Norwegian DKA admissions fell over a decade, but repeat episodes still marked a high-risk subgroup
Admissions for diabetic ketoacidosis among adults with type 1 diabetes declined steadily over 10 years at a Norwegian university hospital, even as the surrounding catchment population grew. The drop points toward better background diabetes care, wider use of modern glucose monitoring and insulin delivery technologies, and faster intervention before metabolic decompensation reaches the point of hospitalization.
The decline did not erase the clinical burden. Recurrent admissions remained common, psychosocial vulnerability clustered tightly around repeat events, and mortality after the index episode stayed clinically meaningful over five years.
A 10-year fall in admissions did not eliminate severe disease
Between 1 August 2012 and 31 August 2022, 178 adults contributed 306 admissions. Median age at admission was 34 years, with an interquartile range of 22 to 51 years, and median HbA1c was 88 mmol/mol, indicating very poor glycemic control in many patients at the time of presentation. The study period captured a broad era of diabetes technology, yet DKA remained a recurring reason for emergency admission in a population already living with type 1 diabetes.
Monthly admission counts decreased by an estimated 6.9% per year. The rate ratio was 0.931, with a 95% confidence interval of 0.896 to 0.968 and a p value below 0.001. Over a decade, that pattern is hard to dismiss as random fluctuation. It suggests that improvements in outpatient diabetes management, patient education, and access to technology may be reducing at least some severe metabolic crises before they require inpatient treatment.
The study design matters here because the hospital served a large and growing population. In 2022, the catchment area covered about 288,000 residents, including about 232,000 aged 16 or older. Even against that expanding denominator, the admission trend moved downward.
Insulin omission, pump failure and psychiatric illness clustered around admissions
The most frequent precipitating cause of ketoacidosis was forgotten or deliberately omitted insulin doses. In plain clinical terms, the pathophysiology is unforgiving: when insulin delivery stops, lipolysis accelerates, ketone production rises, and dehydration compounds the metabolic collapse. In a disease defined by absolute insulin deficiency, missed insulin remains a direct route to emergency care.
A history of psychiatric illness appeared in 40% of admissions in which insulin had been forgotten or intentionally omitted. That overlap is clinically important because it links metabolic instability to mental health, treatment burden and self-management capacity. The admission is not just an endocrine event; it often reflects disrupted care, emotional distress or competing life circumstances that make insulin adherence difficult.
Among admissions among people with pre-existing type 1 diabetes, 83 involved insulin pump use, and 24 were attributed to pump failure. Pump therapy reduces injection burden and can improve glycemic control, but it also introduces a failure mode that can trigger rapid insulin deprivation. In type 1 diabetes, that can translate into ketoacidosis within hours rather than days.
Recurrent DKA tracked with psychosocial vulnerability
Repeated admissions were concentrated in a minority of patients. Forty participants, or 22.5%, had more than one admission during the study period. Recurrence was not random. Psychiatric illness and receipt of disability benefits were associated with experiencing more than one lifetime admission, with adjusted odds ratios of 2.827 and 4.941, respectively.
The signal is clinically uncomfortable but familiar. Recurrent DKA often sits at the intersection of diabetes self-management, mental illness, social stressors and fragmented follow-up. Disability benefit status may reflect severe underlying illness, functional limitations or broader socioeconomic vulnerability. None of those factors alone explains recurrence, but together they identify a population that standard acute care discharge planning may fail to reach.
Prior diabetic ketoacidosis, suspected treatment non-adherence and psychiatric illness were also associated with higher readmission rates during follow-up. In other words, the first episode often was not the last. Once DKA recurs, the next admission becomes more likely unless care is built around adherence barriers, mental health needs and practical support for insulin access and monitoring.
Mortality stayed low in hospital but substantial over five years
In-hospital mortality was 1.6%, which is low in a high-resource setting and consistent with modern emergency and critical care management of ketoacidosis. The short-term survival data should not obscure the longer arc. Five-year cumulative mortality reached 14.7%.
Alcohol or substance misuse was associated with mortality during follow-up, with a hazard ratio of 2.499 and a 95% confidence interval of 1.191 to 5.245. Higher age at the index admission also carried a strong mortality association, with a hazard ratio of 1.874 per 10 years. Those are not abstract epidemiologic signals. They identify a cohort that enters the hospital with a metabolic emergency and leaves with a materially elevated risk of death over the next several years.
The mortality finding broadens the meaning of a DKA admission. The episode is often treated as an acute reversible crisis, yet the follow-up data suggest it should function as a marker of future vulnerability. Age, substance misuse and recurrent metabolic instability appear to capture much of that risk.
Biochemical recovery was measured, but the larger problem was what happened after discharge
The study also tracked time to biochemical normalisation and length of stay, using the first arterial blood gas sample as the anchor for metabolic recovery. Those endpoints matter because they reflect the immediate response to fluids, insulin and electrolyte replacement. Still, the more consequential pattern emerged outside the ward: recurrent admissions and post-discharge mortality.
For clinicians, that means the admission itself should trigger a broader review of diabetes management and psychosocial risk. A discharge plan focused only on insulin titration will miss patients whose next decompensation is driven by psychiatric illness, non-adherence, alcohol or substance misuse, or pump-related failure. The biological insult is acute, but the drivers are often chronic.
The cohort also included a substantial number of admissions in which ketone values were unavailable or below the usual threshold but the clinical picture still supported ketoacidosis. Those cases were retained after chart review when the treating endocrinologist’s assessment, presentation and response to treatment were consistent with DKA. This approach reflects real-world practice, where documentation is imperfect and the bedside diagnosis often rests on the full clinical context rather than a single lab value.
Targeted follow-up after DKA looks more justified than ever
The findings support a practical reading of the data: fewer admissions overall, but persistent high-risk subgroups that need more than routine outpatient follow-up. Psychosocial vulnerability, especially psychiatric illness and substance misuse, sat close to the center of both recurrence and mortality. Disability benefits and prior DKA identified additional risk layers.
The study cannot explain which specific intervention would best prevent the next admission, and it does not establish causality for the associations it found. It does show, with a decade of hospital data, that declining admission counts can coexist with a stubborn burden of repeat ketoacidosis and delayed mortality. For adults with type 1 diabetes who have already had DKA, the next step is not just better glucose control. It is structured follow-up that can detect adherence failure, mental health problems and substance misuse before another crisis develops.