Blindness in insulin-treated diabetic patients with age at onset < 30 years

J Chron Dis Vol. 40, No. 3. pp. 215-220, 1987 Printed in Great Britain. All rights reserved

Copyright

BLINDNESS IN INSULIN-TREATED PATIENTS WITH AGE AT ONSET A. K. ‘Department

SJPILIE’*

and A.

0021-9681/87 $3.00 + 0.00 :c 1987 Pergamon Journals Ltd

DIABETIC <30 YEARS

GREENS

of Ophthalmology, Odense University Hospital, DK-5000 Odense, *University Institute of Clinical Genetics, Odense, Denmark

Denmark

and

(Received in revised form 8 May 1986)

Abstract-The occurrence of blindness was evaluated in a population-based group of Danish patients with insulin-treated diabetes diagnosed before the age of 30 years (N = 727), identified by means of insulin prescriptions. The study comprised a retrospective cross-sectional investigation, a longitudinal observation during the subsequent approximately 8 years, and a cross-sectional ophthalmological examination of all patients still alive at the end of the 8 year observation period. Prevalence rates of registered blind at base line were 3.4 and 2.6% for men and women, respectively. The overall incidence rate for blindness was found to be approximately 1.0 per 100 person years. At the ophthalmological examination 88% of blind patients were registered by the Danish Society for the Blind. The cause of blindness in the majority of patients was proliferative retinopathy. Blindness was found to be a significant problem in insulin-dependent diabetes, with a 5c-80 times higher risk of blindness than the background population. Blindness

Insulin-treated

Diabetes

Age at onset

INTRODUCTION

MATERIALS

Blindness caused by diabetic complications is associated with extensive medical and social problems [l]. Information on the frequency of blindness among diabetic patients is, in general, based on registers. Therefore, comparisons between countries may be invalid because of non-uniform criteria for blindness and differences between guidelines for registration. For example, in Denmark it is mandatory to register children and young adults below the age of 18 years with a visual acuity < 6/18 in the best eye; in contrast, registration of persons above the age of 18 years is only carried out by the Danish Society for the Blind at the request of the blind person, the criterion being vision corresponding to 6/60 or less in the best eye. The present report presents the epidemiologic features of blindness in a population-based group of Danish patients with insulin-treated diabetes, diagnosed before the age of 30 years. *Author

Retinopathy

for correspondence. 215

AND METHODS

The patient group The basic study material comprised all known insulin-treated diabetic patients with diagnosis established before the age of 30 years in the Fyn County, Denmark as of 1 July 1973. The patient group (413 male and 314 female patients) was identified by means of insulin prescriptions and evaluated to be 298% complete [2]. Data on name, date of birth, address, some clinical characteristics (age at diagnosis, insulin dose) were available [2, 31. The distribution of the patients according to sex, age and duration of diabetes as of 1 July 1973 is shown in Figs 1 and 2. Study methods The study fell into three parts: (1) A retrospective, cross-sectional investigation with collection of information on ophthalmological data and registered blind as of 1 July 1973 or as close as possible to this date. Data were traced and obtained from vari-

216

A. K. SJC~LIEand A. GREEN Number 601

n

N=L13

qd

N= 314 HE

3 0-L

5-9

lo-14

15-19 20-21

25-29

30.3L

35-39

10-L&

L5-49

50-54

55-59

60-6L

65-69

70'

Age lyearsl

Fig. 1. Seven hundred

and twenty seven insulin treated diabetic patients with age at diagnosis Distribution according to age and sex as of 1.7.1973.

ous sources as indicated in Fig. 3, all medical records being scrutinized by one of us (AKS). (2) Further, information on registered blindness as well as changes in eye status, particularly concerning retinopathy, covering the subsequent approximately 8 years from 1 July 1973, was collected. For those dead within this 8 year period the same information was recorded, as well as date of death. Data were traced and obtained simultaneously with the collection of data for the retrospective investigation (described above), supplemented with information on survival status from hospital records, public registries and the Fyn County Medical Computing Department. (3) A cross-sectional ophthalmological examination of all patients still alive and resident in Denmark as of 1 June 198 1. This part included an interview, an ophthalmological examination, a general clinical examination including mea-

i 30 years

surement of blood pressure, sampling of blood for glycosylated haemoglobin (Hb-A,), and blood urea, as well as urine tests for albumin, all of which were carried out by one of us (AKS) apart from a few blood samples. Methods of analysis The occurrence of blindness among the diabetic patients was expressed in terms of prevalence (with reference to the initial sampling data as well as to the time of the cross-sectional ophthalmological examination approximately 8 years after), and of the incidence (based on the patients recorded as newly blind during the follow-up period). The denominator of the incidence rates comprised the estimated number of person years at risk, observed among the patients during follow-up within age and sex groups. Withdrawal from the patient population at risk took place at the time of death,

