Original Article. 386 ª 2008 Blackwell Publishing Ltd

Transcription

1 Original Article doi: /j x Gender differences and temporal variation in the incidence of type 1 diabetes: results of 8012 cases in the nationwide Diabetes Incidence Study in Sweden J. Östman 1,G.Lönnberg 2, H. J. Arnqvist 3, G. Blohmé 4, J. Bolinder 1, A. Ekbom Schnell 5, J. W. Eriksson 6, S. Gudbjörnsdottir 7, G. Sundkvist 8 & L. Nyström 2 From the 1 Department of Endocrinology, Metabolism and Diabetes, Karolinska University Hospital, Huddinge, Stockholm; 2 Department of Public Health and Clinical Medicine, Epidemiology and Public Health Sciences, Umeå University, Umeå; 3 Department of Endocrinology, University Hospital, Linköping; 4 Department of Internal Medicine, Södersjukhuset, Stockholm; 5 Department of Medical Sciences, University Hospital, Uppsala; 6 Department of Medicine, University Hospital, Umeå; 7 Diabetes Centre, Sahlgrenska University Hospital, Göteborg; and 8 Department of Endocrinology, Malmö University Hospital, Lund University, Malmö, Sweden Abstract. Östman J, Lönnberg G, Arnqvist HJ, Blohmé G, Bolinder J, Ekbom Schnell A, Eriksson JW, Gudbjörnsdottir S, Sundkvist G, Nyström L (Karolinska University Hospital, Huddinge, Stockholm; Umea University, Umeå; University Hospital, Linköping; Södersjukhuset, Stockholm; University Hospital, Uppsala; University Hospital, Umeå; Sahlgrenska University Hospital, Göteborg; and Lund University Malmo University Hospital, Malmö; Sweden). Gender differences and temporal variation in the incidence of type 1 diabetes: results of 8012 cases in the nationwide Diabetes Incidence Study in Sweden J Intern Med 2008; 263: Objectives. To establish the gender difference amongst newly diagnosed type 1 diabetic patients aged years, considering age at diagnosis, temporal trend and seasonal variation at time of diagnosis. Study design. A population-based prospective study with a mean annual population at risk of 2.3 million. Setting. All departments of medicine, endocrinology and paediatrics and primary health care units in Sweden. Subjects. Incident cases of diabetes aged years at diagnosis Measure instrument. Basic characteristics of patients at diagnosis were reported by the diagnosing doctor on a standardized form. Level of ascertainment was estimated at 80 90%. Results. Amongst all incident cases (n = 8012), 74% was diagnosed with type 1 diabetes. The mean annual incidence rate of type 1 diabetes was , in men and in women The incidence of type 1 diabetes decreased slowly by increasing age but was in all age groups higher in men, yielding an overall male female ratio of 1.8. In both genders the incidence of type 1 diabetes decreased in average of 1.0% per year. A seasonal pattern with significantly higher incidence during January March and lower during May July was seen in both genders. Conclusions. A clear male predominance of type 1 diabetes was seen in all ages. The temporal trend and the seasonal pattern was similar in men and women. Hence, internal factors related to the gender rather than differences in the exposure to environmental factors seem to explain the consistent male female bias in the postpubertal risk of developing type 1 diabetes. Keywords: diabetes, gender, incidence, register, temporal, young adults. 386 ª 2008 Blackwell Publishing Ltd

2 Introduction Type 1 diabetes is a chronic disease that still leads to long-term complications despite improvements in treatment according to multicentre clinical trials [1, 2] as well as population-based observational studies [3, 4]. To prevent the development of type 1 diabetes it is necessary to clarify its aetiology and pathogenesis. A considerable number of studies in children have been reported [5]; however, since type 1 diabetes may frequently appear amongst adults [6] we started in 1983 a nationwide prospective registration of cases with newly diagnosed diabetes in the 15- to 34-year age groups, the Diabetes Incidence Study in Sweden [7, 8]. In this report, we focus on the findings during the first 20 years with regard to age at diagnosis, gender, temporal