1760-P: Under- and Overdosing Insulin in the Hospital: A Retrospective Cohort Study of Dose Deviation and Patient Outcomes

Introduction and Objective: Glycemic abnormalities during hospitalization are common and linked to morbidity and mortality, yet little is known about how deviations from guideline-recommended insulin dosing affect outcomes. In this study we examined in non-critically ill hospitalized adults with type 2 diabetes, deviation from an expected insulin dose, compared with patients receiving a dose within guideline-based range (as stated in the ADA standards of care page S343) affected mortality, length of stay, or dysglycemia. Methods: We conducted a retrospective cohort study of adults with type 2 diabetes hospitalized at an academic medical center (2010-2025) who received ≥24 hours of subcutaneous insulin. Expected total daily dose (TDD) was calculated as 0.4 units/kg/day; Observed-TDD was derived from medication records. Dose deviation was classified as undertreatment (<75% of expected), within-range (±25%), or overtreatment (≥125%). Outcomes included length of stay, 30-day mortality, hypoglycemia, and hyperglycemia. Associations were evaluated using multivariable Cox regression and appropriate comparative tests. Results: Among 5,684 patients (median age 72 years), 24% received guideline-concordant dosing, 34% were undertreated, and 42% were overtreated. Undertreatment was associated with higher adjusted 30-day mortality (aHR 1.55, 95% CI 1.27-1.89). Overtreatment showed longer length of stay but no significant mortality increase (aHR 1.21, 95% CI 0.99-1.62). Overtreated patients had less hyperglycemia, while hypoglycemia incidence did not differ between groups. Conclusion: In hospitalized adults receiving insulin, substantial deviations from expected weight-based dosing are common. Undertreatment is independently associated with increased mortality, whereas overtreatment appears safe but prolongs hospitalization. Closer alignment with guideline-based insulin dosing may improve inpatient outcomes. Disclosure E.Y. Feigin: None. H. Zaitoon: None. M. Lalovich: None. R.A. DeFronzo: Advisory Panel; Ended; AstraZeneca. Research Support; Current; AstraZeneca. Advisory Panel; Current; Novo Nordisk. Research Support; Current; Eli Lilly and Company. Advisory Panel; Current; Corcept Therapeutics. Speaker's Bureau; Current; Corcept Therapeutics. Consultant; Current; Alnylam Pharmaceuticals, Inc. Advisory Panel; Current; Regeneron Pharmaceuticals Inc., Aardvark.