ObjectiveTo evaluate the efficacy of tunneling short-term central venous catheters to prevent catheter-related infections.Data SourcesMEDLINE, EMBASE, conference proceedings, citation review of relevant primary and review articles, personal files, and contact with expert informants.Study SelectionFrom a pool of 225 randomized, controlled trials of venous and arterial catheter management, we identified 12 relevant trials and included seven of these trials in the analysis.Data ExtractionIn duplicate, independently, we abstracted data on the population, intervention, outcomes, and methodologic quality.Data SynthesisTunneling decreased bacterial colonization of the catheter by 39% (relative risk of 0.61; 95% confidence interval [CI] of 0.39 to 0.95) and decreased catheter-related sepsis with bacteriologic confirmation by 44% (relative risk of 0.56; 95% CI of 0.31 to 1) in comparison with standard placement. The majority of the benefit in the decreased rate of catheter-sepsis came from one trial at the internal jugular site (relative risk of 0.30, 95% CI of 0.10 to 0.89) and the reduction in risk was not significant when the data from five subclavian catheter trials were pooled (relative risk of 0.71, 95% CI of 0.36 to 1.43). Tunneling was not associated with increased risk of mechanical complications from placement or technical difficulties during placement. However, this outcome was not rigorously evaluated.ConclusionsTunneling decreases central venous catheter-related infections. However, current evidence does not support routine tunneling until its efficacy is evaluated at different placement sites and relative to other interventions. (Crit Care Med 1998;26:1452-1457)