key: cord-0915426-fffrnhbx
authors: Pedersen, Ole Birger; Nissen, Janna; Dinh, Khoa Manh; Schwinn, Michael; Kaspersen, Kathrine Agergård; Boldsen, Jens Kjærgaard; Didriksen, Maria; Dowsett, Joseph; Sørensen, Erik; Thørner, Lise Wegner; Larsen, Margit Anita Hørup; Grum-Schwensen, Birgitte; Sækmose, Susanne; Paulsen, Isabella Worlewenut; Frisk, Nanna Lond Skov; Brodersen, Thorsten; Vestergaard, Lasse Skafte; Rostgaard, Klaus; Mølbak, Kåre; Skov, Robert Leo; Erikstrup, Christian; Ullum, Henrik; Hjalgrim, Henrik
title: SARS-CoV-2 infection fatality rate among elderly retired Danish blood donors - A cross-sectional study
date: 2020-10-26
journal: Clin Infect Dis
DOI: 10.1093/cid/ciaa1627
sha: 7e1c78e84a86c94c9da79297c2d0c9820af0e30f
doc_id: 915426
cord_uid: fffrnhbx

BACKGROUND: Despite the vast majority of individuals succumbing to infection with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) are elderly, infection fatality rate (IFR) estimates for the age group 70 years older are still scarce. To this end we assessed SARS-CoV-2 seroprevalence among retired blood donors and combined it with national COVID-19 survey data to provide reliable population-based IFR estimates for this age group. METHODS: We identified 60,926 retired blood donors age 70 years or older in the rosters of three region-wide Danish blood banks and invited them to fill in a questionnaire on COVID-19 related symptoms and behaviours. Among 24,861 (40.8%) responders, we invited a random sample of 3,200 individuals for blood testing. Overall, 1,201 (37.5%) individuals were tested for SARS-CoV-2 antibodies (Wantai) and compared to 1,110 active blood donors age 17-69 years. Seroprevalence 95% confidence intervals (CI) were adjusted for assay sensitivity and specificity. RESULTS: Among retired (age 70 years or older) and active (age 17-69 years) blood donors, adjusted seroprevalences were 1.4% (95% CI: 0.3%-2.5%) and 2.5% (95% CI: 1.3%-3.8%), respectively. Using available population data on COVID-19 related fatalities, IFRs for patients age 70 years or older and for 17-69 years were estimated at 5.4% (95% CI: 2.7%–6.4%) and 0.083% (95% CI: 0.054%-0.18%), respectively. Only 52.4% of SARS-CoV-2 seropositive retired blood donors reported having been sick since the start of the pandemic. CONCLUSION: COVID-19 IFR in the age group above 69 years is estimated to be 65 times as high as the IFR for people age 18-69 years.

Since early in the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic it has been clear that co-morbidities such as diabetes mellitus, chronic cardio-vascular, -obstructive pulmonary and -kidney diseases are common among patients who are hospitalised for coronavirus disease 2019 (COVID-19) and among patients who succumb to the infection. [1] [2] [3] The comorbidities accumulating among deceased COVID-19 patients are mostly prevalent in the elderly population. [1] Because of comorbidity and age-dependent frailty, the COVID-19 infection fatality rate (IFR) presumably is higher among older than among younger adults, but solid data to support this assumption are lacking. IFR may be approximated by relating age-specific number of COVID-19 deaths with corresponding measures of SARS-CoV-2 prevalence. [4] [5] [6] So far, only few studies have assessed how SARS-CoV-2 has spread among the elderly in the general population and with mixed results. [7] [8] [9] [10] While a Spanish [8] survey found no strong association between age and SARS-CoV-2 seroprevalence, both a Swiss and a US investigation generally reported lower seroprevalence in the age-group above 65 years than among younger adults. [7;10] The aim of this study was to determine SARS-CoV-2 seroprevalence in an elderly retired blood donor population in Denmark and to estimate the associated IFR for this age group. This knowledge will help tailor public health policies mitigating the impact of the pandemic.

A c c e p t e d M a n u s c r i p t 5

We contacted 60,926 elderly retired blood donors in three out of the five Danish administrative regions (the Danish Capital Region, the Zealand Region, and the Central Denmark Region). We have previously found that these regions had the lowest (0.7%) and highest (3.0%) seroprevalence of SARS-CoV-2 antibodies among active blood donors age 17-69 years. [4] The 60,926 retired blood donors aged 70 years or older were invited to complete a digitalized questionnaire on COVID-19 symptoms and risk factors between May 16 and May 25, 2020. Within two weeks of the invitation, 24,861 (40.8%) donors had filled in the questionnaire.

