Glycemic Control and Lens Transparency inPatients With Type 1 Diabetes Mellitus

SATOSHI KATO, MD, AZUSA SHIOKAWA, MD, HARUMI FUKUSHIMA, MD,JIRO NUMAGA, MD, SHIGEHIKO KITANO, MD, SADAO HORI, MD,

TADAYOSHI KAIYA, MD, AND TETSURO OSHIKA, MD

● PURPOSE: To assess quantitatively the cumulativeeffect of hyperglycemia on lens transparency in patientswith juvenile type 1 diabetes mellitus.● METHODS: Subjects were 30 patients (30 eyes) with type1 diabetes mellitus who had well-documented records onthe duration of diabetes mellitus and condition of glycemiccontrol from the onset. They were 35 years of age oryounger (mean, 26.0 years), had a history of type 1 diabetesmellitus at least 5 years (mean, 8.4 years), had correctedvisual acuity of 20/20 or better, and showed no clinicallyapparent cataract on slit-lamp examination. Twenty-oneeyes of 21 subjects served as age-matched normal controls.They were 35 years of age or younger (mean, 25.7 years),had no diabetes mellitus, had corrected visual acuity of20/20 or better, and showed no signs of cataract on slit-lampexamination. The degree of lens opacity was quantifiedusing the anterior eye segment analysis system based on theScheimpflug principle. An index was created to representthe cumulative effect of long-term glycemic control (hyper-glycemic accumulation) by multiplying the average hemo-globin A1c value and the number of months from the onset.● RESULTS: The patients with diabetes mellitus exhibitedsignificantly greater degree of lens opacity than thenormal controls (P 5 .017, Mann-Whitney U-test).Among the patients with diabetes mellitus, the lensopacity was greater in eyes with retinopathy than thosewithout retinopathy (P 5 .011). Multiple regressionanalysis revealed that only the index of hyperglycemicaccumulation significantly correlated with the degree oflens opacity (P 5 .042).● CONCLUSION: Accumulated effect of hyperglycemia isrelated to the lens transparency in patients with diabetes.

(Am J Ophthalmol 2001;131:301–304. © 2001 byElsevier Science Inc. All rights reserved.)

D IABETES MELLITUS IS A RISK FACTOR FOR CATA-

ract. Epidemiologic studies on the relationshipbetween diabetes mellitus and cataract have

shown that there are several factors contributing to cata-ract development, such as juvenile-onset diabetes mellitus,long duration of diabetes mellitus, advanced age at thetime of clinical diagnosis, advanced retinopathy, treatmentwith diuretics, and poor control of blood sugar level.1–8

Recent development and clinical application of theScheimpflug image camera system have enabled quantita-tive measurement of lens opacity.9 Several studies evalu-ated lens transparency in eyes of diabetic patients.10–12

These previous studies, however, included elderly pa-tients,10–12 and thus their results might have been con-founded with the influence of age-related cataract, makingit difficult to assess the role of diabetes mellitus per se inthe development of cataract. In addition, the duration ofdiabetes mellitus and past record of glycemic control areusually poorly documented in patients with type 2 diabetesmellitus. Therefore, it is difficult to elucidate the accumu-lated effect of hyperglycemia on lens opacity in subjects,including patients with type 2 diabetes mellitus.

The present study was conducted to investigate quanti-tatively whether long-term control of blood glucose levelaffects lens transparency in patients with type 1 diabetesmellitus for whom the duration of diabetes mellitus and thecondition of glycemic control from the onset had been welldocumented. The patients had corrected visual acuity of20/20 or better and no signs of clinical cataract onslit-lamp examination.

SUBJECTS AND METHODS

SUBJECTS WERE 30 PATIENTS WITH TYPE 1 DIABETES MELLI-

tus who visited the Diabetes Center of Tokyo Women’sMedical University. They were 35 years or younger(26.0 6 6.0 years; mean 6 SD). Of the 30 patients, 14

Accepted for publication Sep 14, 2000.From the Department of Ophthalmology, School of Medicine, Uni-

versity of Tokyo (Drs Kato, Fukushima, Numaga, and Oshika), Tokyo;the Department of Ophthalmology, Diabetes Center, Tokyo Women’sMedical University (Drs Shiokawa, Kitano, and Hori), Tokyo; and KaiyaEye Clinic (Dr Kaiya).

