You are here

Article Text

Article menu
Epidemiology/Health services research
Future number of people with diagnosed type 1 diabetes in Germany until 2040: an analysis based on claims data
  1. Dina Voeltz1,2,
  2. Ralph Brinks3,4,
  3. Thaddäus Tönnies3,
  4. Annika Hoyer1
  1. 1Biostatistics and Medical Biometry, Medical School OWL, Bielefeld University, Bielefeld, Germany
  2. 2Department of Statistics, Ludwig Maximilians University Munich, Munchen, Germany
  3. 3Institute for Biometrics and Epidemiology, German Diabetes Center, Leibniz Institute for Diabetes Research at Heinrich-Heine-University Duesseldorf, Duesseldorf, Germany
  4. 4Chair for Medical Biometry and Epidemiology, Witten/Herdecke University, Faculty of Health/School of Medicine, Witten, Germany
  1. Correspondence to Dr Dina Voeltz; dina.voeltz{at}uni-bielefeld.de

Abstract

Introduction We aim to project the number of people with diagnosed type 1 diabetes in Germany between 2010 and 2040.

Research design and methods We first estimate the age-specific and sex-specific incidence and prevalence of type 1 diabetes in Germany in 2010 using data from 65 million insurees of the German statutory health insurance. Then, we use the illness-death model to project the prevalence of type 1 diabetes until 2040. We alter the incidence and mortality underlying the illness-death model in several scenarios to explore the impact of possible temporal trends on the number of people with type 1 diabetes.

Results Applying the prevalence from 2010 to the official population projections of Germany’s Federal Statistical Office yields a total number of 252 000 people with type 1 diabetes in Germany in 2040 (+1% compared with 2010). Incorporating different annual trends of the incidence and mortality in the projection model results in a future number of people with type 1 diabetes between 292 000 (+18%) and 327 000 (+32%).

Conclusions For the first time in Germany, we provide estimates for the incidence, prevalence, and number of people with diagnosed type 1 diabetes for the whole German population between 2010 and 2040. The relative increase of the people with type 1 diabetes ranges from 1% to 32% in 2040 compared with 2010. The projected results are mainly influenced by temporal trends in the incidence. Ignoring these trends, that is, applying a constant prevalence to population projections, probably underestimates future chronic disease numbers.

  • epidemiology
  • incidence
  • diabetes mellitus, type 1
  • chronic disease

Data availability statement

Data are available in a public, open access repository. Data and source file for estimation of the incidence of type 1 diabetes (T1D) and projection of the prevalence and number of people living with T1D in future Germany are freely available on Zenodo (see reference and link below). Prevalence input is based on a huge German claims data set and is uploaded as aggregated data. The population projections are publicly avilable on the website of the German Federal Statistical Office. Voeltz D, Brinks R, Tönnies T, et al. Future number of people with diagnosed type 1 diabetes in Germany until 2040: Zenodo 2022. Available at: 10.5281/zenodo.6799293 or https://zenodo.org/record/6799293%23.ZCvgG4TP02w

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

https://doi.org/10.1136/bmjdrc-2022-003156

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    WHAT IS ALREADY KNOWN ON THIS TOPIC

    • The few studies investigating the current and future population with type 1 diabetes in Germany are limited to certain ages, cohorts, and ignore temporal trends in disease-related rates.

    WHAT THIS STUDY ADDS

    • Based on data from 65 million insurees of all German statutory health insurances, we estimate an increase in the German population with type 1 diabetes ranging from 252 000 (+1% compared with 2010) to 327 000 (+32%) people in 2040 which is mainly driven by the incidence.

    • The peak of the age-specific prevalence is projected to shift toward older ages.

    HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

    • Germany will possibly face greater demand for diabetes-related education, healthcare, and medical resources.

    Introduction

    Worldwide, chronic disease burden rises dramatically.1 As one of the most common chronic diseases in Germany, diabetes mellitus causes suffering among diagnosed people as well as tremendous costs for the overall population.1–4 Without a structured self-management plan, monitoring blood glucose level, physical activity, diet and, most important, without daily insulin treatment, people with type 1 diabetes (T1D) are at high risk of premature death.5 Although defined as autoimmune disease with an onset at ages younger than 35 years,6 7 the disease is nowadays prevalent among the whole age range.1 7 8 In 2009 in Germany, estimates of the prevalence of T1D for girls and boys younger than 19 years were 0.17% to 0.19%, respectively. The prevalence among women and men aged between 20 and 79 years was estimated between 0.28% and 0.39% and among elderly, that is, older than 80 years, between 0.47% and 0.50%.2 9 Reports of the total number of people with T1D in Germany range from 256 000 to 373 000 in 2009 and 2016, respectively.2 9 Approximately 7200 people per year are newly diagnosed.2 Similar to trends observed over the past decades, the T1D incidence and prevalence are expected to rise further.1

    Despite the high relevance of T1D and its considerable health and economic consequences,5 representative studies that investigate the population-wide current and future disease burden in Germany are sparse.1 2 The generalizability of previous studies to Germany is impaired due to their focus on particular age ranges and/or their restriction to data from selected health insurances, and cohorts.10 11 Additionally, although there is evidence that diabetes type-specific interventions help in reducing complications and mortality,6 many existing analyses do not distinguish the diabetes types.11 12 This makes the need for type-specific analyses of the current and future burden of T1D even more compelling.3

    In this article, we provide nationally representative estimates of the age-specific and sex-specific T1D incidence in Germany in 2010. Furthermore, we project the age-specific and sex-specific prevalence and the number of people diagnosed with T1D from 2010 to 2040.

