Background: Catheter-related bloodstream infection (CRBSI) causes substantial morbidity and mortality, but few randomized, controlled studies have been conducted to guide therapeutic interventions.
Methods: To determine whether linezolid would be noninferior to vancomycin in patients with CRBSI, we conducted an open-label, multicenter, comparative study. Patients with suspected CRBSI were randomized to receive linezolid or vancomycin (control group). The primary end point was microbiologic outcome at test of cure 1-2 weeks after treatment, as assessed by step-down procedure. The first analysis population was complicated skin and skin structure infection (cSSSI) in patients with suspected CRBSI; patients with CRBSI were analyzed if noninferiority criteria (lower bound of the 95% confidence interval [CI] not outside -15%) were met.
Results: Noninferiority criteria were met for cSSSI (microbiologic success rate for linezolid recipients, 89.6% [146 for 163 patients]; for the control group, 89.9% [134 of 149]; 95% CI, -7.1 to 6.4) and CRBSI (for linezolid recipients, 86.3% [82 of 95]; for the control group, 90.5% [67 of 74]; 95% CI, -13.8 to 5.4). The frequency and severity of adverse events were similar between groups. Mortality rates were 10.4% for linezolid recipients (28 of 269 patients) and 10.1% for control subjects (26 of 257) in the modified intent-to-treat population (i.e., all patients with gram-positive baseline culture) through test of cure, and they were 21.5% for linezolid recipients (78 of 363) and 16.0% for the control group (58 of 363; 95% CI, -0.2 to 11.2) for all treated patients through poststudy treatment day 84.
Conclusions: Linezolid demonstrated microbiologic success rates noninferior to those for vancomycin in patients with cSSSIs and CRBSIs caused by gram-positive organisms. Patients with catheter-related infections must be carefully investigated for the heterogeneous underlying causes of high morbidity and mortality, particularly for infections with gram-negative organisms.