Reprint requests to Satoshi Kato, MD, Division of Ophthalmology,Tokyo University Branch Hospital, 3-28-6 Mejirodai, Bunkyo-ku, Tokyo,112-8688 Japan; fax: 181-3-3943-3749; e-mail; katou-s@ka2.so-net.ne.jp

© 2001 BY ELSEVIER SCIENCE INC. ALL RIGHTS RESERVED.0002-9394/01/$20.00 301PII S0002-9394(00)00804-7

were men and 16 were women. All eyes had correctedvisual acuity of 20/20 or better and showed no signs ofcataract on slit-lamp examination. Eyes with any otherocular diseases were excluded from the study. Patients hada history of diabetes mellitus at least 5 years, during whichperiodical checkup of glycosylated hemoglobin A1c

(HbA1c) had been performed. The patients were includedonly if information were available on the duration ofdiabetes mellitus and the average level of HbA1c from theonset. Patients’ background data are summarized in Table1. One eye of each patient was randomly selected for thefollowing analysis. The refractive error averaged 21.9 62.2 diopters.

These patients with diabetes mellitus were further di-vided into two groups depending on the presence ofdiabetic retinopathy. Thirteen eyes had diabetic retinopa-thy, and 17 eyes had no retinopathy. The mean age was29.8 6 3.0 years in the retinopathy group and 23.1 6 6.2years in the nonretinopathy group. The HbA1c values atthe time of examination were 7.8% 6 1.3% in theretinopathy group and 7.9% 6 1.7% in the nonretinopa-thy group. No eye had been treated with retinal photoco-agulation.

Twenty eyes of 21 subjects, nine men and 12 women,served as age-matched normal controls. They were 35 yearsor younger, had no diabetes mellitus, had corrected visualacuity of 20/20 or better, and showed no signs of cataracton slit-lamp examination. The average age was 25.7 6 6.6years, and refractive error was 21.8 6 2.0 diopters.

The intensity of lens scattering was measured with theanterior eye segment analysis system (EAS-1000; Nidek,Gamagori, Japan).9 A slit image of the lens was obtainedafter the pupil was maximally dilated with 0.5% tropicam-ide and 0.5% phenylephrine hydrochloride. Slit images ofthe lens in the vertical plane were taken from the temporalside at the position of 0 degrees for the right eyes and at180 degrees for the left eyes. The power of light source was200 W. Analysis was performed with the equipped axialdensitometer only after a clear slit image extending fromthe anterior capsule of the lens to the posterior capsule was

obtained. The light-scattering intensities of all pixels alongthe optic axis of the lens from the anterior to the posteriorcapsule were added. This sum was used to represent thedegree of lens opacity of that eye.11–13

To examine the effect of long-term glycemic control, wemultiplied the average HbA1c value from the onset by thenumber of months of diabetes mellitus history. This pa-rameter was used as an index of hyperglycemic accumula-tion.

Multiple regression analysis was performed to investigatethe relation between various variables and the degree oflens opacity. Variables tested included the index of hyper-glycemic accumulation, patient age, gender, the presenceof diabetic retinopathy, and HbA1c at the time of exami-nation.

RESULTS

THE LIGHT-SCATTERING INTENSITY OF THE LENS WAS

6350.3 6 1761.0 (mean 6 SD) photon count in thecontrol group and 7657.5 6 1565.1 photon count in thediabetes mellitus group. There was a statistically significantdifference between groups (P 5 .017, Mann-WhitneyU-test). Among the diabetes mellitus group, eyes withretinopathy (8579.3 6 1654.9 photon count) showedsignificantly higher intensity than those without retinop-athy (6952.6 6 1080.4 photon count; P 5 .011). Theindex of hyperglycemic accumulation was significantlylarger in the retinopathy group (154.8 6 50.9) than in thenonretinopathy group (90.9 6 55.7).

The degree of lens opacity in patients with diabetesshowed a significant positive correlation with the index ofhyperglycemic accumulation (P 5 .0033, Spearman’s cor-relation coefficient, Figure 1) and age (P 5 .0024, Figure2). Multiple regression analysis revealed that only theindex of hyperglycemic accumulation significantly corre-lated with the degree of lens opacity (P 5 .042, Table 2).

