Time Trends in the Incidence of Diabetic Ketoacidosis Leading to Hospital Admission Among Adults With Type 1 Diabetes: A Nationwide Danish Register Study Free

Diabetic ketoacidosis (DKA) is a major acute metabolic complication of type 1 diabetes characterized by metabolic acidosis, ketosis, and, for most cases, hyperglycemia. Although mortality rates from DKA are declining (1), DKA still impacts morbidity and mortality in young ages and remains one of the biggest contributors to loss of life-years associated with type 1 diabetes (2). Therefore, monitoring the incidence of DKA is important in evaluating current diabetes management and to plan for health services and delivery. In Denmark, ∼30,000 people have type 1 diabetes, which constitutes ∼10% of the total diabetes population. The annual incidence of DKA in Denmark was estimated at 12.9 per 100,000 during the period 1996–2002 (3) in a population including people with type 1 and type 2 diabetes.

Even though the treatment of type 1 diabetes has improved, estimates of the incidence of DKA among individuals with type 1 diabetes and development over time are still important. The incidence of DKA in type 1 diabetes was previously shown to increase after early childhood, plateau in early adulthood, and thereafter decrease (4–8). Throughout life, the incidence is proven highest in girls and women (5–8). The increasing use of insulin pump therapy has contributed to better glycemic control of type 1 diabetes over the last decade. Also, insulin pump therapy is known to increase the risk of DKA (9) but has also been associated with a lower risk of DKA among young people with type 1 diabetes (10,11) and in observational studies was shown not to increase the risk of DKA in large diabetes clinics with sufficient support and patient education (12).

A recently published study of a large Danish cohort of adults with type 1 diabetes showed a marginally increased risk of DKA hospitalization. In this cohort with 243,601 person-years (PY) of observation from 2010 to 2020, 38,823 (16%) of these were insulin pump-treated PY (13). Although the small increase in risk is of no greater clinical significance, with only a 1.8% increase in risk compared with multiple daily injections users, the increased use of insulin pump therapy and the benefit-risk profiles of the potential new treatment with sodium–glucose cotransporter inhibitors (SGLTi), which includes an increased risk of DKA, amplifies the importance of elucidating the current incidence of DKA in type 1 diabetes as well as monitoring the incidence. Also, although a life-threatening condition, DKA is potentially preventable. Understanding time trends in DKA in type 1 diabetes is a first step toward evaluating implemented prevention strategies and guiding the quality improvement work related to DKA risk management in type 1 diabetes. Therefore, this study aimed to quantify the risk of a first event of DKA and describe the trend of DKA incidence among adults with type 1 diabetes in Denmark to evaluate our current management of diabetes and the risk of DKA.

This is a retrospective population-based cohort study using Danish national administrative health registers and databases. The registers are nationwide and cover all residents. A unique civil registration number is given to all Danish residents at birth or at immigration and is recorded in the Danish Civil Registration System (14). Therefore, it is possible to cross-link registers and databases at the individual level and obtain complete follow-up.

The study population consists of adults (aged ≥18 years) living in Denmark between 1996 and December 2019 and identified as individuals with type 1 diabetes based on data from the Danish health care registers and databases containing diabetes-defining information. A detailed description of the identification of individuals with type 1 diabetes can be found in a study from our group from 2020 (15). In short, the cohort was identified from the Registry of Medicinal Products Statistics (16), the Danish Adult Diabetes Database (DADD) (17), the Danish National Patient Register (NPR) (18), the Danish National Health Service Register (19), and the Danish Clinical Quality Assurance Database for Screening of Diabetic Retinopathy and Maculopathy (DiaBase) (20). Individuals were classified as having diabetes with a proxy for a diabetes diagnosis date being the earliest date of any of the following: 1) first occurring diagnosis of diabetes (International Classification of Diseases-8 codes: 249 and 250, ICD-10 codes: E10 and E11, with the exclusion of gestational diabetes) in the Danish National Patient Register (valid from 1977); 2) first occurring use of diabetes podiatry in Danish National Health Service Register (valid from 1990); 3) first date of purchase of any antidiabetes medication (Anatomical Therapeutic Chemical Code [ATC] A10xxx) in the Registry of Medicinal Products Statistics (valid from 1995); 4) earliest mentioned date of diagnosis in DADD; or 5) earliest date of an eye examination recorded in DiaBase (valid from 2009).

