Length of stay and hospital costs among high-risk patients with hospital-origin Clostridium difficile-associated diarrhea.

OBJECTIVE: Hospital-onset Clostridium difficile-associated diarrhea (HO-CDAD) has been associated with longer length of stay (LOS) and higher hospital costs among patients in general. The burden of HO-CDAD is unknown among patients who may be at particular risk of poor outcomes: older patients, those with complex or chronic conditions (renal disease, cancer, inflammatory bowel disease [IBD]), and those with concomitant antibiotic (CAbx) use during treatment for CDAD. RESEARCH DESIGN AND METHODS: A retrospective analysis (2005-2011) of the Health Facts® database (Cerner Corp., Kansas City, MO) containing comprehensive clinical records from 186 US hospitals identified hospitalized adult patients with HO-CDAD based on a positive C. difficile toxin collected >48 h after admission. Control patients were required to have total hospital LOS ≥2 days. Separate logistic regression models to estimate propensities were developed for each study group, with HO-CDAD vs controls as the outcome. Differences in LOS and costs were calculated between cases and controls for each group. RESULTS: A total of 4521 patients with HO-CDAD were identified. Mean age was 70 years, 54% were female, and 13% died. After matching, LOS was significantly greater among HO-CDAD patients (vs controls) in each group except IBD. The significant difference in LOS ranged from 3.0 (95% CI = 1.4-4.6) additional days in older patients to 7.8 (95% CI = 5.7-9.9) days in patients with CAbx exposure. HO-CDAD was associated with significantly higher costs among older patients (p < 0.001) and among those with renal impairment (p = 0.012) or CAbx use (p < 0.001). LIMITATIONS: Missing cost data and potential misclassification of colonized patients as infected. CONCLUSIONS: Renal impairment, advanced age, cancer, and CAbx use are associated with significantly longer LOS among HO-CDAD patients, with CAbx users being the most resource intensive. Early identification and aggressive treatment of HO-CDAD in these groups may be warranted.

Association of blood glucose levels with in-hospital mortality and 30-day readmission in patients undergoing invasive cardiovascular surgery.

OBJECTIVE: This study aimed to evaluate the association of mean and maximum blood glucose (BG) levels with in-hospital mortality and 30-day hospital readmission among patients in the intensive care unit (ICU) undergoing invasive cardiovascular (CV) surgery. RESEARCH DESIGN AND METHODS: The retrospective database analysis consisted of data from 3132 patients from 17 hospitals who underwent an invasive CV surgery during 1/2000-12/2006. Patients with hyperglycemia were identified based on serum BG levels recorded from 12 hours prior to and 24 hours after ICU admission. Separate logistic regression models were used to examine the association of mean and maximum BG levels to in-hospital mortality and 30-day readmission, adjusting for patient demographics, comorbidities and laboratory values. RESULTS: The adjusted odds ratio (OR) for in-hospital mortality was 1.07 (95% CI: 1.01-1.12; p < .001) for every 0.56-mmol/L increase in mean BG, and OR = 1.06 (95% CI: 1.03-1.08, p < .001) for every 0.56-mmol/L increase in maximum BG. Mean BG was not associated with 30-day readmission while maximum BG had a borderline association: OR = 1.02 (95% CI: 1.00-1.03, p = .06). LIMITATION: The results are not generalizable to all cardiovascular surgical patients since only those undergoing invasive procedures were included in the study. CONCLUSIONS: Higher mean and maximum BG levels were associated with increased risk of in-hospital mortality but not with 30-day readmission. Further research is needed to identify optimal BG targets and the effects of avoiding extreme hyperglycemia on patient outcomes.

Economic and clinical impact of inpatient diabetic hypoglycemia.

OBJECTIVE: To assess the clinical and economic impact of hypoglycemia that develops during hospitalization of patients with diabetes. METHODS: In this retrospective cohort study, data from 70 hospitals were used to identify the first inpatient encounter for adult patients with diabetes. Patients were included if all blood glucose measurements were 70 mg/dL or higher during the first 24 hours and their primary discharge diagnosis was for a condition other than hypoglycemia. Those who developed laboratory evidence of hypoglycemia (blood glucose <70 mg/dL after 24 hours) were compared with patients whose blood glucose values were all 70 mg/dL or higher. An alternative definition of hypoglycemia (blood glucose <50 mg/dL after 24 hours) was also evaluated. We adjusted for potential confounders with multivariate models. RESULTS: Hypoglycemia had an adverse effect on all outcomes among more than 100,000 diabetic patients. After adjustment, patients with diabetes who developed hypoglycemia had higher charges (38.9%), longer lengths of stay (3.0 days), higher mortality (odds ratio, 1.07; 95% confidence interval, 1.02-1.11), and higher odds of being discharged to a skilled nursing facility (odds ratio, 1.58; 95% confidence interval, 1.48-1.69) than diabetic patients without hypoglycemia (P<.01 for all). In all cases, using the lower threshold (<50 mg/dL) to define hypoglycemia resulted in similar findings with a larger magnitude of differences. CONCLUSIONS: Although a direct causal relationship cannot be inferred, these study findings suggest the importance of carefully maintaining euglycemia during hospitalizations. Whether the observed worse outcomes were due to hypoglycemia itself or whether they were a marker of worse outcomes due to other causes requires further research.