Type 1 diabetes in 2017: global estimates of incident and prevalent cases in children and adults

Aims/hypothesis Data on type 1 diabetes incidence and prevalence are limited, particularly for adults. This study aims to estimate global numbers of incident and prevalent cases of type 1 diabetes in 2017 for all age groups, by country and areas defined by income and region. Methods Incidence rates of type 1 diabetes in children (available from 94 countries) from the IDF Atlas were used and extrapolated to countries without data. Age-specific incidence rates in adults (only known across full age range for fewer than ten countries) were obtained by applying scaling ratios for each adult age group relative to the incidence rate in children. Age-specific incidence rates were applied to population estimates to obtain incident case numbers. Duration of diabetes was estimated from available data and adjusted using differences in childhood mortality rate between countries from United Nations demographic data. Prevalent case numbers were derived by modelling the relationship between prevalence, incidence and disease duration. Sensitivity analyses were performed to quantify the impact of alternative assumptions and model inputs. Results Global numbers of incident and prevalent cases of type 1 diabetes were estimated to be 234,710 and 9,004,610, respectively, in 2017. High-income countries, with 17% of the global population, accounted for 49% of global incident cases and 52% of prevalent cases. Asia, which has the largest proportion of the world’s population (60%), had the largest number of incident (32%) and prevalent (31%) cases of type 1 diabetes. Globally, 6%, 35%, 43% and 16% of prevalent cases were in the age groups 0–14, 15–39, 40–64 and 65+ years, respectively. Based on sensitivity analyses, the estimates could deviate by ±15%. Conclusions/interpretation Globally, type 1 diabetes represents about 2% of the estimated total cases of diabetes, ranging from less than 1% in certain Pacific countries to more than 15% in Northern European populations in 2017. This study provides information for the development of healthcare and policy approaches to manage type 1 diabetes. The estimates need further validation due to limitations and assumptions related to data availability and estimation methods. Graphical abstract Supplementary Information The online version contains peer-reviewed but unedited supplementary material available at 10.1007/s00125-021-05571-8.


Methods
This section describes data input and modelling techniques used for producing the global estimates of type 1 diabetes incidence and prevalence.
In this study, type 1 diabetes is defined as absolute insulin deficiency of unknown cause and a consequent need for insulin treatment for survival.
Overview of the estimation process Information on incidence and prevalence of type 1 diabetes in all age groups is lacking for most countries whereas incidence rates in children have been reported for many countries. To obtain prevalence estimates the first step is to model the incidence rates in older age groups using the rate assumed for children. The second step is to estimate prevalence from incidence under specified assumptions of life expectancy, i.e. the mean duration from onset of type 1 diabetes until certain age levels and, ultimately, death. The mean duration depends not only on the age at onset but also on the conditions for survival in a given country, particularly concerning access to insulin treatment and other types of healthcare needed to manage type 1 diabetes. Therefore, the mean duration for a given age at onset has been adjusted by a penalty function. The penalty function assigns high values to countries with poor circumstances and low values to countries with good circumstances, as described below.
To estimate prevalence from incidence the general principle of Prevalence = Incidence  Mean Duration has been used. The equation assumes a situation of epidemiological 'steady state', i.e. that the annual number of new cases equals the annual number of deaths from the patient population of persons with type 1 diabetes. With increasing incidence rates and improving prognosis in type 1 diabetes this assumption is violated. To our knowledge there are no data available to quantify this bias for various types of countries but because the attained prevalence population includes cohorts of patients diagnosed in periods with lower incidence levels and poorer prognosis the numbers of prevalent cases in our report may represent overestimates.
The principle of estimating prevalence from incidence and duration has been applied in two different contexts for a given country: First, to obtain an estimate of the total prevalence of type 1 diabetes from the summation of prevalence estimates stratified by age at onset. Second, to stratify the total prevalence by attained age.
ESM Table 1 contains the country-specific incidence rates (age group 0-14 years) used, together with reference to the data sources [4 -66] and the relevant country-specific characteristics.
Age grouping Throughout the analysis age has been grouped by four categories: 0-14; 15-39; 40-64; 65+ years as a compromise between the desire to provide estimates for narrowly-defined agegroups versus the lack of sufficiently detailed epidemiological data on type 1 diabetes.
Due to paucity of data on sex-specific incidence and mortality rates in type 1 diabetes, the present analysis has been performed without stratification by sex.
Incidence rates of type 1 diabetes in adult age groups Due to the lack of data on incidence rates of type 1 diabetes beyond childhood we have for this analysis used Danish data for the year 2017, established as an update of a previously published dataset established on the basis of Danish central health registers [67]. In brief, the population of diabetes (regardless of type) has been ascertained from prescription data on antidiabetic drugs and admissions due to diabetes from hospital registrations. Patients with type 1 diabetes have been identified if insulin treatment was commenced within one year after the first registration qualifying for diabetes (the within one year part of the criterion was applied only for patients with encounters relevant to diabetes after 1995 due to restrictions in data availability). For the years 2016 and 2017, the annual number of incident cases of type 1 diabetes in Denmark was approximately 900 in a population of about 6.5 million. The number of prevalent cases was approximately 26,000 in both 2016 and 2017. The incidence rates for age groups beyond childhood have been scaled relative to that applied for children (aged 0-14 years). Using the Danish age-specific incidence rates for the years 2016 and 2017 scaling values have been obtained as shown in ESM Table 2 and applied globally. The effects of applying alternative scaling factors are investigated in the sensitivity analyses.
