Jugular versus femoral short-term catheterization and risk of infection in intensive care unit patients. Causal analysis of two randomized trials.

RATIONALE: When subclavian access is not possible, controversy exists between the internal jugular and femoral sites for the choice of central-venous access in intensive care unit patients. OBJECTIVES: To compare infection and colonization rates of short-term jugular and femoral catheters. METHODS: Using data from two multicenter studies, we compared femoral and internal jugular for the risks of catheter-related bloodstream infection, major catheter-related infection, and catheter-tip colonization. We also compared the rates of dressing disruption and skin colonization. We used marginal structural models with inverse probability of treatment weighting to adjust on indication bias. MEASUREMENTS AND MAIN RESULTS: We included 2,128 patients (2,527 catheters and 19,481 catheter-days). We found no difference in catheter-related bloodstream infection (internal jugular 1.0 vs. femoral 1.1 per 1,000 catheter-days; hazard ratio [HR], 0.63 [0.25-1.63]; P = 0.34), major catheter-related infection (internal jugular 1.8 vs. femoral 1.4 per 1,000 catheter-days; HR, 0.91 [0.38-2.18]; P = 0.34), and colonization (internal jugular 11.6 vs. femoral 12.9 per 1,000 catheter-days; HR, 0.80 [0.25-1.63]; P = 0.15). However, colonization was higher with femoral for female (HR, 0.39 [0.24-0.63]; P < 0.001) and, at the significance limit, catheter maintained for more than 4 days (HR, 0.73 [0.53-1.01]; P = 0.05). The absence of benefit of internal jugular before Day 5 was related to a higher skin colonization at the internal jugular site for catheters removed before Day 5. After the fourth day, dressing disruption became more frequent with femoral catheters and may explain the subsequent risk of catheter colonization. Differences in cutaneous and catheter colonization between internal jugular and femoral was suppressed by the use of chlorhexidine-impregnated dressings. CONCLUSIONS: Femoral and internal jugular accesses lead to similar risks of catheter infection. Internal jugular might be preferred for female, nonchlorhexidine-impregnated dressings users, and when catheters are left in place more than 4 days. Both sites are acceptable when a subclavian approach is not feasible. Clinical trial registered with www.clinicaltrials.gov (NCT00417235 and NCT01189682).