    Methods

    Based on published data from 65 million insurees of the German statutory health insurance,8 9 13 14 we first estimated the age-specific and sex-specific incidence of diagnosed T1D in Germany in 2010 for all ages between 0 and 100 years. Using mathematical relations between prevalence, incidence, and mortality based on the illness-death model (IDM), we projected the future age-specific prevalence of T1D between 2010 and 2040. In this second step, we compared several scenarios regarding possible temporal trends in the incidence and mortality. To calculate the number of people with diagnosed T1D in each year, we finally multiplied the projected prevalence to the projected distribution of the German population issued by the German Federal Statistical Office (FSO).14 The following paragraphs provide more details about these three steps.

    All methods were implemented using the free statistical software R, V.4.1.0 (The R Foundation for Statistical Computing). The source code and data for running the analysis are published in the open-access repository Zenodo.15 We only used published data on aggregated level.

    Definition of type 1 diabetes

    To define T1D, we refer to the International Classification of Diseases (ICD) coding provided in the underlying claims data which is commonly used to specify a particular type of diabetes.9 10 13 The aggregated prevalence data, provided by the German Institute for Medicine Documentation and Information in 2012, contains information from about 65 million policyholders from the inpatient and outpatient sector insured for at least 360 days per year by one of the German statutory health insurances. This represents >80% of Germany’s overall resident population. Accordingly, the data can be considered a unique and suitable information source that enables nationwide representative estimates of diagnosed diabetes in Germany.9 10 For instance, the Robert Koch Institute features the same data for estimating the overall incidence of diabetes in Germany.16 17 Generally, the ICD-code E10.- specifies a T1D diagnosis. Since in some cases, the data report multiple diagnoses for the same individual, the combination of the ICD-codes E14.- (unspecified diabetes mellitus) and E10.- also resulted in the assignment to T1D. Any other ICD coding or unclear code combination were excluded.

    Estimation of the age-specific T1D incidence in 2010

    To estimate the age-specific and sex-specific T1D incidence in Germany in 2010, we applied a partial differential equation (PDE), which describes the relation between prevalence, incidence and mortality as a function of age and calendar time (see online supplemental material).9 18 This requires prevalence input data for 2 years, information on the general mortality, and on the mortality rate ratio (MRR), that is, the ratio of the mortality rates of people with versus without T1D. The aforementioned claims data9 provide aggregated, age-specific and sex-specific information on the number of people in Germany along with their respective diabetes status (ie, diagnosed with T1D vs no T1D) for the two consecutive years 2009 and 2010. Knowing that the prevalence is the proportion of a particular population found to be affected by a certain condition (here, T1D vs no T1D) at a specific time (here, 2009 and 2010), we analogously calculated the overall and the age-stratified T1D prevalence for each sex for 2009 and 2010 using our input data.9 To do so, we divided the number of T1D diagnoses in Germany by the total German population number for both years on overall and age-specific level, respectively. We obtain the general mortality of the German population as of 2010 from the population projections of the German FSO.14 Due to unavailability of the T1D-related MRR in Germany, we align with other diabetes-related work,9 10 and refer to age-specific and sex-specific estimates from Denmark as the two countries’ MRRs are claimed highly comparable.8 9 19 20

    Supplemental material

    Due to inconsistencies in coding of diagnoses,9 we approximated the course of the T1D incidence in Germany for ages 35+ years (see online supplemental material). For that purpose, and due to a lack of appropriate alternative data from Germany, we used information from the Danish National Diabetes Register. In Denmark, the T1D incidence rate in ages 35+ years decreases logarithmically with age. This resembles the German diabetes burden estimated in previous studies,12 where the proportion of T1D among all diabetes types peaks among the youth and decreases with increasing age. Since the erroneous coding occurred among older ages and since often latent autoimmune diabetes in adults (LADA) commonly showing in patients over 35 years is diagnosed as T1D, the chosen age cut-off seems reasonable.21

    Projection of the age-specific T1D prevalence and future number of people with T1D

    We use the aforementioned PDE to project the age-specific and sex-specific prevalence of T1D in Germany between 2010 and 2040 (see online supplemental material). Solving the PDE returns the age-specific prevalence for each year depending on the incidence and MRR.18 Since information on current and future trends of the incidence and mortality are scarce in Germany,10 we assumed five hypothetical scenarios that have been discussed in previous projections of T1D and T2D in Germany.1 3 10 22 Scenario 1, a ‘simple’ sex-specific and age-specific prevalence extrapolation, assumes a constant age-specific prevalence between 2015 and 2040 and neither reflects on the incidence nor on the mortality. In scenarios 2–5, we alter the temporal trends of the incidence and MRR. We consider scenario 2 as baseline scenario, assuming constant incidence and mortality. In the future, it is likely that the mortality rate among people with T1D will decrease faster than among people without T1D due to probable advancements in medical care.10 20 Consequently in scenarios 3–5, we anticipate that the MRR decreases by 2% per year as observed in Denmark.10 20 Furthermore, scenario 3 assumes a constant incidence rate, while scenarios 4 and 5 assume an annual increase and decrease of 0.5%, respectively.

