- Email: [email protected]
Epidemiology of insulin-dependent diabetes mellitus in Canterbury, New Zealand
Diabetes Research and Clinical Practice, 3 (1987) 21-29
Elsevier
21
DRC00099
Epidemiology of insulin-dependent diabetes mellitus in Canterbury, New Zealand D . R . M a s o n * , R.S. S c o t t a n d B.A. D a r l o w 1 Department of Medichle, Christchurch Clinical School of Medicine. Christchurch, and i Department of Paediatrics. Christchurch Hospital, Christchurch, New Zealand
(Received 8 May 1986, accepted 8 August 1986)
Key words: Insulin-dependentdiabetes; Epidemiology;Incidence; Prevalence
Summary A prospective study of incidence and prevalence of insulin-dependent diabetes mellitus in persons under 20 years was conducted over a 4-year period (1 February 1982-1 February 1986) for the Canterbury Hospital Board (total population 342,000) area in New Zealand. A central register for the area was established at the beginning of the study period. Degree of ascertainment was close to 100%. Average annual incidence was 11.7 persons per 100,000 (females: 10.6 per 100,000; males: 12.7 per 100,000) with no significant sex difference or temporal trends. Incidence peaks were seen for both sexes in the pubertal ages (females: 11 years; males: 13 years), with minor peaks occurring for both sexes in the pre-school ages. Age of onset was significantly younger in females than males. A seasonal variation in incidence was seen for males, with peaks in late autumn and mid-winter. 5.7% of the new diabetics had a first-degree relative with insulin-dependent diabetes mellitus. Islet cell cytoplasmic antibodies were detected in 68% of new diabetics and in 0% of age- and sex-matched healthy controls. Thyroid, gastric and adrenal auto-antibodies were seen more frequently in diabetics than in controls, but this difference was not significant. Prevalence of insulin-dependent diabetes on 1 February 1982 was 1.00 per I000 and 1.05 per 1000 on 1 February 1986. The insulin-dependent diabetes mellitus incidence characteristics noted for the Canterbury Hospital Board area are similar to those reported for European and North American populations.
Introduction Address for correspondence: Dr. R.S. Scott, Department of Medicine, Christchurch Clinical School of Medicine, P.O. Box 4345, Christchurch, New Zealand. * Present address: Department of Zoology, University of Canterbury, Christchurch, New Zealand.
The results of a retrospective study of the incidence of juvenile diabetes mellitus in New Zealand were reported in 1980 [1]. Information was obtained from hospital admission data stored at the Department of Health, National Statistics Centre. Com-
0168-8227/87/$03.50 © 1987 Elsevier SciencePublishers B.V. (BiomedicalDivision)
22 pared with several other studies [2--9], the age-related incidence showed an altered distribution; specifically, there was an absence of childhood peaks and a sustained relatively higher incidence in adolescence. In addition, in contrast to other reports [2,4,5,8,9] no regular seasonal trends in incidence were seen. We have undertaken a prospective study based on a register begun on 1 February 1982 of all new cases of insulin-dependent diabetes mellitus (IDDM) in the age group 0-19 years. The population upon which the register is based corresponds to 10% of the total New Zealand population under 20 years of age. The present investigation covers a 4-year period, until 1 February 1986, and is aimed at answering the following questions: (1) What is the annual incidence of IDDM in the registry area and is it stable? (2) What is the prevalence of IDDM in the 019-year age group in the registry area at the beginning and end of the 4-year study period? (3) Do the sex, age, and seasonal distributions of incidence conform with those previously reported for New Zealand or other countries? (4) To what extent do I D D M and non-insulindependent diabetes (NIDDM) occur in first-degree relatives of subjects diagnosed IDDM over the 4year study period? (5) What percentage of the newly diagnosed diabetic subjects have islet cell cytoplasmic antibodies (ICA) and/or other organ-specific auto-antibodies at diagnosis?
Methods
Subjects A register of new insulin-dependent diabetic subjects in the 0-19-year age group was begun for the Canterbury Hospital Board (CHB) area [10] on 1 February 1982 (Fig. 1). All children aged 0-12 years, living in this area, are admitted to the Christchurch Hospital upon diagnosis of diabetes, whereas adolescents and adults are usually treated on an ambulatory basis from the Christchurch Dia-
"$.
o
,
,oo ~m
Fig. I. Map of the South Island of New Zealand showing the Canterbury Hospital Board (hatched) area. The inset shows a map of New Zealand with latitude and longitudeindicated.
betes Education Centre, in-patient care being organised only for those with significant metabolic derangement. The possibility that a newly diagnosed diabetic subject may not have been identified by these detection methods is very unlikely. The population base is well defined and diabetes treatment is centralized; all individuals presenting with diabetes and requiring insulin are referred to the Paediatric/Adolescent or Adult Diabetes Services in Christchurch; independent annual surveys (see below) identifying insulin users have not revealed any 'missed cases' during the 4 years of the survey. Thus ascertainment was probably 100%. A standardised form was used to record the patient's name, date of birth, sex, address, physician, date of diagnosis, presenting symptoms, presenting biochemical results, previous history of viral infection, family history of diabetes mellitus, date
23 insulin started, and date blood sample taken for measurement of immunological markers. Only insulin-treated subjects were included in the register. A typical history characterised by excessive thirst and frequency of urination was seen in all but two cases. These two patients were included in the register since their attending physician considered insulin treatment was necessary in both cases. One of these individuals had a random plasma glucose level < 11.1 mmol/1, but an oral glucose tolerance test was diagnostic of diabetes and subsequent elevated glucose levels prompted commencement of insulin. Presenting hyperglycaemia (> 11.1 mmol/l) was present in the remaining 98% of cases; ketonuria and weight loss were documented in 75% and 61% of cases respectively. The under 20 year population in the CHB area corresponds to 10% of the total population for this age group in New Zealand. Breakdown of the total population of the area by age and sex was obtained from the 1981 Census. In the 0-19-year age group there was an average annual fall in population of < 2% for the previous 5 years, and population projection figures for the following 5 years suggest a similar trend. Because of the stability of the population, incidences were calculated using the 1981 population figures. The prevalence of IDDM in under 20-year-olds as of 1 February 1982 and at 1 February 1986 was also determined using the 1981 Census figures. All diabetic subjects in this age group who were living in the CHB area on 1 February 1982 and had been
diagnosed prior to this date, were identified using a combination of Hospital (in- and out-patient), Diabetes Education Centre and Dietary Department records. Individuals diagnosed as having IDDM between 1 February 1982 and 1 February 1986 were identified through the prospective study measuring incidence. During the 4-year study period patients were retained in the prevalence register until date of relocation outside the CHB area, death, or reaching 20 years of age. Ascertainment for incidence and prevalence was assessed by annual independent surveys of individuals obtaining insulin from all pharmacies in the CHB area. To further evaluate presenting characteristics of IDDM in new diabetics, serum from these subjects was screened for the presence of immunological markers (see below) reported to be associated with the disorder. In order to directly relate the presence of immunological markers to the risk of developing IDDM, the frequency of these markers in non-diabetic controls was also determined. Age- and sexmatched healthy control subjects were selected from the population at large (Table 1). Anyone with a personal or family history of diabetes mellitus or other endocrine disorders was excluded from the control group.
