Determinants of altered sensorium at presentation with diabetic ketoacidosis.

AIM

The aim of this paper was to determine the prevalence of altered sensorium and their determinants in diabetic ketoacidosis (DKA).

METHODS

Retrospective review of medical records for diabetic ketoacidosis admissions over a two year period.

RESULTS

A total of 94 admissions for DKA in 89 patients were reviewed. Majority of admissions were female related. The mean age of the subjects was 39.1±18.5 years. The prevalence of altered sensorium defined as Glasgow coma scale (GCS) below 15 was 47.9%. In Model 1, systolic blood pressure (SBP) <90 mmHg (odds ratio [OR] 17.5 95% Confidence Interval (CI) 2.1-142.9; P=0.008 and calculated effective serum osmolality >320 mosmol/Kg (OR 3.6 95% CI 1.4-12.2; P=0.043 were independently associated with altered sensorium. In Model 2 where serum osmolality was substituted with uncorrected serum sodium, SBP <90 mmHg OR 19.6 95% CI 2.2-100; P=0.007, uncorrected serum sodium >150 mmol/L OR 18.5 95% CI 2.1-100; P=0.0009 and white cell count >25 x 106/L OR 3.6 95% CI 1.03-13.5; P=0.05 were independently associated with altered sensorium.

CONCLUSION

Systolic blood pressure <90 mmHg, calculated effective serum osmolality >320 mosmol/Kg, uncorrected serum sodium >150 mmol/L and white bood count >25 x 106/L were independently associated with altered sensorium.