    We obtain the number of people with T1D in Germany between 2010 and 2040 for all ages between 0 and 100 by multiplying the projected prevalence from the PDE with the number of people from the population projections of the FSO.14 The FSO releases several variants wherein they assume different future birth rates, life expectancies at birth, and migration. We consider a rather realistic version (variant B1L2M1) which assumes declining fertility (B1: birth rate of 1.4 children per woman), moderate development in life expectancy (L2: life expectancy of 84.4 and 88.1 years for men and women, respectively), and low long-term net migration (M1: 147 000 people per year). This aligns with Tönnies et al10 and Voeltz et al,23 who also examine this variant in their projections of type 2 diabetes (T2D) in Germany.

    Sensitivity analyses

    Model assumptions and data sparsity might hamper the projection of future prevalence. Precise information about temporal trends of the incidence and mortality in Germany are lacking. Since these rates can emerge very heterogeneously, we assessed further, relatively extreme scenarios which may cover possible unexpected developments. First, we evaluated an annual decrease in the incidence of 5% as observed among older ages in Denmark between 1996 and 2016.8 In contrast to a lower incidence, researchers postulated that the SARS-CoV-2 virus might damage cells in the pancreas that produce insulin, which may notably increase the risk of new-onset T1D.24 25 Although data on a relation between SARS-CoV-2 and T1D remain heterogeneous,25 we assessed a 5% annual increase in the incidence as observed in European children younger than 5 years between 2005 and 2020.26 In addition, we performed a Monte Carlo simulation with 500 bootstraps to account for possible sampling error of the input values and to calculate 95% CIs (see online supplemental material). Furthermore, we assess the impact of alternative variants of the FSO population projections (see online supplemental material).

    Results

    Estimated incidence

    Figure 1 visualises the age-specific and sex-specific estimates of the German T1D incidence in 2010. Across all ages, the T1D incidence is estimated higher for men than women. For both sexes, the incidence peaks between 5 and 15 years of age with approximately 0.17 for men and 0.12 for women per 1000 person-years. For all following ages, the estimated course declines.

    Figure 1
    Figure 1

    Estimated age-specific incidence of type 1 diabetes (T1D) in Germany among women (red) and men (blue) in 2010.

    Projected prevalence

    Figure 2 shows the estimated age-specific and sex-specific prevalence for Germany in 2009 and 2010. For any age below 80 years, the prevalence of T1D is higher among men compared with women. Among the elderly, the relationship between prevalence and sex does not remain that obvious.

    Figure 2
    Figure 2

    Prevalence of type 1 diabetes (T1D) among women (red lines) and men (blue lines) in Germany in 2009 and 2010.

    Figure 3 shows the projected age-specific and sex-specific T1D prevalence for five scenarios of the incidence and MRR from 2010 to 2040 assuming moderately increasing life expectancy. For almost any age, the prevalence of T1D for men is slightly higher than for women. Particularly noteworthy is the shift in the course of the prevalence. In 2040, the prevalence is estimated to peak at around 40 years of age for both sexes, while the prevalence decreases among older age groups compared with 2010. Compared with scenario 1 assuming a constant prevalence, results for all other scenarios suggest a substantial increase in the overall prevalence for men and women (online supplemental figure S2). The different variants of the FSO had no remarkable influence on the prevalence projection (online supplemental figure S4).

    Figure 3
    Figure 3

    Projected age-specific and sex-specific prevalence of type 1 diabetes (T1D) in Germany for variant B1L2M1 of population projections of the German Federal Statistical Office. Blue lines depict the prevalence among men in 2010, 2020, 2030 and 2040, red lines visualise the prevalence among women in 2010, 2020, 2030 and 2040. The different line types represent the scenarios, that is, scenario 1, constant age-specific prevalence from 2010; scenario 2, constant T1D incidence rate (IR), constant mortality rate ratio (MRR); scenario 3, constant IR, MRR −2% per year; scenario 4, IR −0.5% per year, MRR −2% per year; scenario 5: IR +0.5% per year, MRR −2% per year.

    Figure 4
    Figure 4

    Projected number of type 1 diabetes (T1D) cases (in thousands) for men and women in Germany in 2010 and 2040 for variant B1L2M1 of the German Federal Statistical Office for scenarios 1–5. Scenario 1, constant age-specific prevalence from 2010; scenario 2, constant incidence rate (IR), constant mortality rate ratio (MRR); scenario 3, constant IR, MRR −2% per year; scenario 4, IR −0.5% per year, MRR −2% per year; scenario 5: IR +0.5% per year, MRR −2% per year.

