Seroprevalence of SARS-CoV-2 Antibodies in Employees of Three Hospitals of a Secondary Care Hospital Network in Germany and an Associated Fire Brigade: Results of a Repeated Cross-Sectional Surveillance Study Over 1 Year

Healthcare workers (HCWs) are playing a vital role in the current SARS-CoV-2 pandemic. This study investigated how infection spreads within three local hospitals and an associated fire brigade in Germany by testing employees for the presence of SARS-CoV-2 IgG antibodies over one year. The three observational periods corresponded to the initial three pandemic waves: first wave: June–September 2020, second wave: October 2020–January 2021, and third wave: February–June 2021. We analysed 3285 serum samples of 1842 employees, which represents 65.7% of all employees. Altogether, 13.2% employees were seropositive: 194/1411 HCWs (13.7%) and 49/431 non-HCWs (11.4%) with a clear increase of seroprevalence from the first (1.1%) to the second (13.2%) and third (29.3%) pandemic wave. HCWs presumably had an additional occupational risk for infection in the second and third wave due to an increase of infection pressure with more COVID-19 patients treated, showing possible weak points in the recommended infection prevention strategy.

The current SARS-CoV-2 pandemic is a worldwide challenge for the medical sector. Healthcare workers (HCWs) are at specific risk for SARS-CoV-2 [8], especially if they are inadequately protected [9,10]. Serological testing of specific antibodies against SARS-CoV-2 has commonly been used to investigate infections of HCWs [11]. An average seroprevalence rate of 8% [12] and 8.6% [13] in HCWs were reported worldwide before the era of vaccination. Since January 2021, the possibility of vaccination has become an add-on to the personal protection and infection control measures.
Data from German HCWs are available from a variety of hospitals, but nearly all published data focus on the first pandemic wave [14][15][16][17][18][19][20] (Table 1). Two hospitals reported data until December 2020 [21,22], but no information for HCWs in Germany is available for the year 2021 so far.  Employees of three hospitals of a secondary care hospital network and an associated fire brigade [39], actual data Our study reports the course of seroprevalence of IgG antibodies against SARS-CoV-2 among employees of three local hospitals of a secondary care hospital network in Germany and an associated fire brigade, an institution mainly responsible for patient transport to and between hospitals, over 1 year (June 2020-June 2021). In addition, we evaluate if HCWs had an additional occupational risk for SARS-CoV-2 infection overtime. Finally, the results are interpreted in the epidemiological context of the local incidence, respectively.

Study Design
The study was a single center study conducted at the St. Vincenz Hospital Datteln (VHD) with 316 beds and the main departments: internal medicine, surgery, gynaecology, obstetrics, and urology. We looked separately at three time periods according to the pandemic waves: first wave: June-September 2020, second wave: October 2020-January 2021, and third wave: February-June 2021. The VHD belongs to the Vestische Caritas Kliniken GmbH hospital network. Within the study time of one year, two other hospitals of the secondary care hospital network took part: St. Laurentius Stift Waltrop (LSW) with 172 beds (geriatric and psychiatric department), and the children's hospital Vestische Kinder-und Jugendklinik (VKJ) Datteln with 244 beds. An associated fire brigade, as institution mainly responsible for patient transport to and between hospitals, took part in the second and third pandemic wave. Furthermore, we investigated 40 employees of associated residential care homes for the elderly.

Enrolment and Data Management
All employees were invited to take part. Written informed consent included a questionnaire and agreement on providing a blood sample (not exceeding 9 mL of venous blood).
In the questionnaire personal data and questions for clinical symptoms 2 months prior testing were recorded. Additionally, we asked for exposure to confirmed COVID-19 cases, results of previous polymerase chain reaction (PCR), or previous serology. In January 2021, the question for COVID-immunization status was added.
Pseudonymized blood samples were sent to our central laboratory for testing of antibodies against SARS-CoV-2. Data from pseudonymized questionnaires were collected and processed with MS Excel 2010 (Microsoft Corporation, Redmond, WC, USA).

SARS-CoV-2 Antibody Testing
Presence of SARS-CoV-2 antibodies was investigated with the chemiluminescence-based immunoassay Elecsys, Anti-SARS-CoV-2 (Roche, Basel, Switzerland). The immunoassay targets recombinant nucleocapsid protein and was carried out according to manufacturer's instructions. Sensitivity and specificity as provided by the manufacturer was high (≥99%) after 14 days post PCR confirmation. Participants with positive test results were regarded as SARS-CoV-2 seropositive. Re-testing was offered to all participants during the test period.
Available vaccines did not interfere with the SARS-CoV-2 antibody test we used in the study as they do not contain parts the nucleocapsid protein, but parts of the spike protein.