Individuals were classified as having type 1 diabetes if the criteria for diabetes and any of the following criteria were met: 1) purchase of insulin before age 30 years; 2) classified as type 1 diabetes in the most recent >50% of the individual’s DADD records; or 3) not classified in DADD but with the most recent >50% of the records from NPR being classified as type 1 diabetes. An individual could not be classified as having type 1 diabetes if they had no recorded date of insulin purchase.

Access to and use of the data was granted by the Danish Data Protection Agency, and all data were anonymized. According to Danish law, ethical approval and patient consent is not required for register studies.

Data on hospitalization with a DKA diagnosis was ascertained from the Danish National Patient Register (18). We defined DKA based on the specific ICD-10 codes DE101, DE111, DE121, DE131, or DE141, which were given by the treating physician. In Denmark, all individuals with type 1 diabetes developing DKA are admitted to the hospital. All admissions trigger reporting to national patient registers where ICD-10 codes are used to indicate diagnoses and comorbidities.

For follow-up, death and emigration was obtained by linkage to the Central Person Register (14). Data for this analysis was obtained from 1 January 1996 to 31 December 2019.

Individuals were followed from the date of type 1 diabetes diagnosis to a first event of DKA, death, or end of follow-up (31 December 2019). As we were interested in DKA incidence after diabetes diagnosis, we did not include DKA 1 month before and after type 1 diabetes diagnosis or DKA events before the age of 18 years. The follow-up time was split along current age in 0.5-year intervals. We modeled the incidence rate of DKA by sex, age, and calendar time using Poisson regression for time-split data. Age and calendar time were included in the models as natural splines with three and four knots, respectively. To adjust for different age effects for men and women, an interaction term between age and sex was fitted in the model. A complete-case analyses approach was used. Statistical analyses were performed using R 4.0.2 software (R Foundation for Statistical Computing, Vienna, Austria; www.R-project.org). The Epi-package 2.44 was used to apply the models (21).

The cohort consisted of 24,718 adults with type 1 diabetes, and 58% (n = 14,457) were men. Of the total cohort, 24% (n = 6,728) were diagnosed with type 1 diabetes before the age of 25 years (including diagnosis before the age of 18 years [n = 4,373]). There were a total of 2,268 episodes of DKA. Age at first event of DKA, occurring after the age of 18, was distributed with 32% in the ages of 18 to 25 years, 31% in ages 25 to 40 years, and only 6% in ages 70 to 100 years, as shown in Table 1.

The incidence rate of first event of DKA per 100 PY decreased with older age for both men and women with type 1 diabetes, as shown in Fig. 1 (left panel). From the age of ≥35 years, the decline in incidence rates were less pronounced. From 20 to 30 years of age, the DKA incidence rate decreased from 3.27 to 1.40 per 100 PY, from 30 to 40 years of age from 1.40 to 0.83 per 100 PY, and from 40 to 80 years of age from 0.83 to 0.38 per 100 PY (the incidence rates given are for men).

From 1996 to 2008, the incidence rate of DKA increased for all age-groups, with a subsequent minor decrease in incidence rate thereafter until end of follow-up (Fig. 1, right panel). From select points on the incidence curves, we saw that from 1996 to 2008, the incidence rates increased from 1.91 to 3.77 per 100 PY for a 20-year-old, from 0.82 to 1.62 per 100 PY for a 30-year-old, from 0.49 to 0.96 per 100 PY for a 40-year-old, and from 0.22 to 0.44 per 100 PY for an 80-year-old man living with type 1 diabetes. From 2008 to 2020, the incidence rates decreased from 3.77 to 3.27 per 100 PY for a 20-year-old, from 1.62 to 1.40 per 100 PY for a 30-year-old, from 0.96 to 0.83 per 100 PY for a 40-year-old, and from 0.44 to 0.38 per 100 PY for an 80-year-old man living with type 1 diabetes.

Based on nationwide registers, we have quantified the risk of incident DKA according to age and described the trend of DKA among adults with type 1 diabetes in Denmark. We have demonstrated that 1) the incidence rates for both men and women decrease from 18 to 80 years of age and 2) the incidence rate of DKA increased for all age-groups between 1996 and 2008, with a subsequent minor decrease in incidence rate thereafter until end of follow-up.

