Catheter-related bloodstream infections in a nephrology unit: Analysis of patient- and catheter-associated risk factors

Background: Central venous catheter use is rising in chronic and acute hemodialysis. Catheter-related bloodstream infections are a major complication of central venous catheter use. This article examines clinical factors associated with catheter-related bloodstream infections incidence. Methods: In this retrospective, single-center study, 413 patients undergoing extracorporeal treatments between 1 February 2014 and 31 January 2017 with 560 central venous catheters were recruited. Clinical parameters, such as gender, age, kidney disease status, diabetes, immunosuppression, and vintage dialysis, were collected at study entry. An incidence rate ratio (95% confidence interval) was calculated to assess the association between catheter-related bloodstream infections incidence rate and each clinical variable/central venous catheter type. Significant associations at the univariate analyses were investigated with multivariate Cox models. Results: During a cumulative time of 66,686 catheter-days, 54 catheter-related bloodstream infections (incidence rate: 0.81) events occurred. Gram negative bacteria were more frequent in patients with age < 80 years (16 (36%) vs. 0, p = 0.02). At the univariate analyses, male sex (incidence rate ratio: 1.9 (1.1-3.5), p = 0.03), age < 80 years (incidence rate ratio: 2.4 (1.1-5.5), p = 0.016) and acute kidney injury (incidence rate ratio: 5.6 (3.1-10), p < 0.0001) were associated with higher catheter-related bloodstream infections incidence rate. Compared with tunneled jugular central venous catheter, higher catheter-related bloodstream infections incidence rate was associated with non-tunneled jugular (incidence rate ratio: 6.45 (2.99-13.56), p < 0.0001) and non-tunneled femoral (incidence rate ratio: 12.90 (5.87-27.61), p < 0.0001) central venous catheter use; tunneled femoral central venous catheter was associated with higher non-significant incidence rate (incidence rate ratio: 2.45 (0.93-5.85), p = 0.07). The multivariate analyses showed that acute kidney injury (hazard ratio: 3.03 (1.38-6.67), p = 0.006), non-tunneled (hazard ratio: 3.11 (1.30-7.41), p = 0.01) and femoral (hazard ratio: 2.63 (1.36-5.07), p = 0.004) central venous catheter were associated with higher catheter-related bloodstream infections incidence rate. Conclusion: Central venous catheter characteristics and acute kidney injury are independently associated with higher catheter-related bloodstream infections rate.