Diabetes, cardiovascular disease, selected cardiovascular disease risk factors, and the 5-year incidence of age-related cataract and progression of lens opacities: the beaver dam eye

Diabetes, Cardiovascular Disease, SelectedCardiovascular Disease Risk Factors, and the5-Year Incidence of Age-Related Cataract and

Progression of Lens Opacities:The Beaver Dam Eye Study

BARBARA E. K. KLEIN, MD, RONALD KLEIN, MD, AND KRISTINE E. LEE, MS

● PURPOSE: To describe the relationships of dia-betes mellitus, cardiovascular disease, and selectedcardiovascular disease risk factors to cumulativeincidence of age-related cataract and to progressionof lens opacities over a 5-year interval.● METHODS: A follow-up examination of the Bea-ver Dam Eye Study cohort was performed 5 yearsafter the baseline evaluation. Ages at the censusprior to baseline ranged from 43 to 84 years of age.Protocols for examination, lens photography, andgrading were the same for both examinations.● RESULTS: Age at baseline was the most signifi-cant characteristic associated with incidence ofnuclear, cortical, and posterior subcapsular cata-ract in those without diabetes (P < .001) for allcataracts. The positive association of age withcataract was found for nuclear and cortical cata-ract in the worse eye (P < .04) but not posteriorsubcapsular cataract in those with diabetes. Pro-gression of nuclear sclerosis was common, occur-ring in about 70% of subjects when considering

either eye. Incident cortical and posterior subcap-sular cataracts (P < .001 for worse eye for eachlesion) and progression of cortical and posteriorsubcapsular opacities were more common in thosewith diabetes (P < .001 for either eye for eachlesion).

Increased glycated hemoglobin level was associ-ated with increased risk of nuclear and corticalcataracts in those with diabetes. Relationships ofrisk factors to posterior subcapsular cataracts,especially among those with diabetes, were often inthe expected direction but lacked significance pos-sibly due to small samples.● CONCLUSIONS: Diabetes mellitus is associatedwith incidence over 5 years of cortical and poste-rior subcapsular cataract and with progression ofmore minor cortical and posterior subcapsular lensopacities. These changes may be related to level ofglycemia. Cardiovascular disease and its risk fac-tors have little effect on incidence of any age-related cataract. (Am J Ophthalmol 1998;126:782–790. © 1998 by Elsevier Science Inc. Allrights reserved.)

R ELATIONSHIPS BETWEEN CARDIOVASCULAR

disease and its risk factors and cataract typeshave been examined in observational studies

in many different populations.1–7 The strength andsignificance of some of these findings have varied.Associations among diabetes mellitus, blood pres-

Accepted for publication July 16, 1998.From the Department of Ophthalmology and Visual Sciences,

University of Wisconsin Medical School, Madison, Wisconsin.This research is supported by National Institutes of Health grant

EYO6594 (Drs BEK Klein and R Klein) and, in part, by the Researchto Prevent Blindness (Dr R Klein, Senior Scientific InvestigatorAward).

Reprint requests to Barbara E. K. Klein, MD, MPH, Department ofOphthalmology and Visual Sciences, University of Wisconsin-Madi-son, 610 North Walnut Street, 460 WARF, Madison, WI 53705-2397;fax: (608) 263-0279.

© 1998 BY ELSEVIER SCIENCE INC. ALL RIGHTS RESERVED.782 0002-9394/98/$19.00PII S0002-9394(98)00280-3

sure, serum lipids, and cataract were found incross-sectional analyses of data from the Framing-ham Eye Study but have not been found in allstudies of these relationships.3,5 Temporal relationsof risk factor and disease usually cannot be deter-mined from prevalence data. We have the opportu-nity to explore relationships of cardiovasculardisease variables and diabetes as potential risk fac-tors for cataract in incidence data from the BeaverDam Eye Study.

MATERIALS AND METHODS

A PRIVATE CENSUS OF THE POPULATION OF BEAVER

Dam, Wisconsin, was performed from 1987 to 1988in order to identify all persons 43 years of age andolder.8 Persons 43 to 84 years of age were invited forthe study evaluation, which was performed during a2.5-year period beginning March 1, 1988. Tenets ofthe Declaration of Helsinki were followed, institu-tional human experimentation committee approvalwas granted, and an informed consent was signed byeach subject. During the study visit, standard mea-surements and questionnaires were administered.All subjects identified at the initial census wereinvited for the second examination. Only thoseparticipating at the first and second examinations(n 5 3,684) supply information pertinent to thecurrent article.

The follow-up examinations occurred 4.8 (60.4)years after the first evaluations and were performedsuch that participants were seen in approximatelythe same order as in the prevalence examinations.