Number 100 90 80 70 60

r

N=413

0

N:314

qQ

0"

r

50 10 30 23 10 0

1 0-L

5-9

i

5-19

20-24

25-29

30-34

35-39

LO-44

65-49

50-54

Duration iyeors,

Fig. 2. Seven hundred and twenty seven insulin treated diabetic patients with age at diagnosis Distribution according to duration of diabetes and sex as of 1.7.1973.

< 30 years.

Blindness

in Insulin-treated

Table 1 shows the age distribution of the whole material and of the registered blind patients, the mean duration of diabetes in the groups, as well as age and sex specific prevalence rates of blindness. With respect to clinical characteristics, all patients were insulin treated at base line, and insulin dependency was confirmed by means of C-peptide determination on surviving patients at follow-up.

Ophthalmologists

’ N=

Fyn Steno

34 I

MedIcoI Records County Hospltols Memorial Hospital Hvldsre Hospital

217

Diabetics

I

Incidence of blindness during follow -up

N-9

Fig. 3. Tracing of data regarding ophthalmological examinations for the retrospective cross-sectional investigation of 727 insulin treated diabetic patients with age at diagnosis >30 years, *From this group (N = 5) no further information concerning eye data was obtained.

emigration of registered blindness, assuming that these events occurred randomly within each of the years covered by the total follow-up period [4]. Patients who were blind at base line (1 July 1973) were excluded from the analysis of incidence. Blindness is defined as corrected vision corresponding to <6/60 in the best eye. RESULTS

Prevalence of blindness at base line As of 1 July 1973, 14 men (3.4%) and 8 women (2.6%) were registered as blind (no statistically significant sex difference).

During the approximate 8 year follow-up, 18 men and 19 women were registered as newly blind by the Danish Society for the Blind. The distribution according to age, sex, and duration of diabetes at registration for these 37 persons is shown in Table 2. Eighty percent of the patients registered were less than 50 years of age at registration. Age and sex specific incidence rates of blindness are shown in Table 3. The incidence rates were highest in the age group 4&59 years for both men and women. The overall incidence rates were 7.6 and 10.2 per 1000 person years for men and women, respectively. The cumulative incidence rate of blindness for specified age intervals can be estimated on the basis of the age and sex specific incidence rates [5]. These rates, also shown in Table 3, are interpretable as the risk of becoming blind during the age interval specified, given not being blind at the beginning of the interval. For example, the risk for a male diabetic patient, current age 20 years, of becoming blind before the age of 30 years is about 5%, while his risk of becoming blind before the age of 60 years is about 30%. It should be stressed that these values are rough average estimates only as no account has been taken to the variation of the duration of diabetes between the patients.

Table 1. The retrospective investigation of 727 insulin treated diabetic patients with age at diagnosis < 30 years. Registered blind according to age, duration of diabetes and sex, compared to the total material Females

Males Mean duration (years)

No. of reg. blind

Mean duration (years)

Blind % of total

Mean duration (years)

No. of reg. blind

Age 1.7.73 (yr)

No. of patients

Ck29 30-39 4&49 5&59 60+

209 94 61 34 15

7.7 14.8 22.6 32.2 44.4

0 6 3 3 2

26.0 31.0 31.3 37.0

(--) (6.4) (4.9) (8.8) (13.3)

147 77 46 30 14

8.0 17.4 26.7 34.6 40. I

0 1 3 3

Total

413

14.9

14

29.8

(3.4)

314

17.0

1.7.73

1.7.73

No. of patients

1.7.73

Mean duration (years) I .7.73

Blind % of total

1

27.0 29.7 35.0 32.0

(-) (1.3) (6.5) (10.0) (7.1)

8

31.6

(2.6)

218

A. K. SJBLIE and A. GREEN

1 July 1973-l on registration.

Table 2. Newly registered blind during the period according to current age, sex and duration of diabetes F: Females Duration &I9

