To assess, whether S. aureus nasal colonization is a risk factor for infections in patients with durable ventricular assist device (VAD).

Prospective, single-centre, cohort study (i) ascertaining S. aureus nasal colonization status of patients admitted for VAD-implantation and detecting time to first episode of VAD-specific or -related infection according to International Society for Heart and Lung Transplantation criteria during follow-up and (ii) comparing whole genomes of S. aureus from baseline colonization and later infection.

Among 49 patients (17 colonized, 32 non-colonized), S. aureus VAD-infections occurred with long latency after implantation (inter quartile range 76–217 days), but occurred earlier (log-rank test P = 0.006) and were more common (9/17, 52.9% vs. 4/32, 12.5%, P = 0.005; incidence rates 2.81 vs. 0.61/1000 patient days; incidence rate ratio 4.65, 95% confidence interval 1.30–20.65, P = 0.009) among those nasally colonized with S. aureus before implantation. We found a similar but less pronounced effect of colonization status when analysing its effect on all types of VAD-infections (10/17, 58.8% vs. 7/32, 21.9%, P = 0.01). These findings remained robust when adjusting for potential confounders and restricting the analysis to ‘proven infections’. 75% (6/8) of paired S. aureus samples from colonization and VAD-infection showed concordant whole genomes.

In patients with durable VAD, S. aureus nasal colonization is a source of endogenous infection, often occurring months after device-implantation and affecting mostly the driveline. Hygiene measures interrupting the endogenous route of transmission in VAD-patients colonized with S. aureus long-term may about half the burden of infections and require clinical scrutiny.