Procedures at the follow-up were the same as atthe baseline examination. Details have been previ-ously published9–12 and are summarized below.Blood pressures were measured according to theHypertension Detection and Follow-Up Programprotocol.13 Subjects were weighed and measured intheir street clothes after removing their shoes.Pupils were pharmacologically dilated after an ocu-lar examination.

Participants responded to an interview at the timeof the study evaluation. Participants were askedwhether they had ever been told by their doctor thatthey had any of the following: angina, a heart attack,a transient ischemic attack, a stroke, diabetes, and/or

hypertension. We also asked whether they had coro-nary bypass surgery or angioplasty. Lifetime smokinghistory was obtained.14 Patients were queried regardingcurrent and past alcohol use.15 All subjects were askedto permit venipuncture. Serum obtained from thesamples was tested for total cholesterol,16 high-density-lipoprotein cholesterol,17 glycated hemoglobin,18 anduric acid.19

Photographs of the lenses were taken with twodifferent cameras: a slit-lamp camera and a retroil-lumination camera.20 Grading procedures for thelens were based on detailed codified decision rules.Graders were masked to subject identity, personalcharacteristics, and the presence and/or severity oflens lesions at the first examination. Scores fornuclear sclerosis were based on comparisons withstandard photographs. The scale has five steps ofseverity based on opacity of the nucleus. Scores forcortical and posterior subcapsular cataracts werebased on weighted estimates of degree of opacity oflens area as defined by a circular grid, divided intoeight “pie-wedged” peripheral areas and a centralcircular area overlaid on the photograph. Prevalentcases of cortical cataract were those with opacity of5% or more of the lens “surface.” Posterior subcap-sular opacity was defined as 5% or more of a gridsegment. The classification of prevalent cataractcorresponded to lens opacity of sufficient severity tomeet the criteria that a clinical ophthalmologistwould be comfortable labeling as a cataract. Hori-zontal and vertical pupil diameters were measuredbefore grading cortical and posterior subcapsularopacities. Means of horizontal and vertical diame-ters were used in the analyses concerning pupil size.

During the course of grading, there was a contin-uous quality control procedure to monitor inter- andintragrader variability. There was no evidence ofmarked or consistent change in variability duringthe course of prevalence or follow-up study orbetween the two study periods. Reproducibility ofgradings using these systems was published prior toreporting the prevalence findings.20 In general, re-producibility was good and was similar for intra- andintergrader comparisons.

Any eye with evidence of direct trauma wasexcluded from analyses. Any eye with a lesion thatconfounded the grading of specific lens lesions wasexcluded from consideration for that lens lesion (for

DIABETES, CARDIOVASCULAR DISEASE, CATARACTS, AND LENS OPACITIESVOL. 126, NO. 6 783

example, the presence of pseudoexfoliation changesthat interfered with photography/grading of corticalopacities). Eyes that had undergone cataract surgerywere excluded from analyses for risk factors forspecific cataract lesions. Estimates of incidence orprogression endpoints were based on having corre-sponding gradable subfields at both visits.

Diabetes was defined as a previous history ofdiabetes treated with either insulin, oral hypoglyce-mic agents, or diet. Newly diagnosed diabetes (atbaseline) was defined as no previous medical historyof diabetes in the presence of elevated glycatedhemoglobin at the baseline examination.21 A posi-tive history of heart attack, angina, stroke, or use ofdigoxin or nitroglycerin medications was includedas a case of cardiovascular disease. An individualwas identified as a smoker if he or she reportedsmoking at least 100 cigarettes in their lifetime.Heavy drinkers were those who reported that therehad been a time in their life when they had imbibedfour or more alcoholic drinks each day. Hyperten-sion was defined as a systolic blood pressure of 160mm Hg or more or a diastolic blood pressure of 95mm Hg or more at the time of the examination or ahistory of hypertension and current use of medica-tion for hypertension. Pulse pressure was the differ-ence between systolic and diastolic blood pressures.

We considered two different approaches to eval-uate the lens endpoints during the interval betweenthe two examinations: incidence and progression.Incidence refers to the classification of lens opacityof sufficient severity to meet the criteria that aclinical ophthalmologist would be comfortable la-beling as a cataract. These categories were the sameas those used to define prevalent cases in priorpublications.22 The specific definitions used to clas-sify such opacities were nuclear sclerosis level of fouror five in the grading scheme, which specifies fivelevels of increasing severity, cortical opacities in-volving 5% or more of visible lens, and posteriorsubcapsular opacity involving 5% or more of any ofnine segments defined by a circular grid. Progressionwas defined as any increase in severity or involve-ment of the lens by opacity of any severity includingabsent to present. In the analyses of progression foreach type of opacity, the following definitions wereused: nuclear sclerosis was defined as an increase ofone level; cortical and posterior opacities were

defined as an increase of 0.75 or more after squareroot transformation of the data. A square roottransformation was used to account for the increasein the variability of the estimated area of involve-ment as the base amount increased. For example,the square root transformation of the differences inseverity requires that for a baseline value of 5%involvement, 8.9% of the lens would need to beinvolved at follow-up in order to consider that theopacity had progressed; for a baseline value of 10%involvement, at least 15.3% involvement wouldneed to be involved at follow-up to consider thatthe opacity had progressed.