Outcomes
The primary aim of the study was to assess the course of seroprevalence of SARS-CoV-2 antibodies in hospital employees and an associated fire brigade during the first pandemic year using an IgG detecting immunoassay. Preliminary results of the first observation period are already published [39]. Secondary aims were: (i) The detection of possible differences among. (ii) The description of clinical symptoms 2 months before testing.
(iii) Correlation of the results with the number of COVID-19 positive patients treated, and to the local incidence.

Statistical Analysis
In descriptive analyses, participant's demographics, professions, symptoms, and other attributes of COVID-19 exposure were determined and compared for the whole cohort, and stratified by antibody test result, using absolute and relative frequencies. Clinical characteristics and test results were compared by Pearson's chi-squared test or Fisher's exact test. Risk factors for a positive antibody test result were estimated by univariate logistic regression, giving odds ratios and 95% confidence intervals versus the reference level for each main category of the characteristics. We applied a significance level of 0.05. Further, to assess influence of number of contacts also inside hospitals, we conducted a bivariable logistic regression with the two predictor variables institution (VHD, LSW, and VKJ) and profession. Data were analysed with the statistical software R [40].

Characteristics of the Study Participants
Overall, from June 2020 to June 2021, 1842 of 2804 employees (65.7%) of the three hospitals VHD, LSW and VKJ with together 732 beds and an associated fire brigade took part in the study. We investigated 80.6% (1177/1460) of the employees of the two hospitals with regular adult care (VHD and LSW) and 41.8% (522/1250) of employees of the children's hospital VKJ. Additionally, 99.0% (103/104) employees of an associated fire brigade took part in the study. Baseline characteristics, such as age, profession, and working area for the observation period are shown in Table 2. The 40 participated employees of associated residential care homes for the elderly are included in Tables 2 and 3, but not considered separately in the results and discussion section.

Total Seroprevalence within the Year of Observation
Over the observational period of one year, the seroprevalence rate against SARS-CoV-2 increased continuously from 1.1% in the first, to 13.2% in the second, and to 29.3% in the third pandemic wave, respectively. Altogether, 360 of 2990 tests (12.0%) were seropositive in the three observational periods, including multiple tests of employees from wave to wave. In Figure 1 the number of already known seropositive employees and newly diagnosed seropositive employees is illustrated for the three time periods. Taking multiple tests out, 243/1841 employees were tested seropositive at least once within the year of observation which represents a total seroprevalence of 13.2%. There were differences between the four working areas (hospitals and fire brigade): 133/858 of employees (15.5%) were positive in VHD, 59/319 (18.5%) in LSW, 35/522 (6.7%) in VKJ, and 13/103 (12.6%) in an associated fire brigade, respectively. Figure 1. Seroprevalence of SARS-CoV-2 IgG antibodies from June 2020 to June 2021. Any employee in a given sector could volunteer to participate in the study at any time, so that we did not test all positive employees in the following observational period again (lack of follow-up).

Seroprevalence Associated with Age
We categorised three age groups: 16-25 years (n = 301), 26-40 years (n = 527), and >40 years (n = 999) ( Table 2). The background for the classification into these groups was the assumption that participants might have different composition of their households (e.g., <25 years: less children, 26-40 years: young children, >40 years: older children) and consequently different risks for acquiring SARS-CoV-2 infection outside the hospital. In our statistical analysis we saw a significant lower risk of infection in the group >40 years (OR 0.65, 95% CI 0.46; 0.94) ( Table 2) and, if we look at the three different observation periods separately, in the third pandemic wave (OR 0.59, 95% CI 0.37; 0.95) (Supplementary Table S1).
Following our hypothesis, employees in this age group perhaps had no or older children living in their households than the youngest age group resulting in less contacts. Furthermore, children were tested regularly in schools while in preschools tests were voluntary resulting perhaps in more infection control especially in the third pandemic wave. However, we unfortunately did not collect data on household composition.

Low-Risk and Intermediate-Risk Non-HCWs
Altogether, 49 of 431 tested non-HCWs (11.4%) were seropositive: 36/328 employees (11.0%) working in low-risk areas with no contact to patients at all, and 13/103 employees (12.6%) working in the fire brigade with intermediate-risk while taking care of patients during transports (Table 2).

Risk According to Profession and Institution
Employees of the two hospitals of adult care (VHD and LSW) had SARS-CoV-2 infections in employees working regularly with patients (MDs, nurses, care workers, therapists) and working without patients, summarised as other professions (e.g., kitchen, administration, cleaning service). In the children's hospital (VKJ) employees with no contact to patients had no SARS-CoV-2 IgG antibodies in our study (Supplementary Table S2). The differences in employees working with patients compared to others was statistically significant especially for nurses (OR 1.64, 95% CI 1.09; 2.55) and care workers (OR 2.07, 95% CI 1.21; 3.54) (Supplementary Table S2). Additionally to the profession, employees in the two hospitals of adult care had a significant higher risk of SARS-CoV-2 infection compared to employees of the children's hospital ( Figure 2).

