [8
]. In some trials, free oxygen radicals have been identified as one of the most important causes of cataract formation [10
]. A fair number of cohort studies have focused on assessing the association between the intake of several specific nutrients in the diet and cataract formation [11
]. In some of those studies, an antioxidant-rich diet was reported to potentially delay cataract progression [20
]; other studies have reported that healthier diets could prevent cataract formation [21
]. Some nutritional supplements, including different antioxidants [26
], lutein/zeaxanthin [27
] and omega-3 [31
] and omega-6 [32
] fatty acids, have been suggested to reduce cataract progression.
The Mediterranean diet (MedDiet) is recognized as one of the healthiest dietary patterns and has proven beneficial for several health outcomes [33
]. However, no randomized trial to date has assessed the long-term effect of a MedDiet on the risk of cataract development in a large population of elderly subjects. We hypothesized that two MedDiets, one enriched with extra-virgin olive oil (EVOO) and another with mixed nuts, rather than a low-fat control diet could decrease the incidence of cataract surgery in a population at a high risk of cardiovascular disease (CVD) enrolled to the Prevención con Dieta Mediterránea (PREDIMED) trial.
2.1. PREDIMED Trial
The analysis was conducted within the framework of the PREDIMED study (www.predimed.es
], a parallel-group, randomized, primary CVD prevention trial in persons at high risk of CVD. The main results of the trial related to the primary cardiovascular endpoint have been published elsewhere [36
]. The study was conducted in accordance with the Declaration of Helsinki. The Institutional Review Board of the respective recruitment centers approved the study protocol and all participants gave their informed consent. Trial Registration: clinicaltrials.gov Identifier: ISRCTN35739639.
Participants were men (age range: 55–80 years) or women (age range: 60–80 years) initially free of CVD. The inclusion criteria were the presence of either type 2 diabetes or ≥3 major cardiovascular risk factors, including current smoking (>1 cigarette/day during the last month), hypertension (systolic BP ≥140 mmHg or diastolic BP ≥90 mmHg or under antihypertensive medication), LDL cholesterol ≥160 mg/dL or receiving lipid-lowering therapy, HDL cholesterol ≤40 mg/dL in men or ≤50 mg/dL in women, body mass index (BMI) ≥25 kg/m2
, and a family history of premature coronary heart disease [35
The exclusion criteria were a previous history of CVD (i.e., a previous medical diagnosis of acute myocardial infarction, stroke, or peripheral arterial disease), any severe chronic illness, immunodeficiency or human immunodeficiency virus positivity, illegal drug or alcohol abuse, history of allergy to olive oil or nuts, and low predicted likelihood of changing dietary habits according to the Prochaska and DiClemente stages of change model [35
In total, 7447 participants were enrolled in the PREDIMED study. The selection process started by extracting the names of potential participants from the records of around 200 primary care centers (PCCs) affiliated with 11 Spanish teaching hospitals between October 2003 and January 2009. The clinical records of these participants were then individually reviewed to exclude those who did not meet the eligibility criteria. Potential participants were approached by the PCCs via a telephone call or during their clinical visits. If the candidates were interested in participating, a face-to-face interview was scheduled. During this interview, the purpose and characteristics of the study were explained, and signed informed consent was obtained from willing participants. A brief explanation of the study, including the possibility that they might receive free allowances of EVOO or nuts for the duration of the trial, was given at this first visit [35
2.3. Randomization and Intervention
Once entered into the study, the participants were randomly assigned in a 1:1:1 ratio to one of the following three intervention groups: (1) MedDiet supplemented with EVOO (MedDiet + EVOO); (2) MedDiet supplemented with mixed nuts (MedDiet + Nuts); or (3) a control diet (a low-fat diet according to the American Heart Association guidelines applicable as of 2002). The two groups allocated the MedDiets received intensive education to follow the MedDiet and supplemental foods at no cost. EVOO (1 L/week for the participant and his/her family) was provided to the MedDiet + EVOO group, and mixed nuts (30 g/day; 15 g walnuts, 7.5 g hazelnuts, and 7.5 g almonds) were provided to the MedDiet + Nuts group. The participants in the control group did not receive education on the MedDiet, but were adviced to follow a low-fat diet [35
Randomization was performed centrally by means of a computer-generated random-number sequence. Investigators and members of all committees were blinded to the treatments assigned to individual participants. In the present analysis, our main objective was to determine the effect of the three dietary interventions on the incidence of cataract surgery.