Assay of immunological markers Serum samples were obtained within one month of diagnosis from 52•54 of the new diabetic subjects. Samples from controls were obtained at the time of
TABLE 1 DETAILS CONCERNING AGE AND SEX OF NEW DIABETIC AND CONTROL SUBJECTS Females
n Mean age (years) Age range (years)
Males
Diabetic subjects
Control subjects
Diabetic subjects
Control subjects
24 10. I 1-19
24 10.1 1-19
30 13.2 2-19
30 13.0 2-19
24
<
z~
I
z~
I=
z~
m
z~ I
© [.<
L~
Z
Z E~
z~
ILl ,-1 <
Z < <
~
~1
e~
~.~ ~v I'--1
I~
'9
-i
25 3O
Gastric parietal cell and adrenal antibodies were also detected by the indirect immunofluorescent technique (described above), using cryostat sections of mouse stomach and human adrenal tissue respectively. A FITC-conjugated anti-human IgG (South Pacific Immunological Laboratories, New Zealand) was used at 1/75 dilution. Thyroid autoantibodies (anti-thyroglobulin and anti-thyroid microsomal) were determined by tanned red cell agglutination using a commercial kit (Fujizoki, Japan).
o O
20
3
o I
I
I
I
I
0
2
4
6
8
I
I
I
I
I
I
10 12 14 16 18 20
AGE (years)
Fig. 2. Risk of developing IDDM by age and sex. ( 0 - - 0 ) males; ( & - - & ) males.
fe-
selection for the study. Sera were stored at - 2 0 " C from one week to 18 months before testing. ICA were detected by indirect immunofluorescence using frozen sections of blood group 0 human pancreas obtained from cadaver kidney donors [11]. All serum samples were tested both undiluted and diluted 1/2 in phosphate-buffered saline (PBS). The end-point titration of positive samples was determined. Fluorescein isothiocyanate (FITC) -conjugated rabbit anti-human IgG (Tago, California) was used at 1/15 dilution. Preparations were read independently using a Leitz Ortholux II microscope by two observers who were not informed of the identity of the samples.
Statistical analysis Statistical significance was assessed using the chisquared test, or in the case of sex-related variations in the age at onset of IDDM, the Mann-Whitney test was used. Seasonal variations were studied by the chi-squared test using totals for each individual month, and totals for 3-month periods (December-February; March-May; June-August; September-November).
Results
Incidence and risk of IDDM During the 4-year study period, 54 individuals under 20 years of age developed IDDM. Table 2 gives numbers of cases and incidences for each year by sex in addition to giving the total and mean values. During the study period there were no significant differences in the annual incidence rates. The risk (cumulative incidence rate; Fig. 2) of developing IDDM in the CHB area by the age of 19 years
TABLE 3 AGE- AND SEX-SPECIFIC PREVALENCE OF IDDM I February 1986
1 February 1982
Females Males Total
Population < 20 years
No. of cases
Prevalence per 1000
No. of cases
Prevalence per 1000
56,569 58,854 115,423
54 62 116
0.95 1.05 1.00
57 64 121
1.01 1.09 1.05
26 S I
50--
7--
40-o o o
6--
°o 3 0 - .~. 5--
Z ___ 20-Z
~D 4 ~
10--
3
O--
Z
rr LU rn 3 - -
I
I
I
I
I
I
I
0
2
4
6
8
10
12
I
I
14 16
I
I
-
\
i oooqJ~%ooooooeo%~%
2--
/
18 20
AGE (years)
Fig. 3. Age-specific incidence of IDDM. O-
I
I
I
I
I
1
I
I
I
I
I
I
J
A
S
0
N
D
J
F
M
A
M
J
was similar for males and females, 2.1 per 1000 and 2.4 per 1000 respectively.
Prevalence The prevalence rates on 1 February 1982 and 1 February 1986 for the 0-19-year age group are shown in Table 3. As can be seen, prevalence has also remained stable over this time period. Sex-, age- and season-related variations in #zcidence Although the mean annual incidence was higher for males than females (Table 2) this was not significant. In fact, during the first 2 years there was an excess of females. Age of onset was significantly (P < 0.001) lower for females than males. Incidence reached a maximum (at 11 years for girls, and 13 years for boys) during the pubertal period (Fig. 3). It then decreased rapidly but rose again to a second smaller peak which occurred at 16 years for both males and females. A small 'childhood' peak was observed at 2-3 years of age. Both sexes showed this trend. Significant seasonal variation of onset of I D D M (Fig. 4) was seen for males (P < 0.025), but not for females or for the combined population when analysed using totals for individual months (11 degrees of freedom). However, if the combined population was examined using four 3-month periods
MONTH
Fig. 4, Seasonal variation of onset of IDDM. (----) males; (O--O) combined data.
(oooo)
females;
(see Methods) the level of significance was improved, with a P value approaching 0.05 (chisquared value 7.8; 3 degrees of freedom).
