Albumin and Transthyretin as Risk Factors for Cataract: Title and subTitle BreakThe POLA Study FREE

Cataract is the leading cause of blindness, accounting for 50% of blindness worldwide.¹ Although the problem is especially critical in developing countries, it affects the more developed countries as well. For instance, cataract surgery has become the most frequent surgical procedure in people 65 years or older in the United States, with an estimated $3.4 billion cost to Medicare in 1991.² An increased understanding of the etiology of cataract may lead to the development of nonsurgical strategies to delay or prevent cataract. In the United States, the National Eye Institute, Bethesda, Md, estimates that a 10-year delay in the onset of cataract would result in a 50% reduction in the prevalence of cataract.¹

Nutritional factors probably play an important role in cataractogenesis. Recently, most of the attention has been focused on the possible role of antioxidant micronutrients.³ However, previous studies in animals and humans have suggested that other nutritional factors may be important.^{4 - 5} In particular, several studies have indicated that protein energy malnutrition, as indicated by low protein intake, low albumin-globulin ratio, or low body mass index (BMI), may be associated with increased risk for cataract.^{6 - 9}

In the present study, we analyze the associations of cortical, nuclear, and posterior subcapsular (PSC) cataracts with albumin and transthyretin, 2 important markers of the body protein pool.¹⁰ Since albumin and transthyretin are negative acute-phase proteins, inflammation down-regulates their synthesis rate. To assess the relative contribution of inflammation and malnutrition in albumin and transthyretin levels, 2 biological markers of inflammation (high-sensitivity C-reactive protein [hs-CRP] and orosomucoid) were also measured.

The Pathologies Oculaires Liées à l’Age (POLA) Study is a prospective study aimed at identifying the risk factors of age-related eye diseases (cataract, age-related macular degeneration). The methods of this study have been published elsewhere.¹¹ Briefly, inclusion criteria were (1) being a resident of Sète (South of France) and (2) being 60 years and older. According to the 1990 population census, there were almost 12 000 eligible residents, from which our objective was to recruit 3000 participants. The population was informed of the study through the local media (television, radio, and newspapers). We also contacted 4543 residents individually by mail and telephone, using the electoral roll. Between June 1995 and July 1997, we recruited 2584 participants.

This research followed the tenets of the Declaration of Helsinki. Participants gave written consent for participation in the study. The design of this study was approved by the ethical committee of the University Hospital of Montpellier, Montpellier, France.

Four ophthalmologists performed the ophthalmologic examinations. This examination included a recording of ophthalmologic history (lens extraction and the year of extraction); a measure of best-corrected far visual acuity in the right and left eyes; after pupil dilation, a quantitative assessment of nuclear, cortical, and PSC lens opacities at slitlamp examination according to the Lens Opacities Classification System III (LOCS III)¹² ; and one 50° color photograph centered on the macular area in each eye.

The type and degree of lens opacification were graded at slitlamp examination following the procedures of the LOCS III.¹² This system, which is based on standard photographs, provides decimal, nearly continuous grades for nuclear opalescence (ranging from 0-6.9, using 6 standards), nuclear color (ranging from 0-6.9, using 6 standards), cortical opacities (ranging from 0-5.9, using 5 standards), PSC opacities (ranging from 0-5.9, using 5 standards).

As in the other publications from the POLA Study,^{13 - 15} the presence of cataract was defined as nuclear color or nuclear opalescence grades of 4 or higher for nuclear opacities, cortical opacity grade of 4 or higher for cortical opacities, and PSC opacity grade of 2 or higher for PSC opacities. This level of opacification corresponded to significant visual impairment in most participants.

Participants were classified as having a single type of cataract (nuclear, cortical, PSC) when only 1 type of opacity was present. The nuclear cataract group, for instance, consisted of participants with only nuclear cataract in both eyes or only nuclear cataract in one eye and no cataract in the other eye. The mixed cataract group consisted of participants with various combinations of nuclear, cortical, and PSC cataracts in one or both eyes. All other subjects were considered free of cataract (nuclear opalescence, nuclear color, and cortical opacity grades <4 and PSC grade <2 in both eyes).

