Longitudinal SARS-CoV-2 seroepidemiological investigation among healthcare workers at a tertiary care hospital in Germany

Following the emergence and rapid global spread of severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2), the World Health Organization (WHO) declared the coronavirus disease 2019 (COVID-19) outbreak a global pandemic on 11 March 2020 [1]. In Germany, the first laboratory-confirmed SARS-CoV-2 case was reported on 27 January 2020 [2]. Cases of SARS-CoV-2 in Germany started to significantly increase in early March 2020 (‘first wave’), with a peak of approximately 35,000 reported cases in calendar week 14, leading to wide-reaching impacts on the German health system. Although cases of SARS-CoV-2 declined in Germany during the following summer months, increasingly diffuse infection chains and outbreak settings were seen. In October 2020, cases of SARS-CoV-2 again started to increase (‘second wave’), and the mean number of occupied intensive care unit (ICU) beds grew from 807 to 5643 in calendar week 13–53 [2].

In many settings, the burden of the COVID-19 pandemic has stressed the capacity of hospitals and ICUs. In Berlin, Germany, a standardised approach was developed to distribute COVID-19 patients requiring invasive ventilation across hospitals in order to optimize the flow and care of these at-risk patients (“SAVE-Model”) [3]. This includes a ‘first-level’ coordinating hospital and 16 ‘second-level’ specialised hospitals with the capacity to treat critically-ill COVID-19 patients. The remaining 60 ‘third-level’ hospitals with emergency services are designated to provide intensive care for non-COVID-19 patients.

However, the capacity of health service delivery and the COVID-19 pandemic response is largely dependent on the availability and protection of health care workers (HCWs). HCWs may have high-risk occupational exposures to SARS-CoV-2 infectious material. Hence, effective infection prevention and control (IPC) measures such as appropriate use of personal protection equipment (PPE) and early detection of SARS-CoV-2 infection among HCWs are critical to prevent and control nosocomial transmission [4‐7]. A meta-analysis of 49 studies including 127,480 HCWs in North America, Europe, Africa and Asia estimated an overall seroprevalence of SARS-CoV-2 of 8.7% among HCW [8]. However, studies in Stockholm and London found that HCW seropositivity was considerably higher than that of the general population, although such findings can be closely related to local outbreak dynamics [9, 10].

By 3 April 2020, 25 SARS-CoV-2 HCW cases had been detected across different departments at a tertiary care hospital in Berlin, Germany and reported to the local public health department. An outbreak was declared, and the regional and federal public health service were invited to assist in further response measures. This prompted systematic hospital-wide staff screening and an intensified outbreak investigation. By repeatedly assessing the SARS-CoV-2 serostatus of the HCWs and associated factors, we aimed to elucidate the chain of infection and prevent future transmission of disease in this health care setting.

A total of 1944 HCWs participated in the May–June and/or December 2020 cross-sectional surveys (see Fig. 1). Among these, a longitudinal sample of 756 (38.9%) HCWs were included in both surveys. Approximately 70% of participants were women, and the median age was around 40 years. Approximately half were physicians or nurses, and more than 18% of HCWs were determined to have a high or moderate occupational risk to SARS-CoV-2 (Table 1).

We present a repeated seroepidemiological investigation at a large tertiary care hospital in the local context of the COVID-19 outbreak in Berlin, Germany. A robust tiered-testing strategy was applied to assess the presence of antibodies against SARS-CoV-2. Seropositivity estimates were found to be lower than other HCW studies internationally, although all those in the longitudinal sample who tested seropositive in the first survey remained seropositive by ELISA and NT in the follow-up 6 months later. In the risk factor analysis, seropositivity was found to be associated with increasing occupational risk and male gender.

Seroprevalence estimates of 1.2% in May/June 2020 and 4.6% in December 2020 found in our investigation were lower than the meta-analysis by Galanis et al. reporting an overall pooled prevalence of 8.7% among all HCWs globally as well as 8.5% among HCWs in Europe. In contrast, other HCW studies in Germany have found similar estimates to our study, ranging from 1.6 to 5.1% across varying regions, HCW populations and time periods from March to July 2020 [14‐20]. Korth et al. found a similar increase in seroprevalence from 2.2% in March–May 2020 to 5.1% in October–December 2020 at a tertiary facility in Western Germany [17]. Lower estimates in Germany could reflect differences in existing health care system infrastructure and implementation of IPC measures as well as a lower SARS-CoV-2 incidence in the population, particularly in the ‘first wave’, compared to other countries. A population-based study in Berlin, Germany, estimated a similar seroprevalence of 4.4% in the community in November–December 2020 [21]. Differences across serosurveys could also be due to varying performance characteristics of serological assays with different targets as highlighted by the utility evaluation of serological assays by Heffernan et al. [22].

