Incorporation of Nutritional Factors and Insulin Resistance into Insulin Sliding Scales in the Non-Critical Care Inpatient Setting: A ?Before After? Study

Objective: Current guidelines emphasise the importance of effective glucose control in medical inpatients outside critical care. Traditional glucose-adapted insulin sliding scales (SSCs) may inaccurately estimate insulin requirements, resulting in hypoglycaemia, hyperglycaemia, or both. We retrospectively performed a “before/after” analysis investigating whether an SSC also incorporating carbohydrate intake and estimated insulin resistance (iSSC) improves glucose control relative to that with conventional SSCs. Methods: We compared glucose control during the initial 120 inpatient hours in patient groups with hospital diagnoses of lower respiratory tract infection or an acute cardiac condition and diabetes mellitus as comorbidity, one group treated during 2010, after introduction of an iSSC, and an historical control group treated during 2008, with an SSC as the standard of care. Mean glucose levels, hypoglycaemic and hyperglycaemic episodes, glucose variability, and hospital outcomes (length-of-stay, all-cause mortality, intensive care unit admission) were compared using multivariate linear regression analysis adjusted for potentially important confounders. Results: Of 215 included patients, 59.5% (n=128) were in the iSSC group, 40.5% (n=87) in the historical control group. Relative to controls, iSSC patients had consistently greater probability of effective glucose control and better hospital outcomes; however, these differences were insufficiently powered to attain statistical significance (p ≥ 0.114). Conclusions: Our results suggest that incorporating nutritional factors and insulin resistance into an SSC may improve glucose control and clinical outcomes in the everyday non-critical care inpatient setting. Due to the small patient sample and borderline significant results, we endorse further large, prospective, randomised controlled studies conclusively answering this question.