Incidence of childhood type 1 diabetes mellitus in Yamanashi Prefecture, Japan, 1986‐2018

Abstract Introduction Several studies have examined the incidence of childhood T1DM in Japan from the 1970s onwards, but none have been long‐term studies using registration data. We estimate the incidence of childhood type 1 diabetes mellitus (T1DM) from 1986 to 2018 in Yamanashi Prefecture, Japan. Methods We began a population‐based, long‐term study of childhood T1DM in 1986 involving every hospital paediatrics department in Yamanashi Prefecture. In the Prefecture, every child newly diagnosed with T1DM is referred to a hospital, and therefore, almost 100% of new patients aged <15 years are registered. We calculated the incidence of T1DM among children aged <15 years from 1986 to 2018. All cases met the Japan Diabetes Society diagnostic criteria and were tested for T1DM‐related autoantibodies whenever possible. Results Ninety‐nine patients (44 boys and 55 girls) were newly diagnosed with T1DM. The annual incidence among 5‐ to 9‐year‐olds increased by 5.35% over the study period (95% confidence interval 2.34%‐8.35%, p = .0005), and there was a trend towards increasing 3‐year incidence (15.52% increase, p = .0516). There were also trends towards increasing annual and 3‐year incidence among 0‐ to 14‐year‐olds. However, there were no changes over time in annual or 3‐year incidence in the 0–4 year or 10–14 year age groups. Conclusions The incidence of T1DM in Yamanashi Prefecture increased among children aged 0‐14 years over the study period, with the most significant increase occurring among 5‐ to 9‐year‐olds. These data suggest that the number of children aged <15 years with T1DM is gradually increasing in one of the local prefectures in Japan, Yamanashi Prefecture and that the age of onset is decreasing.


| INTRODUC TI ON
It has been reported that the incidence of T1DM among children aged <15 years is increasing in various regions globally. The

European Community Concerted Action Programme in Diabetes
(EURODIAB) reported that the incidence of T1DM among children aged <15 years in Europe increased between 1989 and 2003, and if this trend continued, that incidence would double between 2005 and 2020. 4 The incidence of T1DM in this age group from 2011 to 2013 in Kuwait was 2.3 times higher than that from 1995 to 1997. 5 In Croatia, there was a 5.87% increase in the incidence of T1DM among children aged <15 years between 2004 and 2012 (p < .001), 6 and the incidences of T1DM among Jewish and Arabic children aged <18 years in Israel increased by 3.6% and 4.0%, respectively, between 1997 and 2010. 7 In Bauru, Brazil, the incidence of T1DM among children aged <15 years increased at a rate of 3.1% per year, resulting in an overall increase of 2.5fold between 1986 and 2015. 8 In Korea, the incidence of T1DM in this age group increased from 1.36/100 000 person-years from 1995 to 2000 to 3.19/100 000 person-years from 2012 to 2014. 9 In Uzbekistan, the incidence of T1DM among children aged <15 years increased from 1.5/100 000 person-years in 1998 to 3.8/100 000 person-years in 2014, with an average annual rate of increase of 5.6% (p < .001). 10 To assess temporal trends in the incidence of T1DM among children in Japan, we registered all children newly diagnosed with T1DM in Yamanashi Prefecture from 1986 to 2018.

| ME THODS
Our study group was created in 1986 with the co-operation of the paediatric departments of all hospitals in Yamanashi Prefecture (10 paediatric departments in total). All children aged <15 years newly diagnosed with T1DM were registered.
Our registration system includes the following three components: (i) urinalysis conducted in all school students, (ii) information shared by study group hospitals regarding cases of new-onset T1DM and (iii) monitoring of the numbers of patients with new-onset T1DM who are registered to receive medical aid for 'Specific Chronic Pediatric Diseases'. To avoid missing new-onset cases, we asked all patients registered in the present study to register with the medical aid project.
In Japan, urinalysis is performed for all primary and junior high school students once annually in the spring under the School Health and Safety Act. The purpose of urinalysis is to screen for diseases such as nephropathies and diabetes mellitus. Students whose urine tests are positive for glucose are required to visit a clinic or hospital in Yamanashi Prefecture, and those who require hospitalization are inevitably referred to the hospitals that comprise our study group.
Every child with new-onset T1DM is required to be referred to hospital in Yamanashi Prefecture. The major cities in Yamanashi The 3-year intervals were as follows : 1986-1988, 1989-1991, 1992-1994, 1995-1997, 1998-2000, 2001-2003, 2004-2006, 2007-2009 in raw form and following stratification by sex, age and age group (0-4 years, 5-9 years and 10-14 years). The 3-year standardized incidence among 0-to 14-year-olds was calculated by dividing the sum of the incidences among the three age groups, each multiplied by the appropriate standard population, by the sum of the standard populations. We used Segi's world standard population to calculate 3-year standardized incidences among 0-to 14-yearolds as follows: 3-year standardized incidence among 0-to 14-year-olds =

| RE SULTS
The number of patients with new-onset T1DM receiving a healthcare subsidy for 'Specific Chronic Pediatric Diseases' in Yamanashi Prefecture and the number of study registrations were identical. The total number of T1DM cases was 99 (44 boys and 55 girls; female/ male ratio 1.25) over the study period. The age and sex distributions of children with T1DM were bimodal. There was an initial minor peak at age 4-5 years followed by a second larger peak at age 10-12 years ( Figure 1). The numbers of cases and incidences for each age group during each 3-year period are shown in Table 1.
There were no changes in the 3-year incidences among 0-to 4-yearolds (p = .4781) or 10-to 14-year-olds (p = .8019; Figure 2). When incidences were stratified by sex, no changes were identified over the study period.

The age and sex distributions of children with T1DM in Yamanashi
Prefecture between 1986 and 2018 did not differ from those re-  The increase in incidence among children aged 0-14 years was not statistically significant because the age group with the highest incidence (10-to 14-year-olds) showed a relatively small increase; the numbers of cases among 5-to 9-year-olds showed a greater increase but were still too small to significantly impact overall incidence.
In contrast to the present findings, Onda  The trends observed in our age group data were slightly different from those in Europe. In Europe, areas with lower incidence showed higher increases in T1DM incidence among 0-to 4-year-olds. Japan has one of the lowest childhood T1DM incidences in the world 4  Prefecture. The trends observed in the present study were very similar to those identified in previous studies conducted in Japan and Korea. 1,2,9 Collectively, these findings suggest that the incidences of T1DM among 5-to 9-year-olds and among children as a whole might be increasing in Yamanashi Prefecture and that the age of onset may be gradually decreasing, as it is in Europe.
In Japan, as in other countries, it is important to conduct surveillance for chronic diseases such as T1DM in childhood. These data provide a complete picture of chronic disease in childhood and are an important basis for the government to improve medical and welfare support systems for children.

E TH I C S S TATEM ENT
The study was approved by the Institutional Review Board of the School of Medicine of Yamanashi University (IRB number: 2020-2215).

PATI ENT CO N S ENT S TATEM ENT
All study participants provided informed consent obtained in the form of opt-out.

ACK N OWLED G EM ENTS
We thank Mark Cleasby, PhD, from Edanz Group (https://en-authorservi ces.edanz group.com/ac) for editing a draft of this manuscript.

F I G U R E 3
Three-year incidences of type 1 diabetes among 0-to 14-year-old children in Yamanashi Prefecture between 1986 and 2018. The average increase in incidence per 3-year interval was 3.45% (p = .5856, n = 98, Poisson regression test)