Hyperglycemia treatment is associated with longer length of hospital stay in colorectal cancer surgery patients

Abstract

6133 Background: Diabetes mellitus (DM) is associated with longer hospital stays in some medical and surgical inpatient populations. Aggressive control of blood sugar may prevent complications and decrease hospital length of stay (LOS). Methods: We conducted a retrospective study of 519 patients at UT MD Anderson Cancer Center (UTMDACC) who had major surgery for colon or rectal cancer in calendar years 2000–2003. Patient data extracted from the UTMDACC Institutional Database included demographics, admission and discharge dates and diagnoses, surgical procedures, and diabetes medication use during the hospitalization. Known DM was defined as present if the patient had a diabetes diagnosis prior to or at admission; hyperglycemia treatment was defined as receipt of a medication for diabetes (insulin or oral medication) during the index hospitalization. Chi-square and t tests were performed to assess associations between patient characteristics and long LOS, and multiple logistic regression was used to identify independent predictors of hospital LOS at or greater than the 75th percentile for the study population (long LOS). Results: The mean age of the study population was 60.4 years (median 61, range 18–91). Known DM was present in 10.4% of cases; the same percentage received hyperglycemia treatment during their hospital stay, although not all were known diabetics. Mean LOS was 8.9 days; median 7 days and the 75th percentile 9 days. 50% of patients treated for hyperglycemia had long LOS (27 of 54 cases, p< 0.0005). In a logistic regression model controlling for patient demographic and clinical characteristics and the occurrence of post-operative complications, hyperglycemia treatment was an independent predictor of long LOS (odds ratio 4.1, 95% confidence interval 1.6, 10.3). Conclusions: Hyperglycemia treatment is associated with longer LOS in patients undergoing surgery for colon or rectal cancer at UTMDACC. Further studies should determine whether patients at risk for long LOS can be identified prospectively such that they might benefit from an intervention to reduce their LOS. [Table: see text]