Pre-operative risk factors for driveline infection in left ventricular-assist device patients

Abstract

Aims: Implantation of left ventricular-assist devices (LVAD) to treat end-stage heart failure is of increasing relevance due to donor shortage. Infections of the driveline are common adverse events. LVAD infections can lead to high urgency listings for transplantation. However, transplantation in patients with infection leads to worse post-transplantation outcomes. This study aims to evaluate specific risk factors for driveline infections at the time of implantation.

Methods and results: Four hundred forty-one patients receiving either Heartmate II or Heartware system from August 2009 to October 2013 were assessed. An expert committee sorted patients into four different groups concerning the likeliness of infection. Twenty-eight (6%) of discussed infection cases were judged as secured, 33 (7%) as likely, 18 (4%) as possible, and 20 (4%) as unlikely. The remaining 342 (78%) subjects showed either no signs of infection at all times (329 [75%]) or developed signs of infection in a second observation period within 1 year after ending of the first observation period (13 [3%]). For a better discriminatory power, cases of secured and likely infections were tested against the group with no infection at all times in a Cox proportional hazard model. Among all variables tested by univariate analysis (significance level P < 0.15), only age (P = 0.07), LVAD-type (P = 0.12), need for another thoracic operation (P = 0.02), and serum creatinine value (P = 0.02) reached statistical significance. These were subsequently subjected to multivariate analysis to calculate the cumulative risk of developing a drive infection. The multivariate analysis showed that of all the potential risk factors tested, only the necessity of re-thoracotomy or secondary thoracic closure had a significant, protective effect (hazard ratio [95% CI] = 0.45 [0.21-0.95]; P = 0.04).

Conclusion: This single-centre cohort study shows that driveline infections are common adverse events. The duration of support represents the major risk factor for LVAD driveline infections.

Keywords: PET/CT; infection; left ventricular-assist device; risk factor.

Figure 1

Grouping of the entire cohort concerning likeliness of infection according to a specialist panel. For statistical analysis, the secured driveline infection group (A1 + A2) was tested against the secured non‐driveline infection group (B1).

Figure 2

Overview of the bacteria involved leading to left ventricular‐assist devices driveline infections

Figure 3

Cumulative hazard for developing a LVAD driveline infection dependent on length of time after implantation

Figure 4

Negative correlations between the necessity for another thoracic operation immediately or a few days after implantation and LVAD driveline infections during the time period (excluded major surgical complications or those who may have been lost)

Figure 5

Negative correlations between creatinine level >1.3 mg/dl and LVAD driveline infections during the time

Figure 6

Exemplary images of a LVAD driveline infection (arrow) in a 59‐year‐old male patient detected by ¹⁸F‐fluorodeoxyglucose PET (A) coronal and (D, E) axial layering) and after image fusion with a low dose CT ((C) axial layering). A higher uptake of ¹⁸F‐fluorodeoxyglucose can be seen around the driveline in the abdominal wall. (B) low dose CT. (D) attenuation corrected ¹⁸F‐fluorodeoxyglucose PET. (E) non‐attenuation corrected ¹⁸F‐fluorodeoxyglucose PET