Objectives: Hypokalaemia is a potential complication of diabetic ketoacidosis (DKA). We assessed the impact of a revised DKA protocol on the prevalence of hypokalaemia during treatment at our tertiary centre.
Methods: Retrospective audits were conducted on adult patients (≥18 years) with DKA (pH<7.35, blood glucose>11mmol/L and capillary ketones>1.0mmol/L) admitted between Jan 2016 – December 2016 (cohort 1) and September 2020 – July 2021 (cohort 2) following implementation of a revised DKA protocol with a faster intravenous potassium replacement rate on the wards.
Results: Cohort 1 consisted of 59 patients (mean age 43.5±20.4yrs; 88% type 1 and 12% type 2 diabetes) with 68 episodes of DKA. Cohort 2 included 58 patients (mean age 42.8±20.8yrs; 83% type 1 and 17% type 2 diabetes) with 68 DKA episodes. Hypokalaemia (serum potassium <3.5mmol/L) occurred in 5.9% at presentation in cohort 1 and 2.9% in cohort 2 (P=0.408). In the first 48 hours of treatment, 54.4% and 38.2% developed hypokalaemia in cohorts 1 and 2, respectively (P=0.103). Hypokalaemia was mild (3.0-3.4mmol/L) in 39.7% vs 34.6%, moderate (2.5-2.9mmol/L) in 10.3% vs 6.7%, and severe (2.0-2.4mmol/L) to critical (<2.0mmol/L) in 4.4% vs 0% (P=0.568). Intensive care unit (ICU) admission occurred in 36.8% in cohort 1 (median ICU length of stay (LOS) 2 days) and 41% in cohort 2 (median ICU LOS 1 day). In both cohorts, median hospital LOS was 3 days, resolution of DKA within 24 hours (pH >7.3 with capillary ketone <1.0mmol/L) occurred in >90%, and no arrhythmia or in-hospital mortality was observed.
Conclusions: The introduction of a revised DKA protocol appears to have mitigated the rate of severe and critical hypokalaemia during treatment. However, the prevalence of hypokalaemia during treatment remains high, with a 10 to 15-fold increase in the first 48 hours, highlighting the need to remain vigilant of this complication.