Occurrence of diabetes mellitus in first-degree relatives 9.4% (5/53) of subjects diagnosed in the 4-year study period had a first-degree relative with either N I D D M or IDDM (Table 4). One subject who is an adopted child was excluded as no family history was available. Immunological market's Islet cell cytoplasmic antibodies. 67.9% (36/53) of subjects diagnosed in the study period (65% of females and 70% of males) were positive for ICA. One subject with a negative result was excluded since the sample was not obtained until 30 months after diagnosis. The mean ages for both ICA-positive females and males (9.6 and 13.2 years respectively) were similar to those for the total I D D M
27 TABLE 4 DISTRIBUTION OF FIRST-DEGREE RELATIVES WITH DIABETES New diabetic subject Sex
Age
F M M M M
11 4 9 14 16
Diabetic relative
Type of diabetes of relative
Mother Brother Father Father Mother Brother
NIDDM IDDM NIDDM IDDM IDDM NIDDM
population (10.4 and 13.2 years respectively). The ICA end-point titres for positive sera ranged from 1/1 (undiluted) to 1/64 (1.56% v/v plasma). In contrast, none of the control subjects were positive for ICA.
Other organ-specific auto-antibodies. Five new diabetic subjects were positive for one of the other four organ-specific antibodies (thyroid microsomah two; thyroglobulin: one; gastric parietal cell: one; adrenal: one). One new diabetic subject, who had thyrotoxicosis in addition to IDDM, was positive for antibodies to both thyroglobulin and thyroid microsomal antigen. Therefore, 11.1% of new diabetics (two female, four male; age range 11-15 years) had at least one organ-specific auto-antibody. Of these, 67% were positive for ICA. Three (5.6%) of the age- and sex-matched control subjects (two female, one male; age range 1316 years) were positive for one auto-antibody (thyroid microsomal: two; adrenal: one). No control subject had more than one auto-antibody. The frequency of occurrence of these four auto-antibodies in control subjects was not significantly less than that observed for diabetic subjects.
Discussion
This is the first report of a prospective study of incidence and of prevalence of IDDM within New
Zealand. Although the register of under 20-year-old diabetic subjects established in the CHB area is based on a population corresponding to only 10% of the total population for this age group in New Zealand, it has the advantage of high ascertainment. Centralised treatment of diabetic subjects at diagnosis in the CHB area and external validation of cases are the primary factors responsible for the completeness of detection. In the Northern Hemisphere at least, the incidence of IDDM shows a trend to increase in countries in more northern regions, and these geographical differences cannot be explained by variability in precision of survey methods used. The average annual incidence (11.7 per 100,000; range 10.413.0) in this New Zealand study is comparable to rates reported from other countries of similar latitude: Denmark [4], ages 0-29, 13.2; Scotland [8], ages 0"19, 13.8; Pittsburgh, U.S.A. [12], ages 0-19, 14.7; Toronto, Canada [13], ages 0-18, 9.0; Montreal, Canada [6], ages 0-16, 8.8; Netherlands [14], ages 0-19, 10.9; U.K. [3], ages 0-15, 7.7. It is also similar to that reported (10.4 per 100,000) for a retrospective study based on the 0-19-year-old population of New Zealand Ill. There was, however, no evidence for a temporal increase in incidence, a trend which has been reported for some populations [8,9,15], but not for others [1,4,6,7]. The risk of developing diabetes is greater for males than for females, but this is not significant. A similar finding has previously been reported [1,3,4,7,8,13] and was statistically significant in at least one population [9]. In the present study this trend is due entirely to a higher incidence for males than females over the age of 13 years. This is in contrast to a number of other studies [4-9] where a higher incidence was found not only for adolescent males, but also for boys in the youngest age group (0-4 years). Dahlquist et al. [9] have suggested the higher risk for males could be either an expression of a higher frequency of genetic risk factors in males or of a higher susceptibility to infections [16,17]. The age-related incidence pattern differed from that of Crossley and Upsdall [1]. A distinct inci-
28 dence peak was seen for both males (13 years) and females (11 years); a finding similar to that noted in other studies [3,4,6,7,9]. In contrast to the sustained higher level of incidence reported by Crossley and Upsdall between ages 11-19 years, we observed a fall in attack rate for both males and females, with a secondary peak for both sexes at 16 years of age. A minor pre-school peak, seen for both males and females, has also been reported by others [3,4,6,8]. Although the major incidence peak is often referred to as the 'pubertal' peak, an association with pubertal status has not been objectively established, and in fact there has been at least one report of an earlier peak for males than females
[8]. There is a seaonal trend, similar to that observed by several other groups [2-6,8,9,19] with higher incidence rates in autumn and winter, and lower rates in spring and summer. In our study this trend was the result of significant seasonal variation for males. A difference between males and females for seasonrelated incidence has also been noted by others [6,8,9,12,20]. Dahlquist et al. [9] have suggested that if seasonal variation is dependent on an infectious disease pattern, as hypothesised [21], then a more pronounced seasonality would in fact be expected for males because of their higher susceptibility to infections [16,17]. It is possible that a delay in seeking medical care during the summer vacation may have contributed to the lower summer incidence rate observed. In an attempt to minimise the influence of this, seasonal variation was studied by the chi-squared test using the total numbers of cases for four 3-month periods. By including February with December and January it was hoped that the effect on incidence of delayed diagnosis during the schools' summer vacation (which finishes at the end of january) would be largely overcome. When examined in this way, the seasonal trend for the combined male plus female population was not diminished. The percentage of recently diagnosed I D D M subjects who had first-degree relatives with I D D M (5.7%) was approximately half that of previous reports [5,6,9]. However, the small number of subjects
in our study may account for this discrepancy. The frequency of occurrence of islet cell antibodies in new diabetics (68%) and in healthy control subjects (0%) is in good agreement with earlier reports [22-24]. Other organ-specific auto-antibodies were detected more frequently in diabetics than control subjects; this difference was smaller than noticed by some researchers [24-26], and was not significant. The results reported here provide no evidence to suggest that aetiological factors operative in I D D M in the CHB area of New Zealand differ from those in Europe and North America. Our data does not confirm the atypical findings of Crossley and Upsdall [1] with respect to age- and sex-related incidence.