In order to compare the symptoms and behaviour of the older retired blood donors with younger active donors, we also mailed the questionnaire electronically to 75,934 participants in the Danish Blood Donor Study from all Danish Regions age 18-69 years on May 28, 2020. [11] A total of 24,227 (31.9%) invitees responded to this questionnaire.

From among the retired blood donors aged 70 years or older, who completed the questionnaire, we invited a random sample of 3,203 individuals living within geographical proximity of 12 bleeding sites (two in the Central Denmark Region, six in the Zealand Region, and four in the Capital Region) for serological SARS-CoV-2 antibody measurements. The number of retired blood donors invited was determined by capacity for blood sampling in the different regions.

A total of 1,201 (38%) of the invited retired blood donors showed up for testing between June 2 and June 19, 2020 ( Figure 1 ). Characteristics of the cohort in the different steps are summarized in Table   1 .

For comparison, we also tested a random sample of 360, 250, and 500 active blood donors (age 17-69 years) who had given blood between June 1 and 12, 2020 in the Central Denmark Region and A c c e p t e d M a n u s c r i p t 6 between June 22 and June 26, 2020 in the Zealand and Capital Regions, respectively. This testing was done after anonymization of samples.

All 1,201 elderly participants gave informed consent to participate in the Danish Blood Donor Study before samples were taken for SARS-CoV-2 antibody measurements. The study was approved by the Zealand Regional Committee on Health Research Ethics (approval number: SJ-740) and the data protection agency of the Capital Region (P-2019-99). According to Danish legislation, analysis of anonymous material does not require consent.

The questionnaire included the following items: symptoms of disease during the previous three months including fever, symptoms from nose (sneezing, affected sense of smell and/or taste), airways (sore throat, coughing, shortness of breath), and abdominal discomfort (diarrhoea, vomiting); changes of behaviour including hand washing, using handkerchief, sneezing in the elbow, avoiding handshake, wearing facemask, avoiding to hug people, reduced use of public transport, avoiding crowded places, and staying at home; history of previous PCR test for COVID-19 and the test result. The questions on disease symptoms were graded on a four-point scale ranging from no symptoms (1), yes, a little (2), yes, a lot (3), and yes, very much (4). The respondents were also allowed answering "I do not know". For analysis, we dichotomized the scale into none (1) versus any (2) (3) (4) ; if the response was "I do not know" this answer was omitted from the analysis. 

The samples for this study were collected between June 2 and June 19, 2020. Public surveillance data on number of individuals tested for SARS-CoV-2, hospitalised for COVID-19 and dying with COVID-19 are updated daily from Statens Serum Institute (SSI). [12] We retrieved population statistics from Statistics Denmark based on the Danish population in first quarter 2020. [13] 

Self-reported risk factors and symptoms were reported as percentages with 99.8% CI because of multiple testing (exact confidence intervals) and for differences between groups we reported the risk ratios (RR). To compare age and sex compositions between groups we used Mann Whitney U test and Chi squared tests. We used the Rogan Gladen estimate to calculate the true prevalence. CI were derived by 10 8 -sample percentile bootstrapping independently of sampling sensitivity, specificity and apparent prevalence using posterior binomial distributions based on the observations. Prevalences were reported as percentages with 95% CI. The analysis was performed in RStudio 1.2 and R 4.0.2 using the EpiR package to adjust measured seroprevalence for sensitivity and specificity of the diagnostic assay as well as weighing the estimate based on population size of the bleeding sites' recruitment areas. The weights of the geographical areas were based on the number of inhabitants in the municipalities (exact weights can be found in Supplementary Table 1) .

A c c e p t e d M a n u s c r i p t 8

The age and sex distributions of the different selection steps in the present investigation are summarized in Table 1. A total of 24,861 (40.9%) retired blood donors age 70 years or older and a total of 24,227 (31.9%) active blood donors age 17-69 years answered the questionnaire on health behaviours and COVID-19 symptoms ( Table 2 ). Adherence to official recommendations varied between 40% (staying at home) and 95% (frequent hand washing) among the retired blood donors. For the same items, adherence among the younger active donors varied between 31% (staying at home) and 91%

(frequent hand washing) with estimates generally being lower than among the elderly retired donors (Table 2) . By June 2020, use of facemasks was neither recommended or mandatory except under certain circumstances and therefore infrequently reported by either group of blood donors (1.8%).