trend and seasonal variation. Patients and methods Study design This is a prospective study of all incident cases of diabetes aged years in Sweden diagnosed The yearly mean population at risk is has been slightly modified. During the first period ( ) information about the circumstances of detection (presence of clinical symptoms or opportunistically health check-up), and maximum blood glucose level at diagnosis were obtained. Since 1992, information regarding ketoacidosis, ketonuria, duration of certain symptoms (weight loss, fatigue, polyuria, polydipsia and coma) and maintenance treatment (insulin, oral antidiabetics or diet alone) are registered. Diagnostic criteria During the first period diabetes was diagnosed according to the criteria recommended by the World Health Organization (WHO) [9, 10], as described in detail previously [7]. Since 1992 we have used the diagnostic criteria recommended by the American Diabetes Associations expert group [11] and which in this clinical setting are identical with those currently provided by WHO [12]. This change in the diagnostic criteria has most likely not affected the incidence trend because in patients with classical symptoms of diabetes, the fasting blood glucose levels only exceptionally is in the range of 7.0 mmol L )1 [7]. Data generation process Since 1 January 1983, subjects with newly diagnosed diabetes in the 15- to 34-year age groups in the Diabetes Incidence Study in Sweden (DISS) are registered on a standardized form. The data entered are the patients civic number, name and address, date of diagnosis, basis for diagnosis (fasting and nonfasting blood or plasma glucose, oral glucose tolerance test), height and weight, name and address of the reporting department and doctor and the reporting doctor s clinical classification of the type of diabetes (type 1, type 2, unclassified or secondary). This study is based on all cases with type 1, type 2 and unclassified diabetes according to the reported doctor s clinical classification. According to clinical practice, severe hyperglycaemia, ketosis, low or normal body weight and short duration of symptoms often are considered to indicate type 1 diabetes. During the 20 years, the standardized form Data collection The investigation is conducted on a nationwide basis in collaboration with a contact person (a reporting doctor and or a diabetes nurse) at all, currently 120, departments of internal medicine, endocrinology and paediatrics in Sweden. The reporting doctors receive annually a list of patients registered since the start of the study and a newsletter about its progress. A website ( provides information of the study. The study was approved by the Ethics Committee of the Karolinska Institute in Stockholm. Permission to use the Swedish national civic registration number was granted by the Swedish Data Inspection Board. Level of ascertainment By use of a computer-based patient administrative register as a second source, the level of ascertainment ª 2008 Blackwell Publishing Ltd Journal of Internal Medicine 263;

3 during was estimated at 78% in women and at 79% in men in the two southernmost counties, covering 9.2% of the population at risk [13]. For type 1 diabetes the level was 86%. A similar study in the county of Västerbotten in the northern Sweden, covering 2.9% of the population at risk, found no trend in the level of ascertainment during The level of ascertainment for type 1 diabetes was 91%. Finally, the median level of ascertainment for type 1 diabetes for six hospitals using the software diabase for their patients was 82%. The time period for the ascertainment studies varied depending on when diabase was introduced at the hospital and when the study was carried out, thus the later performed ascertainment studies covered almost the whole study period. Altogether, the ascertainment rate did not change during the study period. Level of ascertainment was assessed using the two-sample capture recapture method [14]. Statistical analysis All data were collected and analysed at the Department of Public Health and Clinical