Acknowledgements The authors gratefully acknowledge the assistance of Ms. L.J. Brown, Ms. F. McCone, Ms. L. Taylor, Mr. R.P. Weir and the staff of the Department o f Immunopathology, Christchurch Hospital. We thank Ms. J. Nicol for typing the manuscript. This work was supported by grants from the Medical Research Council of New Zealand, the National Children's Health Research Foundation, the New Zealand Diabetes Association, the Canterbury Medical Research Foundation and N o v o Australia (Pty) Ltd. Preliminary results have been presented at the Annual Meeting of the New Zealand Society for the Study of Diabetes, Auckland, October 1984.
References Crossley, J.R. and Upsdall, M., The incidence of juvenile diabetes mellitus in New Zealand, Diabetologia, 18: 29-34, 1980. Gamble, D.R., Epidemiologicaland virologicalobservations on juvenile diabetes, Postgrad. Med. J., 50: 538-543, 1974. Bloom, A., Hayes,T.M. and Gamble, D.R., Registerof newly diagnosed diabetic children, Br. Med. J., 3: 580-583, 1975. Christau, B., Kromann, H., Andersen. O.O. et al., Incidence, seasonal and geographicalpatterns of juvenile-onsetinsulindependent diabetes mellitus in Denmark, Diabetologia. 13: 281-284, 1977.
29 5 Sterky, G., Holmgren, G., Gustavson, K.H. et al., The incidence of diabetes mellitus in Swedish children 1970-1975, Acta Paediatr. Scand., 67: 139-143, 1978. 6 West, R., Belmonte, M.M., Colle, E., Crepeau, M.P., Wilkins, J. and Poirier, R., Epidemiologic survey of juvenileonset diabetes in Montreal, Diabetes. 28: 690-693, 1979. 7 La Porte, R.E., Fishbein, H.A., Drash, A.L. et al., The Pittsburgh insulin-dependent diabetes mellitus (IDDM) registry. The incidence of insulin-dependent diabetes mellittls in Allegheny County, Pennsylvania (1965-1976), Diabetes, 30: 279-284, 1981. 8 Patterson, C.C., Thorogood, M., Smith, P.G., Heasman, M.A., Clarke, J.A. and Mann, J.L., Epidemiology of type I (insulin-dependent) diabetes in Scotland 1968-1976: evidence of an increasing incidence, Diabetologia, 24: 238-243, 1983. 9 Dahlquist, G., Blom, L., Holmgren, G. et al., The epidemiology of diabetes in Swedish children 0-14 years - - a sixyear prospective study, Diabetologia. 28: 802-808, 1985. 10 Scott, R.S., Brown, L.J. and Clifford, P.C., Use of health services by diabetic persons. II. Hospital admissions, Diabetes Care, 8: 43-47, 1985. 11 Marner, B., Lernmark, A., Nerup, J., Molennar, J.L., Tuk, C.W. and Bruining, G.J., Analysis of islet cell antibodies on frozen sections of human pancreas, Diabetologia, 25: 93-96, 1983. 12 Tajima, N., La Porte, R.E., Hibi, I., Kitagawa, T., Fujita, H. and Drash, A.L., A comparison of the epidemiology of youth-onset insulin-dependent diabetes mellitus between Japan and the United States (Allegheny Country, Pennsylvania), Diabetes Care, 8: 17-23, 1985. 13 Ehrlich, R.M., Walsh, L.J., Falk, J.A., Middleton, P.J. and Simpson, N.E., The incidence of type 1 (insulin-dependent) diabetes in Toronto, Diabetologia. 22: 289-291, 1982. 14 Vaandrager, G.J., Bruining, G.J., VeenhoR F.J. and Drayer, N.M., Incidence of childhood diabetes in the Netherlands: a decrease from north to south over North-western Europe?, Diabetologia. 27: 203-206, 1984. 15 Green, A. and Andersen, P.K., Epidemiological studies of diabetes mellitus in Denmark: 3. Clinical characteristics and
16
17 18
19
20
21 22
23
24
25
26
incidence of diabetes among males aged 0 to 19 years, Diabetologia. 25: 226-230, 1983. Washburn, T.C., Medearis, D.N. and Childs, B., Sex difference in susceptibility to infections, Pediatrics, 35: 57-64, 1965. Michaels, R.H. and Rogers, K.D., A sex difference in immunological responsiveness, Pediatrics. 47: 120-123, 1971. MacMillan, D.R., Kotoyan, M., Zeidner, D. and Hafezi, B., Seasonal variation in the onset of diabetes in children, Pediatrics, 59:113-115, 1977. Fleegler, F.M., Rogers, K.D., Drash, A., Rosenbloom, A.L., Travis, UB. and Court, J.M., Age, sex and season of onset of juvenile diabetes in different geographic areas, Pediatrics, 63: 374--379, 1979. Fishbein, H.A., La Porte, R.E., Orchard, T.A., Drash, A.L., Kuller, L.H. and Wagener, D.K., The Pittsburgh insulin-dependent diabetes mellitus registry. Seasonal incidence, Diabetologia. 23: 83-85, 1982. Gamble, D.R. and Taylor, K.W., Seasonal incidence of diabetes mellitus, Br. Med. J., 3: 613-633, 1969. lrvine, W.J., McCallum, C.J., Gray, R.S. et al., Pancreatic islet-cell antibodies in diabetes mellitus correlated with the duration and type of diabetes, co-existent autoimmune disease, and HLA type, Diabetes. 26: 138-147, 1977. Bottazzo, G.F., Dean, B.M., Gorsuch, A.N., Cudworth, A.G. and Doniach. D., Complement fixing islet-cell antibodies in type I diabetes: possible monitors of active betacell damage, Lancet, i: 668-672, 1980. Neufeld, M., MacLaren, N.K., Riley, W.J. et al., Islet cell and other organ-specific antibodies in U.S. caucasians and blacks with insulin-dependent diabetes mellitus, Diabetes. 29: 589-592, 1980. Riley, W.J., Winer, A. and Goldstein, D., Coincident presence of thyro-gastric autoimmunity at onset of Type I (insulin-dependent) diabetes, Diabetologia. 24: 418-421, 1983. Betterle, C., Zanette, F., Pedini, B. et al., Clinical and subclinical organ-specific autoimmune manifestations in Type I (insulin-dependent) diabetic patients and their first-degree relatives, Diabetologia. 26: 431-436, 1984.