Introducing the penalty function The estimation model adjusts for the fact that life expectancy in type 1 diabetes (mean duration from onset of type 1 diabetes) depends on the general living conditions for a given country by introducing a 'penalty' (Pen). High and low values of Pen indicate poor and good circumstances, respectively. Based on recommendations from World Health Organization, rates of childhood mortality (CM) in children under 5 years old have been used as input to the penalty function. Country-specific values of CM have been obtained from data referring to 2017 as published by the United Nations Inter-Agency Group for Child Mortality Estimation (IGME) [68] and reproduced in ESM Table 1. The value of Pen for a given country is estimated as Pencountry = CMcountry (per 1000 livebirths)/130 The constant 130 has been selected because it is just above the highest CM rate reported by UN for 2017. Formulated this way, the Pen-value increases with increasing CM. For example, taking the value of CM for Sierra Leone at 110.5 per 1,000 gives a Pen-value at 0.85. In contrast, for a country like New Zealand, with a CM at 5.3 per 1,000, the Pen-value is 0.041.
Estimating duration of type 1 diabetes As explained above, the basic principle for the estimation of prevalence is to use agespecific incidence and average duration from diagnosis given age at diagnosis. The Penvalue is applied to the estimation of mean durations in type 1 diabetes by: Mean durationcountry = Maximum mean duration  (1 -Pencountry). The maximum mean duration represents the mean remaining life-years expected for a patient with type 1 diabetes living under optimal circumstances, from disease onset. Mortality experiences from type 1 diabetes in Denmark during the years 2015-2017 have been used according to an example from Scotland [69]. Within each of the age groups 0-14, 15-39, 40-64, and 65+ years, the mean age at onset was obtained from the Danish incident cases registered for the period 2015-2017. The corresponding estimated remaining lifeyears from the mean onset ages were obtained from the data as shown in ESM Fig.1. Finally, since a Pen-value at 0.033 has been assigned to Denmark (ESM Table 1), the Danish remaining life-years, given mean age at onset within each of the four age groups) were divided by 1-0.033 (=0.967) to obtain the maximum mean durations of type 1 diabetes at 60.4, 43.4, 21.1 and 8.5 years for each of the age groups 0-14, 15-39, 40-64, and 65+ years (ESM Table 3).
A global minimum mean value has been set at 0.5 year reflecting the assumed short survival (regardless of age) in persons diagnosed with type 1 diabetes in countries without access to insulin treatment and other healthcare support for persons with type 1 diabetes.
Estimating prevalent case numbers Prevalence numbers by age at onset It is assumed that the country-specific prevalence for patients with a given onset age group is estimated by Prevalenceonset age group = Incidenceonset age group  Durationonset age group applied for the age groups 0-14, 15-39, 40-64, and 65+ years. Age-specific incidence rates are those scaled from the age-group 0-14 years for the country concerned (see ESM Table  1) and presented in ESM Table 2. Duration is obtained as the maximum mean duration for each age at onset group (see ESM Table 3), adjusted for the penalty assigned to the country concerned (ESM Table 1). ESM Table 4 shows the steps in estimating prevalence by age at onset using New Zealand as an example country.
Prevalence numbers by attained age To derive prevalence estimates by attained age from prevalence estimates by age at onset we apply the principle that, assuming a state of epidemiological equilibrium, the total number of life-years produced by a person with type 1 diabetes can be allocated to successive age intervals from disease onset. Estimates of cumulative survival at attained age have, for each age group at onset, been derived from the Danish data described above. After adjusting for the penalty value 0.033 assigned to Denmark the default values used for maximum mean cumulative are summarized in ESM Table 5. The age interval 65+ years is not shown because it is open-ended with a cumulative survival at 0 by definition. For the same reason age at onset group 65+ years is not shown. EMS Fig. 2 illustrates the principles of allocating prevalent case numbers to groups by attained age using the data in EMS Table 4 and using New Zealand as an example.
The principle is that each person with type 1 diabetes represents one lived life-year, and the fractions of this life-year that are spent in each interval after onset are estimated as the respective areas under the curve as shown in Figure 2. For each group defined by age at onset, the number of incident cases is multiplied by the fraction surviving for every age interval after onset of the disease, providing the contributions to the number of prevalent cases by attained age from each of the age at onset groups. For the open-ended oldest attained age interval, the maximum mean duration will vary between countries depending on circumstances. Furthermore, the number of prevalent cases in this age group is obtained by the total number of prevalent cases not accounted for in the younger intervals.
The calculations are further explained in Table 5 using New Zealand as an example (penalty value: 0.041). Note that inconsistencies in numbers occur due to rounding errors. Standardization of prevalence estimates For age standardization of the country-specific overall prevalence proportions the WHO Standard Population for the period 2000-2015 was used, with the corresponding percentage distribution by age groups 0-14, 15-39, 40-64, and 65+ years: 26.1%, 39.4%, 26.3%, and 8.2%, respectively. Table 7 specifies the profiles of the four sensitivity analyses performed separately with the core model as the reference and with focus on the penalty function, mean disease duration, incidence scaling over age at onset groups, respectively. In one of the analyses, the penalty function was not applied ('ignoring penalty function'), in another, life expectancy was reduced ('reduced mean duration') and followed by two different sets of scaling values of incidence rates from childhood to alder age groups ('changed incidence scaling 1' and 'changed incidence scaling 2').