Incidence and progression data are reported forboth eyes. In addition, incidence data are reportedfor the worse eye, and progression is reported whenit occurred in either eye. Wisconsin Storage andRetrieval23 software system was used to processsubject files. SAS was used to calculate chi-squarestatistics and logistic regression.24 Tests of trendwere calculated by the Mantel-Haenszel method.25

RESULTS

COMPARISONS OF PARTICIPANTS AT THE FOLLOW-UP

examination with nonparticipants are given in Tables1 and 2. There were significant differences betweenparticipants and live nonparticipants in mean age,total serum cholesterol, and total- to HDL-cholesterolratio. There were many more differences betweenparticipants and those who died prior to follow-up.Those who died were, in general, older and sicker atbaseline and were more likely to have had nuclear andcortical cataract than participants.

Incidence of nuclear cataract and progression ofnuclear sclerosis by sex and age for those with andwithout diabetes is seen in Table 3. Incidenceincreased with age in all groups. The only instancefor which the age effect was not significant was inleft eyes of diabetic men (P 5 .20). For all others,the age effect was significant (P # .03 for allgroups). In those free of diabetes, women were morelikely to have nuclear cataract in right eyes thanmen (P , .001), although the effect of sex for theleft eye was of borderline significance (P 5 .07). Inthose with diabetes, the difference between thesexes did not reach significance. There was no

AMERICAN JOURNAL OF OPHTHALMOLOGY784 DECEMBER 1998

significant difference in incidence of nuclear cata-ract by diabetes status overall (P 5 .92 and .44 forright and left eyes, respectively). Progression was acommon finding, but there was no consistent effectof age, sex, or diabetes status on progression ofnuclear opacities.

Incident cortical cataract increased significantlywith age in all groups (P # .04 in all groups; Table4). There was no significant difference between thesexes. People with diabetes had higher incidences atall ages, and the overall differences were statisticallysignificant (P # .05 for each analysis). Progressionwas common in all groups, increased significantlywith age in most groups, and was not significantlyaffected by sex. Those with diabetes were at in-creased risk of progression (P # .05).

Incident posterior subcapsular cataract was theleast frequent cataract event (Table 5). In thosewithout diabetes, there was a significant effect of age(P , .001 for all comparisons) but no difference bysex. In those with diabetes, the relationship to agewas not significant, and there was also no differenceby sex. Overall, those with diabetes were at in-creased risk of incident disease (P # .01). Forprogression, there was a consistent effect of age inthose without diabetes. Except for right eyes in men,the age effect was significant (P # .01). There wasno significant effect of sex. In those with diabetes,for most comparisons, the association of age toprogression was not statistically significant. Personswith diabetes were more likely to have progressedthan those without diabetes (P # .008).

TABLE 1. Comparison of Continuous Variables at Baseline of Participants With Nonparticipants at the 5-Year Follow-up:Beaver Dam Eye Study, 1993–1995

Variable

Participant Live, Nonparticipant Dead

n Mean SD n Mean SD P value n Mean SD P value

Age 3,684 60.4 10.5 686 62.7 11.8 ,.001 556 72.2 9.4 ,.001

Body mass index 3,667 28.9 5.4 681 28.9 5.5 .51* 530 28.0 5.3 .09†

Cholesterol

Total 3,675 233.0 43.5 683 239.0 44.6 .002* 549 231.0 47.5 .07†

HDL 3,670 52.4 17.5 683 52.3 18.0 .85* 550 48.3 17.2 #.001†

Ratio 3,670 4.9 1.9 683 5.1 2.0 .03* 549 5.4 2.5 #.001†

Uric acid 3,675 5.9 1.7 683 6.0 2.5 .57* 550 6.6 2.0 #.001†

Glycated hemoglobin 3,672 6.0 1.5 683 6.1 1.6 .18* 548 6.7 2.1 #.001†

HDL 5 high-density lipoprotein; SD 5 standard deviation.

*Age-adjusted comparison of live nonparticipant with participants.†Age-adjusted comparison of deceased with participants.