Medicine, Umea University. Population data from Statistics, Sweden were used to calculate the annual gender-specific incidence rate per and 95% confidence intervals (CI) using Fisher s formula. The linear regression after log-transformation was used to analyse the change in the incidence of diabetes over time. Multiple pairwise comparisons (Tukey procedure) were used to analyse the significance of the fluctuations during the period. Seasonal variation was analysed using several methods. The departure from a uniform occurrence throughout the year was analysed using Freedman s test [15]. Edwards s test [16] was used to identify a single annual peak and a single trough with 6 months between the two. Further Ratchet s circular scan test [17], which is based on the maximum number of events in two or three consecutive months, was used to identify peaks of 2 3 months length. Finally, Hewitt s rank sum test [18] was used to identify peaks of 4 6 months length. Results During the 20-year period , 8012 subjects in the 15- to 34-year age group were diagnosed with diabetes (type 1, type 2 and unclassified, secondary diabetes excluded), yielding a mean annual incidence rate of (95% CI: ; Table 1), in men and in women. The majority of subjects were classified as type 1 diabetes, 77% amongst men and 68% amongst women. Thus, the mean annual incidence in type 1 diabetes was for men and for women. Ketoacidos was present at time of diagnosis in 9.7% amongst men and in 10.3% amongst women. There was no change over the recorded period ( ). The annual incidence of type 1 diabetes showed marked fluctuations (Fig. 1). Multiple pairwise comparisons revealed that the incidence was significantly higher during 1983 and 1984 ( ) than during 1996 ( ) (P < 0.01) and 1997 ( ) (P < 0.05). These findings were restricted to men and more prominent in the older age groups. Linear regression analysis using log-transformed data revealed a relative drop in incidence rate, in men with 0.95% (95% CI: )1.58% to )0.32%) Table 1 Mean annual incidence rate per and 95% confidence intervals (CI) of diabetes mellitus by gender and type (mean annual population at risk: women and men) Gender Type of diabetes Number of cases Per cent Annual incidence 95% CI Women Unclassified Total Men Unclassified Total All Unclassified Total ª 2008 Blackwell Publishing Ltd Journal of Internal Medicine 263;

4 Fig. 1 Temporal trend of the incidence of type 1, type 2 and unclassified diabetes in Sweden in the age group years and in women with 0.87% (95% CI: )1.56% to )0.17%). Table 2 shows the incidence of type 1 diabetes by gender, 5-year age groups and 5-year periods of diagnosis. Compared with the first 5-year period the last one showed a 10% lower age-adjusted incidence, similarly amongst women and men. The mean annual incidence rate per of subjects with type 1 diabetes decreased markedly by increasing age, in males from 19.2 to 13.0, and in females from 10.4 to 6.5 in the 15- to 19- and 30- to 34-year age group, respectively. The male predominance in subjects with Incidence/ Type 1 Type 2 Unclassified Year type 1 diabetes increased by age, from 1.8 to 2.0, whilst the overall male-to-female ratio being 1.8, and only a slight male preponderance was noticed amongst subjects with type 2 diabetes. A seasonal pattern of the diagnosis of type 1 diabetes was demonstrated in Fig. 2. The most prominent finding was a 3-month peak of incidence during January March in both genders (P < 0.005). Amongst men this finding was consistent in the three youngest 5-year age groups but not in the years old, in whom the peak was from November to January (P < 0.05). Amongst women the 3-month peak during Table 2 Number of cases with type 1 diabetes and mean annual incidence per by 5-year age groups in the years group in Sweden during four 5-year periods Mean annual incidence by 5-year periods Gender Age group Number of cases Mean annual incidence 95% CI Women Men ª 2008 Blackwell Publishing Ltd Journal of Internal Medicine 263;

