(Vienna, 11 December 2020) – Between the start of the pandemic and the middle/end of October 2020, around 349,000 people (statistical variation range 282,000 to 420,000 people) in Austria had been infected by SARS-CoV-2 and had produced antibodies to it. 61% of these infections had not been officially identified. This was the result of an extrapolation of the random sample data for infections so far, based on the formation of antibodies (seroprevalence) as part of an Austria-wide COVID 19 prevalence study conducted by Statistik Austria and the Medical University of Vienna on behalf of the Federal Ministry for Education, Science and Research and in collaboration with the Austrian Red Cross.
"Between the middle and end of October, 4.7% of the permanent Austrian population aged 16 and over had antibodies to SARS-CoV-2 and had therefore been infected at some point since the start of the pandemic. Three out of every five infections had not been officially identified," says Tobias Thomas, General Director of Statistik Austria.
"The generally low prevalence of neutralising antibodies underlines how long it would take to achieve so-called herd immunity within the Austrian population, given the necessary measures for slowing the spread of the virus and without a vaccination programme," add Medical University of Vienna project leaders, Robert Strassl and Lukas Weseslindtner.
At the middle/end of October, seroprevalence of SARS-CoV-2 infection was 4.7%
Blood samples were taken from 2,229 people between 12 and 14 November 2020. Neutralising antibodies, that is to say antibodies that protect against infection with SARS-CoV-2, were found in 92 samples. Based on scientific studies, it is safe to assume that virus-specific antibodies can be reliably detected after a period of around three weeks. Accordingly, the test period from 12 to 14 November 2020, reflects the number of infections in Austria up until the middle/end of October.
Extrapolation of the random sample data showed that around 4.7% (95% CI: minimum 3.8%, maximum 5.6%) of the population living in private households and aged 16 and over had undergone a SARS-CoV-2 infection by the middle/end of October 2020 and therefore have antibodies in their blood.
A comparison with the data recorded by the Epidemiological Reporting System (EMS) shows that around 61% of these cases had not been officially identified. It is striking that the proportion of asymptomatic people is very high, particularly in this group (26 out of 57 people reported no symptoms or just one symptom between the middle of February and the middle of October). Accordingly, the vast majority of these people did not think they had been infected, even though they were found to have antibodies (44 out of 57 people rate this possibility as slight).
Seroprevalence tends to be higher in Western Austria (Tirol, Vorarlberg, Salzburg, Upper Austria, 5.7%, 95% CI 4.1-7.4%) than in Eastern Austria (Vienna, Burgenland, Lower Austria, 3.8%, 95% CI 2.7-4.8%).
In the middle of November, the prevalence of SARS-CoV-2 infection was 3.1%
The first interim results have already been presented for the proportion of infected people in the period from 12 to 14 November 2020. These results have now been validated against the EMS data. The extrapolated total for the number of SARS-CoV-2 infections from 12 to 14 November 2020 is 233,000 (3.1% of the population living in private households and aged 16 and over). Allowing for statistical variation, this figure is between 195,000 and 261,000 (95% CI: 2.6% – 3.5%) infected people in the middle of November.
Details of the methodology, definitions:
COVID-19 prevalence: Proportion of the population aged 16 and over in private households, in which SARS-CoV-2 can be detected at the time of testing.
PCR tests: In order to identify current SARS-CoV-2 infections, nose and throat swabs were taken and analysed by PCR (Polymerase Chain Reaction) (2,263 people). In addition, blood samples (2,229 people) were taken and serologically analysed for antibodies to establish whether the tested subjects had already been infected in the past.
Seroprevalence: Proportion of the population that has already undergone an infection and developed antibodies.
Study design: Out of a random sample size of 7,823 people living in private households and aged 16 and over, around 2,711 people completed a questionnaire by the end of October. 2,504 of these arranged an appointment for comprehensive coronavirus testing (nose and throat swab, rapid antibody test and blood sample) at one of the 53 Red Cross testing centres. The tests were conducted all over Austria between 12 and 14 November.
Detection of antibodies: The test procedure for detecting antibodies was very thorough: first of all, two highly sensitive, laboratory-based serological methods were used to check the blood for SARS-CoV-2-specific antibodies. Both methods, an ELISA (Enzyme-Linked ImmunoSorbent Assay) and a CLIA (ChemiLuminescent ImmunoAssay), are characterized by their high level of specificity and sensitivity. Moreover, the methods used different target antigens and determined different "types" of antibodies. This significantly increases the sensitivity of testing. In order to guarantee maximum testing specificity and in order to be able to estimate pre-existing (antibody-mediated) immunity to SARS-CoV-2 in the population, every borderline or positive sample from first-line testing was reconfirmed by a neutralisation test. The neutralisation test allows SARS-CoV-2-neutralising antibodies to be detected with nearly 100% specificity (exclusion of false positive results in conventional binding tests). A rapid antibody test was additionally performed immediately after taking the blood sample, in order to evaluate its usefulness for seroprevalence studies of this kind. Laboratory-based antibody testing and the antibody-positive samples validated by neutralisation test were used as a reference for this, in order to evaluate the sensitivity of the rapid antibody test.
Extrapolation: The random sample data of the COVID-19 prevalence study was extrapolated according to the criteria of age, gender, federal state, degree of urbanisation, household size, education, citizenship, pre-existing conditions (benchmark data of the health survey), risk area (according to the relative proportion of positive cases at community level) and the number of people known to have already tested positive for SARS-CoV-2 prior to the test period according to EMS data. The confidence interval (CI) was estimated using the bootstrap method (rescaled bootstrap for stratified multistage sampling).