Participants who had already had bilateral lens extractions formed a separate group (bilateral cataract surgery). Participants with unilateral lens extraction (n = 100) were classified according to their other eye.

Biological measurements were made from fasting blood samples performed at home on the morning of the examination. They included measurements in plasma (cholesterol; triglycerides; vitamins A, E, and C; and glutathione peroxidase) and in red blood cells (superoxide dismutase). Measurement of plasma glutathione peroxidase concentration was performed by the enzyme-linked immunoassay BIOXYTECH pl GPx-EIA (OXIS International SA, Portland, Ore). Red blood cell superoxide dismutase activity was measured by a spectrophotometric assay (BIOXYTECH SOD-525; OXIS International SA).

In 2001, we performed a new series of biochemical measurements. The plasma samples were kept frozen at −80°C for approximately 5 years before these measurements were performed. Plasma albumin, transthyretin, and orosomucoid concentrations were determined by immunoturbidimetric methods, while hs-CRP concentration was determined by latex-enhanced immunoturbidimetric method using reagents from Olympus (Rungis, France) on an Olympus AU2700 biochemistry analyzer.¹⁶

Data were collected by trained study personnel who were unaware of cataract status. A standardized interview was performed to assess sociodemographic variables (eg, marital status, educational level), medical history (eg, treated hypertension, cardiovascular diseases, diabetes mellitus), all medications currently used, and lifestyle factors (eg, smoking, physical exercise, sunlight exposure). The interviewer then measured height, weight, waist and hip circumferences, and systolic and diastolic blood pressures.

Participants were considered as having high education if they had reached at least the end of high school. Diabetes mellitus was defined as self-reported history of diabetes mellitus confirmed by current antidiabetic therapy and/or a fasting plasma glucose level of 126 mg/dL or higher (≥7 mmol/L). History of cardiovascular disease was defined as a history of myocardial infarction, stroke, or angioplasty. Hypertension was defined as known treated hypertension confirmed by current use of antihypertensive drugs and/or systolic blood pressure of 160 mm Hg or higher and/or diastolic blood pressure of 95 mm Hg or higher. Body mass index was defined as weight in kilograms divided by the height in meters squared. Sunlight exposure was estimated using lifetime residential history.¹⁵

Lens examination was lacking in both the right and left eye in 25 participants (1%); in 13 cases, it was because of lack of dilation (8 refused, 5 had contraindications), and in 12 cases, it was because of technical failure. We excluded 1 case of traumatic cataract. Thus, cataract status could be determined in 2558 participants (99%) of 2584.

Among the 2558 subjects with available cataract status, 60 (2.3%) had some missing data in the initial set of biochemical variables (eg, lipids, antioxidant vitamins), mainly because of refusal of blood sampling or technical failure. In addition, 144 subjects (5.6%) had some missing data in the biochemical measurements performed in 2001. This is mainly because of missing tubes in the plasma collection because of insufficient quantity of blood sampling.

Among the 2354 subjects with available cataract and biochemical data, 133 (5.6%) had some missing interview data, leaving 2221 subjects for the statistical analyses.

For each biochemical variable of interest, we determined the 20th and 80th percentile values in all subjects. This formed 3 groups (low quintile, middle quintiles, high quintile). The reference was taken as the highest quintile for albumin and transthyretin levels (since we expected an increased risk in subjects with low values of albumin and transthyretin) and as the lowest quintile for hs-CRP and orosomucoid (since we expected an increased risk in subjects with high values of hs-CRP and orosomucoid). Age- and sex-adjusted odds ratios (ORs) (and their 95% confidence intervals [CIs]) were obtained by logistic regression, with the type of cataract as the dependent variable and age, sex, and the 2 nonreference quintile groups as the independent variables. Potential confounding variables were added to the model to obtain multivariate ORs. For each type of cataract, all variables that have previously been identified as significant risk factors in the POLA Study were considered as potential confounders.^{13 - 15} Tests for trend were performed by entering the biochemical variable in the logistic regressions as a 3-category variable instead of 2 independent variables.