TABLE 1. Characteristics of Patients With Type 1Diabetes Mellitus

Age (years; 30 patients) 26.0 6 6.0

Duration of diabetes mellitus (years) 13.9 6 6.8

Hemoglobin A1c at the time of examination (%) 7.8 6 1.5

Mean hemoglobin A1c from the onset of

diabetes mellitus (%)

8.4 6 1.3

Diabetes retinopathy (30 eyes)

No retinopathy 17

Mild background retinopathy 5

Moderate background retinopathy 8

Mean 6 SD.

FIGURE 1. Relationship between the degree of lens opacityand hyperglycemic accumulation in 30 eyes of 30 patients withtype 1 diabetes mellitus (Spearman’s correlation coefficient,P 5 .0033).

AMERICAN JOURNAL OF OPHTHALMOLOGY302 MARCH 2001

DISCUSSION

IN THIS STUDY, THE SUM OF LIGHT-SCATTERING INTENSITIES

at all pixels from the anterior to posterior capsule of the lenswas used as an index of lens opacity. Because the intensity ofa beam of light decreases as it passes through the lens, it ispossible that the light intensity scattered by the posteriorportion of the lens is unduly low and our formula may notaccurately represent the opacity of entire lens. However, thesubjects in this study did not have apparent cataract onslit-lamp examination, and the analysis was performed onlywhen a clear image extending to the posterior capsule wasobtained. Therefore, the reduction of light intensity throughthe lens was thought to be negligible.

The diabetic patients displayed significantly greater lensopacity than the normal controls, even though all patientswith diabetes mellitus had visual acuity of 20/20 or betterand none of them had clinically apparent cataract. Thisresult is in accordance with those of previous studies thatexamined the intensity of lens fluorescence in patientswith diabetes.14–16 The decreased lens clarity with diabetesmellitus has been attributed to the substances produced bylate reaction of the Amadori rearrangement in which thelysine residues of lens proteins react with glucose by meansof a glycation, or nonenzymatic glycosylation, under per-sistent elevation of serum glucose concentrations.17 Anal-ysis of the components of lens proteins by quasielastic lightscattering demonstrated degeneration of lens proteins inyoung patients with diabetes.18

Among the patients with diabetes mellitus, eyes withdiabetic retinopathy showed higher light-scattering inten-sity of the lens than those without retinopathy. Previousstudies reported that the severity of retinopathy was aninfluential factor on the degree of cataract and its progres-sion to stages requiring surgery.2,6 These previous andcurrent findings also suggest that cumulative effects ofhyperglycemia somehow play a role in the increases of lensopacity, because the development of retinopathy is closelyrelated to hyperglycemic accumulation. Hyperglycemic

accumulation represents both the duration of diabetesmellitus and the condition of glycemic control.

Although age was not found to be a significant factor inthe multiple regression analysis, a significant positivecorrelation was found between age and lens opacity in thesingle correlation test. There are two possible explanationsfor this. First, the light-scattering intensity of transparentlenses increases with aging even in normal subjects, in-cluding young population.19,20 Second, as the age in-creases, the duration of diabetes mellitus usually increasesin patients with type 1 diabetes mellitus, leading to greaterindex of hyperglycemic accumulation, which was hereinfound to be the factor most relevant to the lens opacity.

Limitations of the current study include relatively smallsample size. To clarify the effect of long-term control ofblood glucose level, patients were included in the studyonly if they had well-documented records on the durationof diabetes mellitus and condition of glycemic control fromthe onset. Accordingly, patients with type 2 diabetesmellitus had to be excluded. Subjects were also confined tothose who were 35 years or younger, had corrected visualacuity of 20/20 or better, and showed no clinically appar-ent cataract on slit-lamp examination. Second, this was aretrospective study. To evaluate various factors other thanthose included herein, a prospective study in larger popu-lation would be needed.

The current study was conducted in patients with type 1diabetes mellitus who were younger than 35 years old. Wefound that cumulative effect of hyperglycemia contributesto the increased opacity of the lens. In cases of olderpatients with type 2 diabetes mellitus, the situation may bemore complicated because of the combined influence ofage and hyperglycemia. Nevertheless, the current studysuggests that long-term glycemic control influences thelens transparency in patients with diabetes.

REFERENCES

1. Skalka HW, Prchal JT. The effect of diabetes mellitus anddiabetic therapy on cataract formation. Ophthalmology1981;88:117–124.

