Although the mortality rate associated with DKA has significantly reduced, hospitalizations for DKA are increasing. Identifying contributing factors for DKA admission could help target prevention efforts.

Aims: To explore risk factors for DKA admissions, investigate clinical characteristics of DKA among hospitalized patients, and describe inpatient DKA management and discharge plan.

Methods: We performed a retrospective analysis of adult patients’ medical records for DKA admissions from April-December 2017. Admission records from the year prior to 2017 review time period were also reviewed.

Results: Of 200 patients, 79 (39.5%) had pre-existing T1D, 94 (47%) had pre-existing T2D, 12 (6%) had new-onset DM, and 15 (7.5%) had other type of diabetes. Metabolic acidosis was more severe in patients with T1D than in those with T2D. Most patients (61.8%) were admitted to ICU. Mean ICU stay was 1.6±3.1 days, and mean hospital stay was 5.4±9.0 days. Insulin (93%) and fluid (95%) infusion were used for initial DKA treatment. Sixty-four (35%) had ≥1 hypoglycemic event, and 29 (14.5%) had a recurrent DKA event during hospitalization. Mortality rate was 2.5% (n=5); 1 patient death (0.5%) was directly related to DKA. Prior DKA admission was observed in 62 (31%) with mean of 2.1±3.9 admissions and were more common in African Americans, patients 18-39 years of age, and those with a history of bipolar disorder or pancreatitis. Nonadherence to home insulin regimen was the predominant precipitating cause of prior DKA admissions. Majority of patients (84.5%) were discharged to home, of whom 41% received medication delivery, 64.2% had a scheduled follow-up appointment, and 21% had a home healthcare referral.

Conclusions: Although patients with T1D had more severe acidosis, DKA admission in patients with T2D is as common as in those with T1D. Nonadherence is the leading precipitating factor of DKA readmissions. Future efforts should focus on reducing DKA recurrence and admission among vulnerable populations.


A. Liu: None. K. Carmichael: None. M. Schallom: None. C. Arroyo: None.

Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at