TABLE 2. Comparison of Categorical Variables at Baseline of Participants With Nonparticipants at the 5-Year Follow-up:Beaver Dam Eye Study, 1993–1995

Variable

Participant Live, Nonparticipant Dead

n % n % P value n % P value

Sex, female 3,684 43.2 686 40.8 .70* 556 52.5 ,.001†

Diabetes, yes 3,684 8.9 686 8.9 .82* 556 21.0 ,.001†

Cardiovascular disease 3,631 16.0 673 20.5 .12* 547 43.1 ,.001†

Hypertension 3,684 34.5 686 39.5 .13* 555 50.6 ,.001†

Nuclear cataract 3,550 13.9 645 19.2 .54* 465 39.4 ,.007†

Cortical cataract 3,557 13.3 639 19.1 .07* 458 31.9 .04†

Posterior subcapsular 3,561 4.2 638 5.8 .34 457 9.2 .24

*Age-adjusted comparison of live nonparticipants with participants.†Age-adjusted comparison of deceased with participants.


In order to determine whether the characteristicsof interest were associated with the risk of incidentcataract, after taking into account other relevantcharacteristics, we performed multivariate analyses.There was little difference between incident cata-ract and progression of opacities in the patterns ofassociation with these factors. Therefore, we presentanalyses for incidence only. Our analyses consideredrisk factors listed in Table 6 for those without andwith diabetes. In each model, we considered age,sex, smoking status, heavy drinking, and hyperten-sion (the latter three because of associations foundin the prevalence data). The variables actually

appearing in the final model are indicated in Table6. In those without diabetes, a change in body massindex of 1 kg/m2 was associated with a 5% change inrisk of posterior subcapsular cataract. In those withdiabetes, a 1.0% difference in glycated hemoglobinwas associated with an average of 15% change inrisk of nuclear cataract. A significant associationwas also found between incident cortical cataractand glycated hemoglobin, a 1.0% difference inglycated hemoglobin being associated with a 12.0%change in risk in those with diabetes.

Cataract surgery was performed on 226 people(70 both eyes, 78 right eyes only, 78 left eyes only)

TABLE 3. Nuclear Sclerosis by Age, Sex, and Diabetes at the 5-Year Follow-up: Beaver Dam Eye Study, 1993–1995

Incident Nuclear Cataract Progression of Nuclear Sclerosis

Right Left Worse Right Left Either

n % n % n % n % n % n %

Without Diabetes

Women

43–54 yrs 605 3.5 599 4.2 588 6.0 606 46.9 601 49.4 591 71.4

55–64 yrs 437 13.3 435 10.3 419 16.7 454 54.9 453 53.2 443 75.9

65–74 yrs 330 34.6 329 34.0 281 42.0 397 49.1 408 46.1 376 71.0

751 yrs 71 40.9 74 36.5 55 58.2 123 31.7 126 35.7 106 55.7

P ,.001 P ,.001 P ,.001 P 5 .13 P 5 .01 P 5 .04

Men

43–54 yrs 533 2.1 523 2.1 520 3.3 533 47.7 523 46.7 520 69.4

55–64 yrs 388 7.5 385 8.1 372 11.6 396 51.5 392 46.7 382 73.0

65–74 yrs 249 23.3 250 30.0 220 38.6 291 44.0 289 53.2 271 73.1

751 yrs 56 39.3 57 42.1 45 51.1 84 41.7 82 45.1 77 66.2

P ,.001 P ,.001 P ,.001 P 5 .22 P 5 .30 P 5 .66

Women versus men* P ,.001 P 5 .07 P 5 .006 P 5 .53 P 5 .73 P 5 .78

With Diabetes

Women

43–54 yrs 29 3.5 29 0.0 29 3.5 29 58.6 29 48.3 29 75.9

55–64 yrs 44 13.6 45 15.6 41 19.5 47 51.1 46 56.5 44 72.7

65–74 yrs 41 29.3 38 36.8 35 37.1 50 48.0 48 50.0 46 69.6

751 yrs 10 40.0 6 83.3 4 75.0 16 50.0 17 41.2 12 50.0

P 5 .001 P ,.001 P ,.001 P 5 .45 P 5 .61 P 5 .16

Men

43–54 yrs 22 9.1 23 8.7 21 14.3 22 63.6 23 52.2 21 81.0

55–64 yrs 50 6.0 50 10.0 49 12.2 50 50.0 50 46.0 49 76.0

65–74 yrs 31 25.8 32 25.0 29 38.0 32 40.6 33 54.6 31 71.0

751 yrs 3 33.3 4 0.0 3 33.3 3 66.7 4 25.0 3 66.7

P 5 .03 P 5 .20 P 5 .02 P 5 .21 P 5 .84 P 5 .38

Women versus men* P 5 .48 P 5 .12 P 5 .76 P 5 .87 P 5 .67 P 5 .57

With diabetes versus without

diabetes†

P 5 .92 P 5 .44 P 5 .80 P 5 .41 P 5 .59 P 5 .81

*Age-adjusted.†Age- and sex-adjusted.