Age (yr)

M

F

20-29 M F

2C29 3&39 4W9 50f

2 1 I 0

3 0 0 0

4 4 12

Total

4

3

Prevalence of blindness examination

I

10

2 3 3 10

at the ophthalmological

Causes of blindness Ophthalmoscopy was performed when possible at the ophthalmological examination; however, opacities of the ocular media precluded ophthalmoscopy in a few cases, and records from previous examinations were employed in these patients. The cause of blindness, as judged

Males

Females

(yr)

3&39 M F 0 1

I

Total

40+ M

F

M

F

0 I 2

0 0 0

0 0

3 6 6 3

5 4 6 4

4

1

2

18

19

I

I111

At the time of the ophthalmological examination about 8 years after the initial date of sampling, 627 patients were alive and resident in Denmark. Of these, 577 accepted the invitation to be examined (participation rate 92%). The best corrected visual acuity was measured on all patients. Vision of >6/9 in the best eye was observed in 290 men (89.2%) and 227 women (90.1%). Among the patients, 20 men (6.2%) and 14 women (5.6%) were blind, a total of 34 patients (5.9%). Thus, only 15 men (4.6%) and 11 women (4.4%) had moderate visual loss in the best eye. Of the 34 blind patients, 30 (88%) were registered as blind by the Danish Society for the Blind.

Table 3. Incidence

June 1981 M: Males,

3

from the procedure described, in the 34 patients recorded as blind was retinopathy and/or neovascular glaucoma in all but two patients. Of these two, one had bilateral amblyopia caused by congenital cataract, the other bilateral juvenile cataract (the latter patient developed panophthalmia and subsequent phthisis after cataract-operation on one eye and refused operation of the other eye). DlSCUSSlON

There is little doubt that in general registers of blind are incomplete, as some of the blind have aversion to registers of any type. Another, perhaps more important aspect is that many elderly blind do not consider it worthwhile going through the necessary examinations or treatment for reduction in vision, in order to be registered. Others may be unaware of the practical advantages of being members of the Danish Society for the Blind. Caird et al. [6], based on the survey of Sorsby [7] of registered blind in England and Wales, calculated that 2.8% of all diabetics were blind. The incidence rate for blindness in diabetics, all age groups, is stated by Caird et al. to be 0.2

rates and estimated cumulative incidence rates for registered blindness patients. Observation period: 1 July 1973~ I June 1981

Age interval (yr)

No. of person years

No. of blind patients

2Ck29 3&39 40-49 5&59 60-69 Total

747 746 491 291 81 2356

4 5 6 3 0 18

2&29 3ck39 40-49 so- 59 60-69 Total

550 539 410 221 137 1857

5 4 6 3

1 19

Incidence rate (per 1000 person years)

among

insulin

treated

diabetic

Estimated cumulative incidence rate (%) to end of age interval from age 20 years

30 years

40 years

50 years

5.4 6.7 12.2 10.3 0 7.6

5.2 11.4 21.6 29.2 29.2

6.5 17.2 25.3 25.3

II.5 20.2 20.2

9.8 9.8

9.1 7.4 14.6 13.6 7.3 10.2

8.7 15.2 26.8 36.1 40.6

7.2 19.8 30.0 34.9

13.6 24.6 29.9

12.7 18.8

Blindness

in Insulin-treated

so, it is possible to obtain an estimate assuming that the rates for blindness in the population of England and Wales, as given by Sorsby, are applicable to the Danish population. Given this, Table 4 shows the observed number of blind in each age group compared to the expected number; the latter number assumes that the prevalence of blindness in diabetics is similar to that of the background population. It can be seen that the observed numbers are exceedingly high, 50 to 80 times higher than expected, up to 70 years of age; thereafter the present ophthalmological material contained only a few diabetics, none of whom were blind. The cause of blindness in the present material was, in the majority of cases, proliferative retinopathy, although maculopathy was also present in a few of these patients. This is in contrast to older diabetics with non insulin dependent diabetes, in whom maculopathy is the dominating cause of blindness from retinopathy, and in whom cataract and macular degeneration are frequent causes of reduction in vision. The present investigation is a population based study, and is in part based on retrospective data. Unfortunately, systematic baseline data regarding retinopathy as of 1.7.1973 was not available, and a precise characterization of the retinal status of the patients at that time, therefore, is impossible. Neither was systematic data on other clinical characteristics, such as blood sugar levels, blood pressure or degree of proteinuria available. We conclude, from this study, that blindness is a significant problem in insulin-dependent diabetes. Since the cause of blindness in the great majority of blind patients is proliferative retinopathy, the means of reducing the problem is by prevention and treatment of diabetic

