Re-Gap or Not? Predicting the Likelihood of Anion Gap Reopening in Diabetic Ketoacidosis Patients

Introduction: There is currently scant literature that has looked at specific variables associated with anion gap re-opening (AGRO) in diabetic ketoacidosis (DKA). Anecdotally, AGRO after transition from continuous insulin infusion to a subcutaneous insulin regimen is associated with increased length of ICU and hospital of stay, increased hospital resource utilization and potentially worse outcomes. The purpose of this study is to identify factors that contribute to AGRO in DKA patients to aid in predicting the likelihood of AGRO and further guide treatment decisions in this patient population.


Methods: This was a retrospective, multicenter case control study which included adult patients admitted with DKA between May 1, 2022, and September 30, 2024. Patients were treated with the 2-bag insulin infusion method and grouped based on whether their anion gap re-opened to >12 within 24 hours of closure (anion gap ≤12). Data on demographics, comorbidities, home medications, lab values, and inpatient treatments were analyzed. Logistic and Cox proportional hazards models were used to evaluate predictors of AGRO.


Results: A total of 783 patients were included: 549 without AGRO and 234 with AGRO. Baseline beta-hydroxybutyrate (BHB) was significantly associated with AGRO (OR 1.02; 95% CI 1.00–1.03; p=0.042). Higher baseline anion gap (HR 0.99; CI 0.97-1.00; p=0.037) and presence of CKD (HR 0.79; CI 0.63-1.00; p=0.047) were associated with slower anion gap resolution. It was observed that AGRO group had longer ICU stays (43.4 vs. 41.8 hours), hospital stays (4.94 vs. 4.10 days), and time to bicarbonate correction (19.35 vs. 17.41 hours).


Conclusions: With elevated BHB levels significantly associated with the development of AGRO as well as underlying CKD and higher baseline anion gap being predictive of slower anion gap normalization in our study, we have identified at least 3 areas to focus on to ensure best possible care for patients. These findings highlight the importance of early identification of patients at increased risk for AGRO. Enhanced monitoring and individualized therapeutic strategies, including extended insulin infusion duration and optimized transition protocols may be areas of future study to curb AGRO incidence, reduce hospital duration, and improve overall clinical outcomes.