in the interval between the baseline and follow-upexaminations. In our attempts to evaluate whetherthe surgery (causing subjects to be ineligible forincidence or progression) was systematically influ-encing our findings, one investigator (B.E.K.) re-viewed medical charts of affected subjects toascertain the particular cataract type that precededthe surgery. We then combined the doctor’s diag-nosis as to type of cataract prior to surgery with ourdiagnoses from grading of subjects’ lenses and as-sessed the “total” incidence rates. There was noevidence of a substantial change in the findingsregarding risk factors and specific cataract types(data not shown).

DISCUSSION

THE NEGATIVE ASSOCIATION OF HIGHER LEVELS OF

glycated hemoglobin with nuclear and cortical cat-aract in those with diabetes was perhaps not surpris-ing. A direct relation of glycated hemoglobin tocataract was not found in the prevalence data;however, all three lens opacities were associatedwith duration of diabetes.26 It may be that theadditional 5 years’ duration was necessary to observethe effect of glycemia. In data from the WisconsinEpidemiologic Study of Diabetic Retinopathy, levelof glycemia at baseline was an important predictorof cumulative incidence of cataract surgery.27 Dobbs

TABLE 4. Cortical Opacities by Age, Sex, and Diabetes at the 5-Year Follow-up: Beaver Dam Eye Study, 1993–1995

Incident Cortical Cataract Progression of Cortical Opacities


n % n % n % n % n % n %

Without Diabetes

Women

43–54 yrs 607 2.0 608 2.3 597 2.7 610 6.6 612 7.8 604 12.3

55–64 yrs 428 6.8 428 11.2 407 12.8 451 21.7 454 22.0 442 32.6

65–74 yrs 328 14.6 334 11.7 295 15.9 393 31.0 399 29.8 366 43.2

751 yrs 84 23.8 82 22.0 63 31.8 123 37.4 133 38.4 111 53.2

P ,.001 P ,.001 P ,.001 P ,.001 P ,.001 P ,.001

Men

43–54 yrs 528 2.0 525 1.3 515 2.5 530 10.0 530 8.3 521 16.1

55–64 yrs 380 7.9 376 6.9 359 10.0 400 18.8 398 15.8 391 27.9

65–74 yrs 260 16.5 257 13.2 230 18.3 288 34.4 288 28.8 269 46.1

751 yrs 58 17.2 57 12.3 51 19.6 75 28.0 77 33.8 70 47.1

P ,.001 P ,.001 P ,.001 P ,.001 P ,.001 P ,.001


With Diabetes

Women

43–54 yrs 28 3.6 28 10.7 27 7.4 30 13.3 30 26.7 30 30.0

55–64 yrs 42 14.3 43 18.6 40 22.5 49 30.6 48 35.4 48 47.9

65–74 yrs 40 25.0 36 30.6 34 38.2 47 40.4 44 40.9 43 53.5

751 yrs 11 27.3 7 42.9 4 0.0 16 37.5 18 55.6 12 58.3

P 5 .01 P 5 .02 P 5 .04 P 5 .02 P 5 .05 P 5 .04

Men

43–54 yrs 23 0.0 24 0.0 23 0.0 23 17.4 24 8.3 23 17.4

55–64 yrs 47 10.6 46 13.0 44 13.6 50 32.0 50 36.0 49 46.9

65–74 yrs 27 14.8 26 23.1 25 32.0 30 36.7 32 40.6 29 51.7

751 yrs 3 33.3 3 66.7 3 66.7 4 50.0 6 16.7 4 50.0

P 5 .03 P 5 .001 P ,.001 P 5 .09 P 5 .09 P 5 .02


With diabetes versus without

diabetes†

P 5 .05 P #.001 P ,.001 P 5 .002 P ,.001 P #.001

*Age-adjusted.†Age- and sex-adjusted.


and associates28 reported significant associationsbetween glycated hemoglobin and both nuclear andcortical lens changes. Whether the relationship isdue to direct glycation of lens proteins or is relatedto aldose reductase or other metabolic processes isbeyond the scope of this observational study. Dia-betes poses serious burdens to the eye due toretinopathy and glaucoma, as well as cataracts.Although the data from this observational study arenot equivalent to findings from a clinical trial inwhich the outcome of primary interest is cataractand the test intervention is tight control, one mightbe willing to infer that there may be a direct causal

relationship such that decreasing glycemia mightsubstantially reduce the risk of cataract. This, cou-pled with the finding of lower risk of retinopathy inthose with lower levels of glycemia, has importantimplications for care because glycemia can be con-trolled, even if not easily so.

