Risk Factors for Incident Cortical and Posterior Subcapsular Lens Opacitiesin the Barbados Eye Studies

Objective  To evaluate risk factors for the 4-year incidence of cortical and posteriorsubcapsular (PSC) lens opacities.

Design  Population-based cohort study with 85% participation at 4-year follow-up.

Participants  Three thousand one hundred ninety-three black participants of the BarbadosEye Studies, Barbados, West Indies, of whom 2040 and 2954 were free of corticaland PSC lens opacities, respectively, at baseline.

Methods  The standardized protocol at baseline and follow-up included an interview,anthropometric and blood pressure measurements, and ophthalmic measurementsincluding slitlamp lens grading, fundus photography, and an ophthalmologicexamination. Factors associated with incident cortical and PSC opacities (LensOpacities Classification System II, ≥2) were evaluated by logistic regression.

Main Outcome Measure  Relative risks (RRs) with 95% confidence intervals.

Results  The 4-year incidence of cortical lens opacities was 22.2% (452/2040);the factors increasing risk were older age, female gender (RR = 1.3), lowsocioeconomic status (RR = 1.4), and a history of diabetes mellitus (RR =2.4), while aspirin use was associated with a lower RR (RR = 0.2; 95% confidenceinterval, 0.1-0.8), a result based on small numbers. The 4-year incidenceof PSC opacities was lower at 3.3% (97/2954), and risk also increased withage and a history of diabetes mellitus (RR = 2.9). A dose-response relationshipwas evident between incident opacities and increased levels of glycosylatedhemoglobin at baseline, with the highest risk of cortical (RR = 3.60; 95%confidence interval, 2.23-5.81) and PSC (RR = 4.93; 95% confidence interval,2.69-9.05) opacities at more than an 11.5% glycosylated hemoglobin level.

Conclusions  Diabetes mellitus and hyperglycemia are major modifiable risk factorsfor the development of cortical and PSC lens opacities in this African-descentpopulation with a high rate of diabetes mellitus. Prevention and improvedcontrol of diabetes mellitus are likely to reduce the burden of cataract.The finding of a reduced incidence of cortical lens opacities in aspirin usersmerits further investigation, given its potential for cataract prevention.

As the leading cause of blindness worldwide, age-related cataract hasmajor public health significance. In recognition of its importance, the WorldHealth Organization designated cataract as the priority condition for a globalinitiative to eliminate avoidable blindness.¹ Manyepidemiological studies have attempted to identify potentially modifiablecataract risk factors, with the aim of developing preventive interventions.Current evidence suggests that cataract is multifactorial in origin, withdifferent associations reported for opacities in different regions of thelens.^2-18

Populations of African origin are known to have high prevalences oflens opacities; among black populations in the United States and the Caribbean,cortical cataract is particularly more frequent than in white persons.^{4,10,13,15,17} Possiblecauses for the high cataract risk in black populations have been examinedin cross-sectional studies, but there is still a dearth of longitudinal, population-baseddata available. We have previously reported factors predictive of incidentnuclear cataract in black participants of the Barbados Incidence Study ofEye Diseases¹⁸ such as age, female gender,iris color, myopia, leaner body mass, and topical intraocular pressure (IOP)–loweringmedications. The current article evaluates associations with incident corticaland posterior subcapsular (PSC) lens opacities in this population. Our purposewas to determine baseline characteristics associated with the developmentof new opacities after 4 years of follow-up.

The Barbados Eye Studies, funded by the National Eye Institute, Bethesda,Md, are a series of population-based investigations of the prevalence, incidence,and risk factors for major causes of visual loss in a preponderantly African-originpopulation. The baseline study, the Barbados Eye Study (BES) (1987-1992)¹⁹ measured prevalence and risk factors for major eyediseases in a random sample of Barbadian-born citizens, aged 40 to 84 years,with 84% participation. Surviving members of the cohort were invited to returnfor a 4-year (SD = 5 months) follow-up visit in the Barbados Incidence Studyof Eye Diseases (1992-1997).^20,21 Atotal of 3427 persons participated (85% of the eligible population). Studymethods have been described in detail elsewhere.^10,18,19,21 Insummary, the study protocol at baseline and follow-up comprised an interview,blood pressure and anthropometric measurements, and ophthalmic measurementsincluding best-corrected visual acuity, visual fields, IOP, and color stereofundus photography. A blood sample was drawn for glycosylated hemoglobin (GHb)measurement.