Elsevier
21
DRC00099
Epidemiology of insulin-dependent diabetes mellitus in Canterbury, New Zealand D . R . M a s o n * , R.S. S c o t t a n d B.A. D a r l o w 1 Department of Medichle, Christchurch Clinical School of Medicine. Christchurch, and i Department of Paediatrics. Christchurch Hospital, Christchurch, New Zealand
(Received 8 May 1986, accepted 8 August 1986)
Key words: Insulin-dependentdiabetes; Epidemiology;Incidence; Prevalence
Summary A prospective study of incidence and prevalence of insulin-dependent diabetes mellitus in persons under 20 years was conducted over a 4-year period (1 February 1982-1 February 1986) for the Canterbury Hospital Board (total population 342,000) area in New Zealand. A central register for the area was established at the beginning of the study period. Degree of ascertainment was close to 100%. Average annual incidence was 11.7 persons per 100,000 (females: 10.6 per 100,000; males: 12.7 per 100,000) with no significant sex difference or temporal trends. Incidence peaks were seen for both sexes in the pubertal ages (females: 11 years; males: 13 years), with minor peaks occurring for both sexes in the pre-school ages. Age of onset was significantly younger in females than males. A seasonal variation in incidence was seen for males, with peaks in late autumn and mid-winter. 5.7% of the new diabetics had a first-degree relative with insulin-dependent diabetes mellitus. Islet cell cytoplasmic antibodies were detected in 68% of new diabetics and in 0% of age- and sex-matched healthy controls. Thyroid, gastric and adrenal auto-antibodies were seen more frequently in diabetics than in controls, but this difference was not significant. Prevalence of insulin-dependent diabetes on 1 February 1982 was 1.00 per I000 and 1.05 per 1000 on 1 February 1986. The insulin-dependent diabetes mellitus incidence characteristics noted for the Canterbury Hospital Board area are similar to those reported for European and North American populations.
Introduction Address for correspondence: Dr. R.S. Scott, Department of Medicine, Christchurch Clinical School of Medicine, P.O. Box 4345, Christchurch, New Zealand. * Present address: Department of Zoology, University of Canterbury, Christchurch, New Zealand.
The results of a retrospective study of the incidence of juvenile diabetes mellitus in New Zealand were reported in 1980 [1]. Information was obtained from hospital admission data stored at the Department of Health, National Statistics Centre. Com-
0168-8227/87/$03.50 © 1987 Elsevier SciencePublishers B.V. (BiomedicalDivision)
22 pared with several other studies [2--9], the age-related incidence showed an altered distribution; specifically, there was an absence of childhood peaks and a sustained relatively higher incidence in adolescence. In addition, in contrast to other reports [2,4,5,8,9] no regular seasonal trends in incidence were seen. We have undertaken a prospective study based on a register begun on 1 February 1982 of all new cases of insulin-dependent diabetes mellitus (IDDM) in the age group 0-19 years. The population upon which the register is based corresponds to 10% of the total New Zealand population under 20 years of age. The present investigation covers a 4-year period, until 1 February 1986, and is aimed at answering the following questions: (1) What is the annual incidence of IDDM in the registry area and is it stable? (2) What is the prevalence of IDDM in the 019-year age group in the registry area at the beginning and end of the 4-year study period? (3) Do the sex, age, and seasonal distributions of incidence conform with those previously reported for New Zealand or other countries? (4) To what extent do I D D M and non-insulindependent diabetes (NIDDM) occur in first-degree relatives of subjects diagnosed IDDM over the 4year study period? (5) What percentage of the newly diagnosed diabetic subjects have islet cell cytoplasmic antibodies (ICA) and/or other organ-specific auto-antibodies at diagnosis?
Methods
Subjects A register of new insulin-dependent diabetic subjects in the 0-19-year age group was begun for the Canterbury Hospital Board (CHB) area [10] on 1 February 1982 (Fig. 1). All children aged 0-12 years, living in this area, are admitted to the Christchurch Hospital upon diagnosis of diabetes, whereas adolescents and adults are usually treated on an ambulatory basis from the Christchurch Dia-
"$.
o
,
,oo ~m
Fig. I. Map of the South Island of New Zealand showing the Canterbury Hospital Board (hatched) area. The inset shows a map of New Zealand with latitude and longitudeindicated.
betes Education Centre, in-patient care being organised only for those with significant metabolic derangement. The possibility that a newly diagnosed diabetic subject may not have been identified by these detection methods is very unlikely. The population base is well defined and diabetes treatment is centralized; all individuals presenting with diabetes and requiring insulin are referred to the Paediatric/Adolescent or Adult Diabetes Services in Christchurch; independent annual surveys (see below) identifying insulin users have not revealed any 'missed cases' during the 4 years of the survey. Thus ascertainment was probably 100%. A standardised form was used to record the patient's name, date of birth, sex, address, physician, date of diagnosis, presenting symptoms, presenting biochemical results, previous history of viral infection, family history of diabetes mellitus, date
23 insulin started, and date blood sample taken for measurement of immunological markers. Only insulin-treated subjects were included in the register. A typical history characterised by excessive thirst and frequency of urination was seen in all but two cases. These two patients were included in the register since their attending physician considered insulin treatment was necessary in both cases. One of these individuals had a random plasma glucose level < 11.1 mmol/1, but an oral glucose tolerance test was diagnostic of diabetes and subsequent elevated glucose levels prompted commencement of insulin. Presenting hyperglycaemia (> 11.1 mmol/l) was present in the remaining 98% of cases; ketonuria and weight loss were documented in 75% and 61% of cases respectively. The under 20 year population in the CHB area corresponds to 10% of the total population for this age group in New Zealand. Breakdown of the total population of the area by age and sex was obtained from the 1981 Census. In the 0-19-year age group there was an average annual fall in population of < 2% for the previous 5 years, and population projection figures for the following 5 years suggest a similar trend. Because of the stability of the population, incidences were calculated using the 1981 population figures. The prevalence of IDDM in under 20-year-olds as of 1 February 1982 and at 1 February 1986 was also determined using the 1981 Census figures. All diabetic subjects in this age group who were living in the CHB area on 1 February 1982 and had been
diagnosed prior to this date, were identified using a combination of Hospital (in- and out-patient), Diabetes Education Centre and Dietary Department records. Individuals diagnosed as having IDDM between 1 February 1982 and 1 February 1986 were identified through the prospective study measuring incidence. During the 4-year study period patients were retained in the prevalence register until date of relocation outside the CHB area, death, or reaching 20 years of age. Ascertainment for incidence and prevalence was assessed by annual independent surveys of individuals obtaining insulin from all pharmacies in the CHB area. To further evaluate presenting characteristics of IDDM in new diabetics, serum from these subjects was screened for the presence of immunological markers (see below) reported to be associated with the disorder. In order to directly relate the presence of immunological markers to the risk of developing IDDM, the frequency of these markers in non-diabetic controls was also determined. Age- and sexmatched healthy control subjects were selected from the population at large (Table 1). Anyone with a personal or family history of diabetes mellitus or other endocrine disorders was excluded from the control group.