It appears that interval surgery has had little influ-ence on the relationships we investigated here. If weassume that ophthalmologists in practice evaluatecataracts in ways that are similar to the grading ofphotographs (and that they record these evaluations inthe patients’ charts), then there is likely to be minimalbias as to risk factor status regarding the diagnosis of

TABLE 5. Posterior Subcapsular Opacities by Age, Sex, and Diabetes at the 5-Year Follow-up:Beaver Dam Eye Study, 1993–1995

Incident Posterior Subcapsular Cataract Progression of Posterior Subcapsular Opacities


n % n % n % n % n % n %

Without Diabetes

Women

43–54 yrs 612 1.3 610 0.7 601 1.7 615 0.7 612 0.8 604 1.3

55–64 yrs 446 2.9 453 2.4 431 3.5 450 2.0 456 1.1 435 2.1

65–74 yrs 385 6.2 386 4.9 350 7.4 389 3.1 391 4.6 358 6.2

751 yrs 106 9.4 113 7.1 87 9.2 110 4.6 114 4.4 91 6.6

P ,.001 P ,.001 P ,.001 P 5 .001 P ,.001 P ,.001

Men

43–54 yrs 529 1.3 524 1.0 513 1.6 535 1.7 527 1.1 521 2.3

55–64 yrs 385 2.1 395 2.5 374 2.4 393 2.3 399 2.5 382 2.4

65–74 yrs 261 4.2 281 4.3 237 6.3 270 2.6 284 3.2 247 3.6

751 yrs 70 5.7 63 17.5 59 15.3 70 5.7 65 12.3 61 9.8

P 5 .002 P ,.001 P ,.001 P 5 .07 P ,.001 P 5 .01


With Diabetes

Women

43–54 yrs 30 3.3 28 3.6 28 3.6 30 0.0 29 0.0 29 0.0

55–64 yrs 49 12.2 48 8.3 46 17.4 51 5.9 49 8.2 48 10.4

65–74 yrs 44 15.9 40 7.5 36 11.1 45 13.3 41 4.9 36 11.1

751 yrs 14 14.3 14 14.3 10 20.0 14 14.3 15 13.3 10 20.0

P 5 .14 P 5 .30 P 5 .29 P 5 .02 P 5 .17 P 5 .05

Men

43–54 yrs 23 4.4 22 0.0 22 4.6 23 4.4 22 0.0 22 4.6

55–64 yrs 49 2.0 49 6.1 47 6.4 49 2.0 49 6.1 47 6.4

65–74 yrs 27 14.8 30 10.0 25 16.0 27 11.1 30 10.0 25 12.0

751 yrs 4 0.0 6 0.0 4 0.0 4 0.0 6 0.0 4 0.0

P 5 .22 P 5 .34 P 5 .33 P 5 .41 P 5 .34 P 5 .55


With diabetes versus without diabetes† P #.001 P 5 .010 P #.001 P #.001 P 5 .008 P ,.001

*Age adjusted.†Age- and sex-adjusted.


cataract prior to surgery, as far as we can judge from ourdata. Because we cannot be certain of the comparabil-ity of the criteria for diagnosis of cataract of differentsurgeons, and also between surgeons and our gradings,we present only the risk estimates that are derivedfrom our gradings. It appears, however, that our esti-mates may be representative of the population expe-rience regarding cataracts whether or not surgery hasbeen performed.

In prevalence data from the same population, wefound few significant relationships between cardio-vascular disease and its risk factors examined in thisarticle and cataracts. Because of the possible effectsof selective survival, longitudinal evaluation wasindicated. The data we use here were collectedabout 5 years after the prevalence survey. We didfind an association of body mass index to incidentposterior subcapsular cataract in those without dia-betes. A positive association was found in theprevalence data but was not significant. It is possiblethat cataract surgery prior to the first examination

was responsible for the lack of statistical significancein the prevalence study. Although the relationshipis now significant, the increase in risk associatedwith body mass index is small. We conclude that thefactors investigated had little influence on incidentcataract or progressed lens opacities in personswithout diabetes. For cardiovascular disease diag-noses, we relied on the report of the participant.There may be error in such reporting, which may beobscuring a relation of cardiovascular events tocataract; however, we have chosen those eventsthat would likely be remembered accurately, thusattempting to minimize the variability in the his-toric data.