Lens gradings were based on the Lens Opacities Classification SystemII²² and performed with direct reference tophotographic standards at the slitlamp, under maximum dilatation with tropicamide.Cortical, PSC, and nuclear lens opacities were defined by a Lens OpacitiesClassification System II score of 2 or higher. Four-year incidence of corticalor PSC lens opacities was defined as the development of the specific opacitytype in at least 1 eye, among persons without that opacity in both eyes atbaseline. The incidence for each opacity type, for example, cortical, didnot consider if other opacity types, for example, nuclear or PSC coexistedin the same individual. Independent replicate gradings among different graders,evaluated throughout both the BES¹⁰ and BarbadosIncidence Study of Eye Diseases²¹ data collectionperiods, demonstrated high intergrader agreement.

As previously reported,¹⁸ potential baselinerisk factors investigated included age, gender, socioeconomic status (SES),iris color, skin pigmentation, body mass index and waist-hip ratio, use ofnutritional supplements (such as multivitamins or cod liver oil), use of corticosteroidsand aspirin therapy prior to evaluation, smoking and alcohol use, sunlightexposure, use of ocular medications, and a family history of cataract. Diabetesmellitus (DM) was defined by self-reported physician-diagnosis and GHb levels²³ were also evaluated (the level of GHb was not measuredin the first months of the study; therefore, results were unavailable for19% of the participants). Hypertension was defined as systolic blood pressureof 140 mm Hg or higher and/or diastolic blood pressure of 90 mm Hg or highermeasured by random zero sphygmomanometer at the clinic visit, and/or a historyof antihypertensive treatment. Ocular variables included IOP and myopia, definedas a spherical equivalent less than −0.5 diopter. The effects of ocularvariables were evaluated in the eye in which lens changes of interest developed.If neither eye or both eyes developed opacities, evaluations were based onthe worse eye.

Age- and gender-adjusted univariate logistic regression analyses wereused to evaluate possible factors associated with incident cortical and PSCopacities separately. Significant (P <.10) variablesfrom these analyses were then retained for multivariate logistic regressionanalyses. The final models included age, gender, and the significant variablesrelated to individual risk of incident cortical and posterior subcapsularcataract. Factors with P <.05 were consideredas statistically significant. The results are presented as relative risks(RRs) and 95% confidence intervals (CIs), based on the results from logisticregression models.

Of the 3427 persons who participated in Barbados Incidence Study ofEye Diseases, 3193 (93%) reported their race as black. As described previously,¹⁸ 2609 of the black participants had no nuclear opacitiesat baseline and of these, 9.2% (241/2609) developed nuclear opacities after4 years of follow-up. A higher percentage developed cortical lens opacities,22.2% (452/2040), while there were few participants with incident PSC lensopacities, 3.3% (97/2954).

Table 1 compares the baselinecharacteristics of persons with and without incident cortical opacities. Theincident group was, on average, older (P <.001)with more female participants (P = .02). The groupdeveloping new cortical opacities had a higher proportion of persons of lowerSES with a positive history of DM and elevated mean GHb levels. In contrast,the group not developing cortical opacities reported more regular aspirinuse (2.6% vs 0.7%). After age-gender adjustment, borderline significant differencesbetween the groups were found for an IOP of more than 21 mm Hg and for IOP-loweringtreatment. The association was not significant, however, in the model thatincluded other significant variables from Table 1.

Table 2 provides data onfactors associated with risk of incident cortical lens opacities, based onlogistic regression results. Compared with those aged 40 through 49 years,risk increased from 3-fold at age 50 through 59 years to 12-fold at age 70years and older. Female gender (RR = 1.3) and lower SES (RR = 1.4) modestlyincreased risk, while persons with a history of DM were almost 2½ timesmore likely to develop cortical opacities. Regular use of aspirin productsprior to evaluation was associated with an 80% reduction in incident corticalopacities (RR = 0.2), although the use of such products was infrequent inthis population.

Table 3 presents the baselinecharacteristics of the groups with and without incident PSC lens opacities.While the distribution of variables examined varied between the groups, agreater proportion of the group with incident PSC lens opacities reporteda positive history of DM, which was consistent with the higher mean GHb leveldocumented in this group.

As indicated in Table 4,the risk of new PSC lens opacities also increased with higher age at baseline,with RR increasing from 1.9 at age 50 through 59 years to 12.1 at age 70 yearsand older. Unlike cortical lens opacities, there was no relationship betweenincident PSC lens opacities and gender; however, a self-reported history ofDM was associated with increased risk of new PSC lens opacities (RR = 2.9).