    Projected number of people with T1D

    Figure 4 and table 1 show the projected number of people with T1D in Germany in 2010 and 2040 and approximate CIs for several scenarios of the incidence and MRR. Coherently, as for the prevalence, the number of men diagnosed with T1D is higher than the number of women throughout the overall projection horizon. Assuming that the prevalence remains as in 2010 (scenario 1), overall numbers are projected to increase by 1% to a total of 252 000 people with T1D in 2040. For scenarios 2–5, the number of people with T1D in Germany in 2040 is projected to increase by 44 000 people (+18%) to 78 800 people (+32%) compared with 2010. Compared with a relatively linear course for scenario 1, the course of the estimated case numbers for scenarios 2–5 resembles a J-shaped growth pattern in the first few years and then grows linearly for men and women alike. The FSO population variants had minor impact (see online supplemental material).

    View this table:
    Table 1

    Number of people with T1D (in thousands) in Germany in 2010 and 2040 for variant B1L2M1, 95% CIs for case numbers in 2040 as well as absolute and relative change comparing 2010 vs 2040

    Sensitivity of the results

    Additionally, we assessed an annual change of the incidence of −5% and +5%. Comparing 2010 and 2040, we obtain a decrease of the population with T1D of −27 500 (−11%) and an increase of almost 300 000 (+121%), respectively. Possible sampling error of the input values had minor impact and led to deviations of 5%–7% for women and men (see online supplemental material).

    Discussion

    The aim of this study was to project the number of people with T1D in Germany between 2010 and 2040. For this purpose, we first estimated the age-specific and sex-specific incidence in 2010. Then, we projected the age-specific and sex-specific prevalence and the number of individuals with T1D for each year until 2040. Overall, our results indicate a general increase in the future burden of T1D. Comparing different scenarios of disease-specific rates, we found considerable differences in the projected population with T1D. In scenario 1, we estimated an increase in the population with T1D of 1% in 2040 compared with 2010. In this scenario, we simply applied the constant sex-specific and age-specific prevalence of T2D from 2010 to future population distributions projected by the German FSO. Although this approach is commonly used for chronic disease projections,23 it appears that ignoring temporal trends in the prevailing epidemiological situation likely leads to misleading results since the future development of a disease is influenced by various disease-specific aspects. More precisely, disregarding potential changes in mortality and incidence as is done in scenario 1 may provoke an underestimation of future T1D cases in Germany.10 23 Contrastingly, reflecting on changes in epidemiological indicators such as mortality and incidence likely produces more reliable projection results.23 This was done in scenarios 2–5, where the use of the PDE enables to additionally account for disease-specific information when modeling the future prevalence of T1D. Considering scenarios 2–5, which take temporal trends in the incidence and mortality into consideration, results ranged from 44 000 (+18%) to 78 800 (+32%) additional people with T1D in 2040. Unfortunately, there are no data available to validate our projected results and/or pinpoint the most plausible scenario. Nonetheless, the general increase observable across all scenarios highlights the importance of intensified diabetes-related public health management and the need for more resources allocated to diabetes healthcare. Moreover, it emphasizes the need for research on primary prevention of T1D and effective interventions to reduce its incidence.27

    Environmental and lifestyle changes may explain the rapid increase of the number of people with T1D worldwide.1 5 28 The relatively short projection period makes changes in the genetic background of populations an unlikely cause. For instance, lifestyle factors that lead to rapid growth and large gains in weight particularly in the first years of life are more likely causing the increasing incidence. Factors such as accelerated autoimmune processes contribute to a growing T1D incidence in childhood.22 Besides, factors that initiate the autoimmune destruction of beta cells are also discussed, for example, fetal and neonatal factors. Another hypothesis for the increase in the incidence is the reduced exposure to microbial antigens.22 Overall, to gain a better understanding of drivers of the T1D burden, future studies could include these environmental factors in the evaluation to shed light on possible causes.

    With regard to the methodological approach, our results underline the advantage of the IDM as theoretical background as it allows to incorporate temporal trends in disease-specific rates which impact the epidemiological trends of a disease.10 23 Comparing multiple future scenarios, we showed that the incidence rate is a major driver of future numbers. A blunt extrapolation, that is, applying the observed prevalence in 2010 to the population projections, is unlikely to project accurate future numbers.10 23 Ignoring disease-specific rates probably leads to underestimation of an upcoming disease situation, and consequently, underestimates the importance of public health management, healthcare facilities, education of specialists, and the like.

    In addition to the increasing T1D burden, our projection indicates a shift of the prevalence peak toward older ages. Concomitant, although historically believed, T1D may not remain a disease of childhood. In line with Ostrauskas et al,29 our results contradict a shift of the onset of T1D toward younger ages as explanation for the increasing temporal trend among children.