A systematic review by Post et al. reported that SARS-CoV-2 antibody dynamics have been well described in the acute phase but there remains limited evidence on longer-term patterns. The review found that IgG peaked at weeks 3–7 post-symptom onset and persisted for at least 8 weeks [23]. We demonstrated that all seven participants in our longitudinal sample who tested seropositive in the first survey remained seropositive by ELISA and NT in the follow-up 6 months later. Although those who self-reported COVID-19-like symptoms in the last 14 days had significantly higher odds of being seropositive, we could not determine the exact rate of seroconversion in asymptomatic versus symptomatic cases since not all participants provided a NP/OP swab for PCR testing. The proportion of seropositive HCWs self-reporting a past positive PCR test result increased from 44.4% in the first survey to 85.7% in the second survey, which could reflect the increase in availability and access to staff screening and testing from the first to second COVID-19 wave.

The Euroimmun ELISA ratios were well correlated with the NT titres in the first survey but less correlated in the second survey. Such findings could be explained by the use of different Euroimmun test lots in each survey, although not all samples were tested by NT which could have biased results. An in-house validation study found that the level of the ratio values obtained in the Euroimmun test lots varied from April to September 2020 [24]. We aimed to address this by applying a tiered testing algorithm with a lower Euroimmun test cut-off followed by a verification assay and NT. Our sensitivity analysis found no substantial differences between our results and those of only the Euroimmun ELISA ratios adjusted for test performance. However, validated lot test performance data is only available for some of the lots (Additional file 1: section 3). Furthermore, the kinetics of the antibody response over time and the difference between the timeframe from infection before the first survey (February–May 2020) and second survey (February–December 2020) may have also affected the correlation of the ELISA ratios and NT titres at the different surveys [25].

In the risk factor analysis, we found that increasing occupational risk based on working site and profession was clearly associated with higher odds of SARS-CoV-2 seropositivity. This is consistent with other HCW studies which have shown an association between SARS-CoV-2 seropositivity and working in COVID-19 wards, ICUs and/or as frontline HCWs [9, 10, 26, 27]. This is likely multifactorial due to hospital versus community exposures, varying implementation of IPC and PPE measures, and differences in behaviour or awareness. In the first survey, few HCWs reported close contact to a confirmed case of COVID-19 outside of the hospital and most reported adequate use of PPE and hand hygiene. However, a little less than half reported close contact to colleagues without a medical mask outside of patient care. This may have contributed to transmission but was only assessed in the first survey subset and these practices likely changed during the course of the pandemic response due to hospital-wide mandatory masking policies. We also found that males had 2 times the odds of SARS-CoV-2 seropositivity. This finding varies across HCW studies [26, 28‐30]. It may be due to differences in behaviour or other biological factors, as a higher seropositivity (4.9%) was also found among males compared to females (3.8%) in the population-based study in Berlin, Germany [21].

Several limitations should be considered. Participation was voluntary, so this may have introduced bias in the study sample. However, more than half of HCWs participated in the study and a balanced distribution of type of profession and working site was included. If staff were on sick leave during the two surveys, they would not have been included in the study, although COVID-19 related sick leave was not noted on the days of the surveys. It is unknown what proportion of HCWs who were infected did not mount a detectable antibody response or in whom it had waned by the time of testing. In the first survey, the number of seropositive HCWs was small so the significance of associated factors could not be reliably assessed. Overall, questionnaire responses could have been affected by recall and social desirability biases.

In conclusion, this investigation provides insight into the burden of infection among HCWs in the context of the local COVID-19 outbreak in Berlin, Germany. Findings highlight the need for effective IPC measures particularly among HCWs with the highest occupational risk to minimize SARS-CoV-2 infection in the present COVID-19 wave. The investigation also adds to the body of evidence on SARS-CoV-2 antibody dynamics over time and provides further reflection on serology testing strategy considerations. Such testing considerations continue to be important for ongoing serosurveys in the current COVID-19 pandemic in order to continue to assess duration of immunity as well as vaccine responses according to risk groups, type of vaccine, and SARS-CoV-2 variants of concern. This is true, for example in Germany, whose outbreak situation remains dynamic including a vaccination coverage of approximately 69% (as of 2 December 2021) and rapid detection of the latest variant of concern Omicron following its designation by WHO in the week prior [31].