Assay of immunological markers Serum samples were obtained within one month of diagnosis from 52•54 of the new diabetic subjects. Samples from controls were obtained at the time of
TABLE 1 DETAILS CONCERNING AGE AND SEX OF NEW DIABETIC AND CONTROL SUBJECTS Females
n Mean age (years) Age range (years)
Males
Diabetic subjects
Control subjects
Diabetic subjects
Control subjects
24 10. I 1-19
24 10.1 1-19
30 13.2 2-19
30 13.0 2-19
24
<
z~
I
z~
I=
z~
m
z~ I
© [.<
L~
Z
Z E~
z~
ILl ,-1 <
Z < <
~
~1
e~
~.~ ~v I'--1
I~
'9
-i
25 3O
Gastric parietal cell and adrenal antibodies were also detected by the indirect immunofluorescent technique (described above), using cryostat sections of mouse stomach and human adrenal tissue respectively. A FITC-conjugated anti-human IgG (South Pacific Immunological Laboratories, New Zealand) was used at 1/75 dilution. Thyroid autoantibodies (anti-thyroglobulin and anti-thyroid microsomal) were determined by tanned red cell agglutination using a commercial kit (Fujizoki, Japan).
o O
20
3
o I
I
I
I
I
0
2
4
6
8
I
I
I
I
I
I
10 12 14 16 18 20
AGE (years)
Fig. 2. Risk of developing IDDM by age and sex. ( 0 - - 0 ) males; ( & - - & ) males.
fe-
selection for the study. Sera were stored at - 2 0 " C from one week to 18 months before testing. ICA were detected by indirect immunofluorescence using frozen sections of blood group 0 human pancreas obtained from cadaver kidney donors [11]. All serum samples were tested both undiluted and diluted 1/2 in phosphate-buffered saline (PBS). The end-point titration of positive samples was determined. Fluorescein isothiocyanate (FITC) -conjugated rabbit anti-human IgG (Tago, California) was used at 1/15 dilution. Preparations were read independently using a Leitz Ortholux II microscope by two observers who were not informed of the identity of the samples.
Statistical analysis Statistical significance was assessed using the chisquared test, or in the case of sex-related variations in the age at onset of IDDM, the Mann-Whitney test was used. Seasonal variations were studied by the chi-squared test using totals for each individual month, and totals for 3-month periods (December-February; March-May; June-August; September-November).
Results
Incidence and risk of IDDM During the 4-year study period, 54 individuals under 20 years of age developed IDDM. Table 2 gives numbers of cases and incidences for each year by sex in addition to giving the total and mean values. During the study period there were no significant differences in the annual incidence rates. The risk (cumulative incidence rate; Fig. 2) of developing IDDM in the CHB area by the age of 19 years
TABLE 3 AGE- AND SEX-SPECIFIC PREVALENCE OF IDDM I February 1986
1 February 1982
Females Males Total
Population < 20 years
No. of cases
Prevalence per 1000
No. of cases
Prevalence per 1000
56,569 58,854 115,423
54 62 116
0.95 1.05 1.00
57 64 121
1.01 1.09 1.05
26 S I
50--
7--
40-o o o
6--
°o 3 0 - .~. 5--
Z ___ 20-Z
~D 4 ~
10--
3
O--
Z
rr LU rn 3 - -
I
I
I
I
I
I
I
0
2
4
6
8
10
12
I
I
14 16
I
I
-
\
i oooqJ~%ooooooeo%~%
2--
/
18 20
AGE (years)
Fig. 3. Age-specific incidence of IDDM. O-
I
I
I
I
I
1
I
I
I
I
I
I
J
A
S
0
N
D
J
F
M
A
M
J
was similar for males and females, 2.1 per 1000 and 2.4 per 1000 respectively.
Prevalence The prevalence rates on 1 February 1982 and 1 February 1986 for the 0-19-year age group are shown in Table 3. As can be seen, prevalence has also remained stable over this time period. Sex-, age- and season-related variations in #zcidence Although the mean annual incidence was higher for males than females (Table 2) this was not significant. In fact, during the first 2 years there was an excess of females. Age of onset was significantly (P < 0.001) lower for females than males. Incidence reached a maximum (at 11 years for girls, and 13 years for boys) during the pubertal period (Fig. 3). It then decreased rapidly but rose again to a second smaller peak which occurred at 16 years for both males and females. A small 'childhood' peak was observed at 2-3 years of age. Both sexes showed this trend. Significant seasonal variation of onset of I D D M (Fig. 4) was seen for males (P < 0.025), but not for females or for the combined population when analysed using totals for individual months (11 degrees of freedom). However, if the combined population was examined using four 3-month periods
MONTH
Fig. 4, Seasonal variation of onset of IDDM. (----) males; (O--O) combined data.
(oooo)
females;
(see Methods) the level of significance was improved, with a P value approaching 0.05 (chisquared value 7.8; 3 degrees of freedom).