In addition to a DM history, we also evaluated the relationship betweenGHb values and the development of lens opacities. Table 5 demonstrates that higher levels of GHb significantly increasedthe risks of the 2 lens opacity types: each percentage increase in GHb levelat baseline was associated with a 16% and 23% increased risk of incident corticaland PSC lens opacity, respectively. In analyses based on GHb categorized into4 groups, incidence and risk of both opacity types progressively increasedabove a GHb level of greater than 8%. Persons with a GHb level higher than11.5% had a greater than 3-fold risk of developing either lens opacity type,with RR being 3.60 for cortical and 4.93 for PSC lens opacities.

This study determined baseline characteristics associated with the riskof developing cortical and PSC lens opacities. As expected, increased agewas related to risk of both lens opacity types. Diabetes mellitus also increasedthe risk of both cortical and PSC opacities, as did higher values of GHb atbaseline, while female gender, lower SES, and myopia were related to increasedrisk of cortical lens opacities. Another interesting observation was the verylow risk of cortical lens opacities among regular users of aspirin, giventhe conflicting findings among studies evaluating associations between aspirinand cataract.^{4,12,16,24-36} Adiscussion of these results follows.

As widely documented,^2,3 agewas the principal risk factor related to the development of cortical and PSClens opacities in the BES population. It is likely that age is representativeof duration of exposure to many factors leading to oxidative damage of thelens, such as UV radiation and low levels of antioxidant nutrients.^{4,6,9,11,15-17}

Consistent with cross-sectional data, lower SES was related to increasedrisk of incident cortical lens opacities in this study.^11,13 The40% increased risk seen when comparing the low and high SES groups was identicalto the measure of effect in cross-sectional comparisons in this study population.⁹ While a recent review suggests a possible independentassociation between a low educational level and cataract,³ confoundingis still likely to account for this association in many populations. As anexample, a longer duration of education was related to a higher frequencyof nutritional supplement use (including vitamins) in the BES population atbaseline.⁹

Data from several cross-sectional studies including the BES indicatehigher risk of cortical cataract among women.^{4,6,9,11,13,16,17} Thusat baseline, women in the BES had a 40% increased risk of cortical lens opacitiescompared with men,⁹ a measure of effect similarto that observed at 4-year follow-up (Table3). Evaluation of associations between gender and incident cataractwould assist our understanding of whether there are truly gender-related cataractogenicfactors such as hormonal influences,³⁷ ratherthan such observations being due to confounding.

In the BES prevalence phase, hypertension, DM, and central obesity oftenaccompanied cortical lens opacities.⁸ Whilehypertension and obesity were not related to incident lens opacities, DM anda high GHb level at baseline increased the risk of cortical lens opacities2- to 3-fold. Furthermore, each unit increase in GHb level was associatedwith a 16% increased risk of cortical lens opacity over 4 years. Klein etal⁷ similarly reported a 12% increased riskof cortical cataract per unit increase in GHb level over a 5-year period inpersons with DM evaluated in the Beaver Dam Eye Study (BDES), Beaver Dam,Wis. While the higher risk in the overall black BES population might be partlyexplained by differences in diagnostic criteria for cortical cataract andlaboratory methods for evaluation of GHb levels, the prevalence of self-reportedDM in the BES was almost twice that in the BDES (17.5% vs 8.9%).³⁸

The ability to detect risk factors for incident PSC was compromisedby the few cases developing during the 4-year period. Gender was unrelatedto PSC, but DM was strongly predictive, increasing risk about 3-fold. In particular,high GHb values at baseline (>11.5%) increased risk almost 5-fold (Table 5). The BDES also reported significantassociations between DM and increased risk of incident PSC opacities.⁷

The role of DM and hyperglycemia in cataractogenesis has been clearlyestablished,^2,3 the underlyingmechanisms being the direct toxic effect of sugar alcohols formed throughthe aldose reductase pathway on lens fibers.³⁹ However,studies that document opacity type consistently link hyperglycemia to cortical,PSC, or mixed cortical and PSC lens opacities.³ Dataavailable from the BDES⁷ and this study¹⁸ indicate that DM also increases the risk of incidentnuclear cataract, consistent with associations with the level of baselineGHb also seen in this study (odds ratio = 1.13, 95% CI, 1.06-1.20). Thesefindings can contribute to clinical and public health strategies for the managementof patients with DM.

