We previously showed that a proactive inpatient diabetes service decreased adverse glycemia (AG) and hospital acquired infections (RAPIDS: ADA2017, 231-OR). To further focus our proactive care on high-risk inpatients, we investigated clinical risk factors associated with AG. We analyzed multiple clinical variables in 643 consecutive inpatients with diabetes or new hyperglycemia (random BG ≥11.1mmol/L). Capillary BG from day 2 of admission until discharge (censored at day 14) were analyzed. AG was defined as BG <4 or >15mmol/L on any day and recurrent AG (RAG) was defined as AG on ≥2 days. A split-sample multivariable logistical regression was performed with internal validation. The patient characteristics included 87% type 2 diabetes, 33% insulin-treated and mean HbA1c 7.6%. AG and RAG occurred in 278 (43%), and 176 (27%) patients respectively. Pre-hospital factors (sulphonylurea or insulin treatment, HbA1c, Charlson index) and in-hospital factors (dysglycemia on day 1, length of stay) were independently associated with both AG and RAG (Table). Glucocorticoid treatment was associated with RAG but not AG. A model using these multiple variables accurately identified AG (ROC-AUC 0.88). Age, diabetes type, creatinine and admission unit were not associated with either AG or RAG. This study identified multiple key clinical risk factors associated with adverse glycemia, and may be used to better concentrate efforts for inpatient diabetes care.


M. Kyi: None. J.E. Reid: None. A. Gorelik: None. S.S. Kumar: None. A. Galligan: None. L.M. Rowan: Speaker's Bureau; Self; AstraZeneca, Sanofi, Novo Nordisk Inc.. A.J. Nankervis: None. K.A. Marley: None. D.M. Russell: None. P.R. Wraight: None. P.G. Colman: None. S. Fourlanos: None.

Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at http://www.diabetesjournals.org/content/license.