5 Type 1 Men Type 1 Women Type 2 Men Type 2 Women No of cases/month Month of diagnosis Fig. 2 Seasonal variation in clinical onset of type 1 and type 2 diabetes by gender in Sweden in the age group years January March was significant (P < 0.05) only in the 25- to 29-year age group. The lowest incidence was found during May July, both in men, particularly in years old, and in women (P < 0.005). The incidence of type 2 diabetes showed a similar seasonal pattern, but only in men, with a 3-month peak in November January (P < 0.005) and a nadir in May July (P < 0.025). Discussion This 20-year prospective, nationwide study from Sweden showed that type 1 diabetes amongst young adults has a marked male predominance with an overall 1.8 male-to-female ratio. Independently of gender, the incidence of type 1 diabetes has decreased during the observation period. The male excess was independent of a clear seasonal variation in the incidence, with a peak in January March and a nadir in May July. Gender The current study establishes the marked male predominance of type 1 diabetes in young adults (15 34 years) previously observed in Sweden [7, 8]. This has been confirmed in another large registry [19] and also extensively reviewed [20]. A statistical analysis of the sex ratio in type 1 diabetic children undertaken for 76 populations showed in the 10- to 14-year age group a male excess more often in populations with high incidence (> ) and in the European countries [21]. A similar analysis of studies performed in young adults, years, shows diverging results. Amongst 21 populations with high or intermediate incidence ( ), a M F ratio between 1.2 and 2.1, was reported in Sweden [7, 8], Belgium [19], Croatia [22], Denmark [23], Norway [24], New Zealand [25], Estonia [26], Catalonia [27], Sardinia [28], Pavia [29] and Bucharest, Leicestershire, Lithuania and West Yorkshire [30]. The only exceptions were Antwerp [30] and the Maltese Islands [31], in which regions the M F ratio was about 1, however, the number of cases was small, around 30. A male predominance has been observed also amongst young adults, in Allegheny (only whites) [32], the Netherlands [33], Scotland [34] and Lombardy [35], but not in Oxford [36]. In populations with low incidence (< ), there was also a significant male excess in Slovakia [30] and Turin [37], but not in Poland [38]. Thus, a male predominance has been demonstrated in the majority of 24 populations and a female excess in none of them. Furthermore, the male excess was not related to the level of the incidence. The M F ratio amongst all antibody-negative subjects was elevated, 1.4, thus confirming the findings in the Belgian Registry [39]. 390 ª 2008 Blackwell Publishing Ltd Journal of Internal Medicine 263;

6 In a sample of Caucasian cases from the USA, UK and Sardinia the male predominance of type 1 diabetes was largely restricted to DR3 X [40]. We are currently evaluating this question by relating HLA antigens to gender and clinical onset in the to 2002-year DISS cohort. Besides a genetic basis, the gender bias may be attributable to a difference in the immune response, because this is regulated by sexual hormones [41]. Therefore, it is of interest to note that the gender difference starts immediately after puberty. Type 1 diabetes incidence in girls peaks at age of 12 and in boys at age of 14 [42]. Studies on potential differences between genders regarding autoimmune mechanisms might be of interest in this context. Of interest is also the twofold higher prevalence of autoantibodies against insulin (IAA) in men than in women with newly diagnosed type 1 diabetes between 15 and 21 years of age, but with no difference between the genders in childhood [43]. Temporal trend A declining incidence of type 1 diabetes was observed in both genders, thus confirming and extending the data reported for , based on the two nationwide registries, the Swedish Childhood Diabetes Registry and DISS [42]. The previous study displayed a shift to onset at younger age for the 7- to 14-year-old children born during the 1980s and for the 0- to 6-year-old children born during the 1990s. In the 15- to 34-year age group the incidence