2.4. Follow-Up and Adherence
At baseline and during the yearly follow-up visits, all participants underwent clinical evaluations with their general practitioners. In addition they completed a 14-item questionnaire to assess adherence to the MedDiet; a 77-item general questionnaire about lifestyle variables, educational achievement, history of illnesses (including cataract surgery), and medication use; a 137-item validated food-frequency questionnaire [36
]; and a validated Spanish version of the Minnesota Leisure-Time Physical Activity Questionnaire.
Biomarkers of adherence to the MedDiet interventions were measured in a random sample of PREDIMED participants during the first 5 years of follow-up, and included urine hydroxytyrosol concentrations and plasma α-linolenic acid proportions, which are reliable biomarkers of EVOO and walnut intake, respectively [36
]. General practicioners and laboratory technicians were blinded to the participants’ intervention groups.
2.5. Outcome Measures
The outcome we investigated was the occurrence of cataract surgery at any time throughout the study period. Cataract surgery (externally confirmed by an independent Adjudication Committee blinded to the intervention and to the dietary habits of the participants) was a prespecified secondary outcome of the PREDIMED trial. Patients with cataract surgery in any eye present at baseline were excluded from the analyses. For the assessment of incident cataract surgery during the follow-up period, the participants visited their general practitioners yearly for clinical evaluations and completed several questionnaires, including a general medical questionnaire recording changes in the health status and, specifically, if they had undergone cataract surgery. The cataract surgery outcome was defined by the medical diagnosis made by an ophthalmologist, and was explicitly reported in the medical charts. The occurrence of cataract surgery was also confirmed by periodically reviewing the computer-based records of the corresponding PCCs. These reports and all relevant documentation, including medical records made by the ophthalmologist, were sent to the PREDIMED members of the Clinical Adjudication Event Committee, who were blinded to the interventions. As cataract extraction was a predefined secondary endpoint in the trial, the Adjudication Committee reviewed the medical charts comprehensively for potential cataract extraction, and only definitively confirmed cases were included in this analysis. Cases of traumatic cataracts and cases that appeared after another intraocular surgery, such as vitrectomy or glaucoma surgery, were also excluded. In cases of bilateral surgery in the same patient, only the first event was considered in our time-to-event analyses. Our analyses focused on a subset of 5802 participants from the PREDIMED trial who had not undergone cataract surgery at baseline: 1998, 1914, and 1890 participants were allocated to the MedDiet + EVOO, MedDiet + Nuts, and control diet groups, respectively. The remaining 1645 participants enrolled in the PREDIMED study were excluded because they had undergone cataract surgery before the beginning of the trial.
2.6. Statistical Analysis
Baseline differences between the three dietary intervention groups were tested using analysis of variance or chi-squared tests, and results were expressed as means ± SD or numbers (percentages), respectively. The normality of variables was examined by using the Kolmogorov-Smirnov test. All analyses were performed on the basis of an intention-to-treat principle.
Person-time of follow-up was calculated as the interval between the randomization date and the earliest date of the follow-up contact at which a new cataract surgery was recorded, or the last visit or death, whichever came first.
We used unadjusted, age- and sex-adjusted, and multivariable time-dependent Cox proportional hazard models to assess the effect of the two MedDiet interventions on the incidence of cataract surgery in comparison with the control group. Hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated using the control group as the reference. In multivariable models, we adjusted the estimates for the following confounders: age, sex, baseline type 2 diabetes, baseline hypertension, baseline BMI (4 categories), and smoking status (3 categories). All models were stratified by recruitment center. Prespecified interactions with sex, age, or baseline type 2 diabetes were tested using the likelihood ratio test in fully adjusted models. A fully adjusted multivariable analysis was repeated after both the MedDiet groups were merged into a single category for comparison with the control group. The assumption of proportional hazards was tested by analyzing the scaled Schoenfeld residuals, and it was not violated (p > 0.50). The test for time-varying covariates also suggested that the assumption of proportional hazards was met. We also used the Kaplan-Meier method to graphically estimate the cumulative incidence of cataract surgeries. A 2-tailed p value < 0.05 was considered statistically significant. In addition, we performed prespecified subgroup analysis within the strata of age, sex, and BMI. Analyses were performed using IBM SPSS Statistics for Windows Version 19.0 (IBM Corp., Armonk, NY, USA) and Stata 12.0 (StataCorp, College Station, TX, USA).
