Setting
The Balgrist University Hospital prospectively registers all moderate to severe infections, including many postoperative complications, since July 2018. The immediate pre-lockdown period witnessed no specific Covid-19 policies besides the promotion of the "respiratory etiquette" and HH [5,6,7,8,9]. During the lockdown, the authorities banned all elective surgeries and visitors in hospitals. They implemented social distancing and home office for healthcare workers (HCW) at risk [2]. Hence, during the lockdown, only emergency patients were treated, or patients with multiple co-morbidities transferred from other hospitals to release capacities for their severely ill Covid-19 patients. Importantly, a mandatory mask use was introduced only after the first epidemic wave, as was Contact Tracing and the post-exposition quarantine for asymptomatic HCW with close contact to Covid-19 positive persons.
Data collection and study criteria
This study followed the ethical principles of the Helsinki Declaration and approval of the project was obtained from the Cantonal Ethics Commission. A one-year study period (October 1, 2019–October 31, 2020) was arbitrarily chosen, was divided into three periods and included all surgeries performed in the operation theater: pre-lockdown period with 2688 interventions from October 1, 2019 to March 15, 2020; the Covid-19 lockdown period with 230 surgeries from March 16, 2020 to April 26, 2020 (Fig. 1); and a post-lockdown period with 2873 interventions from April 27, 2020 to October 31, 2020. The most important reason regarding the choice for this precise study period was the stability of the surgical teams, lasting on average one academic year. Similarly, the one-year study period, six months before and six months after the lockdown, guaranteed the continuity of the operating personnel. Additionally, for this chosen period, the accuracy of the perioperative prophylactic antibiotic regimens could be verified. Finally, the two control periods (before and after lockdown) reduced the selection bias compared to only one control period (Fig. 1).
All adult patients undergoing orthopedic surgery at our institution were included. A general consent form allowed the registration of healthcare data for scientific usage signed by all included patients. The following exclusion criteria were used: adolescent patients, patients without agreeing to the general consent, and surgeries with a minimal active follow-up of less than 30 days. Database closure was on November 30, 2020. On this time, Switzerland already entered the second pandemic wave.
Literature search
A literature review was performed to compare our orthopedic results of deep SSI, HAI and HH performances during the first lockdown for pandemic Covid-19 with available publications from other surgical and non-surgical specialties. For the literature search in English language, the following MeSH terms in PubMed and on the internet were used: "nosocomial infection", "healthcare associated infection", "surgical site infections", "lockdown", "hand hygiene", and "Covid".
Outcome parameters, setting and definitions
The primary outcome was the incidence of deep SSI [10,11,12] after the index surgery. According to internationally accepted norms [10,11,12], SSI is defined as an event acquired in the operating theater manifesting within 30 days after the intervention, with drainage of purulent fluid out of the incision or presence of typical infectious signs (rubor, calor, tumor, dolor), and requiring revision surgery.
Secondary outcomes were the incidence of non-infectious postoperative wound healing disorders, postoperative local complications other than infection or visible wound problems, and the epidemiology of other non-surgical HAI between the three study periods. Wound healing disorders were defined as substantial necrosis, uninfected dehiscence without typical infectious signs and/or hematoma necessitating surgical drainage. Other causes for revision surgery were recurrence of disease, residual symptoms and/or intervention-specific complications. Key HAI were grouped as urinary tract infections, bacterial pneumonia, Covid-19 disease, and bloodstream infections [4]. Nosocomial infections attributed to other clinics were excluded. Additionally, the following parameters known to be associated with SSI [11] were documented: sex, age, body mass index (BMI), American Society of Anesthesiology (ASA)-Score, diabetes, date, types (including primary or revision surgery) and localization of index surgery, duration of surgery and length of hospital stay.
On the hospital's level, three collaborators with experience in infection control observed the HH compliance [5] at the beginning (October 2019) and the end of the study period (September 2020). There were no HH observations during the total lockdown. The HH observations were performed at the end of the study period to judge an eventual residual effect five months after the end of the lockdown. Additionally, the accuracy of the protocolled perioperative antibiotic prophylaxis was assessed. According to the intra-hospital standards, perioperative antibiotic prophylaxis was implemented with cefuroxime or with clindamycin (or vancomycin) in case of intolerance. The preoperative surgical skin site disinfection was performed with chlorhexidine or povidone-iodine. All operated patients were seen for a regular postoperative control 4–6 weeks after the index surgery.
Statistical analysis
Group comparisons were performed using the Pearson-χ2 (categorical variables), the Wilcoxon-rank sum-test or the Kruskal–Wallis-test for non-parametric, continuous variables. To adjust for the heterogeneity of the surgeries and the large case-mix imposed by the ban of elective surgeries, three multivariate Cox regression analyses were performed. The three final Cox regression models targeted three different outcomes: "SSI", "wound healing disorders" and "other complications". In the first multivariate model, the group of SSI was compared to all other non-infected surgery episodes. In the second model, non-infectious wound problems were compared to all episodes without infection or wound problems, and in the last analysis, all local complications were compared to all uneventful surgeries. Most variables were analyzed as a continuum, but stratifications were added for the ASA-Score, the duration of surgery and the study period. The cut-off values for these strata relied on the 25%, 50% and 75% percentiles of the distribution of values of that variable. To adjust for the case-mix of the surgeries within the three study periods, we recurred to a multivariate Cox regression analysis by inserting stepwise independent variables with a p value ≤ 0.05 from the univariate analysis into the final model. The software IBM SPSS Statistics 25 and STATA™ (15.0, College Station, USA) were used and p values ≤ 0.05 (two-tailed) were considered significant.