In summary, we have found significant relationsof levels of glycemia with incident nuclear andcortical cataract in persons with older-onset diabe-tes. Because level of glycemia can be manipulated,this finding may have direct preventive utility. Incontrast, we find little evidence of associations ofcardiovascular disease and its risk factors to incident

TABLE 6. Multivariate Relations to Incidence of Cataract in the Worse Eye

Nuclear Cortical Posterior Subcapsular Cataract

OR* (95% CI) P Value OR* (95% CI) P Value OR* (95% CI) P Value

Without Diabetes

History of CVD Yes or no 0.81 (0.58, 1.14) .23 1.16 (0.79, 1.70) .45 1.09 (0.62, 1.90) .77

Body mass index 1 kg/m2 0.99 (0.97, 1.02) .60 1.00 (0.97, 1.03) .82 1.05 (1.02, 1.09) .01

Cholesterol

Total 0.1 mmol/L 1.00 (0.99, 1.01) .76 1.00 (0.98, 1.01) .65 1.00 (0.99, 1.02) .66

HDL 0.1 mmol/L 1.01 (0.98, 1.04) .55 0.99 (0.96, 1.02) .51 1.00 (0.95, 1.05) .99

Ratio 1 0.97 (0.91, 1.04) .39 1.01 (0.94, 1.10) .72 1.01 (0.90, 1.13) .92

Uric acid 10 mmol/L 0.99 (0.98, 1.00) .20 1.00 (0.99, 1.02) .70 1.01 (0.99, 1.03) .53

Glycated hemoglobin 1% 0.99 (0.84, 1.17) .92 1.03 (0.84, 1.26) .80 0.99 (0.74, 1.34) .97

Hypertension Yes or no 0.90 (0.70, 1.16) .43 0.99 (0.74, 1.34) .97 1.41 (0.93, 2.14) .11

Pulse pressure 10 mm Hg 1.00 (0.92, 1.08) .92 1.01 (0.92, 1.11) .78 1.04 (0.91, 1.19) .54

With Diabetes

History of CVD Yes or no 1.70 (0.78, 3.71) .18 1.54 (0.66, 3.64) .32 —

Body mass index 1 kg/m2 1.02 (0.95, 1.09) .60 1.08 (1.01, 1.16) .02 —

Cholesterol

Total 0.1 mmol/L 0.99 (0.96, 1.03) .58 0.98 (0.94, 1.02) .26 —

HDL 0.1 mmol/L 0.92 (0.82, 1.04) .17 0.97 (0.86, 1.09) .60 —

Ratio 1 1.06 (0.87, 1.28) .57 0.95 (0.76, 1.18) .61 —

Uric acid 10 mmol/L 0.98 (0.95, 1.02) .35 1.02 (0.98, 1.06) .30 —

Glycated hemoglobin 1% 1.15 (1.03, 1.28) .01 1.12 (1.00, 1.25) .04 —

Hypertension Yes or no 0.78 (0.37, 1.62) .50 1.38 (0.64, 2.97) .41 —

Pulse pressure 10 mm Hg 1.06 (0.87, 1.29) .54 1.02 (0.83, 1.25) .86 —

CVD 5 cardiovascular disease; HDL 5 high-density lipoprotein; OR (95% CI) 5 odds ratio (95% confidence interval).

*Adjusted for age, sex, smoking, and heavy drinking.


cataract and, therefore, find no direction for sugges-tions of preventive intervention.

REFERENCES

1. Clayton RM, Cuthbert J, Duffy J, et al. Some risk factorsassociated with cataract in SE Scotland: a pilot study. TransOphthalmol Soc UK 1982;102:331–336.

2. Glynn RJ, Christen WG, Manson JE, Bernheimer J, Hen-nekens CH. Body mass index: an independent predictor ofcataract. Arch Ophthalmol 1995;113:1131–1137.

3. The Italian-American Cataract Study Group. Risk factorsfor age-related cortical, nuclear and posterior subcapsularcataracts. Am J Epidemiol 1991;133:541–553.

4. Jahn CE, Janke M, Winowski H, vonBergmann K, Leiss O,Hockwin O. Identification of metabolic risk factors forposterior subcapsular cataract. Ophthalmic Res 1986;18:112–116.

5. Kahn HA, Leibowitz HM, Ganley JP, et al. The Framing-ham Eye Study II: association of ophthalmic pathologywith single variables previously measured in the Framing-ham Heart Study. Am J Epidemiol 1977;106:33–41.

6. Klein BEK, Klein R, Lee KE. Cardiovascular disease,selected cardiovascular disease risk factors, and age-relatedcataracts: the Beaver Dam Eye Study. Am J Ophthalmol1997;123:338–346.