The characteristics of the participants in the POLA Study (with complete data for this analysis) are presented in Table 1. Among the 2221 subjects, 143 (6.4%) were classified as having PSC cataract only; 75 (3.4%), cortical cataract only; 127 (5.7%), nuclear cataract only; 130 (5.8%), mixed cataract; and 111 (5.0%), cataract surgery. The group with mixed cataract was composed of 20 participants with cortical and PSC cataract, 17 with nuclear and cortical cataracts, 63 with nuclear and PSC cataract, and 30 with all 3 types. The remaining 1635 subjects were classified as having no cataract and constituted the control group in all statistical analyses. The prevalence of the different types of cataract was similar in men and women, except for cortical cataract, which was about twice as frequent in women. As expected, diabetes mellitus, cardiovascular disease, and smoking were much more frequent in men than in women. Body mass index was similar among sexes (mean, 26.8 kg/m² in men and 26.2 kg/m² in women).

As shown in the Figure, the distribution of plasma albumin levels was skewed to the left. In 6.2% of men and 8.4% of women, the plasma albumin level was lower than 35 g/L. The distribution of transthyretin levels was more symmetrical. Women had much lower transthyretin levels than men (mean ± SD, 0.24 ± 0.05 vs 0.27 ± 0.05; P<.001). In 3.4% of men and 5.8% of women, the plasma transthyretin level was lower than 0.16 g/L. The distribution of hs-CRP and orosomucoid levels was skewed to the right. Most of the subjects had hs-CRP values lower than 5 mg/L. Only 9.6% of men and 6.0% of women had hs-CRP levels higher than 10 mg/L.

As shown in Table 2, after adjustment for age and sex, the lowest quintile of plasma albumin levels was associated with a 2-fold increased risk for mixed cataract (OR, 2.08 [95% CI, 1.08-4.03]), but there were no significant associations with the other types of cataract. After adjustments for potential confounders, the associations of plasma albumin level with mixed cataract persisted (OR, 1.87 [95% CI, 0.95-3.68]; P for trend = .02). Overall, the risk for cataract (any type) was increased by 49% in subjects with low plasma albumin levels (multivariate OR, 1.49 [95% CI, 1.04-2.14]).

As shown in Table 3, low plasma transthyretin levels were significantly associated with increased risk of nuclear and mixed cataract (OR, 2.74 [95% CI, 1.39-5.39] and OR, 2.53 [95% CI, 1.33-4.81], respectively) and remained significant after multivariate adjustment for potential confounders. The association of transthyretin with cataract surgery was close to significance after adjustment for age and sex (OR, 1.89 [95% CI, 0.94-3.81]; P for trend = .06). Multivariate adjustment for potential confounders further weakened this association (OR, 1.79 [95% CI, 0.86-3.76]; P = .11). Overall, the risk for cataract (any type) was increased by 48% in subjects with low plasma transthyretin levels (multivariate OR, 1.48 [95% CI, 1.03-2.13]).

Having both low albumin (<38.28 g/L) and low transthyretin (<0.21 g/L) levels did not add to the risk. Indeed, after adjustment for age and sex, the ORs for nuclear, mixed, and any cataract were, respectively, 1.46 (95% CI, 0.78-2.73), 2.82 (95% CI, 1.62-4.93), and 1.62 (95% CI, 1.15-2.29) in subjects with both low albumin and transthyretin levels, while they were 1.74 (95% CI, 1.09-2.76), 2.20 (95% CI, 1.37-3.56), and 1.39 (95% CI, 1.07-1.81) in subjects with low albumin or low transthyretin levels, by comparison with subjects with albumin levels of 38.28 g/L or higher and transthyretin levels of 0.21 g/L or higher.

We observed no significant associations of cataract with plasma hs-CRP or orosomucoid levels (Table 4 and Table 5, respectively). Further adjustment for hs-CRP and orosomucoid levels did not materially change the associations of albumin and transthyretin levels with cataract (for instance, for any cataract, OR, 1.48 [95% CI, 1.01-2.16] and OR, 1.49 [95% CI, 1.03-2.15] for low vs high levels of transthyretin and albumin, respectively).

