Endocrine Abstracts

Background: A substantially reduced fluid replacement regimen was introduced in the New’ BSPED (2015) compared to ‘Old’ BSPED (2009) guideline for DKA management in children. However, effects of varying fluid replacement regimens are limited and we explored this by comparing outcomes of the two guidelines on the resolution of DKA.

Methods: In a retrospective audit of consecutively admitted patients (age <18 years) to two hospitals in Barts NHS trust with DKA between Jan-2014 and March-2017, we evaluated the resolution time of DKA defined by recovery of acidosis (pH>7.30), ketosis (blood ketones<1.0 mmol/l) or bicarbonate (>18.0 mmol/l) levels. Biochemical parameters before, the nearest to 6 and 12 hours into treatment and at resolution were collected. Effective osmolality was calculated using formula: (2xSodium)+glucose.

Results: Of 82 patients admitted data were available for 44 (transferred (n=12), data unavailable (n=26)). Twenty and 24 patients were managed by the ‘New’ and ‘Old’ guidelines respectively. The median age was 10.1 years (interquartile range, 6.4–13.1), 28 patients (63.6%) were newly-diagnosed and 15 (34.1%) had severe DKA (pH<7.1). Age, DKA severity and proportion of newly-diagnosed patients were similar in both groups. The fluid administration rates were substantially lower (24.0(24.0–39.1) vs 55.0(45.0–69.0)ml/hour, P<0.0001), in the ‘New’ guideline, but frequency of fluid boluses was similar (40% vs 50%, P=0.44). The resolution of DKA evaluated by pH (‘New’ vs ‘Old’, 14.8(7.9–19.2) vs 15.7(7.8–24.7) hours, P=0.72) or ketosis (21.2(12.3–29.8) vs 20.3(12.5–35.6) hours, P=0.59) or bicarbonate levels (15.8(10.4–27.8) vs 20.7(11.9–25.6) hours, P=0.63) were similar. The levels of sodium, potassium, chloride and bicarbonate, pH and effective serum osmolality before, at 6 and 12 hours and resolution, and hypoglycaemia rates were similar. However, the time to decline of glucose levels to 14 mmol/l tended to be lower (4.28(3.41–6.83) vs 6.15(4.29–10.17) hours, P=0.11) and was significantly lower (3.86(3.25–5.43) vs 5.48(4.50–9.75), P=0.018) in mild DKA in the ‘New’ guideline. No patients developed cerebral oedema.

Conclusions: We found that ~50% reduction in fluid replacement in DKA was not associated with significant changes in resolution time, electrolyte levels or osmolality. However, hyperglycaemia was corrected faster in the ‘New’ guideline. Larger studies are important to evaluate the effects on cerebral oedema.