7. West SK, Valmadrid CT. Epidemiology of risk factors forage-related cataract. Surv Ophthalmol 1995;39:323–334.

8. Campbell JA, Palit CD. Total digit dialing for a small areacensus by phone: proceedings of the section on surveyresearch methods of the American Statistical Association.Alexandria, Virginia. Am Stat Assoc 1988:S49–S59.

9. Klein R, Klein BEK, Linton KLP. Prevalence of age-relatedmaculopathy: the Beaver Dam Eye Study. Ophthalmology1992;99:933–943.

10. Klein R, Klein BEK, Lee KE. The changes in visual acuityin a population: the Beaver Dam Eye Study. Ophthalmol-ogy 1996;103:1169–1178.

11. Klein R, Klein BEK. The Beaver Dam Eye Study manual ofoperations. U.S. Department of Commerce, Springfield,VA 22161. NTIS Accession No. PB 91-149823,AS, 1991.

12. Klein R, Klein BEK. The Beaver Dam Eye Study II manualof operations. U.S. Department of Commerce, Springfield,VA 22161. NTIS Accession No. PB95-273827, 1995.

13. Hypertension Detection and Follow-up Program Coopera-tive Group. The hypertension detection and follow-upprogram. Prev Med 1976;5:207–215.

14. Klein BEK, Klein R, Linton KLP, Franke T. Cigarettesmoking and lens opacities: the Beaver Dam Eye Study.Am J Prev Med 1993;9:27–30.

15. Ritter LL, Klein BEK, Klein R, Mares-Perlman JA. Alcoholuse and lens opacities in the Beaver Dam Eye Study. ArchOphthalmol 1993;111:113–117.

16. Allain CC, Poon LS, Chan CGS, et al. Enzymatic deter-mination of total serum cholesterol. Clin Chem 1994;20:470–475.

17. Lopes-Virella MR, Stone P, Ellis S, et al. Cholesteroldetermination in high-density lipoproteins separated bythree different methods. Clin Chem 1977;23:882–884.

18. Klenk DC, Hermanson GT, Krohn RI, et al. Determinationof glycosylated hemoglobin by affinity chromatography:comparison with colorimetric and ion-exchange methodsand effects of common interferences. Clin Chem 1982;28:2088–2094.

19. Fossati P, Prencipe L, Berti G. Use of 3,5-dichloro-2-hydroxybenzenefulfonic acid/4-amenophenazore chromo-gene system in direct enzymic assay of uric acid in serumand urine. Clin Chem 1980;26:222–231.

20. Klein BEK, Klein R, Linton KLP, Magli YL, Neider MW.Assessment of cataracts from photographs in the BeaverDam Eye Study. Ophthalmology 1990;97:1428–1433.

21. Klein R, Klein BEK, Moss SE, Linton KLP. The BeaverDam Eye Study: retinopathy in adults with newly discov-ered and previously diagnosed diabetes mellitus. Ophthal-mology 1992;99:58–62.

22. Klein BEK, Klein R, Linton KLP. Prevalence of age-relatedlens opacities in a population: the Beaver Dam Eye Study.Ophthalmology 1992;99:546–552.

23. Entine S, Holladay D, Olscheske T, Anderson E, TormeyP. Wisconsin Information Storage and Retrieval System.Madison, Wisconsin: University of Wisconsin ClinicalCancer Center, 1981.

24. Hastings RP, editor. SUGI supplemental library user’sguide, Version 6. Cary, NC: SAS Institute, 1990.

25. Mantel N. Chi-square tests with one degree of freedom:extensions of the Mantel-Haenszel procedure. J Am StatAssoc 1963;58:690–700.

26. Klein BEK, Klein R, Wang Q, Moss SE. Older onsetdiabetes and lens opacities: the Beaver Dam Eye Study.Invest Ophthalmol Vis Sci 1993;34:1065.

27. Klein BEK, Klein R, Moss SE. Incidence of cataract surgeryin the Wisconsin Epidemiologic Study of Diabetic Reti-nopathy. Am J Ophthalmol 1995;119:295–300.

28. Dobbs RE, Smith JP, Chen T, Knowles W, Hockwin O.Long-term following of lens changes with Sheimpflugphotography in diabetics. Ophthalmology 1987;94:881–890.

Authors InteractivetWe encourage questions and comments regarding this article via theInternet on Authors Interactivet at http://www.ajo.com/ Questions, com-ments, and author responses are posted.


Documents

Diabetes, cardiovascular disease, selected cardiovascular disease risk factors, and the 5-year incidence of age-related cataract and progression of lens opacities: the beaver dam eye