One-Year Incidence, Demografic and Laboratory Caracteristics of Patients with Diabetic Ketoacidosis at the Emergency Department of a University Hospital in Brazil

Presentation Number: MON 616
Date of Presentation: April 3rd, 2017

Rachel Teixeira Leal Nunes*1, Paulo Ricardo Gessolo Lins2, Igor Gouveia Pietrobom2, Thaís Castanheira de Freitas3, Fernanda Correia Salles2, Pedro Henrique Luiz da Silva2, Ludmila de Andrade Barberino2, Carolina Frade Magalhães Girardin Pimentel Mota2, Alvaro Nagib Atallah2 and Aécio Flávio Teixeira Gois2
1Universidade Federal de São Paulo, São Paulo, BRAZIL, 2Universidade Federal de São Paulo, 3Faculdade de Medicina da Universidade de São Paulo


Diabetic ketoacidosis (DKA) is the most common acute hyperglycemic complication of diabetes mellitus (DM). It is associated with significant morbidity and mortality, with a mortality rate of 2-10%. Hyperglycemia, excessive ketone bodies production, metabolic acidosis, and dehydration characterize DKA. Current diagnostic criteria and classification of DKA are based on measurements of plasma glucose, pH, serum bicarbonate, anion gap, and mental status alterations. Most cases seem to be triggered by cessation of insulin and by infection.

The aims of this study were assessing the demografic and laboratory characteristics of patients with DKA who were admitted to Emergency Department (ED) and investigating the relationship between paraclinical (glucose, anion gap, and serum bicarbonate) and clinical prognosis in patients with DKA.

Methods: The incidence of DKA was studied during a 12-month period (June 1st 2015 - May 30th2016) at the ED of a teaching hospital in Brazil (Hospital São Paulo-Universidade Federal de São Paulo). The data collected included the demographic data, precipitating factors, DKA fluid and insulin therapy regimen, length of hospital stay. and mortality rate.

Results: A total of 52 patients were included in the study (out of 5943 admissions during that period); aged 15-72 years (median age was 28). Of these, 23 (58%) were female. The majority of cases (94%) consisted of moderate or severe episodes (bicarbonate less than 15 mmol/l, arterial pH < 7,24). Newly diagnosed DM accounted for 8 (25%) of the cases; the remainder were previously known type 1 (32) or type 2 (12) DM patients. Mortality rate was low, 5% (3). The main precipitating factors of DKA were infection/sepsis (48%) and insulin treatment cessation/inadequate insulin therapy (33%). 3 patients (5%) evaded treatment despite medical advice an had to be excluded from the final statistical analyses. Mean glucose was 524 (±191) mg/dL, median bicarbonate 6,85 (IQR 9,2) mEq/L; and mean Anion GAP, 26,6 (±9,4)mEq/L. 19 (36%) patients had acute kidney injury at admission. The median time of intravenous insulin infusion therapy was 22 (IQR 12) hours; although 4% (2) patients needed about 72 hours of intravenous insulin to recover from DKA. The median time between admission and discharge was 5 (IQR 4,75) days. However, a chi-squared test showed that neither severity of DKA, diagnosis of sepsis, age, sex, or blood chemistries at admission (sodium, glucose, and anion gap) had been correlated with discharge from hospital after the 5th and 7thday or mortality.

Discussion and conclusions: Infection/sepsis was the main precipitating factor of DKA, data consistent with previous studies. Other precipitating factors include discontinuation/inadequate insulin therapy or new-onset type 1 diabetes. The mortality rate also matches previous studies; clinical and laboratory characteristics were not related to length of stay or mortality rate.