Information about the epidemiology of type 1 diabetes in adults is limited, mostly because many studies categorize those with high levels of blood glucose as having diabetes without giving information on the proportions of individuals with type 1 diabetes versus type 2 diabetes (22). Therefore, there are few studies with complete national population data and the ability to stratify by diabetes type. Here, we present population-based data from a nationwide cohort based on highly valid and complete Danish registries. This allows us to distinguish between different types of diabetes and to focus on the incidence rates of hospitalization for DKA in type 1 diabetes.

In our study, we observed a decrease in DKA incidence rates for both men and women ≥18 years of age. In 2007, similar results were observed in the Danish study by Henriksen et al. (3). Here, they demonstrated that 36% of the 4,807 episodes of DKA occurred in individuals aged <30 years, 27% in individuals aged 30 to 50 years, 23% in individuals aged 51 to 70 years, and 14% in individuals aged >70 years. However, data included all admissions with a discharge diagnosis of DKA regardless of type of diabetes. Decreasing incidence rates with increasing age are also observed in other countries (4,5). In data from three multinational type 1 diabetes registries/audits in Austria, Germany, England, Wales, and U.S., including 49,859 individuals aged <18 years with type 1 diabetes, the risk of DKA increased after early childhood, plateaued in early adulthood, and thereafter decreased with increasing age (6). In contrast to our study, this study included individuals aged <18 years and did not end follow-up at the first event of DKA but included recurrent events also. The decrease in incidence of DKA in adulthood might be based on longer diabetes duration, and thereby a better management of the diabetes, and leaving childhood/young adulthood and, for most, entering a more stable period of life.

In our study, the incidence rate increased from 1996 until 2008 with a subsequent decline for all age-groups until the end of follow-up. These findings contrast with results from similar studies. A recent study from 2021 with a cohort of 38,000 Scottish individuals with type 1 diabetes showed an increase in DKA event rates from 2004 to 2018 (23). That study included all individuals with type 1 diabetes alive in Scotland between 1 January 2004 and 31 December 2018 and included all episodes of DKA, with 72.6% of the cohort never being admitted to a hospital with DKA, 15.1% with one admission, and 12.3% with multiple admissions.

Similar trends were observed in a retrospective cohort study from Finland that aimed to determine the long-term incidence and predictors of DKA in adults with type 1 diabetes of long duration. The study included all hospitalizations and deaths due to DKA between 1996 and 2015, including events that occurred prior to enrollment and baseline visits, in ∼20,000 individuals with type 1 diabetes. In contrast to our study, the study also included individuals with more than one event of DKA. The study observed an increasing incidence rate during the study period (24). Of interest, this study also found that individuals with a least one episode of DKA had a higher frequency of insulin pump use, poorer glycemic control and lipid profile, lower age at onset, lower BMI, and a greater alcohol consumption. In an observational study from the U.S. with data from >20 million individuals, an increase in DKA hospitalization rates was observed for all age-groups between 2009 and 2014, highest among those aged <45 years (25). However, there was no distinction between type 1 and type 2 diabetes in this study.

The incidence rates per calendar year vary from country to country and between different studies. Methodological differences may be contributing to this difference, and in the studies mentioned above, including all DKA events and not just the first event affects the estimates when looking at the incidence of DKA. Also, the differences between health care systems are important to consider. Access to medical care for all might help prevent some of the triggering causes since human factors are one of the biggest contributors to DKA, with some of them preventable (26). Measures regarding better glycemic control have also improved over the last decades, among others, the increasing use of insulin pumps and glucose sensor monitoring, which improve glycemic control (13,27). We can only speculate that the decrease in the incidence rate we have seen since 2008 is a result of better glycemic control, as we previously showed in a large national demography report from 2020 (28), and greater use of glucose monitoring systems.

