Adherence to Diabetic Ketoacidosis Management Protocol: A Paediatric Centre Experience

Abstract

Abstract BACKGROUND: Paediatric diabetic ketoacidosis (DKA) management should be regulated by specific protocols. Following the most recent medical literature, our tertiary pediatric hospital was provided with such a DKA protocol in 2009. OBJECTIVES: Assess the proportion of cases of DKA that are marked by non-adherence to our DKA management protocol (DKAp). Describe the proportion of patients developing hyperchloraemic acidosis (HCl) during acute treatment of DKA. DESIGN/METHODS: We conducted a retrospective study by chart review. Deviations to DKAp were classified as minor or major according to their potential impact on patient security. Non-adherence to protocol was defined as the occurrence of ≥ 1 major deviation or ≥ 2 minor deviations. RESULTS: Seventy-nine patients were included with a mean age of 9,7 years (±4,9). Forty-two (53,2%) were females. Majority of cases were new-onset diabetes (60/79, 75,9%). Within the known diabetic patients (19), the DKA precipitating factor was poor treatment compliance in 10 (52,6%), and insulin pump dysfunction in 4 (21%). Mild DKA (pH 7,21 – 7,30) occurred in 24 (30,4%) patients, moderate DKA (pH 7,11- 7,20) in 30 (38,0%) and severe DKA (pH ≤ 7,10) in 25 (31,6%). Continuous insulin drip was used in 73/79 (92,4%) cases. We observed ≥ 1 major deviation to DKAp in 37/73 (50,7%) cases, mostly failure to monitor hourly neurological status (32/69, 46,4%, 4 missing values) and inappropriate administration of 0,9% NaCl IV bolus in well-hydrated patients (5/73, 7%). Minor deviations (≥2) occurred in 66/73 (90,4%) patients; mostly failure to monitor urine ketones levels every 4 hours (66/73, 90,4%) and blood gas and serum electrolytes every 2 hours (33/73, 45,2%). Overall, non-adherence to protocol occurred in 69 (94,5%) cases. The following complications were observed: 7/73 (9.6%) hypoglycaemia and 1/73 (1.4%) CT scan suspected cerebral edema. No death occurred. Fifty-five patients over 73 (75,3%) experienced HCl, The mean time apparition of HCl was 4:41 hours from the insulin perfusion beginning, using a normal anion gap definition ≤14. The HCl occurrence was associated with a longer insulin perfusion length (with HCl: 17.1±9.1 hours; without HCl: 9.2±5.6 hours, p<0.002). CONCLUSION: Non-adherence to our DKAp was observed in majority of cases. The establishment of a protocol does not guarantee its adherence and surveillance of its application is needed. More studies have to be done to understand the clinical impact of HCl and to suggest an appropriate IV fluid management.