Evaluation of diabetic ketoacidosis cases treated in the pediatric intensive care unit of Bezmialem Vakif University Faculty of Medicine Hospital

Diabetic ketoacidosis (DKA) is the leading cause of mortality and morbidity in children with type 1 diabetes mellitus (T1DM). Mortality is predominantly associated with cerebral oedema; such oedema develops in 0.3–1% of all patient who suffer DKA attacks. However, the aetiology and pathophysiology remain poorly understood, and there is no consensus treatment method. The relationship between the amount and administration rate of fluid used to treat DKA, and the development of brain oedema, remain a matter of debate. Here, we report our DKA treatment and the results thereof. Our patients were admitted to the paediatric intensive care unit of the Istanbul Bezmialem Vakif University Faculty of Medicine Hospital with DKA attacks between 1 December 2014 and 1 December 2018. Of 72 patients, 30 were included in the study; 42 were excluded because some medical information was missing. The hospital records were retrospectively examined in terms of demographic characteristics, serum biochemical values at the time of admission, blood gas levels, the clinical and laboratory courses after treatment, recovery from ketoacidosis, and DKA complications. Half of all DKA patients were newly diagnosed with T1DM; six (40%) were under the age of 6 years. This finding is consistent with the literature; DKA is often the first T1DM symptom at younger ages. At admission, 11 (36.6%) patients were severely dehydrated, 11 (36.6%) moderately dehydrated, and 8 (26.7%) slightly dehydrated. The average extent of dehydration was 7.03 ± 2.75%, and generally ranged withing 6– 8%, as determined on a body weight change basis. Traditionally, DKA patients are viewed as severely dehydrated if the fluid loss at presentation is 10–15%. In one study that measured body weight changes, the average dehydration extent was 5.8–8.7%, similar to our findings. The average blood pH at admission was 7.05 ± 0.12, the lowest was 6.83, and the highest was 7.30. Four (13%) patients presented with mild, 5 (17%) with moderate, and 21 (70%) with severe acidosis. All patients were commenced on intravenous crystallised insulin infusion after liquid bolus therapy. No insulin bolus was administered to any patient. Insulin infusion commenced at 0.1 U/kg/h in all patients and continued for an average of 18.2 ± 5.78 h. On average, after 16 h, the pH exceeded pH 7.30 and the bicarbonate level was over 15 mEq/L. No patient was given intravenous bicarbonate. Serious electrolyte disturbances may be apparent at the time of presentation or during the treatment of DKA. The potassium and phosphorus levels must be closely monitored. We found that, despite insulin treatment and gradual improvement of the acidosis, the average potassium level slowly decreased despite potassium replacement. However, no severe hypokalaemia or hyperkalaemia was detected in any patient. During treatment, potassium levels should be closely monitored every 2 or 4 h, as must the electrographic findings, especially in patients with hypokalaemia or hyperkalaemia. Of our patients, 26 (86.6%) developed varying extents of hypophosphataemia on DKA treatment; 29 required phosphorus replacement during treatment. Serum phosphorus levels must be closely monitored throughout treatment, especially until nutrition recommences, and phosphorus replacement should be dynamic in nature. Many protocols recommend the addition of dextrose to the replacement fluid after the blood sugar falls to within 250–300 mg/dL. In our study, all patients were commenced on replacement fluids with dextrose. The most rapid decrease in the average blood sugar level occurred at the end of the first fluid loading. As the blood sugar level will decrease further with insulin treatment, we recommend that all replacement fluids should contain dextrose. In this way, any sudden osmolarity problem attributable a rapid drop in the glucose level will be prevented and the risk of brain oedema reduced. Excessive daily fluid administration is a risk factor for the development of cerebral oedema. One study found that more than 4,000 mL/m2 /day was such a risk factor. In the present study, the highest level was 3,600 mL/m2 /day, the lowest 2,000 mL/m2 /day, and the average 2,770 ± 465 mL/m2 /day; the literature indicates that these levels are acceptable. The fastest decrease in serum osmolarity occurred from 2 to 8 h; the nadir was attained at 8 h. One (3.3%) patient, younger than 5 years of age, was diagnosed with cerebral oedema at the 4-h treatment point; this was in line with the literature. He presented with severe metabolic acidosis. Rapid hydration was not performed and there was no sudden decrease in serum osmolarity. However, recent studies suggest that, apart from treatment-related factors, factors specific to DKA also contribute to brain damage. Recent hypotheses include cerebral hypoperfusion and neuroinflammation. Patients at risk of brain damage should be meticulously monitored and treatment should not be delayed if brain oedema is suspected.