    Comparison with previous studies

    Worldwide, T1D incidence varies largely. In the great majority of countries, it has steadily increased over the past and its growth is expected to continue. In 2021, the International Diabetes Federation (IDF) estimates that 1 211 900 children and adolescents younger than 20 years were diagnosed with T1D and reports an annual growth of 149 500 children worldwide.5 Direct health expenditures related to diabetes are close to US$1 trillion and are projected to exceed this by 2030. Strikingly, current evidence suggests that the IDF figures are still substantially underestimated.30 Using data from the Childhood Diabetes Registry of Saxony, Germany, Manuwald et al1 report an increase in the T1D incidence among children and adolescents in Germany from 17.1 to 24.7 per 100 000 person-years between 1999 and 2019. Continuing this trend, they project a growth to 34.8 per 100 000 person-years in 2030. Based on medical data from Baden-Wuerttemberg and Saxony and the national diabetes registry of North Rhine-Westphalia, Rosenbauer et al2 report a national T1D incidence among adults of 6.1 per 100 000 person-years between 2014 and 2016. Their prevalence estimate equals 445 and 544 per 100 000 individuals with slightly lower figures for women compared with men. This equals a number of 341 000 adults with T1D in 2016, which is estimated to grow by 4150 adults annually. Notably, these studies neglect relevant age groups, the role of disease-specific rates or are based on limited data. Contrarily, one of the main strengths of our prevalence input data is its completeness in terms of covering all age groups as well as all health insurances in the German statutory health insurance.

    Limitations

    Although the reasons remain unclear, ICD codes are often inaccurately captured in administrative data in Germany (such as our prevalence input data),9 12 13 impeding a valid differentiation of the diabetes types.31 Indeed, research confirms that limited diagnostic accuracy, that is, sensitivity and specificity, from aggregated routine data may lead to unstable or implausible estimation results.32 33 Potential misclassifications and double diagnoses in our setting may have occurred when a diagnosis was changed after more precise specification of a person’s health state by different care providers or clinicians. For instance, and controversially discussed, the ICD coding specifies LADA as E10.-,34–36 although its pathogenesis only gradually resembles the one of T1D. Moreover, since T1D is more common to develop at younger ages, it can be assumed that many unspecific cases among the older population could be attributed to T2D.13 31 Consequently, these false positive diagnoses might explain questionable estimates among older ages. On the contrary, unspecific diabetes diagnosis among children very likely mask a T1D case.13 31 To filter out such inaccuracies observed among older ages and since our focus is on T1D developed during childhood, we adapted the log-linear trend from Denmark to the German T1D incidence in 2010. Including misspecifications, unclear or double diagnoses, or cases masked by the unspecified diabetes coding E14.-, may lead to an additional increase in the prevalence of T1D.13 Alternatively, previous studies advise the additional integration of medication to validate that the coding accurately reflects T1D prevalence rates from 2010,31 since for patients with T1D it is necessary to inject insulin to control blood glucose levels. This method is of help for children and adolescents, in adulthood however it is inapplicable.31 This falls beyond the scope of our article and is left open for future research.

    Since the future course of the incidence and mortality is unknown, we explored the impact of trends in incidence and mortality by considering several scenarios. This highlights the need for long-term, country-specific and diabetes type-specific studies to obtain confidence about future T1D cases. Assuming that medical aid and the treatment of diabetes will improve further, a decrease in the mortality seems likely. Evidence for continuous declines in mortality is found in several countries such as the USA, Canada, Finland, Sweden and Denmark.8 37 38 Contrastingly, the T1D incidence and its development are heterogeneously distributed across countries.1 5 Despite a universally reported growth, evidence from some regions suggests a slowing in this increase.5 28 Against this reporting, pooled estimates for Europe show a continuous annual 3.4% growth, indicating a doubling of the incidence within 20 years.28 In Germany, the T1D incidence among children has increased threefold in the past decades.1 In line with the trend over the past 20 years, other studies assume a 40% increase in the incidence of T1D in children and adolescents for the next 10 years.22 Overall, declining mortality and increasing incidence as in scenario 3 seems most plausible. Considering that the claims data reflect the medical situation a decade ago, it does neither comprise latest happenings such as the SARS-CoV-2 pandemic nor other recent developments that may have affected the disease situation.39 It appears that a SARS-CoV-2 diagnosis is associated with an increased risk of new-onset T1D.24 25 With our scenario in the sensitivity analysis assuming a 5% annual increase in the incidence, we aim to account for extreme events such as the impact of SARS-CoV-2. In that scenario, we project an increase in the future T1D burden in 2040 in Germany of up to 121% additional T1D cases compared with 2010. Evidently, other unexpected events such as the global climate, energy or inflation crisis, future developments of migration, or changes in diagnostic criteria may alter future trends of the T1D incidence and mortality.40–43 Furthermore, there are medical and non-medical issues that make particular ethnic groups more or less vulnerable to diabetes. Routine data and particularly aggregated prevalence data from administrative origin do not provide any identification possibility of the individual’s ethnicity.13 Assessing the impact therefore goes beyond scope of our study, but could be addressed in the future.

    Due to data scarcity in Germany, we referred to information on the incidence and MRR from Denmark for our initial T1D incidence estimation and prevalence projection. The motivation to use Danish data is threefold. First, it is based on the national diabetes register which has a coverage of 100% due to compulsory reporting and which accurately differentiates the diabetes types.20 Second, previous studies concerning diabetes surveillance in Germany refer to Danish data.9 10 Third, for countries that are comparable in terms of their disease burden and healthcare systems, such as Denmark and Germany, deviations of disease-specific rates should remain minimal.9 10 Obviously, modelling the T1D situation over a 30-year horizon from 2010 until 2040 using prevalence data from 2010 and borrowed information from Denmark requires some assumptions. Our analysis and the reliability of our results would be strengthened if more recent, validated data from Germany were available. Moreover, information on T1D directly measured in the German population would be more appropriate to generate reliable national projections. As final remark with regard to an accurate diabetes surveillance in Germany, we claim that more detailed, precise, and timely data regarding diabetes in general and T1D in particular is urgently needed.