In the present study, after adjustment for potential confounders, low plasma levels of albumin and transthyretin, which are markers of protein malnutrition,¹⁰ were associated with increased risk of cataract, in particular, mixed and nuclear cataracts.

Similar to our results, in the Lens Opacities Case-Control Study, plasma albumin level was significantly associated with mixed opacities, while the albumin-globulin ratio was associated with nuclear and mixed cataracts.¹⁷ However, there were no associations of plasma albumin level with cataract in the Italian-American Cataract Study.¹⁸ To our knowledge, there are no previous studies on plasma transthyretin level and cataract. Although the synthesis rate of albumin seems not to be influenced by age up to 80 years,¹⁹ serum albumin and transthyretin levels are affected by nutritional intake and pathological conditions. Daily protein intake and sufficient energy supply are needed to prevent protein from being used as an energy source via gluconeogenesis. However, albumin and transthyretin levels are affected by other factors than nutrition, such as liver disease, renal failure, or inflammation. Indeed, in recent years it has been recognized that inflammation, alone or in combination with a low protein intake, plays a significant role in causing hypoalbuminemia via a cytokine-induced decrease in synthesis rate.^{20 - 22} In the present study, 2 major markers of inflammation were measured, namely, hs-CRP and orosomucoid. In contrast with the previous report from the Physicians' Health Study,²³ none of these showed significant associations with cataract. It is therefore unlikely that the associations of albumin and transthyretin with cataract were due to underlying inflammation. The absence of inflammation in patients with cataract suggested that the observed decrease in albuminemia and transthyretinemia is mainly due to low protein and energy intake.

Other studies have assessed the associations of diet or anthropometric measurements with cataract. In the Blue Mountains Eye Study, high protein intake was associated with a reduced risk of nuclear cataract.⁶ In an Indian case-control study, low protein intake was associated with increased risk of PSC, nuclear, and mixed cataracts.⁷ In this study, protein intake was taken as a surrogate of general good nutrition, since all nutritional variables were highly correlated. In an Italian case-control study, high intakes of meat and cheese were associated with a lower risk of cataract extraction.⁸

Low BMI is also a marker of protein-energy malnutrition. In 4 studies, low BMI was associated with increased risk of nuclear cataract,^{7 ,9 ,24 - 25} while in 4 other studies, including the POLA Study, there were no significant associations between BMI and nuclear cataract.^{14 ,26 - 28} These inconsistencies across studies may be explained by different distributions of BMI. For instance, while underweight is generally defined as a BMI lower than 22 kg/m² in elderly individuals, in the Salisbury Eye Evaluation Project,²⁵ BMI was considered low if lower than 22.5 kg/m², while in the Beaver Dam Eye Study,²⁷ low BMI was defined as lower than 26 kg/m². Moreover, other anthropometric measurements, such as midarm muscle circumference or midarm muscle area, are more sensitive indicators for somatic protein reserve, while BMI is influenced both by protein and fat mass.²⁹ To our knowledge, no epidemiologic studies have assessed the associations of cataract with specific indicators of muscle mass.

Low levels of albumin and transthyretin are associated with increased mortality in elderly individuals.^{30 - 34} Previous studies have shown an increased mortality in subjects with nuclear or mixed cataract after adjustment for a number of potential confounders.^{35 - 37} However, plasma albumin or transthyretin levels were not included in these analyses. It is therefore possible that the increased mortality observed in subjects with nuclear or mixed cataracts was due to their poor nutritional status.

The mechanism by which protein undernutrition may influence cataract development remains to be determined. Low protein intake may induce deficiencies of specific amino acids that are needed to maintain the health of the lens.⁴ In this respect, transthyretin has the highest proportion of essential to nonessential amino acids of any protein in the body.¹⁰ This may explain the stronger association of cataract with transthyretin level than with albumin level.

