Clinical impact of positive blood cultures in intensive care patients
KeywordsBlood Culture Respiratory Failure Prospective Cohort Prospective Cohort Study Initial Diagnosis
Infectious complications are frequent in patients admitted to ICUs, and great effort is made in order to identify possible infecting microorganisms. In this setting, multiple blood cultures are usually collected, although their true value is still to be ascertained.
To ascertain the clinical impact of positive blood cultures in patients admitted to the ICU.
We carried out a prospective cohort study in 340 patients in a general ICU of a tertiary hospital from June 2003 to November 2004. Clinical characteristics and blood cultures from all patients were reviewed. Differences were considered significant when P < 0.05.
One hundred and eighty-one (53.2%) patients were male, and the mean age was 51.9 (12–88) years. The initial diagnosis was respiratory failure in 115 (33.8%), cardiovascular failure in 97 (28.5%), neurological in 41 (12.1%), diabetic ketoacidosis in 27 (7.9%), postoperatory status in 21 (6.2%) and other conditions in 39 (11.5%). The median APACHE score was 17 (range 2–45) and 107 patients (31.5%) died in the ICU. Four hundred and eighty-eight blood cultures were drawn from 201 patients; 156 (31.9%) cultures were positive in 93 (46.2%) patients. Gram-positive cocci were found in 44 (21.9%) patients and Gram-negative bacilli in 25 (12.4%) patients. As compared with patients with negative blood cultures, patients with positive blood culture had a longer ICU stay (32.5 ± 50.5 vs 12 ± 12.3 days, P < 0.001), a higher initial SOFA score (7.2 ± 1.9 vs 5.6 ± 4.4, P = 0.01), a higher total SOFA score (152.2 ± 119.8 vs 62.8 ± 76.5, P < 0.001), a higher maximum SOFA score (11.1 ± 4.1 vs 8.3 ± 5.7, P < 0.001), a higher frequency of central venous catheterization (90.3% vs 71.3%, P = 0.002), urinary catheterization (86% vs 6.8%, P = 0.001), invasive blood pressure monitoring (44.1% vs 27.8%, P = 0.06) and a higher frequency of mechanical ventilation (80.6% vs 50.9%, P < 0.001). Mortality was increased in patients with positive blood cultures (50.5% vs 32.4%, P = 0.009).
Patients admitted to the ICU that develop positive blood cultures are more severely ill than patients with negative blood cultures. The finding of an infecting microorganism in the blood of intensive care patients does not warrant a better prognosis.