    Conclusion

    This is the first study addressing the current and future age-specific T1D incidence and prevalence for the whole German population for all ages from 0 to 100 years between 2010 and 2040. Using the IDM and data from 65 million people in Germany, we show that temporal trends in the incidence and mortality mainly drive the future number of individuals with T1D. Our projection suggests a substantial increase in the incidence, prevalence, and in the number of people diagnosed with T1D ranging from 4000 (+1%) to 78 800 (+32%) additional cases in Germany in 2040 vs 2010.

    Data availability statement

    Data are available in a public, open access repository. Data and source file for estimation of the incidence of type 1 diabetes (T1D) and projection of the prevalence and number of people living with T1D in future Germany are freely available on Zenodo (see reference and link below). Prevalence input is based on a huge German claims data set and is uploaded as aggregated data. The population projections are publicly avilable on the website of the German Federal Statistical Office. Voeltz D, Brinks R, Tönnies T, et al. Future number of people with diagnosed type 1 diabetes in Germany until 2040: Zenodo 2022. Available at: 10.5281/zenodo.6799293 or https://zenodo.org/record/6799293%23.ZCvgG4TP02w

    Ethics statements

    Patient consent for publication

    Ethics approval

    An ethics committee approval was not required since no individual data on humans or animals were involved.

    References

    1. ↵
      1. Manuwald U,
      2. Schoffer O,
      3. Kugler J, et al
      . Trends in incidence and prevalence of type 1 diabetes between 1999 and 2019 based on the childhood diabetes registry of saxony, Germany. PLOS ONE 2021;16:e0262171. doi:10.1371/journal.pone.0262171
    2. ↵
      1. Rosenbauer J,
      2. Neu A,
      3. Rothe U, et al
      . Types of diabetes are not limited to age groups: type 1 diabetes in adults and type 2 diabetes in children and adolescents. J Health Monit 2019;4:29–49. doi:10.25646/5987
      OpenUrl
    3. ↵
      1. Livingstone SJ,
      2. Levin D,
      3. Looker HC, et al
      . Estimated life expectancy in a Scottish cohort with type 1 diabetes, 2008-2010. JAMA 2015;313:37–44. doi:10.1001/jama.2014.16425
      OpenUrlCrossRefPubMed
    4. ↵
      1. Milan V,
      2. Fetzer S,
      3. Hagist C
      . Healing, surviving, or dying?-projecting the German future disease burden using a Markov illness-death model. BMC Public Health 2021;21:123. doi:10.1186/s12889-020-09941-6
    5. ↵
      1. International Diabetes Federation
      . IDF diabetes atlas. Brussels, Belgium, 2021.
    6. ↵
      1. Rawshani A,
      2. Rawshani A,
      3. Franzén S, et al
      . Mortality and cardiovascular disease in type 1 and type 2 diabetes. N Engl J Med 2017;376:1407–18. doi:10.1056/NEJMoa1608664
      OpenUrlCrossRefPubMed
    7. ↵
      1. Palmer JP,
      2. Hampe CS,
      3. Chiu H, et al
      . Is latent autoimmune diabetes in adults distinct from type 1 diabetes or just type 1 diabetes at an older age? Diabetes 2005;54 Suppl 2:S62–7. doi:10.2337/diabetes.54.suppl_2.s62
      OpenUrl
    8. ↵
      1. Carstensen B,
      2. Rønn PF,
      3. Jørgensen ME
      . Prevalence, incidence and mortality of type 1 and type 2 diabetes in Denmark 1996-2016. BMJ Open Diabetes Res Care 2020;8:1–9. doi:10.1136/bmjdrc-2019-001071
      OpenUrlPubMed
    9. ↵
      1. Tamayo T,
      2. Brinks R,
      3. Hoyer A, et al
      . Prävalenz und inzidenz von diabetes mellitus in Deutschland. Deutsches Aerzteblatt Int 2016;113:117–82. doi:10.3238/arztebl.2016.01
      OpenUrl
    10. ↵
      1. Tönnies T,
      2. Röckl S,
      3. Hoyer A, et al
      . Projected number of people with diagnosed type 2 diabetes in Germany in 2040. Diabet Med 2019;36:1217–25. doi:10.1111/dme.13902
      OpenUrlPubMed
    11. ↵
      1. Waldeyer R,
      2. Brinks R,
      3. Rathmann W, et al
      . Projection of the burden of type 2 diabetes mellitus in Germany: a demographic modelling approach to estimate the direct medical excess costs from 2010 to 2040. Diabet Med 2013;30:999–1008. doi:10.1111/dme.12177
      OpenUrlCrossRefPubMed
    12. ↵
      1. Reitzle L,
      2. Ihle P,
      3. Heidemann C, et al
      . Algorithmus Zur unterscheidung von diabetes mellitus typ 1 und typ 2 bei Der analyse von routinedaten. Gesundheitswesen 2022. doi:10.1055/a-1791-0918