Figure 2. Risk of profession and institution for infection.
To assess influence of number of contacts also inside hospitals, we conducted a bivari-able logistic regression with the two predictor variables institution (VHD, LSW, and VKJ) and profession. The latter was recoded regarding patient contact, specifying med-ical doctor, nurse, care worker, therapist, and all other professions without patient con-tact as "others". To analyse if there were excess risk, the children's hospital VKJ and "others" profession were chosen as reference. Results are displayed in a forest plot showing odds ratios.

Clinical Symptoms of Seropositive Employees and Available PCR Results
Independent from this study, employees were tested with PCR according to the recommendations of the Robert Koch Institute (RKI) in case of cold-like symptoms of any severity, exposure to COVID-19 positive persons, and returning from a region at risk [41]. Analysing available information about positive PCR results we identified SARS-CoV-2 infections in employees where both PCR and IgG antibodies were positive, in subjects where PCR was positive with no detectable IgG antibodies, and in subjects with positive PCR with no information about detectable antibodies. Altogether, 93.4% seropositive employees (227/243) had a positive PCR test previously, according to the test criteria of the RKI, and 6.6% seropositive employees (16/243) were missed with this PCR test strategy [41] ( Table 3).

SARS-CoV-2 IgG Antibody Titer
During the time of observation we got exactly one positive SARS-CoV-2 IgG antibody titer of 94 employees (Figure 3a), and follow-up titers in 149 seropositive employees (Figure 3b).

Seroprevalence with Regard to the Treated COVID-19 Patients and the Epidemiological Context
As the incidence of COVID-19 increased in the general population in the region of Recklinghausen, the number of COVID-19 patients increased in the hospitals. There was especially an increase of infection pressure in the second and third wave, and seroprevalence in employees increased accordingly (Figures 1 and 4). The SARS-CoV-2 restrictions measures in the county Recklinghausen followed the nationwide regulations (https: //www.bundesgesundheitsministerium.de/coronavirus/chronik-coronavirus.html). Basic points of these restrictions are summarized in Figure 5. Altogether, 552 COVID-19 patients were treated in the three hospitals (VHD: 380, LSW: 106, VKJ: 66) during the observation period. Interestingly, the LSW had no COVID-19 positive patient in the first infection period (Figure 4). Hence, the eight seropositive HCWs in this hospital observed in the first wave presumably acquired their infection not nosocomial, but more likely during private contacts. Looking at the risk to acquire the SARS-CoV-2 infection privately or at work, we may assume that the participating hospitals were probably no places of infectious spread to employees in the 1 st pandemic wave. Some studies on HCWs in Germany also differentiated between seroprevalence rates according to the risk at work [14,21], and investigated high-risk group HCWs on ICU, the COVID-19 ward and emergency department [19,22,23,31]. All these studies found a low seroprevalence up to 4.36% [18] in