Occurrence of diabetes mellitus in first-degree relatives 9.4% (5/53) of subjects diagnosed in the 4-year study period had a first-degree relative with either N I D D M or IDDM (Table 4). One subject who is an adopted child was excluded as no family history was available. Immunological market's Islet cell cytoplasmic antibodies. 67.9% (36/53) of subjects diagnosed in the study period (65% of females and 70% of males) were positive for ICA. One subject with a negative result was excluded since the sample was not obtained until 30 months after diagnosis. The mean ages for both ICA-positive females and males (9.6 and 13.2 years respectively) were similar to those for the total I D D M
27 TABLE 4 DISTRIBUTION OF FIRST-DEGREE RELATIVES WITH DIABETES New diabetic subject Sex
Age
F M M M M
11 4 9 14 16
Diabetic relative
Type of diabetes of relative
Mother Brother Father Father Mother Brother
NIDDM IDDM NIDDM IDDM IDDM NIDDM
population (10.4 and 13.2 years respectively). The ICA end-point titres for positive sera ranged from 1/1 (undiluted) to 1/64 (1.56% v/v plasma). In contrast, none of the control subjects were positive for ICA.
Other organ-specific auto-antibodies. Five new diabetic subjects were positive for one of the other four organ-specific antibodies (thyroid microsomah two; thyroglobulin: one; gastric parietal cell: one; adrenal: one). One new diabetic subject, who had thyrotoxicosis in addition to IDDM, was positive for antibodies to both thyroglobulin and thyroid microsomal antigen. Therefore, 11.1% of new diabetics (two female, four male; age range 11-15 years) had at least one organ-specific auto-antibody. Of these, 67% were positive for ICA. Three (5.6%) of the age- and sex-matched control subjects (two female, one male; age range 1316 years) were positive for one auto-antibody (thyroid microsomal: two; adrenal: one). No control subject had more than one auto-antibody. The frequency of occurrence of these four auto-antibodies in control subjects was not significantly less than that observed for diabetic subjects.
Discussion
This is the first report of a prospective study of incidence and of prevalence of IDDM within New
Zealand. Although the register of under 20-year-old diabetic subjects established in the CHB area is based on a population corresponding to only 10% of the total population for this age group in New Zealand, it has the advantage of high ascertainment. Centralised treatment of diabetic subjects at diagnosis in the CHB area and external validation of cases are the primary factors responsible for the completeness of detection. In the Northern Hemisphere at least, the incidence of IDDM shows a trend to increase in countries in more northern regions, and these geographical differences cannot be explained by variability in precision of survey methods used. The average annual incidence (11.7 per 100,000; range 10.413.0) in this New Zealand study is comparable to rates reported from other countries of similar latitude: Denmark [4], ages 0-29, 13.2; Scotland [8], ages 0"19, 13.8; Pittsburgh, U.S.A. [12], ages 0-19, 14.7; Toronto, Canada [13], ages 0-18, 9.0; Montreal, Canada [6], ages 0-16, 8.8; Netherlands [14], ages 0-19, 10.9; U.K. [3], ages 0-15, 7.7. It is also similar to that reported (10.4 per 100,000) for a retrospective study based on the 0-19-year-old population of New Zealand Ill. There was, however, no evidence for a temporal increase in incidence, a trend which has been reported for some populations [8,9,15], but not for others [1,4,6,7]. The risk of developing diabetes is greater for males than for females, but this is not significant. A similar finding has previously been reported [1,3,4,7,8,13] and was statistically significant in at least one population [9]. In the present study this trend is due entirely to a higher incidence for males than females over the age of 13 years. This is in contrast to a number of other studies [4-9] where a higher incidence was found not only for adolescent males, but also for boys in the youngest age group (0-4 years). Dahlquist et al. [9] have suggested the higher risk for males could be either an expression of a higher frequency of genetic risk factors in males or of a higher susceptibility to infections [16,17]. The age-related incidence pattern differed from that of Crossley and Upsdall [1]. A distinct inci-
28 dence peak was seen for both males (13 years) and females (11 years); a finding similar to that noted in other studies [3,4,6,7,9]. In contrast to the sustained higher level of incidence reported by Crossley and Upsdall between ages 11-19 years, we observed a fall in attack rate for both males and females, with a secondary peak for both sexes at 16 years of age. A minor pre-school peak, seen for both males and females, has also been reported by others [3,4,6,8]. Although the major incidence peak is often referred to as the 'pubertal' peak, an association with pubertal status has not been objectively established, and in fact there has been at least one report of an earlier peak for males than females
[8]. There is a seaonal trend, similar to that observed by several other groups [2-6,8,9,19] with higher incidence rates in autumn and winter, and lower rates in spring and summer. In our study this trend was the result of significant seasonal variation for males. A difference between males and females for seasonrelated incidence has also been noted by others [6,8,9,12,20]. Dahlquist et al. [9] have suggested that if seasonal variation is dependent on an infectious disease pattern, as hypothesised [21], then a more pronounced seasonality would in fact be expected for males because of their higher susceptibility to infections [16,17]. It is possible that a delay in seeking medical care during the summer vacation may have contributed to the lower summer incidence rate observed. In an attempt to minimise the influence of this, seasonal variation was studied by the chi-squared test using the total numbers of cases for four 3-month periods. By including February with December and January it was hoped that the effect on incidence of delayed diagnosis during the schools' summer vacation (which finishes at the end of january) would be largely overcome. When examined in this way, the seasonal trend for the combined male plus female population was not diminished. The percentage of recently diagnosed I D D M subjects who had first-degree relatives with I D D M (5.7%) was approximately half that of previous reports [5,6,9]. However, the small number of subjects
in our study may account for this discrepancy. The frequency of occurrence of islet cell antibodies in new diabetics (68%) and in healthy control subjects (0%) is in good agreement with earlier reports [22-24]. Other organ-specific auto-antibodies were detected more frequently in diabetics than control subjects; this difference was smaller than noticed by some researchers [24-26], and was not significant. The results reported here provide no evidence to suggest that aetiological factors operative in I D D M in the CHB area of New Zealand differ from those in Europe and North America. Our data does not confirm the atypical findings of Crossley and Upsdall [1] with respect to age- and sex-related incidence.