Alternatively, protein undernutrition may be associated with other nutritional deficiencies. For instance, in the Blue Mountains Eye Study, increased risk of nuclear cataract was found for low protein intake but also for low intakes of niacin, thiamin, riboflavin, and vitamin A.⁶ Low intakes of these vitamins were also associated with increased risk of cataract in the Beaver Dam Eye Study³⁸ and the Lens Opacities Case-Control Study²⁴ ; no information was given for protein intake in these studies. Intakes of protein, niacin, thiamin, and riboflavin are probably intercorrelated, since the major sources of these vitamins are also rich in protein (meat, fish, egg, milk, and cheese). It is therefore difficult to determine which of these nutrients plays a more important role in cataract development. Only randomized trials can answer this question. To date, only 1 randomized trial was performed in this field and showed a reduction of the risk of nuclear cataract in subjects who received niacin and riboflavin supplements by comparison with placebo.³⁹

Transthyretin is normally bound to the retinol-binding protein at a 1:1 molar ratio in physiologic pH and therefore plays a role in vitamin A transportation.¹⁰ In the POLA Study, we previously found a negative association of plasma retinol with nuclear and mixed cataracts and cataract surgery.¹⁴ Because of the close biological and statistical association of retinol with transthyretin (r = 0.60), it is not possible to determine which, of transthyretin and retinol, is more important to lens health. Because of the important collinearity of plasma retinol and transthyretin, we did not include retinol in the multivariate adjustments.

Our study has several limitations. First, our sample underrepresents older persons and overrepresents the middle and upper social classes by comparison with the whole eligible population.¹¹ The subjects of this study may thus be healthier and have different lifestyle habits, in particular concerning diet and physical activity, than the general population. This is likely to have affected the distribution of plasma albumin and transthyretin levels or the prevalence of cataract. However, it is unlikely to have affected the association between cataract and plasma measurements. Moreover, although a selection bias cannot be dismissed, the associations of the different types of cataract with their known risk factors (ie, smoking, diabetes, corticosteroid use, light exposure) are similar to those observed in other studies.^{14 - 15} Another limitation is the statistical power, in particular for cortical cataract (the less prevalent type of cataract in the present study). Indeed, the power to detect an OR of 2 was only 63%, while it was 90%, 86%, 87%, and 80% for PSC cataract, nuclear cataract, mixed cataract, and cataract surgery, respectively. The power for any cataract was logically much higher (98% for detecting an OR of 1.5).

In observational studies, the concern is always about confounding. We have therefore performed multivariate adjustments to take into account all known risk factors for each type of cataract. In addition to age, sex, and educational level, specific factors were used for the different types of cataract. The selected factors were those identified in previous analyses of the POLA Study.^{13 - 15} The associations of albumin and transthyretin with cataract were not affected by these adjustments.

Since this study is cross-sectional, we cannot assume that the low albumin and transthyretin levels preceded the development of cataract. It is possible that participants with cataract modified their diet after developing visual impairment. These results must, however, be confirmed in prospective studies. This will be carried out by our study, with a follow-up visit 3 years after baseline.

The results of the present study are suggestive of an association of protein undernutrition with increased risk of nuclear and mixed cataracts. Previous studies conducted in this field, although limited in number, are globally consistent with these results.

Correspondence: Cécile Delcourt, PhD, INSERM, Unité 500, 39 Ave Charles Flahault, Montpellier CEDEX 5 34093, France (delcourt@montp.inserm.fr).

Submitted for Publication: September 8, 2003; final revision received April 12, 2004; accepted June 21, 2004.

Financial Disclosure: None.

Funding/Support: This study was supported by the Institut National de la Santé et de la Recherche Médicale, Paris, France; the Fondation de France, Department of Epidemiology of Ageing, Paris; the Fondation pour la Recherche Médicale, Paris; the Région Languedoc-Roussillon, Montpellier, France; the Association Retina-France, Toulouse; Rhônes Poulenc, Paris; Essilor, Paris; Specia, Paris; and the Centre de Recherche et d'Information Nutritionnelle, Paris.

Additional Information: Dr Delcourt 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.