In a cohort of adult patients with type 1 diabetes who were hospitalized with diabetic ketoacidosis (DKA) in Edinburgh, Scotland, almost all patients survived the hospitalization, but they had an increased risk of early death, new research shows.

After one hospitalization for DKA, patients had a 10% greater risk of dying within 5 years, but, importantly, they had a 30% increased risk of dying within 6 years after more than four hospitalizations for DKA.

It was also "really quite striking" that the patients had a median age of 31 when they died after multiple DKA hospitalizations, and many died at home, often from "uncertain causes," said Dr Fraser W Gibb (Edinburgh Centre for Endocrinology & Diabetes, Scotland) presenting this study at the American Diabetes Association (ADA) 2015 Scientific Sessions. Therefore, there is an urgent need for greater efforts to manage psychological and medical morbidity in this young, high-risk population and for prospective, larger studies to better identify mortality risk, he said.

This was a "very scary, sobering" presentation, said session chair Dr Kasia J Lipska (Yale University, New Haven, CT).

"Diabetologists need to be more concerned that patients with recurring DKA have a high risk of mortality," she told Medscape Medical News. They should look for substance abuse or mental-health problems that often precipitate DKA in diabetic patients and refer patients to appropriate healthcare providers. "We see these patients who bounce around from hospital to hospital with DKA," she said. "It's a big problem, so I'm glad that this presentation brought that to light."

Dr Guillermo Umpierrez (Emory University, Atlanta, Georgia), who has conducted research in patients with DKA, agrees. "In the United States, and [probably also] in Canada and Europe, those who have recurrent DKA have a high rate of alcohol intake and drug abuse and they're poorly educated and more likely to have comorbidities that precipitate DKA," he told Medscape Medical News. This is what further research needs to determine: "Is it just because you have a DKA episode or is it something else [that increases risk of death]?"

Multiple Causes of DKA

DKA — which is blood glucose of 250 mg/dL or higher with ketonuria, ketonemia, and metabolic acidosis (pH < 7.3 or blood bicarbonate levels < 15 mEq/L — generally occurs in patients with type 1 diabetes but may occur in patients with type 2 diabetes under stress conditions such as trauma, surgery, or infection.

"We diagnose DKA and we treat it, but we also look for why patients went into DKA," Dr Lipska said. DKA may be caused by discontinuing or not taking doses of insulin or it may be triggered by infection, myocardial infarction, pancreatitis, stroke, trauma, alcohol abuse, new onset of type 1 diabetes, or drugs (steroids, thiazides, sympathomimetics).

As reported earlier by Medscape Medical News, the US Food and Drug Administration (FDA) recently warned that sodium-glucose cotransporter-2 (SGLT2) inhibitors used to treat type 2 diabetes may lead to ketoacidosis requiring hospitalization. The agency advised that patients should be evaluated for signs or symptoms of ketoacidosis — difficulty breathing, nausea, vomiting, abdominal pain, confusion, and unusual fatigue and sleepiness — and if acidosis was confirmed, to stop the drug, address the acidosis, and monitor glucose levels.

Mortality with DKA used to be close to 100% in the 1920s before the discovery of insulin, then gradually dropped to around 15% in the 1970s and is now less than 1% in the United States, Dr Umpierrez said. Dr Gibb said the current rate of DKA is the same in Scotland, but little is known about death related to hospitalization for DKA.

What Predicts Death From DKA?

The researchers aimed to investigate clinical features of DKA and differences between patients who had one vs multiple hospitalizations for DKA and to examine predictors and rates of DKA mortality.

Dr Gibb and colleagues retrospectively identified 298 adult patients with type 1 diabetes who were admitted to their hospital for DKA from 2007 to 2014 (a total of 631 admissions). They also obtained death records for these patients until 2014.

The patients had a mean age of 32, and most were 20 to 40 years old when they presented to the hospital emergency department with DKA.

There were no significant gender differences in people presenting with DKA — 164 men (55%) and 134 women (45%) — or in the numbers of hospitalizations — 339 in men (54%) and 289 in women (46%). There were also no significant differences in HbA1c levels in patients.

Most patients (189 patients; 63.4%) had only one hospitalization for DKA, but 27 patients (9.1%) had more than four hospitalizations for DKA.

There were no inpatient deaths during DKA admission. However, 43 patients (14.4%) died during about 5 years of follow-up.

The median time to death was 390 days, "but within that, about a quarter of patients were dying within 2 months of the last admission," Dr Gibb noted. In this subgroup of 10 patients with death within 2 months, eight of the deaths occurred at home. The patients had a median age of 41. The cause of death was "unclear" in five patients and was due to cardiovascular events, poisoning, sepsis, or pulmonary thromboembolism in the other patients. "These are largely unanticipated and arguably preventable deaths with unclear cause of death in most patients," Dr Gibb said.

About one in 10 patients (10.6%) with a single hospitalization due to DKA died within 4.6 years; 18.3% of patients with two to four hospitalizations died within 5 years, and 29.7% of patients with more than four hospitalizations died within 6.4 years (P = .016). Although the follow-up was slightly longer in patients with more than four hospitalizations, with adjustment, this was still a two- to threefold higher risk, Dr Gibb stressed.

The patients who were most likely to die during follow-up had a longer hospital stay during DKA admission (P < .001), higher plasma urea levels (P < .001), and a longer duration of diabetes (P = .004) and were older when they had their last DKA admission (P = .001). HbA1c and social class were not associated with mortality.

Factors associated with a significantly increased risk of dying during follow-up after hospitalization for DKA included psychological issues (odds ratio [OR], 2.26), peripheral neuropathy (OR, 2.56), ischemic heart disease (OR, 4.06), alcohol excess (OR, 2.57) and prior admission to intensive care (OR, 2.46).

On the one hand, inpatient DKA deaths were extremely rare in their center, Dr Gibb summarized. However, the subsequent risk of death was high in patients with type 1 diabetes with multiple episodes of DKA requiring hospitalization, and many of the deaths occurred suddenly in young people.

"It does raise the question of whether greater efforts are required to deal with psychological issues and manage medical risk" in these young, at-risk patients, he concluded.

The authors have no relevant financial relationships.

American Diabetes Association 2015 Scientific Sessions; June 7, 2015; Boston, Massachusetts.Abstract 273-OR



www red DiabetologNytt

2 of 2