Risk factors affecting in-hospital mortality in patients with nosocomial infections.

BACKGROUND/PURPOSE: Nosocomial infection (NI)-associated death is an important issue for both patients and clinicians, and is of emerging importance in public health. This study investigated the factors associated with in-hospital deaths among patients with NI. METHODS: Between July 1, 2002 and June 30, 2003, a total of 1574 patients with NI at National Taiwan University Hospital were enrolled to investigate the factors associated with fatal outcome. Host factors, hospital services, surgical and medical interventions, microbial factors, infection sites, and the treatment and complications of NI were analyzed retrospectively. RESULTS: During the study period, 554 of the 1574 patients died (mortality rate, 28.3%). NI was directly involved in 80.5% of them (n = 446), and over two-thirds (67.9%) of deaths occurred within 2 weeks of NI onset. Sixteen variables were statistically implicated as independent factors significantly associated with mortality. Host factors included higher disease severity (p < 0.0001), liver cirrhosis (p < 0.0001), solid tumors (p < 0.0001), chronic lung disease (p = 0.003), and congestive heart failure (p = 0.005). Hospital and interventional factors included intensive care hospitalization (p = 0.002), longer hospitalization before NI onset (p = 0.004), hemodialysis (p = 0.0003), arterial-line insertion (p < 0.0001), urinary catheterization (p < 0.0001), and central venous catheterization (p = 0.001). Blood stream infections (p < 0.0001), NI due to Candida (p < 0.0001), and multiple (> or = 2) episodes of NI (p < 0.0001) were significant risk factors for death, as were occurrence of NI-associated septic shock (p < 0.0001) and disseminated intravascular coagulation (p < 0.0001). No significant associations of mortality with age, sex, species of bacteria, multi-antibiotic resistant bacteria, regimen for initial treatment, or multiple antibiotic therapy were evident. CONCLUSION: Measures that prevent the occurrence of NI, such as improving the immunity status of the host, removal of catheters as soon as possible, and implementing an infection control program, could reduce the risk of in-hospital deaths attributable to NI.