Background: Diabetic emergencies associated with ketoacidosis (DKA) and a hyperosmolar, hyperglycaemic state (HHS) are both acute life-threatening metabolic disturbances. Traditionally, DKA and HHS have been classified as distinct entities but there is evidence to suggest that patients can present with elements of both conditions.
Aims: To examine the presentation profiles, mortality rates and prognostic factors associated with a fatal outcome for diabetic patients admitted with ketoacidosis and/or hyperosmolarity.
Methods: A retrospective analysis of 312 admissions to an Australian tertiary referral hospital between 1986 and 1999.
Results: Of the patients surveyed, DKA was the diagnosis for 171 presentations (55%), HHS was the diagnosis for 47 presentations (15%) and combined DKA and HHS (DKA-HHS) was diagnosed for 94 presentations (30%). Age at presentation for DKA patients (33+/-1.2 years) was significantly less (P< 0.01) than DKA-HHS patients (44+/-2.4 years). This, in turn, was significantly less (P < 0.01) than HHS patients (69+/-1.7 years). There were 15 deaths for the 312 presentations, resulting in an overall mortality rate of 4.8%. Combined mortality rates according to age at presentation were: (i) 0/134 for patients aged <35 years, (ii) 1/85 (1.2%) for patients aged 35-55 years and (iii) 14/93 (15.0%) for patients aged >55 years. For the three categories of diabetic emergencies, mortality rates were: (i) 2/171 (1.2%) for DKA, (ii) 5/94 (5.3%) for DKA-HHS and (iii) 8/47 (17%) for HHS. For all presentations associated with ketoacidosis - regardless of the degree of hyperosmolarity - the mortality rate was 7/264 (2.7%), however for all presentations with hyperosmolarity regardless of the degree of acidosis - the mortality rate was 13/141 (9.2%). When the associations between age, category of diabetic emergency, serum osmolarity and various other biochemical parameters with mortality were assessed by logistic regression analysis, age and the degree of hyperosmolarity were found to be the most powerful predictors of a fatal outcome. In particular, patients aged >65 years presenting with a serum osmolarity >375 mOsmol/L were at greatest risk. However, in a multivariate analysis only age emerged as a significant independent predictor of mortality (P < 0.01).
Conclusions: The mixed state of ketoacidosis and hyperosmolarity was observed in 30% of presentations for diabetic hyperglycaemic emergencies. Although age and degree of hyperosmolarity both influenced mortality rates, only age was found to be an independent predictor of mortality. The mortality rate for diabetic emergencies associated with ketoacidosis remained low, in keeping with other studies. By contrast, the mortality rate for diabetic emergencies associated with a hyperosmolar state remained considerably higher. This higher mortality will most likely persist because deaths associated with a hyperosmolar state were in elderly patients with significant comorbidity.