M. M. Jamil2, A. Akujuo1, M. Tafen Wandji1, J. Ortiz2 1Albany Medical College,Albany, NY, USA 2University Of Toledo Medical Center,Surgical Education,Toledo, OH, USA
Introduction:
Alcohol abuse has been associated with various cardiac dysfunctions. The effect of alcohol abuse on perioperative mortality and complications after cardiac surgery is unclear.
Methods:
We analyzed the National Inpatient Sample database for patients over 30 years having CABG and/or other cardiac (OC) procedures (valvuloplasty, septal defect repair etc) between 2008 and 2012. Patients were stratified based on presence or absence of alcohol abuse (AA). Chi square/Kruskall Wallis test were used to compare demographics and comorbidity. The effect of alcohol abuse on mortality was estimated as Odd’s ratio and 95% confidence interval, using logistic regression with adjustment for demographics and comorbidities. Length of stay, charges, discharge disposition, and complications were also compared.
Results:
Between 2008 and 2012, a total of 1,482,018 patients (1,129,236 CABG and 550,709 OC) were reviewed. The incidence of AA was 2.7%. Patients in the 30-50 years category had the highest incidence of AA (4.31%). Males had a higher incidence of AA (3.6%) compared to females (0.8%). Blacks had the highest incidence of AA (3.4%) and Asian/Pacific Islanders had the lowest (1.3%). Patients with non-elective admission had a higher incidence (3.5 vs. 2.0%) of AA compared to elective ones. Patients with higher Elixhauser comorbidity index had a higher incidence of AA (2.4% in the first and 3.4% in the fourth quartile of ECI). Mortality among patients with AA was 1.8% compared to 2.9% among ones without AA (p=0.000). Unadjusted Odd’s ratio for mortality for AA was 0.61 (95% CI: 0.52-0.72) and persisted after adjustment for age, sex, elective/ non-elective status and Elixhauser comorbidity index. The incidence of atrial fibrillation was lower among patients with AA (28.73% vs. 32.63%, p=0.000). Postoperative respiratory failure and hepatic encephalopathy were higher among patients with AA (10.63 vs. 6.73%, p=0.000 and 0.22 vs. 0.07%, p=0.000 respectively). Length of stay and charges were slightly higher among patients with AA (median 9 vs. 8 days and USD 168,512 vs. 156,928 respectively).
Conclusion:
Our results appear counterintuitive to traditional views regarding alcohol abuse among cardiac surgery patients. Although alcohol has been associated with cardiac dysfunction, there may be a mechanism for the apparent protective effect in the immediate postoperative period, which merits further investigation.