Nosocomial infection after lung surgery: incidence and risk factors.

ABSTRACT

STUDY OBJECTIVES: To assess the incidence and risk factors for nosocomial infection after lung surgery. DESIGN: Prospective cohort study. SETTING: Service of thoracic surgery of an acute-care teaching hospital in Santander, Spain. PATIENTS Between June 1, 1999, and January 31, 2001, all consecutive patients undergoing lung surgery were prospectively followed up for 1 month after discharge from the hospital to assess the development of nosocomial infection, the primary outcome of the study. INTERVENTIONS: During the hospitalization period, patients were visited on a daily basis. Postdischarge surveillance was based on visits to the surgeon. MEASUREMENTS AND RESULTS: We studied 295 patients (84% men; mean age, 60.9 years), 89% of whom underwent resection operations. Ninety episodes of nosocomial infection were diagnosed in 76 patients, including pneumonia (n = 10), lower respiratory tract infection (n = 47), wound infection (n = 16; one third were detected after hospital discharge), urinary tract infection (n = 9), and bacteremia (n = 8; three fourths were catheter-related bacteremia). Twenty patients had severe infections (pneumonia or empyema), with a mortality rate of 60%. COPD (adjusted odds ratio [OR], 2.70; 95% confidence interval [CI], 1.52 to 4.84), duration of surgery with an increased risk for each additional minute (Mantel-Haenzel chi(2) test for trend, p = 0.037), and ICU admission (OR, 3.69; 95% CI, 1.94 to 7.06) were independent risk factors for nosocomial infection. The use of an epidural catheter was a protective factor (OR, 0.45; 95% CI, 0.22 to 0.95). There were no differences according to the use of amoxicillin/clavulanate or cefotaxime for surgical prophylaxis. CONCLUSIONS: Nosocomial infections are common after lung surgery. One third of wound infections were detected after hospital discharge. The profile of a high-risk patient includes COPD as underlying disease, prolonged operative time, and postoperative ICU admission.