TABLE 2. Results of Multiple Regression Analysis toAssess Factors Relevant to the Degree of Lens Opacity

Standardized

Regression

Coefficent P Value

Index of hyperglycemic accumulation 0.460 .042

Age 0.265 .148

Gender 0.359 .059

Presence of diabetic retinopathy 0.25 .208

Hemoglobin A1c at the time of examination 20.193 .270

R2 5 0.499.FIGURE 2. Relationship between the degree of lens opacityand age in 30 eyes of 30 patients with type 1 diabetes mellitus(Spearman’s correlation coefficient, P 5 .0024).

LENS TRANSPARENCY IN PATIENTS WITH DIABETES MELLITUSVOL. 131, NO. 3 303

2. Klein BEK, Klein R, Moss SE. Prevalence of cataracts in apopulation-based study of persons with diabetes mellitus.Ophthalmology 1985;92:1191–1196.

3. Schwab IR, Dawson CD, Hoshiwara I, et al. Incidence ofcataract extraction in Pima Indians. Arch Ophthalmol 1985;103:208–212.

4. Harding JJ, Harding RS, Egerton M. Risk factors for cataractin Oxfordshire. Diabetes, peripheral neuropathy, myopia,glaucoma and diarrhoea. Acta Ophthalmologica 1989;67:510–517.

5. Klein BE, Klein R, Lee KE. Diabetes, cardiovascular disease,selected cardiovascular disease risk factors, and the 5-yearincidence of age-related cataract and progression of lensopacities: the Beaver Dam Eye Study. Am J Ophthalmol1998;126:782–790.

6. Klein BEK, Klein R, Moss SE. Incidence of cataract surgeryin the Wisconsin epidemiologic study of diabetic retinopa-thy. Am J Ophthalmol 1995;119:295–300.

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

8. Bron AJ, Sparrow J, Brown NAP, et al. The lens in diabetes.Eye 1993;7:260–275.

9. Sasaki K, Sakamoto Y, Shibata T, et al. The multi-purposecamera: a new anterior eye segment analysis system. Oph-thalmic Res 1990;22:3–8.

10. Dobbs RE, Smith JP, Chen T, et al. Long-term follow-up oflens changes with Scheimpflug photography in diabetics.Ophthalmology 1987;94:881–890.

11. Sasaki H, Hockwin O, Kasuga T, et al. An index for humanlens transparency related to age and lens layer: comparison

between normal volunteers and diabetic patients with stillclear lenses. Ophthalmic Res 1999;31:93–103.

12. Shiokawa A, Kato S, Ohara K, et al. Lens transparency usingScheimpflug slit images in young diabetic patients. Jpn J ClinOphthalmol (Japanese) 2000;54:464–468.

13. Tsukuda S, Ninomiya Y, Okano Y, et al. Age-related changesin lens transparency analyzed by EAS-1000 (Nidek). FoliaOphthalmol Jpn (Japanese) 1995;46:22–25.

14. Helve J, Nieminen H. Autofluorescence of the humandiabetic lens in vivo. Am J Ophthalmol 1976;81:491–494.

15. Mota MC, Morgado AM, Matos A, et al. Evaluation of anon-invasive fluorescence technique as a marker for diabeticlenses in vivo. Graefe’s Arch Clin Exp Ophthalmol 1999;237:187–192.

16. Bursell SE, Delori FC, Yoshida A, et al. Vitreous fluoropho-tometric evaluation of diabetics. Invest Ophthalmol Vis Sci1984;25:703–710.

17. Harding JJ. Nonenzymatic covalent posttranslational modi-fication of protein in vivo. Adv Prot Chem 1985;37:247–334.

18. Iida K, Maatsusima H, Shimizu Y, Obara Y. Quasielastic lightscattering study of the living human diabetic lens. AtarashiiGnaka (Japanese) 1998;15:267–270.

19. Kashima K, Trus BL, Unser M, Edwards PA, Datiles MB.Aging studies on normal lens using the Scheimpflug slit-lampcamera. Invest Ophthalmol Vis Sci 1993;34:263–269.

20. Fujisawa K, Sasaki K. Changes in light scattering intensity ofthe transparent lenses of subjects selected from population-based surveys depending on age: analysis through Scheimp-flug images. Ophthalmic Res 1995;27:89–101.

The full-text of AJO is now available online at www.ajo.com. AuthorsInteractivet, currently available in limited form, is undergoing an upgrade.

AMERICAN JOURNAL OF OPHTHALMOLOGY304 MARCH 2001