    13. ↵
      1. Schmidt C,
      2. Reitzle L,
      3. Heidemann C, et al
      . Excess mortality in adults with documented diabetes in Germany: routine data analysis of all insurance claims in Germany 2013-2014. BMJ Open 2021;11:e041508. doi:10.1136/bmjopen-2020-041508
    14. ↵
      1. Federal Statistical Office of Germany
      . 14. coordinated population projection. 2019. Available: https://www.destatis.de/EN/Themes/Society-Environment/Population/Population-Projection/_node.html#sprg479856 [Accessed 21 Jan 2022].
    15. ↵
      1. Voeltz D,
      2. Brinks R,
      3. Tönnies T, et al
      . Future number of people with diagnosed type 1 diabetes in Germany until 2040: Zenodo; 2022.
    16. ↵
      1. Heidemann C,
      2. Paprott R,
      3. Schmidt C, et al
      . Aufbau einer Diabetes-Surveillance in Deutschland – Ergebnisse der ersten Projektphase 2015 – 2019. 2019.
    17. ↵
      1. Robert Koch-Institut
      . Diabetes surveillance. Available: https://diabsurv.rki.de/Webs/Diabsurv/DE/startseite/startseite-node.html [Accessed 13 Sep 2022].
    18. ↵
      1. Brinks R,
      2. Landwehr S,
      3. Icks A, et al
      . Deriving age-specific incidence from prevalence with an ordinary differential equation. Stat Med 2013;32:2070–8. doi:10.1002/sim.5651
      OpenUrlPubMed
    19. ↵
      1. World Health Organization (WHO)
      . European detailed mortality database. Available: data.euro.who.int/dmdb/ [Accessed 11 Aug 2015].
    20. ↵
      1. Carstensen B,
      2. Kristensen JK,
      3. Ottosen P, et al
      . The Danish national diabetes register: trends in incidence, prevalence and mortality. Diabetologia 2008;51:2187–96. doi:10.1007/s00125-008-1156-z
      OpenUrlCrossRefPubMedWeb of Science
    21. ↵
      1. Groop L,
      2. Tuomi T,
      3. Rowley M, et al
      . Latent autoimmune diabetes in adults (LADA) -- more than a name. Diabetologia 2006;49:1996–8. doi:10.1007/s00125-006-0345-x
      OpenUrlCrossRefPubMedWeb of Science
    22. ↵
      1. Rosenbauer J,
      2. Stahl A
      . Häufigkeit des diabetes mellitus Im kindes- und jugendalter in deutschland. Diabetologe 2010;6:177–89. doi:10.1007/s11428-009-0497-7
      OpenUrl
    23. ↵
      1. Voeltz D,
      2. Tönnies T,
      3. Brinks R, et al
      . Future prevalence of type 2 diabetes-A comparative analysis of chronic disease projection methods. PLOS ONE 2022;17:e0264739. doi:10.1371/journal.pone.0264739
    24. ↵
      1. Qeadan F,
      2. Tingey B,
      3. Egbert J, et al
      . The associations between COVID-19 diagnosis, type 1 diabetes, and the risk of diabetic ketoacidosis: a nationwide cohort from the US using the cerner real-world data. PLOS ONE 2022;17:e0266809. doi:10.1371/journal.pone.0266809
    25. ↵
      1. Watson C
      . Diabetes risk rises after COVID, massive study finds. Nature 2022. doi:10.1038/d41586-022-00912-y
    26. ↵
      1. Patterson CC,
      2. Dahlquist GG,
      3. Gyürüs E, et al
      . Incidence trends for childhood type 1 diabetes in Europe during 1989-2003 and predicted new cases 2005-20: a multicentre prospective registration study. Lancet 2009;373:2027–33. doi:10.1016/S0140-6736(09)60568-7
      OpenUrlCrossRefPubMedWeb of Science
    27. ↵
      1. Skyler JS
      . Primary and secondary prevention of type 1 diabetes. Diabet Med 2013;30:161–9. doi:10.1111/dme.12100
      OpenUrlCrossRefPubMed
    28. ↵
      1. Patterson CC,
      2. Harjutsalo V,
      3. Rosenbauer J, et al
      . Trends and cyclical variation in the incidence of childhood type 1 diabetes in 26 European centres in the 25 year period 1989-2013: a multicentre prospective registration study. Diabetologia 2019;62:408–17. doi:10.1007/s00125-018-4763-3
      OpenUrlCrossRefPubMed
    29. ↵
      1. Ostrauskas R,
      2. Žalinkevičius R,
      3. Jurgevičienė N, et al
      . The incidence of type 1 diabetes mellitus among 15-34 years aged Lithuanian population: 18-year incidence study based on prospective databases. BMC Public Health 2011;11:813. doi:10.1186/1471-2458-11-813
    30. ↵
      1. Tönnies T,
      2. Rathmann W,
      3. Hoyer A, et al
      . Quantifying the underestimation of projected global diabetes prevalence by the International diabetes Federation (IDF) diabetes atlas. BMJ Open Diabetes Res Care 2021;9:e002122. doi:10.1136/bmjdrc-2021-002122
    31. ↵
      1. Schmidt C,
      2. Reitzle L,
      3. Dreß J, et al
