Comparative analysis of one versus two-bag methods in the treatment of adult diabetic ketoacidosis

Document Type

Conference Proceeding - Restricted Access

Publication Date


Publication Title

Critical Care Medicine


Introduction/Hypothesis: Diabetic ketoacidosis (DKA) is a significant complication of diabetes mellitus and prompt standardized management is critical. In addition to insulin therapy, the “two-bag method” (2B) of fluid administration has shown clinical benefits for patients with DKA versus the “one-bag method” (1B) including faster correction of acidemia and decreased length of hospitalization. Data evaluating 1B and 2B protocols are lacking in adults. The present study compared the effectiveness and safety of a 1B DKA protocol to a control.

Methods: This retrospective study evaluated DKA patients treated with a protocol-based insulin infusion for ≥12 hrs and 1B fluids (S1) between 2/2019-12/2020 compared to a control of 1B (C1) and 2B (C2) groups as reported in J Emerg Med. 2018;54:593. Effectiveness was defined as time to pH ≥7.3, bicarbonate ≥18 mEq/L, and discontinuation of insulin infusion. Safety was evaluated by occurrences of hypokalemia and hypoglycemia. Data are reported as control (C1 or C2) versus study (S1). SPSS v21.0 was used for analysis.

Results: Of 598 S1 patients screened, 238 were included with a mean age of 45.4±19 years and APACHE-II 15±6. In the S1 group, 79% of patients had moderate to severe DKA. The time to pH ≥7.3 tended to be shorter in C1 (10.8 v 18.1 hr; p=0.07), but not different compared to C2 (p=0.26). Additionally, the time to bicarbonate normalization was not different between S1 and C1, (p=0.17) but improved with C2 (13.4 v 26.3 hr; p=0.01). The mean duration of insulin infusion was shorter in both C1 (21.8 v 36.1 hr; p< 0.01) and C2 (14.1 v 36.1 hr; p< 0.01). Hospital length of stay was shorter in C1 (3.4 v 5.9 days; p< 0.01) and C2 (3.1 v 5.9 days; p< 0.01). Episodes of hypokalemia were similar between C1 and S1 groups (27% v 29%; p=0.72), however this was decreased in the C2 group (16% v 29%; p=0.03). Hypoglycemia was observed more in the S1 group compared to C1 (10% v 30%; p< 0.01) and C2 (3% v 30%; p< 0.01).

Conclusions: Our institution utilizes a standardized insulin infusion and 1B fluid method for clinical management of patients with DKA. Compared to control groups, it was associated with increased duration of insulin infusion, risk of hypoglycemia, and prolonged hospital stay. Institution-specific DKA management protocols should be evaluated to ensure optimal treatment.





First Page