Quantifying the risk of incident DKA and thereby describing the trend of DKA among adults with type 1 diabetes is important. Not only regarding planning for health care delivery but also with the increasing use of insulin pump therapy and glucose sensor monitoring and the benefit and risk profiles of new antidiabetes therapies, there is a need to characterize the current incidence of DKA as well as follow the incidence. The use of insulin pump therapy has grown markedly in Denmark. A large cohort study from Denmark just published, including 26,687 people with type 1 diabetes with 16% having an insulin pump, showed a small increase in the risk of DKA hospitalization in insulin pump users corresponding to 0.52 additional hospitalizations per 1.000 PY (13). Another Danish study from 2018 evaluated the incidence rates of DKA in individuals with type 1 diabetes in Denmark according to treatment with multiple daily injections or an insulin pump. The study found no significant difference in the incidence rate of DKA between the two groups. However, the study did find a difference in incidence rates when larger clinics compared with smaller clinics, indicating the importance of sufficient support and patient education, which might be more achievable in the larger clinics (12). A meta-analysis also just published on the effectiveness of continuous glucose monitoring among people with type 1 diabetes found no effect on the number of DKA events, although the study concluded that they could not confer a concrete conclusion on the effectiveness of continuous glucose monitoring on DKA outcome as DKA incidences were rare (29).

SGLT 1 and 2 inhibitors were approved as an adjunct to insulin treatment in type 1 diabetes in Europe and Japan in 2019. However, this approval was withdrawn again in 2021 (30–32). Because SGLTi were only recently approved for type 1 diabetes and with very limited use, they are unlikely to have contributed to the DKA trend in Denmark. However, if the cardiovascular- and kidney-protective effects seen in high-risk individuals with type 2 diabetes are proven in type 1 diabetes, the use of the SGLT2 inhibitors in type 1 diabetes may increase (33). Our study group recently collected data in a retrospective observational study of 134 people with type 1 diabetes treated with an SGLT2 inhibitor in Denmark and found that none of the participants developed DKA (34).

This study has several strengths: 1) we used reliable data based on highly valid and complete Danish registries; 2) the population was large, and we had many years of data; 3) our ability to validly distinguish between different types of diabetes allowed us to focus on type 1 diabetes; and 4) all analyses are fully documented and can be used for future monitoring to ensure consistent methodology over time.

Limitations include 1) we did not analyze the incidence of recurrent events, which constitutes a large part of the total burden when estimating the incidence of DKA; and 2) in the current work we did not analyze the impact of the increased use of insulin pumps and glucose monitoring over recent years.

The incidence rates of DKA are declining over age and with an overall decline from 2008 for men and women. This may suggest improved diabetes management for individuals with type 1 diabetes in Denmark.

Duality of Interest. F.P. has served as a consultant, on advisory boards, or as an educator for AstraZeneca, Novo Nordisk, Boehringer Ingelheim, Sanofi, Mundipharma, MSD, Novartis, and Amgen and has received research grants to the institution from Novo Nordisk, Boehringer Ingelheim, Amgen, and AstraZeneca. K.N. serves as an adviser to Medtronic, Abbott, Convatec, and Novo Nordisk, owns shares in Novo Nordisk, has received research grants to the institution from Novo Nordisk, Zealand Pharma, Dexcom, and Medtronic, and has received fees for speaking from Medtronic, Novo Nordisk, and Dexcom. D.V. has received research grants from Bayer A/S, Sanofi, Novo Nordisk A/S, and Boehringer Ingelheim and holds shares in Novo Nordisk A/S. P.R. has received consultancy and/or speaking fees (to Steno Diabetes Center Copenhagen) from Abbott, Astellas, AstraZeneca, Bayer, Boehringer Ingelheim, Eli Lilly, Gilead, MSD, Novo Nordisk, and Sanofi, and research grants from AstraZeneca, Bayer, and Novo Nordisk. H.A. was affiliated with Steno Diabetes Center Copenhagen during finalization of the study analysis and is currently employed at Novo Nordisk A/S.

Author Contributions. E.B.S. wrote and edited the manuscript and finalized it for submission. H.A. and B.C. did the statistical work. H.A., E.S., M.E.J., K.N., P.R., F.P., and D.V. contributed to discussion and editing of the manuscript. H.A., B.C., M.E.J., and D.V. collected data. D.V. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were in abstract form at the 59th Annual Meeting of the European Association for the Study of Diabetes (EASD), Hamburg, Germany, 2–6 October 2023.