      . Prävalenz und Inzidenz des dokumentierten Diabetes mellitus – Referenzauswertung für die Diabetes-Surveillance auf Basis von Daten aller gesetzlich Krankenversicherten. Bundesgesundheitsbl 2020;63:93–102. doi:10.1007/s00103-019-03068-9
      OpenUrl
    32. ↵
      1. Brinks R,
      2. Tönnies T,
      3. Hoyer A
      . Impact of diagnostic accuracy on the estimation of excess mortality from incidence and prevalence: simulation study and application to diabetes in German men. F1000Res 2021;10:49. doi:10.12688/f1000research.28023.1
      OpenUrl
    33. ↵
      1. Brinks R,
      2. Tönnies T,
      3. Hoyer A
      . Assessing two methods for estimating excess mortality of chronic diseases from aggregated data. BMC Res Notes 2020;13:216. doi:10.1186/s13104-020-05046-w
    34. ↵
      1. Monge L,
      2. Bruno G,
      3. Pinach S, et al
      . A clinically orientated approach increases the efficiency of screening for latent autoimmune diabetes in adults (LADA) in a large clinic-based cohort of patients with diabetes onset over 50 years. Diabet Med 2004;21:456–9. doi:10.1111/j.1464-5491.2004.01177.x
      OpenUrlPubMed
    35. ↵
      1. Zimmet P,
      2. Turner R,
      3. McCarty D, et al
      . Crucial points at diagnosis. Diabetes Care 1999:B59–63. Available: https://www.proquest.com/openview/861c1cab74e5b7e47de6a947fd92a8de/1?pq-origsite=gscholar&cbl=47715
    36. ↵
      1. Biesenbach G,
      2. Auinger M,
      3. Clodi M, et al
      . Prevalence of LADA and frequency of GAD antibodies in diabetic patients with end-stage renal disease and dialysis treatment in Austria. Nephrol Dial Transplant 2005;20:559–65. doi:10.1093/ndt/gfh662
      OpenUrlCrossRefPubMed
    37. ↵
      1. Gregg EW,
      2. Zhuo X,
      3. Cheng YJ, et al
      . Trends in lifetime risk and years of life lost due to diabetes in the USA, 1985-2011: a modelling study. Lancet Diabetes Endocrinol 2014;2:867–74. doi:10.1016/S2213-8587(14)70161-5
      OpenUrl
    38. ↵
      1. Ringborg A,
      2. Lindgren P,
      3. Martinell M, et al
      . Prevalence and incidence of type 2 diabetes and its complications 1996-2003 -- estimates from a Swedish population-based study. Diabet Med 2008;25:1178–86. doi:10.1111/j.1464-5491.2008.02541.x
      OpenUrlCrossRefPubMedWeb of Science
    39. ↵
      1. Kamrath C,
      2. Rosenbauer J,
      3. Eckert AJ, et al
      . Incidence of type 1 diabetes in children and adolescents during the COVID-19 pandemic in Germany: results from the DPV registry. Diabetes Care 2022;45:1762–71. doi:10.2337/dc21-0969
      OpenUrl
    40. ↵
      1. Zilbermint M
      . Diabetes and climate change. J Community Hosp Intern Med Perspect 2020;10:409–12. doi:10.1080/20009666.2020.1791027
      OpenUrlCrossRefPubMed
    41. ↵
      1. Vallianou NG,
      2. Geladari EV,
      3. Kounatidis D, et al
      . Diabetes mellitus in the era of climate change. Diabetes Metab 2021;47:101205. doi:10.1016/j.diabet.2020.10.003
      OpenUrl
    42. ↵
      1. Seligman HK,
      2. Bindman AB,
      3. Vittinghoff E, et al
      . Food insecurity is associated with diabetes mellitus: results from the National health examination and nutrition examination survey (NHANES) 1999-2002. J Gen Intern Med 2007;22:1018–23. doi:10.1007/s11606-007-0192-6
      OpenUrlCrossRefPubMedWeb of Science
    43. ↵
      1. Anekwe TD,
      2. Rahkovsky I
      . The association between food prices and the blood glucose level of US adults with type 2 diabetes. Am J Public Health 2018;108:S475–82. doi:10.2105/AJPH.2013.301661r
      OpenUrl

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Footnotes

    • Contributors DV performed the analytic calculations, that is, implemented the methods and scenarios and analyzed the data. All coauthors verified the analytical methods, the results, and the main conceptual ideas. Furthermore, DV took the lead in writing the manuscript and, together with AH, was in charge of overall direction and planning. All authors provided critical feedback and helped shape the research, analysis, and manuscript. DV accepts full responsibility for the finished work and the conduct of the study, had access to the data, and controlled the decision to publish.

    • Funding The support for the publication costs was provided by the Open Access Publication Fund of Bielefeld University.

    • Competing interests None declared.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.