per 100 per year, with rising incidence rates with increasing age, the highest occurring after the age of 70 years. In contrast, the incidence of blindness in the present material was highest in the younger age groups from 40 to 59 years, while the overall rate was 0.8 per 100 person years for men and 1.O per 100 person years for women, in other words, four to five times higher than the rate of Caird et al. There was no greater difference between men and women in this respect in the present material, confirming other findings for the age groups below 70 years L&71. The discrepancy between the incidence rates of Caird et al. and the present investigation can, in part, be ascribed to the selection of patient materials. The material of Caird et al. included both tablet and insulin-treated patients with all ages of onset, and thus comprised a very inhomogeneous patient group, whereas the present material appears homogeneous in respect of age at onset and type of diabetes. Furthermore, in our estimates of incidence of blindness, we have included those patients who became blind and died within the follow-up period; in this particular group, the incidence of blindness was 2.7 per 100 person years. Thus, serious bias may be introduced, if such patients are not accounted for in incidence estimates based on historical data. The cross-sectional ophthalmological examination demonstrated a prevalence rate of blindness of 5.9% in our patient group, with a minimum duration of diabetes of 8 years. Prevalence rates for blindness in the general Danish population are unknown, although overall prevalence rates have been stated to be between 0.3 and 0.5% [8]. Therefore, the excess prevalence of blindness in diabetics cannot be calculated directly in the present investigation. Even

Table

4. Relative

prevalence

219

Diabetics

of blindness

among

diabetic

patients

according

to age

Age (yr)

Prevalence of blindness in general population* No. of diabetics at the ophthalmological exam. (present investigation) Expecfed no. of blind diabetics, using prevalence in general population (from Caird) Observed no. of blind diabetics (present inv.) Observed/expected *Prevalence

rates calculated

f&29

3&49

5&69

70f

Total

0.23 x IO-’

0.72 x 1O-3

2.26 x 10-j

12.61 x lo-’

1.77 x IO-3

from Caird

159

0.04

300

0.22

111

0.25

3

18

13

75.0

81.8

52.0

et al. [6], Table A4, Appendix.

7

0.09

0

577

0.60

34 56.7

A. K. SJC~LIE and A. GREEN

220

retinopathy. available

Photocoagulation in

this

geographic

treatment area

from

was the

early 70’s, and large scale treatment of diabetic retinopathy by this means in recent years will undoubtedly change the current statistics of

3.

visual

4.

outcome.

Studies

on the frequency

and

determinants of retinopathy, and of other ocular diabetic manifestations will be presented in subsequent

5.

papers. 6.

REFERENCES 1.

2.

Caird FI: Epidemiology of diabetic retinopathy. In Diabetic Retinopathy, Lynn JR et al. (Eds). New York: Grune & Stratton, 1974. pp. 3545 Green A, Hauge M, Holm NV, Rasch LL: Epidemiological studies of diabetes mellitus in Denmark.

7.

8.

II. A prevalence study based on insulin prescriptions. Diabeiologis 20: 468270, 1981 Green A. Houeaard P: Epidemiological studies of diabetes mellitui in Denmark. 4. Clinical characteristics of insulin-treated diabetes. Diabetologia 25: 231-234, 1983 Doll R, Hill AB: Lung cancer and other causes of death in relation to smoking. Br Med J II: 1071-1081, 1956 Morgenstern H, Kleinbaum DG, Kupper LL: Measures of disease incidence used in epidemiologic research. Int J Epidemiol 9: 97-104, 1980 Caird FI, Pirie A, Ramsell TG: Diabetes and Eye. Oxford: Blackwell Scientific Publications, 1969 Sorsby A: The incidence and causes of blindness in England and Wales 194881962. Reports on Public Health and Medical Subjects. No. 114. London: Her Majesty’s Stationery Office, 1966 Nielsen NV: The prevalence and causes of impaired vision in diabetics. Acta Ophthahnol (Copenh) 60: 677-691, 1982