Acknowledgements The authors gratefully acknowledge the assistance of Ms. L.J. Brown, Ms. F. McCone, Ms. L. Taylor, Mr. R.P. Weir and the staff of the Department o f Immunopathology, Christchurch Hospital. We thank Ms. J. Nicol for typing the manuscript. This work was supported by grants from the Medical Research Council of New Zealand, the National Children's Health Research Foundation, the New Zealand Diabetes Association, the Canterbury Medical Research Foundation and N o v o Australia (Pty) Ltd. Preliminary results have been presented at the Annual Meeting of the New Zealand Society for the Study of Diabetes, Auckland, October 1984.
References Crossley, J.R. and Upsdall, M., The incidence of juvenile diabetes mellitus in New Zealand, Diabetologia, 18: 29-34, 1980. Gamble, D.R., Epidemiologicaland virologicalobservations on juvenile diabetes, Postgrad. Med. J., 50: 538-543, 1974. Bloom, A., Hayes,T.M. and Gamble, D.R., Registerof newly diagnosed diabetic children, Br. Med. J., 3: 580-583, 1975. Christau, B., Kromann, H., Andersen. O.O. et al., Incidence, seasonal and geographicalpatterns of juvenile-onsetinsulindependent diabetes mellitus in Denmark, Diabetologia. 13: 281-284, 1977.
29 5 Sterky, G., Holmgren, G., Gustavson, K.H. et al., The incidence of diabetes mellitus in Swedish children 1970-1975, Acta Paediatr. Scand., 67: 139-143, 1978. 6 West, R., Belmonte, M.M., Colle, E., Crepeau, M.P., Wilkins, J. and Poirier, R., Epidemiologic survey of juvenileonset diabetes in Montreal, Diabetes. 28: 690-693, 1979. 7 La Porte, R.E., Fishbein, H.A., Drash, A.L. et al., The Pittsburgh insulin-dependent diabetes mellitus (IDDM) registry. The incidence of insulin-dependent diabetes mellittls in Allegheny County, Pennsylvania (1965-1976), Diabetes, 30: 279-284, 1981. 8 Patterson, C.C., Thorogood, M., Smith, P.G., Heasman, M.A., Clarke, J.A. and Mann, J.L., Epidemiology of type I (insulin-dependent) diabetes in Scotland 1968-1976: evidence of an increasing incidence, Diabetologia, 24: 238-243, 1983. 9 Dahlquist, G., Blom, L., Holmgren, G. et al., The epidemiology of diabetes in Swedish children 0-14 years - - a sixyear prospective study, Diabetologia. 28: 802-808, 1985. 10 Scott, R.S., Brown, L.J. and Clifford, P.C., Use of health services by diabetic persons. II. Hospital admissions, Diabetes Care, 8: 43-47, 1985. 11 Marner, B., Lernmark, A., Nerup, J., Molennar, J.L., Tuk, C.W. and Bruining, G.J., Analysis of islet cell antibodies on frozen sections of human pancreas, Diabetologia, 25: 93-96, 1983. 12 Tajima, N., La Porte, R.E., Hibi, I., Kitagawa, T., Fujita, H. and Drash, A.L., A comparison of the epidemiology of youth-onset insulin-dependent diabetes mellitus between Japan and the United States (Allegheny Country, Pennsylvania), Diabetes Care, 8: 17-23, 1985. 13 Ehrlich, R.M., Walsh, L.J., Falk, J.A., Middleton, P.J. and Simpson, N.E., The incidence of type 1 (insulin-dependent) diabetes in Toronto, Diabetologia. 22: 289-291, 1982. 14 Vaandrager, G.J., Bruining, G.J., VeenhoR F.J. and Drayer, N.M., Incidence of childhood diabetes in the Netherlands: a decrease from north to south over North-western Europe?, Diabetologia. 27: 203-206, 1984. 15 Green, A. and Andersen, P.K., Epidemiological studies of diabetes mellitus in Denmark: 3. Clinical characteristics and
16
17 18
19
20
21 22
23
24
25
26
incidence of diabetes among males aged 0 to 19 years, Diabetologia. 25: 226-230, 1983. Washburn, T.C., Medearis, D.N. and Childs, B., Sex difference in susceptibility to infections, Pediatrics, 35: 57-64, 1965. Michaels, R.H. and Rogers, K.D., A sex difference in immunological responsiveness, Pediatrics. 47: 120-123, 1971. MacMillan, D.R., Kotoyan, M., Zeidner, D. and Hafezi, B., Seasonal variation in the onset of diabetes in children, Pediatrics, 59:113-115, 1977. Fleegler, F.M., Rogers, K.D., Drash, A., Rosenbloom, A.L., Travis, UB. and Court, J.M., Age, sex and season of onset of juvenile diabetes in different geographic areas, Pediatrics, 63: 374--379, 1979. Fishbein, H.A., La Porte, R.E., Orchard, T.A., Drash, A.L., Kuller, L.H. and Wagener, D.K., The Pittsburgh insulin-dependent diabetes mellitus registry. Seasonal incidence, Diabetologia. 23: 83-85, 1982. Gamble, D.R. and Taylor, K.W., Seasonal incidence of diabetes mellitus, Br. Med. J., 3: 613-633, 1969. lrvine, W.J., McCallum, C.J., Gray, R.S. et al., Pancreatic islet-cell antibodies in diabetes mellitus correlated with the duration and type of diabetes, co-existent autoimmune disease, and HLA type, Diabetes. 26: 138-147, 1977. Bottazzo, G.F., Dean, B.M., Gorsuch, A.N., Cudworth, A.G. and Doniach. D., Complement fixing islet-cell antibodies in type I diabetes: possible monitors of active betacell damage, Lancet, i: 668-672, 1980. Neufeld, M., MacLaren, N.K., Riley, W.J. et al., Islet cell and other organ-specific antibodies in U.S. caucasians and blacks with insulin-dependent diabetes mellitus, Diabetes. 29: 589-592, 1980. Riley, W.J., Winer, A. and Goldstein, D., Coincident presence of thyro-gastric autoimmunity at onset of Type I (insulin-dependent) diabetes, Diabetologia. 24: 418-421, 1983. Betterle, C., Zanette, F., Pedini, B. et al., Clinical and subclinical organ-specific autoimmune manifestations in Type I (insulin-dependent) diabetic patients and their first-degree relatives, Diabetologia. 26: 431-436, 1984.