rate during the period decreased significantly in men, but not in women. An earlier presentation of type 1 diabetes has been reported also in the Belgian Diabetes Registry, but in contrast to the findings in Sweden the incidence did not change in the years old [44]. In the low incidence province of Turin in Italy, however, the incidence increased from 1984 to 1996 not only in children, but also in young adults (15 29 years) [45]. Recently, it has been reported from the UK that the incidence of type 1 diabetes in children has increased steadily from 1978 to 2000, whereas it was stable in young adults during [46]. Whereas weight gain in early infancy is associated with increased risk of developing type 1 diabetes in childhood [4 50] a similar relationship has hitherto not been found in young adults. Actually, in DISS the incidence of type 1 diabetes declined during although the body mass index at time of diagnosis increased during this period in both genders [51]. Furthermore, the prevalence of obesity in the 25- to 44-year age group in Sweden has increased considerably during the period , in males from 4% to 7% and in females from 2% to 6% [52]. Hence, in contrast to some other countries, there is in Sweden no evidence for obesity as a major risk factor for the development of type 1 diabetes in young adults. Since the prevalence of ICA-positive subjects with clinical diagnosis of type 2 diabetes and unclassified diabetes was higher, 35%, in DISS 1987 [53] when compared with in 1998, 22%, [54], the declining incidence of type 1 diabetes in young adulthood most likely is not attributed to a misclassification of diabetes type. A consistent finding in DISS has been positive autoantibodies in 20 25% of subjects with the clinical diagnosis of type 2 and in about 50% of those with unclassifiable diabetes. By use of these figures, it can be calculated that the mean annual incidence rate of type 1 diabetes, year, in our study was underestimated with approximately The reason for sudden increments in the incidence of type 1 diabetes during certain years has not been analysed. A variation in the level of ascertainment cannot be completely ruled out, however, similar temporal, epidemic-like, fluctuations have been observed in a number of long-term studies in childhood diabetes with high detection rate [55 57, for survey see Ref. 58]. Seasonal variation of diagnosis A seasonal pattern in type 1 diabetes incidence was already reported about 80 years ago, when higher rates of acute diabetes were found during the late autumn, winter and early spring [59]. This has been repeatedly confirmed in children [58]. However, no seasonal pattern was observed amongst ª 2008 Blackwell Publishing Ltd Journal of Internal Medicine 263;

7 young adult cases in Zagreb [22], Chile [60], Edinburgh [61], Wisconsin [62], except in men below the age of 19 years, which had the highest onset during the winter months [60 62]. In Denmark, the incidence of type 1 diabetes in the 15- to 29-year age group was reduced during the summer months [23] and no peak was seen. In Norway, the highest number of patients (15 29 years) was reported in January and September and the lowest in July [24]. Both a significant peak during the cold months (January March) and a significant nadir during the summer has previously not been documented, which most conceivable is explained by the low number of incident cases in the majority of studies [19]. The seasonality in men was largely restricted to the lack of the high-risk genotype HLA-DQ2 DQ8. In childhood diabetes, the high incidence during the winter has been taken as an indirect evidence for the role of environmental exposure, such as virus infections, and other triggering or modulating factors [58, 63 65]. The variation of sunshine hours and temperature are other putative factors of impact [63]. In conclusion, in the present 20-year study with more than 5800 patients (15 34 years) registered as type 1 diabetes, the male predominance was strong in all age groups, giving an overall male-to-female ratio of 1.8. A slight, although