Predictors of Insulin Requirements Among Hospitalized Adults Receiving Parenteral Nutrition

Abstract

Objective: The objective of this quality improvement project was to determine factors predictive of parenteral nutrition (PN) insulin therapy. Methods: Patients receiving PN at a tertiary care academic medical center between January 1, 2009, and December 1, 2012, 18 years or older were included. Variables collected included demographics, medical information, and PN-specific data. χ² and Student t tests were used to determine differences between patients who did and did not require PN insulin. Odds ratios (ORs) with 95% confidence intervals (CIs) were used to determine associations between characteristics. Stepwise forward logistic regression was used determine the best predictors of PN insulin. Results: A total of 1388 patients were started on PN. After adjusting for potential confounders, strong associations existed between PN insulin requirements and diabetes mellitus (DM) diagnosis (OR, 8.90; 95% CI, 4.98–15.90, P < .001), overweight/obese status (body mass index ≥25.0 kg/m²) (OR, 2.12; 95% CI, 1.04–4.30, P = .04), intensive care unit (ICU) admission (OR, 1.79; 95% CI, 1.03–3.11, P = .04), blood glucose (BG) on day of PN start >120 mg/dL (OR, 2.32; 95% CI, 1.32–4.05, P = .003), mean BG >180 mg/dL while receiving PN (OR, 6.10; 95% CI, 2.18–17.04, P = .001), and hemoglobin A1c (A1c) ≥5.7% (OR, 3.18; 95% CI, 1.84–5.50, P < .001). Among variables available at PN initiation, DM diagnosis (P < .001), A1c ≥5.7% (P < .001), BG >120 mg/dL on PN start day (P < .001), and ICU admission (P < .001) predicted the need for PN insulin.