Overall, 1,201 (38%) elderly retired donors were tested for antibodies to SARS-CoV-2. Compared with non-invitees and non-attendees who had answered the questionnaire, those who were tested for SARS-CoV-2 antibodies were marginally more compliant with the official precautionary recommendations (Table 3) 12.0%-18.6%), although individually the distribution of specific symptoms were not statistically significantly different between the two groups (Table 3) .

For four out of five officially recommended precautions against spread of disease, self-reported compliance was lower among the elderly blood donors who tested positive for SARS-CoV-2 antibodies than among those who did not have antibodies although this difference was statistically significant only for one recommendation (avoiding hugging or kissing on the cheeks) (54.5% (99.8% CI: 22.3%-84.2%) vs. 87.4% (99.8% CI: 84.2%-90.2%) ( Table 3) . Only concerning staying at home as A c c e p t e d M a n u s c r i p t 9 much as possible, compliance was higher among SARS-CoV-2 positive participants than among SARS-CoV-2 negative participants ( Table 3) .

The adjusted SARS-CoV-2 seropositivity prevalences for men and women combined among those 70 years or older were 2.2% (95% CI: 0.6%-4.1%), 1.9% (95% CI: 0.4%-3.7%), and 0.7% (95%CI: -0.4%-3.0%) in the Capital, Zealand and Central Regions, respectively (Table 4 ). In these three regions we also tested for SARS-CoV-2 antibodies in an anonymized sample of active blood donors aged 17-69 years, yielding corresponding adjusted prevalences of 5.3% (95% CI: 2.3%-9.2%), 2.5% (95% CI: 0.6%-5.3%), and 1.1% (95% CI: -0.3%-2.8%), respectively.

These three regions represent the entire spectrum of regional-level COVID-19 seroprevalences detected in Denmark so far [4;12] . Thus, according to official statistics the two administrative regions not included in this investigations (the North and South Denmark Regions) have slightly lower prevalence of patients treated for COVID-19 than the Central Region [12] .

Assuming that the SARS-CoV-2 sero-prevalence observed for the Central Region in the present investigation also applies to the North and South Regions, and that blood donors are representative of the general population, we estimate an adjusted nation-wide SARS-CoV-2 seroprevalence of 2.2% (95% CI: 1.0%-3.4%) in the age-group 17-69 years and of 1.2% (95% CI: 0.1%-2.4%) among individuals aged 70 years or older in the Danish population. For the entire age group 17 years or older the adjusted seroprevalence was 1.6% (95% CI: 0.58%-2.5%).

On June 22 2020, the official nationwide number of individuals aged 70 years or older registered with SARS-CoV-2 since the beginning of the pandemic was 2,180. The seroprevalence estimated in the present investigation suggest that this number represents corresponds to only 21.6% (95% CI:

10.8%-28.9%) of all COVID-19 cases in the age-group at the time.

As of June 22, the total number of registered deaths in Denmark from COVID-19 was 71 and 542 for the age group below and above 70 years, respectively. Given the estimated seroprevalences these numbers correspond to IFRs of 0.083% (95% CI: 0.054%-0.18%) and 5.4% (95% CI: 2.7%-64.0%) A c c e p t e d M a n u s c r i p t 10 among people aged 17-69 years and aged 70 years or older, respectively. Thus, the IFR among people aged 70 years or older is 65 (95% CI: 40-356) times as high as the IFR among people aged 17-69 years. The IFR for the adult Danish population aged 17 years or older was 0.81% (95% CI: 0.52%-2.2%).

