Endocrine Abstracts

Introduction

Glycemic control during hospitalization has an impact on the length of hospital stay (LOS) and infection rates. This knowledge has led to a growing concern with hyperglycemia and its repercussions.

Objectives

This study aims to evaluate the glycemic profile and diabetes mellitus (DM) treatment paradigm of hospitalized patients in a central hospital by main diagnosis and LOS.

Methods

We performed an observational cross-sectional study on the 9^th October 2020 that included non-critical adult patients admitted to Centro Hospitalar Universitário do Porto (except pregnant/puerperal women), with DM and a minimum of 24 hours of hospitalization. The diagnosis of DM was made by consulting the clinical file and glycemic control was assessed using capillary blood glucose (CBG) values in the previous 24 hours. The main diagnosis of hospitalization was categorized as infection/non-infection and the LOS as A (1–3 days), B (4–7 days), C (8–14 days), D (15–30 days) and E (> 30 days). Statistical analysis: SPSS version 20.0.

Results

A total of 128 patients (57.8% male, mean age 73.7 ± 11.6 years) were included, with medians of LOS of 11.0 (19.0) days and mean CBG of 166.2 (69.4) mg/dl (95.5–392.0). Regarding the diagnosis, 51 (39.8%) had infection and 7 (5.5%) were hospitalized for DM or its complications. The median mean CBG was higher in the group of patients diagnosed with infection [183.7 (65.2) vs 160.1 (75.1) mg/dl, P = 0.099]. A larger proportion of patients without a diagnosis of infection was being treated exclusively with glucose-lowering drugs (6.5% vs 3.9%, P = 0.702) and a lower percentage was insulin-treated (26.0% vs 37.3%, P = 0.239) compared to infected patients. There were also considerable percentages of patients only under sliding scale regimen, especially in the group without infection (44.2% vs 37.3%, P = 0.468). By LOS, the median maximum CBG was higher in groups B to D (P = 0.025) and there were no significant differences in terms of median minimum/mean CBG. The percentage of insulin-treated patients was minimal in group A (9.1%) and maximum in group D (43.3%, P = 0.133).

Conclusion

The diagnosis of infection is common in hospitalized diabetic patients and tends to be associated with higher median CBG values. The rates of exclusive use of the sliding scale regimen are still higher than desirable, particularly in infected patients. LOS between 4 and 30 days also seem to be associated with higher CBG values and there appears to be an increased insulin therapy use as hospitalization time increases.