SARS-CoV-2 antibody prevalence among homeless people, sex workers and shelter workers in Denmark: a nationwide cross-sectional study

Study design and sampling strategy

We conducted a nationwide, cross-sectional seroprevalence study between November 2^nd and 20^th 2020, to determine the prevalence of SARS-CoV-2 antibodies among PEH using a rapid point-of care SARS-CoV-2 antibody test (POCT). In addition, we also included shelter workers at the recruitment sites. Participants were invited to fill in a questionnaire in collaboration with a trained project employee, at the same time as the test was performed. We recruited participants from 21 sites in the four biggest cities in Denmark; Copenhagen, Aarhus, Aalborg and Odense. The recruitment sites were shelters, supervised sites for intravenous drug abusers, food distribution sites, meeting places and day/night cafés. In the week prior to our visit, written information was distributed by shelter workers at the recruitment sites notifying the participants of our project. To ensure a high attendance and inclusion, recruitment sites were visited several times, on different days and at different time of the day, including weekends and evenings. Participants were encouraged to wait around for15 minutes for their test results, but in case they did not want to, they were contacted if the test result was positive. Most came back throughout the day to receive their test result.

SARS-CoV-2 in Denmark

The first confirmed case of SARS-CoV-2 in Denmark was detected on February 27^th, 2020. Since then, there have been more than 246 460 confirmed cases of SARS-CoV-2 in Denmark (2). Overall, more than 12 880 of the Danish SARS-CoV-2 patients were hospitalized between January 2020 and March 2021.

Background population

All Danish blood donations are routinely screened for the presence of SARS-CoV-2 antibodies since October 2020. Blood donations take place in all five Danish administrative regions (17) and donors are 17-69 years old. The seroprevalence estimates from the period between the 1^st and 22^nd of November 2020, are used in this study as proxies for the SARS-CoV-2 infection in the background population.

Point of care test

SARS-CoV-2 antibodies in PEH and shelter workers were detected in whole blood, by use of the OnSite COVID-19 IgG/IgM Rapid Test (CTK Biotech inc., Poway, California, United States of America) according to the manufacturer’s recommendations. This POCT is a single use lateral flow chromatographic immunoassay rapid test, intended for qualitative detection and differentiation of SARS-CoV-2 immunoglobulin M (IgM) and immunoglobulin G (IgG) antibodies. Blood was extracted from the fingertip. Test results were read after 15 minutes by a trained project employee. Participants were categorized as seropositive if they were either IgG and/or IgM positive. The test sensitivity and specificity is 96.86% (95% confidence interval (95% CI:93.66%-98.47%) and 99.39% (95% CI 97.80%-99.83%) respectively(18), as reported from the manufacturer. In house validation of CTK’s POCT showed a slightly lower sensitivity of 90% and a specificity of 100% (19). In contrast, blood donors were screened for seropositivity using an Enzyme-Linked Immunosorbent Assay (ELISA), with a sensitivity of 96.7% (95% CI 92.4-98.6) and specificity of 99.5% (95% CI 98.7-99.8) (Beijing Wantai Biological Pharmacy Enterprise, China).

Questionnaire

Participants were invited to fill in a questionnaire provided at the recruitment site, comprised of questions covering sociodemographic characteristics, physical health, use of drugs and alcohol, co-morbidities, symptom manifestations and the use of personal protective equipment against SARS-CoV-2. Alcohol abuse was defined as drinking more than the national recommendations of <1 or <2 beverages per day for women and men, respectively (22). The questionnaire was filled out with the assistance of a trained project employee. Personal data was collected using a web-based electronic data capture tool (Research Electronic Data Capture, REDCap) (20,21).

Study group

Homelessness was defined as people living rough, in emergency accommodation, in accommodation for the homeless and living in non-conventional dwellings due to lack of housing according to the ETHOS (European Typology on Homelessness and Housing Exclusion) classification (23) established by the FEANTSA (European Federation of National Organizations Working with the Homeless) (24). Sex workers were defined as individuals either having reported to be engaged in sex work in the questionnaire and/or being included at one of the designated safe havens for sex workers; Reden and Reden International. Shelter workers are defined as people who either work or volunteer at the recruitment sites.

Primary outcome

The primary outcome was the proportion of the study population with a positive antibody test (IgG and/or IgM) for SARS-CoV-2.

Ethics

This study was performed as a national surveillance study under the authority task of Statens Serum Institut (SSI; Copenhagen, Denmark, the Danish National Institute for Infectious Disease Control and Prevention which performs the epidemiological surveillance of infectious diseases for the Danish government), hence does not require any formal approval from an ethics committee according to Danish law. This decision was made by the regional Ethics Committee of the Capital Region in Denmark (20057075). This study was carried out in agreement with the Helsinki II declaration. All participation in this study was voluntary. This study was registered with the Danish Data Protection Authorities (P-2020-901). All personal data, obtained in REDCap, was kept in accordance with the general data protection regulation and data protection law stated by the Danish Data Protection Agency.

Statistical analysis

Seropositivity is presented as numbers (n) and percentage (%) with 95% confidence intervals (95% CI). Baseline characteristics and exposures are presented as n (%) for factors and median and interquartile range (IQR)) for numeric variables as appropriate. Answers with “do not know” were classified as missing and answers marked “not relevant” were classified as “no”. We tested for significance using students T-test, Wilcoxon rank sum test or Fischer’s exact test. Significant risk factors of seropositivity were combined in a multivariate logistic regression model including region. P-values <0.05 was considered significant. Analyses were performed using RStudio version 1.2.5001 (25).