Discussion
Looking at the risk to acquire the SARS-CoV-2 infection privately or at work, we may assume that the participating hospitals were probably no places of infectious spread to employees in the first pandemic wave. Some studies on HCWs in Germany also differentiated between seroprevalence rates according to the risk at work [14,21], and investigated highrisk group HCWs on ICU, the COVID-19 ward and emergency department [19,22,23,31]. All these studies found a low seroprevalence up to 4.36% [18] in the first pandemic wave. This is consistent to our results.
Only sparse seroprevalence data of HCWs are available for the second and third wave in Germany. Korth et al. found a seroprevalence rate of 5.1% in 315 HCW between August and December 2020 at the University Hospital Essen, close to our region [21]. Another study in Oberspreewald-Lausitz reported a seroprevalence of 13.3% in 166 HCWs between July and December 2020 in a standard care hospital [35]. During the second and third observational period in our study, the incidence in the general population of the region of Recklinghausen and the number of COVID-19 patients increased markedly. Accordingly, the seroprevalence in employees in our study went up sharply to 13.2% in the second, and to 29.3% in the third wave, respectively, with a significant increase especially in high-risk HCWs regularly working with COVID-19 positive patients. We saw infections spread on regular wards in adult care, even in the geriatric department, resulting also in infections of employees. On the COVID-19 ward in the VHD at that time there were several potential contributors to an additional occupational risk to get infected with SARS-CoV-2: e.g., structural alteration works on the ward with relocation to another ward, a high turnover of patients, and longer retention time of deceased patients on the ward due to a temporary lack of storage capacities. Additionally, infected high-risk HCWs led to personnel shortage with personnel shifting, high workload with less time for correct self-protection and a shortage of manpower to instruct new personnel carefully. Ongoing infections on regular wards in both hospitals of adult care (VHD and LSW) led to an additional occupational risk of infection in intermediate-risk HCWs. In contrast, due to the lower hospitalization rate of COVID-19 in children, no such effects were observed in the children's hospital, nor were there transmissions between staff noticed. Data on nosocomial spread of infections within the two hospitals of adult care during the second and third waves were actually evaluated in a retrospective study to get more information about the weak points in the concept of infection prevention.
The population-based sequential study "MuSPAD" investigated the SARS-CoV-2 seroprevalence of the general not vaccinated population in seven regions in Germany between July 2020 and May 2021. The authors report a low seroprevalence of 1.3-2.8% after the first pandemic wave and an increase up to 4.1-13.1% until May 2021 [42]. From October 2020 to February 2021 a nationwide seroepidemiological study in Germany called "RKI-SOEP-Study" noticed a clear increased risk of SARS-CoV-2 infection for employees of healthcare professions (4.6%) compared to non-healthcare employees (1.8%) [43].
First SARS-CoV-2 seroprevalence data of extraclinical personnel depending on their operational area in the fight against the COVID-19 pandemic was recently reported from Brune et al. in Essen, Germany. The authors detected in 8 of 732 employees of the professional fire brigade and aid organizations in the city area SARS-CoV-2 IgG antibodies in the first pandemic wave which corresponds to a seroprevalence rate of 1.1% [37]. Our study investigated employees of the fire brigade in the second and third wave with a seroprevalence rate of 12.6% (13/103) according to the high incidence in the general population at that time. Unfortunately, we have no information about the number of transported COVID-19 positive patients. Because of the seroprevalence of 11.0% (36/328) in employees with no contact to patients, we expect that the fire brigade had no or at best a minimal additional occupational risk for infection in our study.
The longitudinal course of SARS-CoV-2 antibodies in HCWs after infection was similar to that reported by other authors who investigated infected patients, with the Spike IgG titers showing only modest declines at 6 to 8 months [44].
An important limitation of our study is its convenient sampling design. It is conceivable that employees with a previous COVID-19 diagnosis or symptoms consistent with COVID-19 would be more interested in a measurement of their serum COVID-19 antibodies and follow-up titers. Many employees wanted to know their antibody titer for the decision to get vaccinated. Those employees who were hesitant to get vaccinated, tent to take part in the study repeatedly. There are several reasons for missing following antibody tests. The shortage of antigen tests and antibody tests at the beginning of the pandemic was overcome at the beginning of 2021, so that employees could also get tested ambulant by their general practitioner. For persons who got vaccinated (two vaccines or one vaccine 3 months after infection), the antibody test result and follow-up titer was not relevant any longer.
However, our study also has strengths. It presents data on the course of SARS-CoV-2 seroprevalence based on a cohort of employees of three hospitals and an associated fire brigade. The time period of one year contains a "pre-vaccination" period, and a short period after vaccination had started. We could not demonstrate an effect of the vaccination in this short "post-vaccination" observation period so far, but it will be interesting to investigate the seroprevalence of vaccinated employees again in the following pandemic waves.

Conclusions
The overall seroprevalence in the investigated employees was 13.2% after one year. Community acquired transmission seems to have played a larger role for SARS-CoV-2 infection than professional exposure during the first pandemic wave. However, this resulted from an overall low exposure of hospital employees to COVID-19 positive patients at a time where the region was not a SARS-CoV-2 hotspot. With the increase of infection pressure in the second and third pandemic wave, HCWs had an additional occupational risk for infection, and we observed some hints towards in-hospital transmission. This underlines the need to adapt the concept of infect prevention continuously, especially in situations with structural limitations, such as high workload and personnel shortage in a pandemic, in order to keep hospitals safe places.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/ijerph19042402/s1, Table S1: Age as risk factor for infection in the different observation periods; Table S2: Profession and institution as risk factor for infection. Funding: The study was financed by internal funding.

Institutional Review Board Statement:
Research was conducted in accordance with the declaration of Helsinki and national standards. The study protocol was approved by the local ethics committee: Ärztekammer Westfalen-Lippe and Westfälische Wilhelms Universität Münster (approval no. 2020-478-f-S). The study was registered at the German Clinical Trials Register (DRKS00022941).

Informed Consent Statement:
Informed consent was obtained from all included participants included in the study.

Consent for Publication:
Informed consent was obtained from all included participants under the provision that anony-mised data will be published.

Data Availability Statement:
The datasets and materials used and/or analysed during the current study are available from the corresponding author upon reasonable request.