Regular use of aspirin products at baseline was infrequent in this population(n = 44), being reported by 2.6% of the nonincident cases and only by 0.7%of the incident cases of cortical lens opacities. Aspirin use was, however,associated with a substantially lower risk of incident cortical cataract.There are conflicting reports about the association between aspirin use andcataract. While data from several cross-sectional studies have suggested thataspirin or aspirinlike analgesics may protect against cataract,^12,24-27 othersfound either no association or an increased risk of lens opacities.^4,16,28,29 Datafrom prospective observational studies are also inconclusive. Aspirin usewas associated with a reduced incidence of nuclear cataract after 5 yearsin the BDES while nonsteroidal anti-inflammatory medications were not protectiveas a group and showed a tendency to an increased risk of incident corticallens opacities.¹⁴ Associations between aspirinuse and a tendency to increased likelihood of cataract extraction have alsobeen reported.^30,31 Data fromclinical trials did not demonstrate a reduced frequency of cataract extractionin patients with DM who were treated with aspirin.³² Randomizedtrials conducted among British male physicians³³ andthe US Physicians' Health Study,³⁴ failed todemonstrate clear benefit from aspirin use on incident cataract or cataractextraction. Five-year follow-up of the US Physicians' Health Study trial populationsuggested a possible reduced risk of cataract extraction in the aspirin interventiongroup compared with placebo, while subgroup analyses suggested a lower riskof treatment-related incident PSC lens opacity.³⁵ Longer-termfollow-up failed to confirm benefit from aspirin, and findings instead suggestedan increased risk of cataract in the aspirin-treated group.³⁶ Giventhe few persons using aspirin in the BES, the association between aspirinuse and reduced likelihood of incident cataract must be interpreted with caution.Confounding may also partly explain this association as individuals takingaspirin could have higher SES and might also differ in lifestyle practices.

In summary, several factors differentially affected the developmentof cortical, PSC, and nuclear cataract in this black Caribbean population.Diabetes mellitus and older age at baseline were associated with the developmentof all 3 opacity types, while female gender was predictive of incident corticaland nuclear lens opacities. As previously reported, darker iris color, myopia,low body mass index, and treatment with topical IOP-lowering medications alsoincreased the risk of incident nuclear cataract.¹⁸ Thesefactors, however, were not associated with the development of cortical orPSC lens opacities in 4 years. Low SES at baseline increased the likelihoodof incident cortical cataract but not other types of opacities, while aspirinwas protective.

Diabetes mellitus is a well-established risk factor for cataract, mostoften associated with cortical and PSC cataract. This article provides evidencethat DM and hyperglycemia are major risk factors for the development of corticaland PSC lens opacities in an African-descent population with a markedly highrate of DM. This result has particular public health significance, as corticalcataract is the principal opacity type in this population. One inference ofthe current findings is that improved prevention and treatment of DM mightreduce the burden of cataract.

This study also provides evidence of a reduction in incident corticallens opacities resulting from regular aspirin use, an observation that mustbe interpreted with caution. The result merits further investigation becauseof its potential for prevention.

Corresponding author: M. Cristina Leske, MD, MPH; Department of PreventiveMedicine, L3 HSC Room 086, Stony Brook University, Stony Brook, NY 11794-8036.

Submitted for publication July 10, 2003; final revision received November17, 2003; accepted November 19, 2003.

We thank the Barbados Eye Studies participants and the Ministry of Health,Barbados, for their role in the study.

The members of the Barbados Eye Studies Group are as follows: Principal Investigator: M. Cristina Leske, MD, MPH

Coordinating Center

Stony Brook University, Stony Brook, NY: M.Cristina Leske, MD, MPH; Barbara Nemesure, PhD; Suh-Yuh Wu, MA; Leslie Hyman,PhD; Xiaowei Li, PhD; Shu-Hong Xie, MS; Lixin Jiang, MS; Melinda Santoro;Koumudi Manthani.

Data Collection Center

Ministry of Health, Bridgetown, Barbados, West Indies: Anthea M. S. Connell, FRCS, FRCOphth; Anselm Hennis, MRCPUK, PhD;Ann Bannister, MBBS, DO, MRCOphth; Muthu A. Thangaraj, MBBS, DO; Coreen Barrow;Patricia Basdeo; Kim Bayley; Anthanette Holder.

Fundus Photography Reading Center

The Johns Hopkins University, Baltimore, Md: AndrewP. Schachat, MD; Judith A. Alexander; Noreen B. Javornik, MS; Cheryl J. Hiner;Deborah A. Phillips; Reva Ward; Terry W. George.

Local Advisory Committee

School of Clinical Medicine and Research, Universityof the West Indies, Barbados: Trevor Hassell, MBBS, FRCP, FACC, GCM;Henry Fraser, FRCP, FACP, PhD, GCM; Clive Gibbons, FRCS, FRCP, FRCOphth.