Routine replacement of peripheral intravenous catheters is associated with fewer bloodstream infections

Reference

Buetti N, Abbas A, Pittet D, et al. Comparison of routine replacement with clinically indicated replacement of peripheral intravenous catheters. JAMA Intern Med 2021;181(11):1471-1478.

Synopsis

This study included all hospitalized patients with at least one peripheral venous catheter (PVC) who were admitted to a large, tertiary care hospital system in Switzerland during a four-year period. The primary outcome was a PVC-associated blood stream infection (BSI) defined as an infection that occurred from the day of PVC insertion until 48 hours after catheter removal and a blood culture and PVC tip culture with the same organism. Cultures with common skin contaminant bacteria were only considered clinically significant if a patient had fever/chills or hypotension or had 2 positive cultures from 2 separate blood samples within 48 hours. During the time of the study, the hospital system went from a baseline period of routine replacement of PVCs every 96 hours to an intervention period of clinically indicated PVC replacement only, followed by a reversion period during which routine replacement was re-implemented. A total cohort of 412,631 PVCs in 164,331 patients was identified. Although the monthly number of PVC-days stayed the same throughout the different study periods, the number of newly inserted PVCs decreased during the intervention period, as expected. Additionally, during the intervention, 20% of PVCs stayed in place for 4 or more days as compared with only 10.9% during the baseline period. Overall, the rates of BSI were quite low: 11 in the baseline period, 46 in the intervention period, and 4 in the reversion period. There was a statistically significant higher incidence of BSI during the intervention period than in the baseline period (0.90 BSI per 10,000 catheter-days compared with 0.13 per 10,000 catheter-days; incidence rate ratio [IRR] 7.20; 95% CI 3.65 - 14.22; P < .001). No such difference was seen when comparing the baseline period with the reversion period. Results were confirmed in multiple sensitivity analyses.