Administering Nutritional Insulin Postprandially in the Inpatient Setting Is Associated with Improved Glycemic Metrics

In many hospitals, as in the outpatient setting, nutritional (i.e. rapid-acting) insulin is typically injected before the meal. However, hospitalized patients frequently have poor oral intake due to their acute illness. Nurses must often guess whether to administer or withhold a patient’s preprandial insulin, depending on whether they feel the patient will eat. Guessing incorrectly can be deleterious, as administering insulin without food intake may result in hypoglycemia, while conversely holding insulin unnecessarily can lead to hyperglycemia. The objective of this study was to evaluate whether implementation of a hospital policy to deliver nutritional insulin post-prandially, as compared to pre-prandially, was associated with changes in inpatient glycemic metrics.

Methods:

This was a retrospective cohort study performed at Howard County General Hospital, a member of Johns Hopkins Medicine, using a de-identified data set. In June 2019, hospital policy shifted the timing of all nutritional insulin to be given after at least 50% of the meal was consumed, instead of pre-prandially. Because a full-time endocrine hospitalist was hired to consult on all inpatients with diabetes in August 2018, we restricted the data set to compare the nine months pre-policy (September 2018 to May 2019) vs. nine months post-policy implementation (July 2019 to March 2020). Patients who received at least one unit of insulin and had a glucose measurement were included in the analysis. Primary outcome was the rate of inpatient hypoglycemia, as defined by a glucose ≤70mg/dL. Secondary outcomes included rates of moderate hypoglycemia ( <54mg/dL), severe hypoglycemia (≤40mg/dL), hyperglycemia (mean daily glucose ≥180mg/dL), length of stay (LOS), and 30-day readmission rates (30DRR). Chi-square with Yates’ correction or Student’s t-test were used to analyze the differences between groups.

Results:

Rates of hypoglycemia significantly decreased from 5.9% (592 of 10,023 patient-days) to 5.0% (500 of 9,987 patient-days) post-intervention (p=0.006). Hyperglycemia rates also significantly decreased post-intervention (45.7% vs 42.5%; p< 0.0001). Rates of moderate (1.9% vs 2.1%; p=0.23) and severe hypoglycemia (0.6% vs 0.6%; p=0.98) were not significantly different between groups. LOS and 30DRR were not significantly affected.

Discussion/Conclusion:

Our study found that administering nutritional insulin post-prandially, instead of pre-meal, in the inpatient setting significantly decreased rates of hypoglycemia as well as hyperglycemia. Further study is warranted to evaluate whether this dosing strategy impacts patient experience or nursing comfort/distress regarding insulin dosing and inpatient diabetes.