The overall baseline characteristics and intake of energy and nutrients as well as the consumption of key foods by the study participants in the dietary intervention groups are listed in Table 1
and Table 2
, respectively. The mean age of the participants was 66 years and their mean BMI was 30.1 kg/m2
; 45% of them were men. The participants in all three intervention groups were well matched for age, sex, anthropometric features, and other cardiovascular risk factors (data not shown).
During the follow-up period, 559 of the 5802 participants (9.6%) from the three intervention groups underwent cataract surgery, including 206 (10.3%) from the MedDiet + EVOO group, 174 (9.1%) from the MedDiet + Nuts group, and 179 (9.4%) from the control group. It must be noted that the number of participants who underwent cataract surgery does not correspond with the total number of eyes that underwent surgeries because we considered only the number of participants operated, regardless of whether they underwent surgery in one or both eyes.
We did not observe a reduction in the incidence of cataract surgery in the groups assigned to the MedDiet versus the control group. The observed rates (per 1000 person-years) were 16.9, 17.6, and 16.2 for the MedDiet + EVOO, MedDiet + Nuts, and control groups, respectively (Table 3
). Figure 1
displays the incidence of cataract surgery and the HRs and their 95% CIs for the effect of the two MedDiet interventions in comparison with the control group. Compared with the control group, the MedDiet + EVOO group had an unadjusted HR for the incidence of cataract surgery of 1.02 (95% CI, 0.84–1.25) and the MedDiet + Nuts group had a value of 0.97 (95% CI, 0.78–1.19). When we adjusted for some possible confounders (age, sex, baseline type 2 diabetes, baseline hypertension, baseline BMI, and smoking status), we did not observe a significant difference between the MedDiet groups and the control group, with multivariable-adjusted HR of 1.03 (95% CI, 0.84–1.26) for the MedDiet + EVOO group and 1.06 (95% CI, 0.86–1.31) for the MedDiet + Nuts group. Moreover, no significant difference was observed in the incidence of cataract surgery between the control group and both the MedDiet groups, when combined. In addition, we did not find any meaningful difference between the control group and the MedDiet groups when we analyzed for different subgroups of age, sex, and BMI (Table 4
Our analysis in the setting of the PREDIMED trial, including a middle-aged and elderly population with three or more CVD risk factors, was not able to confirm the initial hypothesis, because after 7.0 years of follow-up, we did not observe any significant differences in the incidence of cataract surgery between the participants assigned to the two MedDiet groups and the control group. The results showed that for both the MedDiet groups, the HRs for the incidence of cataract surgery were near the reference (null) value of 1.0. The results also remained non-significant after adjusting for certain potential confounders and analyzing the different subgroups according to age, sex, and BMI. Our findings are in accordance with those of good-quality trials, such as the Age-Related Eye Disease Study, and prospective epidemiological studies [37
] that did not provide support for a beneficial effect of antioxidants in reducing the risk of cataract formation or extraction. However, we acknowledge that other studies did show a significant reduction in cataract incidence or progression associated with the intake of antioxidants [13
The primary aim of the PREDIMED trial was to assess the role of a MedDiet in the primary prevention of cardiovascular events [36
].The main focus of the intervention was to change the overall dietary pattern, adding to a healthy diet the extra benefit of monounsaturated and polyunsaturated fatty acids and other nutrients provided by EVOO and nut supplements, instead of focusing on changes in single macronutrients or micronutrients. In contrast with the null effect on cataract surgery, both the MedDiets supplemented with EVOO or nuts were beneficial in reducing the risk of major cardiovascular events [36
], when compared to a low-fat diet, thereby demonstrating the protective effects of the MedDiet including EVOO and nuts.
