Article:
Griffey RT, et al. The SQuID Protocol (Subcutaneous Insulin in /diabetic Ketoacidosis): Impacts on ED Operational Metrics. Acad Emerg Med 2023. PMID: 36775281
Background and Results:
This was a prospectively derived, quasi experimental (pre-post) study that took place at an urban academic hospital with over 90,000 annual ED visits. The researchers evaluated whether, in adult patients with mild to moderate DKA, a subcutaneous insulin protocol (SQuID) could reduce ED length-of-stay compared to a traditional IV infusion DKA protocol (Appendix A, B). They retrospectively obtained electronic data in evaluating three outcome measures - fidelity, safety, and operational measures. Adult ED patients that met their criteria for DKA (hyperglycemia, positive ketone test, presence of anion gap) were eligible for the study. Exclusion criteria for the SQuID protocol included pregnancy, concomitant serious infections, concerns for myocardial infarction, altered mental status, active comorbidities (end-stage renal disease, congestive heart failure, active use of immunosuppressants), need for a surgical procedure, or determination by the ED or inpatient team that a patient was too ill for the designated floor. Eligible patients were stratified by severity of DKA (mild, moderate, high) based on a widely used scheme by Kitabchi et al [1].
Figure 2. DKA Severity Index (Kitabchi et al.).
They had 177 individuals that were included in the intervention. 78 received the SQuID protocol and 99 were treated with traditional DKA management of IV fluids and insulin infusion (Figure 1). They reviewed the safety of the intervention by looking at the frequency of hypoglycemic events. They reviewed operational impacts, which was the primary outcome, by comparing ED length-of-stay (EDLOS) between the SQuID cohort and the traditional insulin infusion cohort, as well as looking at the number of ICU admissions between the post-intervention cohort (SQuID + traditional) versus the number of ICU admissions for mild-to-moderate DKA in historical control periods. They found no differences in safety between the SQuID and traditional pathways. They also found that median EDLOS was significantly shorter for the SQuID cohort, compared to the traditional cohort by approximately 3 hours (median EDLOS 8.9 hours in the SQuID cohort versus 11.9 hours in the traditional cohort). They also saw reductions in ICU admissions, but those findings were not statistically significant.
Discussion:
Overall, this study demonstrated that using subcutaneous insulin to treat mild-to-moderate DKA is safe and reduces EDLOS when compared to traditional insulin infusions. They also saw some decrease in ICU admissions, though it was not statistically significant. In the age of ED boarding, EDLOS is a very important metric. Safe and effective clinical pathways that can help patients decrease their EDLOS are more important than ever.
This study did have some limitations. This was a single-center study, which can limit generalizability to smaller community hospitals. This study was a prospectively-derived quasi-experimental (pre-post) study, which may have some confounders in it. A RCT of SQuID versus traditional protocols would be important for further research. The ICU metrics were not statistically significant, so it is hard to infer by how much, if any, this would decrease ICU admissions.
Here at Cook County, almost all DKA patients are treated with insulin infusions and sent to the ICU. Occasionally, we will be asked to close a patient’s anion gap and bridge them to subcutaneous insulin, but this increases EDLOS and, given our ED resources, is not always the best option for the patient or ED staff. Therefore, subcutaneous insulin is not yet ready for widespread use, including in the Cook County ED. However, it may become a viable solution for patients with mild-to-moderate DKA in the future pending additional studies.
Authored by Taylor Wahrenbrock, MD, Kathryn McGregor, MD, and Eric Leser, MD.
References:
Kitabchi AE, Umpierrez GE, Miles JM, Fisher JN. Hyperglycemic crises in adult patients with diabetes. Diabetes Care. 2009; 32(7): 1335-1343.