We analysed patient and catheter characteristics in a retrospective cohort of more than 200,000 PVCs to identify mechanisms of increased PVC-BSI during the second year of the COVID-19 pandemic. We undertook dedicated prospective surveillance of all hospital-onset BSI and identified an increase in PVC-BSI between 2020 and 2021. In 2020, at the beginning of pandemic conditions in Switzerland, absolute COVID case numbers were lower in our institution; in 2021, PVC-BSI increased in line with rising COVID-19 case numbers. In our analysis of patient and catheter characteristics, we did not identify significant differences that would explain this increase, and only a minority of patients with PVC-BSI were admitted to dedicated COVID-19 units. Taken collectively, this may suggest institutional, systems-level factors to be responsible for increased PVC-BSI during pandemic conditions. Possible explanations include diversion of IPC resources and expertise, redeployment of staff to new envionments, multiple rapid changes in IPC practice, and strain on hospital resources, workload and staffing [13].

Very few previous data describing PVC-BSI are available for comparison. Regarding central lines, increased CLABSI during the COVID-19 pandemic, have been reported in single-centre studies and routine surveillance data [5, 6, 14,15,16]. These reports include surveillance data from the United States reported to the Centers for Disease Control and Prevention/National Healthcare Safety Network [5, 15], and from the Dutch national surveillance system PREZIES [17]. In addition, a consortium of seven low and middle income countries reporting data to the International Nosocomial Infection Control Consortium (India, Mongolia, Jordan, Lebanon, Palestine, Egypt, and Turkey) reported increased CLABSI rates after the beginning of the COVID-19 pandemic [14]. Dedicated single centre studies in the United States [6] and Europe [16] have also reported a significant increase in CLABSI rates. However, these did not include PVC, and primarily report data from 2020 only. Additionally, most studies include very limited patient- and catheter-related data. These reports do not investigate specific causative mechanisms for the changes in hospital acquired infections during pandemic conditions.

In our institution, only a minority of PVC-BSI were reported from patients admitted to dedicated COVID-19 wards, or with active COVID-19 infection. This is in contrast to one single-centre cohort study in the United States, which reported increased CLABSI risk in patients with COVID-19, but no difference in CLABSI rates in the non-COVID hospital population [6]. However, other multi-centre studies have described increased non-COVID mortality relating to global effects of pandemic conditions, including hospital strain, with non-COVID mortality rising in proportion to COVID-19 admission rates [13, 18].

We identified several patient groups at elevated risk of PVC-BSI. Admission to ICU and wide-gauge catheter were associated with increased risk of infection. In addition, more than 95% of all PVC-BSI occurred in catheters inserted either in the emergency department, or surgical wards. This is consistent with previous findings of increased PVC-BSI risk with wide-gauge catheters [19]. These data suggest that acutely unwell patients, who may have catheters inserted in emergency conditions, are at higher risk and represent a target group for prevention. However, both ICU admission and wide-gauge catheter were more common in 2020 than 2021, and therefore cannot explain the rise in PVC-BSI over this study period.

Patients with prolonged PVC-dwell time were also disproportionately represented in the PVC-BSI group. Routine PVC change every four days was recommended in our institution over the study period. Adherence to this policy was around 83%, but approximately half of all PVC-BSI occurred in patients with PVC dwell time above four days. Adherence to PVC replacement was stable and does not explain the increase in PVC-BSI over the COVID period, but may be a target to reduce PVC-BSI risk [9].

Our study has several limitations. First, this investigation may be under-powered to detect differences in catheter- or patient-related factors that contribute to changing risk. Second, we describe only PVC-BSI in the first two years of the COVID-19 pandemic. We did not compare pre- and post-pandemic conditions. The study period was determined by a previous observation of increased PVC-BSI rates during the alpha and delta waves in 2021 [7] Pre-pandemic analysis was not possible, as prior to 2020 we conducted an intervention investigating routine versus clinically indicated PVC replacement, which would have confounded our observations [9]. Further analyses that include the pre-pandemic and post-pandemic periods (2022 onward) are needed to better understand the impact of pandemic conditions.

Third, due to low incidence of intravascular catheter infections, we fitted only univariable Cox models to identify PVC-BSI risk factors. Multivariable analysis was not performed due to low event number. This increases the risk of confounding and precludes analysis of interactions between different risk factors; the results of this analysis should therefore be interpreted cautiously.

Finally, potentially contributory hospital factors such as bed occupancy, staff absences, hospital systems changes and compliance with infection prevention measures, were not assessed. We hypothesise that system-wide changes in the first two years of the COVID-19 pandemic contributed to increasing risk of nosocomial infection. In March 2020, the HUG hospital network was restructured to allow admission of large numbers of COVID-19 patients. This included: transformation of medical units into units dedicated to COVID-19 care; diversion of non-COVID-19 patients to other hospitals in the area; and reassignment of health care workers (HCWs) to different care environments. In this setting, burnout, workforce shortages, multiple rapid changes in procedure and loss of specific expertise due to diversion of staff to new roles may all have contributed to increasing rates of nosocomial infection.

This study has several strengths. Dedicated prospective surveillance of all hospital-onset BSIs was conducted throughout the study period. We applied ECDC surveillance definitions with high specificity, and assessed a range of patient-level clinical data and catheter-related data.