Incidence of Blood Stream Infection and Impact on Survival in Patients on Left Ventricular Assist Device Support

Purpose

Bloodstream infection (BSI) is a severe complication in patients with continuous flow left ventricular assist devices (CF-LVADs). In this study, we aimed to describe the incidence of BSI on contemporary LVAD support and its impact on subsequent survival.

Methods

This analysis included adult patients with first time CF LVAD implantation at a single center (June 2005-June 2020). BSI was defined as ≥1 positive blood culture with a likely pathogen (e.g. Staphylococcus aureus) or >1 positive blood culture with a low virulence skin organism (e.g. Staphylococcus epidermidis). Time to BSI was recorded for each patient. Survival-free of BSI was estimated using Kaplan-Meier method and compared among device types. Survival among patients who did and did not develop BSI by 9 months post implant was compared using Cox regression adjusting for medical and social factors.

Results

502 patients were included in the analysis. Over a median follow up period of 1,468 days (IQR 454-3,081 days), 73 (14.5%) developed a BSI. Patients who developed BSI were more likely to have diabetes, implanted cardiac defibrillators, and to have a designation of destination therapy. Median time to BSI was 389 days (IQR 94-997). The most common organisms were enterococci (23%) and streptococci (23%). At the time of BSI, 53% of patients had a WBC count above 11 × 10⁹/L. Compared to the HeartMate 2, patients implanted with the HeartMate 3 had a lower rate of 1-year survival free of BSI (87% vs. 95 %, p=0.007). When comparing survival after 9 months among those with and without BSI, mortality was significantly higher in patients with BSI in both unadjusted (HR =1.96, 95% CI; 1.02 - 3.76) and adjusted models (2.00, 95% CI; 1.03 - 3.88).

Conclusion

BSI continues to occur in the contemporary era of LVAD therapy and is associated with significantly worse survival. The absence of elevated WBC count in LVAD patients with severe infections such as bacteremia raises the question of whether leukocytosis should be infection criteria for UNOS status increase.