significant, decrease in the incidence of type 1 diabetes was seen in both genders. We also noted a seasonal variation, with the highest incidence during January March and the lowest during May July, which was similar in the two genders, even if more prominent in men. To summarize, our findings concerning the temporal trend and seasonal variation of type 1 diabetes suggest that the male excess, which consistently is observed postpubertal, is not related to the exposure to various environmental factors, but rather to endogenous factors, which include, for instance the sensitivity to virus infections. Conflict of interest statement No conflict of interest was declared. Acknowledgements We thank all diabetologists and diabetes nurses for their invaluable efforts to report newly diagnosed subjects with diabetes. DISS has been supported by grants from the Juvenile Diabetes Foundation-Wallenberg Diabetes Research Program (K JD ), the Swedish Diabetes Association, and the Swedish Medical Research Council (7X-14531). References 1 Reichard P, Nilsson BY, Rosenqvist U. The effect of long-term intensified insulin treatment on the development of microvascular complications in insulin-dependent diabetes mellitus. N Engl J Med 1993; 320: The Diabetes Control and Complications Trial Research Group. The effect of intensive treatment on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med 1993; 329: Henricsson M, Nyström L, Blohmé G et al. The incidence of retinopathy 10 years after diagnosis in young adult people with diabetes. Diabetes Care 2003; 26: Svensson M, Sundkvist G, Arnqvist HJ et al. Signs of nephropathy may occur early in young adults with diabetes despite modern diabetes management. Diabetes Care 2003; 26: Onkamo P, Väänänen S, Karvonen M, Tuomiletho J. Worldwide increase in the incidence of Type 1 diabetes the analysis of the data on published incidence trends. Diabetologia 1999; 42: Molbak AG, Christau B, Marner B, Borch-Johnsen K, Nerup J. Incidence of insulin-dependent diabetes mellitus in age-groups over 30 years in Denmark. Diabet Med 1994; 11: Östman J, Arnqvist H, Blohmé G et al. Epidemiology of diabetes mellitus in Sweden. Results of the first year of a prospective study in the population age group years. Acta Med Scand 1986; 220: Blohmé G, Nyström L, Arnqvist HJ et al. Male predominance of Type 1 (insulin-dependent) diabetes mellitus in young adults: results from a 5 year prospective nationwide study of the year age group in Sweden. Diabetologia 1992; 35: World Health Organization Expert Committee. Second Report on Diabetes Mellitus. Technical Report, Series No Geneva, Switzerland: World Health Organization Expert Committee, World Health Organization Expert Committee. Diabetes mellitus. Technical Report, Series No. 742, Geneva, Switzerland: World Health Organization Expert Committee, The Expert Committee on the Diagnosis and the Classification of Diabetes Mellitus. Report of the Expert Committee on the Diagnosis and the Classification of Diabetes Mellitus. Diabetes Care 1997; 20: World Health Organization. Department of Non-communicable Disease Surveillance. Definition, Diagnosis and Classification 392 ª 2008 Blackwell Publishing Ltd Journal of Internal Medicine 263;

8 of Diabetes Mellitus and its Complications: Part 1. Diagnosis and Classification of Diabetes Mellitus. Report of a WHO Consultation. Geneva: World Health Organization, Littorin B, Sundkvist G, Scherstén B et al. Patient administrative system as a tool to validate the ascertainment in the diabetes incidence study in Sweden (DISS). Diabetes Res Clin Pract 1996; 33: Bishop YMM, Fienberg SE, Holland PW. Discrete Multivariate Analysis: Theory and Practice. Cambridge, MA: The MIT Press 1974; Freedman LS. The use of a Kolmogorov-Smirnov type statistic in testing hypothesis about seasonal variation. J Epidemiol Community Health 1979; 33: Edwards JH. The recognition of cyclic trends in incidence data. Ann Hum Genet 1961; 25: Wallenstein S, Weinberg CR, Gould M. Testing of a pulse in seasonal event data. Biometrics 1989; 45: Rogerson PA. A generalization of Hewitt