This Danish study on health behaviour, COVID-19 symptoms, and SARS-CoV-2 antibody prevalence among retired and active blood donors suggests that older adults (here those aged 70 years or more) in general are more likely to adhere to official recommendations to reduce SRAS-CoV-2 transmission than younger adults (here those aged 17-69 years). Consistent with this agedependent difference in guideline adherence, our study also indicated that the SARS-CoV-2 seroprevalence in Denmark is lower (1.2%) among older than among younger adults (2.2%) even though this difference was not statistically significant. In support of the suspected underlying mechanism, questionnaire data showed that older retired blood donors who tested positive for SARS-CoV-2 antibodies were less likely to have adhered to official recommendations to prevent COVID-19 than older retired blood donors testing negative for antibodies. Overall, our data indicate that IFR for COVID-19 among individuals aged 70 years or older is 5.4% (95% CI: 2.7%-64%), i.e. 65 times as high as that estimated for younger adults aged 17-69 years.

In our study, only about half of the older retired blood donors, who tested positive for SARS-CoV-2 antibodies, recollected feeling sick with COVID-19 symptoms since the start of the pandemic. This proportion of apparently asymptomatic infections with SARS-CoV-2 is in the same range as in We surveyed health-related behaviour among the participants in our study. The official Danish recommendations are to 1) stay home if you are sick, 2) to clean your hands regularly, 3) to sneeze in your elbow, 4) to avoid hugging or close contact, and 5) to avoid crowded places. The Danish political interventions towards the COVID-19 pandemic are summarized in Supplementary Table S2. In our investigation, retired blood donors aged 70 years or more who tested positive for SARS-CoV-2 antibodies had been less inclined (54.5%-90.9%) to follow official recommendations than those, who tested negative (76.6%-95.6%) with the exception of staying more at home. Whether this tendency to stay more at home has resulted in a more relaxed attitude towards other precautions cannot be ruled out. This finding may also be an example of reverse causality, i.e., individuals who experience symptoms may be more likely to report that they stayed at home. Regardless, although based on small numbers the association between taking SARS-CoV-2 precautions and testing negative for antibodies is reassuring in terms of their effectiveness.

A c c e p t e d M a n u s c r i p t

As of now there are only few studies on health behaviour and compliance with recommendations and their impact on COVID-19. A previous questionnaire based study from 27 different countries including Denmark reported that elderly people aged 70 year or more were less compliant than people aged 60-70 years. [16] We cannot confirm this finding in our study; rather, older retired donors tended to adhere stronger to official guidelines than younger donors.

Compared with retired blood donors aged 70 years or more, active blood donors aged 17-69 years twice as often reported having been sick with COVID-19 symptoms (21.9% vs. 12.7%), having been tested for SARS-CoV-2 (35.8% vs. 17.1% of those reporting sickness), and having had a positive SARS-CoV-2 test (7.2% vs. 4.3% of those reporting test-results). These differences in both COVID-19

prevalence and in test seeking may reflect age-dependent variation in mobility and in physical distancing also apparent in the present study. Whereas precautions to limit spread of SARS-CoV-19

with regard to hand washing, hugging, and handshaking were followed by large proportions of both active and retired donors, active donors aged 18-69 years did not stay home, avoid public transport or crowded places to the same extent as the retired blood donors aged 70 years or more. It is not surprising that individuals of working age and with children at home face problems isolating themselves from other people. This could expose them to more infectious agents and explain why the active donors report much higher disease rates than the retired donors.

Based on official Danish reports on COVID-19 related deaths, we estimated the IFR to be 5.4% in the age group 70 years or older. This was 65 times as high as the IFR among active blood donors aged 17-69 years. Our IFR estimate among older individuals was only slightly lower than the 6·9% recently been reported from rural settings in Latin America. [17] These estimates for the elderly population are as expected much higher than previously reported all age estimates from Greece [18] , China [19] and cruise ships [20] as well as the one we reported in the younger donor population. [4] The strength of the present study is that we have combined questionnaire data and serological SARS-CoV-2 measurements on a large population of retired Danish blood donors aged 70 years or A c c e p t e d M a n u s c r i p t 13 older. In particular, blood donors are more inclined to participate in public health studies than other Danes (unpublished results), and also familiar with the process of blood sampling, both factors adding to the participation rate. Geographically, we carried out the investigation in three of the five Danish administrative regions to ensure that regional variation in COVID-19 prevalence was represented. Indeed, we found the same geographical variation SARS-CoV-2 seroprevalence in the older age group as the one we have reported among Danish blood donors aged 17-69 years [4] which also mirrors the official incidence of COVID-19 across Denmark. Active blood donors resemble the average general population with the exception of people with low income, people that are marginalized, and men that have had sex with men [21] . We also know that the mortality among blood donors relative quickly approaches that of the general population after cessation of the donation activity (unpublished results).