To our knowledge, this study is the first large randomized trial to assess the effects of a MedDiet on the incidence of cataract. Numerous studies have investigated the association between diet and cataract and have focused on the effects of single micronutrients with antioxidant properties (e.g., vitamins A, B, C, and E; lutein/zeaxanthin; beta-carotene; or selenium). Although several of them have demonstrated an inverse relationship between these micronutrients and the development of age-related cataract or the incidence of cataract surgery, others have reported inconsistent results [38
], showing no consensus about their importance. Instead, other studies have focused on the TAC [8
], taking into account the capacity of all antioxidants in the diet and their synergic effect, demonstrating that dietary TAC was inversely associated with the risk of age-related cataract. Other trials have focused on the effects of single macronutrients such as lipids, and several have reported an inverse association between the intake of polyunsaturated fatty acids, like omega-3 and omega-6 fatty acids, and the risk of cataract [31
]. The MedDiet is considered one of the healthiest dietary patterns and contains all the individual micronutrients that have an antioxidant effect. In our study, we tested the combined benefits of a MedDiet as well as EVOO and nuts that contain significant amounts of omega-3 fatty acids and bioactive compounds (including fiber, minerals, tocopherols, phytosterols, and phenolic compounds) with strong antioxidant effects. However, despite combining the effects of single antioxidants and the benefits provided by EVOO and nuts, the MedDiet used in our study did not show any superiority in reducing the incidence of cataract surgery when compared to a low-fat diet. Lu et al. [40
] observed that saturated and polyunsaturated fats were not associated with the risk of cataract extraction, even though total fat intake was marginally associated with an increased risk of cataracts [31
]. We cannot exclude the fact that the low-fat diet followed in our control group may have masked the potential benefit of the MedDiet.
The present study has some limitations and strengths that should be considered. The first limitation is the type of population participating in the study. As the participants were individuals with either type 2 diabetes or several CVD risk factors, like obesity and smoking, our findings cannot be extrapolated to younger subjects or to other populations. Moreover, these factors lead to the production of proinflammatory and oxidative agents and might play important roles in accelerating cataract formation. Furthermore, having type 2 diabetes could influence the strategy of when to operate a cataract, because in these patients, earlier cataract extraction has been known to contribute to an improved visual outcome [41
]. Second, the timing of cataract extraction depends on multiple factors that can speed up or delay the surgery, including the ophthalmologist’s subjective decision; the patient’s visual acuity, comorbidities, and patient waiting attitude; being a more objective outcome formation of cataract diagnosed by slit lamp examination. Third, to assess the incidence of cataract surgery we considered only one surgery event in every patient, even though the same patient underwent cataract surgery in both eyes. This could have led to the underestimation of the real incidence of cataract surgery in our cohort, because other similar trials have observed an incidence of cataract surgery between 17.7% and 26.8% [42
] , which was very different from our findings. Fourth, previous studies have shown that a high total dietary fat intake is related to a greater risk of developing cataracts [40
], because dietary fat may affect lens cell membrane composition and function, which are related to age-related cataract [44
]. Although there is lack of evidence, a low-fat diet could have a preventive role on the risk of cataract extraction. In our study, this diet could have masked the potential beneficial effect of the MedDiet on the incidence of cataract surgery, because a low-fat diet (control group) could be as effective as the MedDiet (intervention groups) in reducing the incidence of cataract surgery, without showing any difference between the two groups. Fifth, considering the outcome we investigated in this study, it could have been useful the additional presence of a fourth group (second control group) following no specific diet (neither MedDiet nor low-fat diet) to compare it with the three groups of the study to obtain information that could have indicated whether the MedDiet or the low-fat diet had any correlation with the incidence of cataract surgery, as comparing only the study groups we could not find any correlation. Sixth, the assessment of cataract surgery was not the primary endpoint, because the PREDIMED trial was designed to assess the effect of the MedDiet on primary CVD prevention. Lastly, our implicit assumption regarding the induction period for relating the diet to cataract surgery could be regarded as insufficient, even if the induction period was 6–7 years in our study. Given that cataract surgery implies there was a long-standing previous pathophysiological process in the lens, which probably started many years before our recorded date of surgery, our assumption regarding the induction period could provide an alternative explanation to our null results. We admit that we may have reduced our sensitivity in identifying the cataract-related endpoint. However, the advantage of using surgery in our operational definition of this endpoint is its high specificity. Moreover, it is well known in epidemiology that, theoretically, with perfect specificity, the nondifferential sensitivity of disease misclassification will not bias the relative risk estimate [45
The main strengths of our study are its randomized design; long follow-up period relative to previous studies; large sample size; good compliance with the dietary regimens assigned, which is further supported by the MedDiet questionnaire results and changes in the biomarkers of food supplementation; and the control for several potential confounders, which together with the randomization allows us to rule out residual confounding.