s test for seasonality. Int J Epidemiol 1996; 25: Weets I, Kaufmann L, Van der Auwera B et al. Seasonality in clinical onset of Type 1 diabetes in Belgian patients above the age of 10 is restricted to HLA-DQ2 DQ8-negative males, which explains the male to female excess in incidence. Diabetologia 2004; 47: Dahlquist G, Mustonen L, for the Swedish Childhood Diabetes Study Group. Analysis of 20 years of prospective registration of childhood onset diabetes time trends and birth cohort effects. Acta Paediatr 2000; 89: Levý-Marchal C, Patterson C, Green A, on behalf of the EURODIAB ACE Study Group. Variation by age group and seasonality at diagnosis of childhood IDDM in Europe. Diabetologia 1995; 38: Roglic G, Pavlic-Renar I, Sestan-Crnek S et al. Incidence of IDDM during in Zagreb, Croatia. Diabetologia 1995; 38: Christau B, Kromann H, Andersen OO et al. Incidence, seasonality and geographical patterns of juvenile-onset insulin-dependent diabetes mellitus in Denmark. Diabetologia 1977; 13: Joner G, Sövik O. The incidence of type 1 (insulin-dependent) diabetes mellitus in years in Norway Diabetologia 1991; 34: Croesly JR, Upsdell M. The incidence of juvenile diabetes mellitus in New Zealand. Diabetologia 1980; 18: Kalits I, Podar T. Incidence and prevalence of Type 1 [insulindependent] diabetes in Estonia in Diabetologia 1990; 33: Goday A, Castell C, Tresserras R et al. Incidence of Type 1 (insulin-dependent) diabetes mellitus in Catalonia, Spain. Diabetologia 1992; 35: Muntoni S, Songini M. High incidence rate of IDDM in Sardinia. Sardinan Collaboration Group for Epidemiology of IDDM. Diabetes Care 1992; 5: Tenconi MT, Devoti G, Albani I et al. IDDM in the province of Pavia, Italy, from a population-based registry. Diabetes Care 1995; 18: Kyvik KO, Nyström L, Gorus F et al. The epidemiology of Type 1 diabetes mellitus is not the same in young adults as in children. Diabetologia 2004; 47: Schranz AG, Prikatsky V. Type 1 diabetes in the Maltese Islands. Diabet Med 1989; 6: LaPorte RE, Fishbein HA, Drash AL et al. The Pittsburgh Insulin-dependent Diabetes Mellitus (IDDM) Registry: the incidence of insulin-dependent diabetes mellitus in Allegheny county, Pennsylvania ( ). Diabetes 1981; 30: Vaandrager GJ, Bruining GJ, Veenhof FJ, Drayer N. Incidence of childhood diabetes in The Netherlands a decrease from north to south over North-Western Europe? Diabetologia 1984; 27: Patterson CC, Smith PG, Webb J, Heasman MA, Mann JI. Geographical variation in the incidence of diabetes mellitus in Scottish children during the period Diabet Med 1988; 5: Malori G, Gallus G, Garancini P, Repetto F, Micossi P. Identification of the cohort of Type 1 diabetes presenting in Lombardy in : a validated assessment. Diabet Med 1990; 7: Bingley PJ, Gale EAM. Incidence of insulin-dependent diabetes in England: a study in the Oxford region, BMJ 1989; 298: Bruno G, Merletti F, Vuolo A, Pisu E, Giorio M, Pagano G. Sex differences in incidence of IDDM in age-group yr. Diabetes Care 1993; 16: Grzywa MA, Sobel AM. Incidence of IDDM in the province of Rzeszów, Poland, 0- to 29-year-old age-group, Diabetes Care 1995; 18: Weets I, Van Autreve J, Van der Auwera BJ et al. Male-tofemale excess in diabetes diagnosed in early adulthood is not specific for the immune-mediated forms nor it is HLA-DQ restricted: possible relation to increased body mass index. Diabetologia 2001; 44: Cucca F, Goy JE, Kawaguchi Y et al. A male-female bias in type 1 diabetes and linkage to chromosome Xp in MHC-DR3- positive patients. Nat Genet 1998; 19: Whitacre CC, Reingold SC, O Looney PA, the Task Force on Gender. Multiple sclerosis and autoimmunity. Science 1999; 283: Pundziuti-Lyckå A, Dahlquist G, Nyström L et al. The incidence of Type 1 diabetes has not increased but shifted to a younger age at diagnosis in the 0-34 years group in Sweden 1983 to Diabetologia 2002; 45: Williams AJ, Norcoss AJ, Dix RJ, Gillespies KM, Gale EA, Bingley PJ. The prevalence of insulin auto antibodies at the onset of type 1 diabetes is higher in males than females during adolescence. Diabetologia 2003; 46: Weets I, De Leeuw IH, Du Caju MVL et al. The Belgian Diabetes Register: the incidence of type 1 diabetes in the age group 0-39 years has not increased in Antwerp (Belgium) between 1989 and Evidence for earlier disease manifestation. Diabetes Care 2002; 25: Bruno G, Merletti F, Biggeri A et al. Increasing trend of type 1 diabetes in the province of Turin [Italy]. Analysis of age, period ª 2008 Blackwell Publishing Ltd Journal of Internal Medicine 263;

