MON-700 Protective Effects of Smoking? Improved In-Hospital Mortality in Smokers Admitted for Diabetic Ketoacidosis

PURPOSE: To determine the relationship between tobacco smokers and the rate of hospital readmission within 30 days, mortality, morbidity, and health care resource utilization in patients admitted to the hospital in the United States with diabetic ketoacidosis (DKA). METHOD: A retrospective study was conducted using the AHRQ-HCUP National Inpatient Sample for the year 2014. Adults (≥ 18 years) with a principal diagnosis of DKA and a secondary diagnosis of tobacco dependence or active smokers were identified using ICD-9 codes as described in the literature (1). The primary outcome was in-hospital mortality. Secondary outcomes include readmission rate, length of hospital stay (LOS), total hospitalization costs. Propensity score (PS) using the next neighbor method without replacement with 1:1 matching was utilized to adjust for confounders (2). Independent risk factors for readmission were identified using multivariate logistic regression model (3). RESULTS: In total, 186,824 hospital admissions with a primary diagnosis of DKA were identified, of which 32.44% (47,382) had TD. In-hospital morality among the smoking cohort (0.39%, SD 0.03) was lower than the non-smoking cohort (0.32%, SD 0.04) during the first hospitalization. Similar effects were observed after propensity match - 0.33% (SD 0.18) vs 0.27% (SD 0.03). The mortality rate during next hospitalization was also lower in the smoking cohort (0.72%, SD 0.03) in comparison to their counterpart (1.16%, SD 0.01). Smokers had a higher readmission rate of 17.6% (SD 0.57) than non-smokers (9.6%, SD 0.25). The length of stay among smokers and non-smokers were similar after propensity match - 3.12 days (SD 0.03) vs 3.06 days (SD 0.09), p=0.42, respectively. Total hospital cost was also similar between the two groups, $6,898 (SD $82) vs $7,100 (SD $203), p=0.32, respectively. Based on multivariate logistic regression, female and high Charlson comorbidity index were associated with higher 30-day readmission rate; whereas private insurance and high household income were associated with reduced readmission rate. CONCLUSION: Smoking has been associated with improved survival in patients with DKA (4). Previous studies have shown that glucose concentration were significantly lower at fasting and 120 min in current smokers than non-smokers. However, the effects of cigarette smoking on glucose metabolism and insulin resistance are still disputed.