While we assume that the same participation bias applied to active and retired blood donors concerning the questionnaire part of the study, low participation rates warrant caution in the interpretation of our results. Concerning the serological survey, participants invited to blood sampling were required to be free of symptoms of infectious disease and to be able to meet for testing at a bleeding site. Diseased and immobile elderly individuals are therefore likely to be underrepresented in our study population of older retired blood donors. There have been outbreaks at Danish nursing homes that will not be reflected by this study. It is therefore unknown which way this selection has biased the estimate. However, given that the proportion of individuals above the age of 70 years that lives at nursing homes is less than five percent, we assume that this bias will have limited influence on the overall result. Another possible bias is the inadequate antibody responses that are frequently seen among elderly individuals. This phenomenon has primarily been reported in studies on vaccine responses. [22] Whether the antibody response to SARS-CoV-2 infection is as inadequate as has been reported from vaccine studies is uncertain and, thus, it is not possible to quantify the impact of this bias. However, if applicable to SARS-CoV-2 infection and to A c c e p t e d M a n u s c r i p t 14 the present study, we may have underestimated the SARS-CoV-2 seroprevalence among the elderly retired blood donors and consequently overestimated the corresponding IFR for this age segment.

In conclusion, this cross-sectional study on seroprevalence, self-reported disease, and compliance indicate that social distancing reduces the risk of SARS-CoV-2. In the present study, retired blood donors aged 70 years or older was able to self-isolate to a higher extent than the younger age groups, which associated with a lower seroprevalence in the elderly population. However, the IFR among individuals aged 70 years or more is 65 times as high as among younger individuals, which underscores the need for continuous precautions to avoid general spread of SARS-CoV-2.

M a n u s c r i p t 20 Tables: 

M a n u s c r i p t 22 

Age and Multimorbidity Predict Death Among COVID-19 Patients: Results of the SARS-RAS Study of the Italian Society of Hypertension

Prevalence of Comorbidities Among Individuals With COVID-19: A Rapid Review of current Literature

Chronic heart diseases as the most prevalent comorbidities among deaths by COVID-19 in Brazil

Estimation of SARS-CoV-2 infection fatality rate by real-time antibody screening of blood donors

Estimating The Infection Fatality Rate Among Symptomatic COVID-19 Cases In The United States. Health Aff (Millwood )

COVID-19 Infection Fatality Rate Associated with Incidence-A Population-Level Analysis of 19 Spanish Autonomous Communities

Seroprevalence of anti-SARS-CoV-2 IgG antibodies in Geneva, Switzerland (SEROCoV-POP): a population-based study

Prevalence of SARS-CoV-2 in Spain (ENE-COVID): a nationwide, population-based seroepidemiological study

Populationbased surveys of antibodies against SARS-CoV-2 in Southern Brazil

Seroprevalence of Antibodies to SARS-CoV-2 in 10 Sites in the United States

The Danish Blood Donor Study: a large, prospective cohort and biobank for medical research

Danish data on surveillance of COVID-19

Statistics Denmark -population data

GarcÃ-a HH. SARS-CoV-2 in rural Latin America. A population-based study in coastal Ecuador

Estimated Community Seroprevalence of SARS-CoV-2 Antibodies -Two Georgia Counties

Elderly people and responses to COVID-19 in 27 Countries

GarcÃ-a HH. SARS-CoV-2-related mortality in a rural Latin American population

Repeated leftover serosurvey of SARS-CoV-2 IgG antibodies

Effects of policies and containment measures on control of COVID-19 epidemic in Chongqing

SARS-CoV-2 (COVID-19) by the numbers. Elife

Sociodemographic characteristics of Danish blood donors

Factors That Influence the Immune Response to Vaccination

Unless otherwise stated the data presented are % that responded yes (99.8% CI), N. If N is different from the total number it is because not all responded to this item. * Risk ratio of difference in prevalence of symptoms or behaviour between the two age groups, 99.8% CI

* Risk ratio of difference in prevalence of symptoms and behaviour between those that tested positive and those that tested negative for SARS-CoV-2, 99.8% CI and p-value

None of the authors have any conflicts of interest to declare. A c c e p t e d M a n u s c r i p t 24 A c c e p t e d M a n u s c r i p t Figure-1