Introduction and Update of an Inpatient 72-Hour Fast Protocol at the Cleveland Clinic

The 72-hour fast is used in the evaluation of hypoglycemia with concern for insulinoma. In January 2018 we implemented an inpatient protocol for the 72-hour fast at our institution. This includes strict fasting, baseline labs (C-peptide, insulin, proinsulin, beta hydroxybutyrate (BHB), glucose), and fingerstick glucose (FSG) checks every 4 hours. Once FSG is < 60mg/dL, FSG is checked hourly. When FSG is ≤45 mg/dL, the baseline labs are repeated. The fast ends when the blood glucose level is ≤45 mg/dL or the 72-hour mark is reached. In June of 2020 we updated the protocol to include insulin and BHB checks every 12 hours and to end the fast when the BHB level is ≥2.7 mmol/L. The aim of this study is to examine the efficacy of our protocol for identifying insulinoma and assess how we can reduce unnecessary hospitalization time and costs.

Methods:

Since June of 2020, there have not yet been sufficient 72-hour fasts performed for a comparative analysis due to a hold on elective admissions during the COVID-19 pandemic. Therefore, this is a retrospective cohort study of patients who underwent a 72-hour fast at our institution between January 2018 and January 2020. Data was collected through chart review, and descriptive statistics were used.

Results:

From January 2018 to January 2020, there were 34 patients who underwent a 72-hour fast at our institution; 23 (68%) were female with a mean age of 51 (±18 yrs) and BMI of 31 (±7 kg/m2). Of the 34 that underwent the 72-hour fast, 23 (68%) completed the full 72-hour time period, while 11 (32%) terminated the fast for hypoglycemia ≤45 mg/dL. Insulinomas were confirmed in 5 (15%) patients, and all had ended their fast before 48 hours due to hypoglycemia. Reviewing our data, it should be noted that if the fast had been ended at a BHB ≥2.7 mmol/L, no insulinomas would have been missed. This has clinical importance as potentially 12 (35%) of patients could have had their fast end earlier, reducing hospital cost and stress on the patient.

Discussion/Conclusion:

The 72-hour fast remains a standard test in the evaluation of symptomatic hypoglycemia, but incurs inpatient costs and resources. Our data supports the utility of routine BHB measurement with a clause to terminate the fast early for BHB elevations, and we look forward to comparing the data since our protocol update in June of 2020 to further evaluate. Additionally, our data supports a 48-hour fast, which would further reduce expenses and stress on patients, and result in no missed insulinoma diagnoses in our review. However, given the relatively small cohort, this clinical question warrants further study.