IJCS | Volume 32, Nº2, May/June 2019

222 Figure 1 - Frequency of hospital infection by infection site during hospital stay. Incidence (%) Site of infection Respiratory tract Urinary tract Mediastinum Soft tissue Invasive device Hospitalized, waiting for surgery Postoperative Hospital infection (at any time) Lobato et al. Patients undergoing cagb in Pará, Brazil Int J Cardiovasc Sci. 2019;32(3)217-226 Original Article complete atrioventricular block in patients older than 70 years undergoing MRS. We did not observe a relationship of mortality with age. Some studies have reported higher mortality rates among women, as in a study conducted from 2002 to 2010 including 655 patients. 22 Mean admission-to-surgery time was 23 ± 15.9 days, mean intensive care unit stay was 10.5 ± 9.8 days and the mean number of days fromsurgery to hospital discharge/ death was 15.6 ± 14.2 days. The longer waiting time for surgery was associated with higher risk of infectious events during hospitalization, which was probably the main cause of mortality in our study. In a study carried out in Rio de Janeiro, the authors found that the waiting time for MRS had no effect on operative mortality; however, approximately 11.0% of patients died in this period. Factors associated with mortality in these patients were left ventricular ejection fraction < 45% and a waiting time longer than 16weeks. 23 In our population, left ventricular dysfunction was not a predictor of death. Another study conducted at Santa Casa de Limeira , including patients older than 70 years, showed that an intensive care unit stay longer than 48 years was associated with higher mortality, whereas hospital stay was not a predictor of mortality in these patients. 24 With respect to comorbidities, only previous MRS and chronic kidney disease were associated with mortality. Anatomic localization of the arteries affected or postoperative complications (major bleeding, need for transfusion of blood derivatives, complex arrhythmias, stroke and angina) showed no association with mortality. On the other hand, both postoperative infection and kidney injury were associated with higher mortality. Results of a study conducted in a referral center for cardiology care differ from ours, as no difference in mortality was reported among patients who had infection. A study carried out at HCGV from