9 and birth cohort effects from 1984 to Diabetologia 2001; 44: Feltbower RG, McKinney PA, Parslow RC, Stephenson CR, Bodansky HJ. Type 1 diabetes in Yorkshire, UK: time trends i 0-14 and year-olds, age at onset and age-period-cohort modelling. Diabet Med 2003; 20: Johansson C, Samuelsson U, Ludvigsson J. A high weight gain early in life is associated with an increased risk of type 1 (insulin-dependent)] diabetes mellitus. Diabetologia 1994; 37: Hyppönen E, Kenward MG, Virtanen SM et al. Infant feeding, early weight gain, and risk of type 1 diabetes. Diabetes Care 1999; 22: Hyppönen E, Virtanen SM, Kenward MG, Knip M, Åkerblom HK, the Childhood Diabetes in Finland Study Group. Obesity, increased linear growth, and risk of Type 1 diabetes in children. Diabetes Care 2000; 23: EURODIAB Sub Study 2 Study Group. Rapid early growth is associated with increased risk of childhood type 1 diabetes in various European populations. Diabetes Care 2002; 25: Littorin B, Nyström L, Gullberg B et al. Increasing body mass index at diagnosis in young adult people during in the Diabetes Incidence Study in Sweden (DISS). J Intern Med 2003; 254: Östman J, Britton M, Jonsson E. Treating and Preventing Obesity. KgaA,Weinheim: Wiley-VCH Verlag Gmbh & Co, 2004; Landin-Olsson M, Karlsson A, Lernmark Å, Sundkvist G, the Diabetes Incidence Study in Sweden Group. Islet cell and thyrogastric antibodies in 633 consecutive 15- to 34-yr-old patients in the Diabetes Incidence Study in Sweden. Diabetes 1992; 41: Borg H, Arnqvist HJ, Bjork E et al. Evaluation of the new ADA and WHO criteria for classification of diabetes mellitus in young adult people (15-34 yrs) in the Diabetes Incidence Study in Sweden (DISS). Diabetologia 2003; 46: Rewers M, Stone RA, LaPorte RE et al. Poisson regression modelling of temporal variation in incidence of childhood insulin-independent diabetes mellitus in Allgheny county, Pennsylvania, and Wiekopolska, Poland, Am J Epidemiol 1989; 129: Doheel TM, for the Pittsburgh Diabetes Epidemiology Research Group. An epidemic of childhood diabetes in the United States? Evidence from Allegheny County, Pennsylvania Diabetes Care 1993; 16: Libman IM, LaPorte RE, Becker D, Dorman JS, Drash AL, Kuller L. Was there any epidemic of diabetes in non-white adolescents in Allegheny County, Pennsylvania? Diabetes Care 1998; 21: Karvonen M, Sekikawa A, LaPorte R, Tuomiletho J, Tuomiletho-Wolf E. Type 1 diabetes: global epidemiology. In: Ekoe J-M, Zimmet P, Williams R, eds. The Epidemiology of Diabetes Mellitus. Chichester: J Wiley & Sons, Ltd, 2002; Adams SF. The seasonal variation in the onset of acute diabetes. Arch Intern Med 1926; 37: Durruty P, Ruiz F, Garcia de los Rios M. Age at diagnosis and seasonal variation in the onset of insulin-dependent diabetes in Chile (Southern Hemisphere). Diabetologia 1979; 17: Gray RS, Duncan LJP, Clarke BF. Seasonal onset of insulindependent diabetes in relation to sex and age at onset. Diabetologia 1979; 17: Allen C, Palta M, DÁlesio DJD. Incidence and differences in urban-rural seasonal variation of Type 1 (insulin-dependent) diabetes in Wisconsin. Diabetologia 1986; 29: Dahlquist G, Mustonen L. Childhood diabetes time trends and climatological factors. Int J Epidemiol 1994; 23: Gale EAM, Gillespie KM. Diabetes and gender. Diabetologia 2001; 44: Gale EAM. Spring harvest? Reflections on the rise of type 1 diabetes. Diabetologia 2005; 48: Correspondence: Jan Östman, Department of Endocrinology, Metabolism and Diabetes, Karolinska University Hospital, SE Huddinge, Sweden. ( jan.j.c.oestman@solna.mail.telia.com). 394 ª 2008 Blackwell Publishing Ltd Journal of Internal Medicine 263;