Introduction Delay diagnosis is the major cause of Diabetic ketoacidosis (DKA). Children with profound acidosis are at great risk for symptomatic cerebral oedema.

Objective To identify the epidemiological profile, clinical feature, factors related to delayed diagnosis in children with DKA and to analyse the factors associated with prolonged acidosis.

Methods We analysed the records of all children with DKA, admitted to our PICU during January 2004–December 2013. We evaluated clinical features, biochemical profile at admission, 6, 12 and 24 hrs, presence of sepsis, shock, complications and outcome. The severity of DKA was defined by the degree of acidosis: mild (pH = 7.2–7.3), moderate (7.1–7.2) and severe (pH <7.1). Anion gap (AG), delta gap (DG) and delta ratio were calculated. Prolonged acidosis was analysed against various independent factors.

Results Mean age of the patients was 7.06 ± 4.24 years, with misdiagnosis in 32% of cases. By the degree of acidosis, DKA was mild in 16%, moderate in 56% and severe in 28% of cases, with prolonged acidosis (>24 hrs) in 36% of cases. Factors associated with prolonged acidosis were: Na >133 mEq/L (p = 0.01), HCO₃ <4.8 mEq/L (p = 0.03), pH <7.01 (p = 0.01), Cl^- >100 mEq/L (p = 0.02) and AG >25.1 (p = 0.03). HbA₁C, azotemia, DG and misdiagnosis didn’t resulted significative for prolonged acidosis. Three cases are complicated with cerebral oedema. Initial blood glucose or decline in glucose had no association with cerebral oedema. Mortality rate was 8%.

Conclusion Misdiagnosis of diabetes with DKA as consequence, is still high in children in Albania. Clinical and laboratory findings help identifying the patients who require a higher level of intervention.