Characteristics

We recruited a total of 827 participants between the 2^nd and 20^th of November 2020 of which 819 provided their SARS-CoV-2 antibody result as illustrated in figure 1. Participants were recruited from 21 recruitment sites placed in Copenhagen (n = 351), Odense (n = 128), Aarhus (n = 144) and Aalborg (n = 144). Further, 52 participants did not register their test location. The recruited participants included 628 PEH (median age 50.8 (IQR 40.9-59.1) years, 35.5% female) and 191 shelter workers (median age 46.6 (IQR 36.1-55.0) years, 74.5% female). Baseline characteristics of the cohort is shown in table 1. The PEH were older (p<0.001), had a lower BMI (p=0.02) and were more likely to be male (p<0.001) and smoke (p<0.001). Supplementary table 1 shows baseline characteristics of all participants stratified according to seropositivity. The only significant difference was for BMI which was significantly higher among the seropositive group (<0.001).

• Download figure
• Open in new tab

Figure 1: Flow chart reporting the flow of participants through each stage.

Seroprevalence

Of the 819 participants, 55 (6.7%) were seropositive. We found that 43 of 628 (6.8%) PEH and 12 of 191 (6.3%) shelter workers were seropositive, the prevalence in the two groups was not significantly different (p=0.87).

Seroprevalence compared to the background population

In the period between the 1^st and 22^nd of November 2020, the background population (n= 18505) had a 2.9% seroprevalence. This group was characterized by a median age of 43 (IQR 29-54) years and a higher proportion (47.9%) were women. Taken together, the participants in our study (PEH and shelter workers combined) were at a significantly higher risk of seropositivity than the background population (RR 2.2, 95% CI 1.75-2.99, p<0.001). Figure 2 illustrates the regional SARS-CoV-2 seroprevalence in PEH, shelter workers and the background population.

• Download figure
• Open in new tab

Figure 2: Seroprevalence among PEH and shelter workers, compared to the background population

Red: blood donors serving as proxy for the general population; Light blue: shelter workers; Dark blue: people experiencing homelessness.

The seroprevalence among PEH alone was also significantly higher than that in the background population (RR 2.3, 95% CI 1.73-3.14, p <0.001).

The subset of shelter workers (n=191) in general was significantly associated with seropositivity compared to the background population (RR 2.1, 95% CI 1.23-3.72, p=0.02).

Risk factors for attracting infection

Table 2 shows the prevalence of risk factors (drugs, alcohol, sex work) in the study population, stratified by seropositivity. Although sex workers were 1.9 times more likely to be seropositive compared to non sex workers (OR 1.9), none of the lifestyle risk factors were significantly associated to seropositivity.

A total of 285 (45.4%) PEH reported use of drugs, the most commonly used were cocaine (18.6%) and heroin (15.6%). Among the seropositive PEH, heroin (14.5%) was the most common drug while the seronegative PEH were more likely to use cocaine (16.9%). Further, we found that 326 (51.9%) PEH were smokers; of those 326 (51.9%) smoked tobacco and 180 (28.7%) smoked cannabinoids. Other reported smoked substances were cocaine (3.5%) and heroin (2.5%). A total of 307 (51.5%) PEH reported some use of alcohol within the past year.

Of the 191 shelter workers only 3 (1.6%) reported using drugs, all of which were of cannabinoids. Further, 166 (89.2%) reported any use of alcohol within the past year. PEH were significantly more likely to abuse alcohol (p<0.001) and use drugs (p<0.001), compared to the shelter staff.

Sex work

Of the 72 (11.5%) PEH who engaged in sex work, 8 (11.1%) were seropositive. Table 3 illustrates characteristics and risk factors of PEH stratified by sex work. Sex workers were younger, more likely to be female and less likely to report smoking and/or IV drug use than other PEH (all p<0.001). Further, two (1.0%) shelter workers reported having engaged in sex work, of whom one was found to be SARS-CoV-2 seropositive. The seroprevalence among all participants, both PEH and shelter workers, engaging in sex work (n =74) was 12.2%.

For all participants engaging in sex work (n=74), there was a significantly increased risk of seropositivity for IgG antibodies compared to the rest of the study group (RR 2.8, 95% CI 1.26-6.29, p=0.02). However, for the combined seropositivity (IgG and/or IgM antibodies) the difference did not reach significance (RR 2.0, 95% CI 1.00-3.86, p=0.08).

We included a total of 33 shelter workers working at designated safe havens for sex workers, of whom 4 (12.1%) were found to be seropositive. Shelter workers at designated safe havens for sex workers had a far greater infection rate than other shelter workers, however, the difference was not significant (RR 2.4, 95% CI 0.53-10.38, p=0.13). Further, there was no difference in seropositivity between shelter workers at designated safe havens (12.1%) and all sex workers (12.2%) (RR 1.3, 95% CI 0.47-3.75, p=0.75). When combining all participants who reported sex work or were recruited at designated safe havens, shelter workers and sex workers (n=107), we found a significantly increased risk of seropositivity (IgG and/or IgM antibodies) compared to other participants (RR 2.1, 95% CI 1.16-3.75, p=0.02). In a multivariate logistic regression model of region of stay, being a sex worker or working at a designated safe haven remained a significant risk factor of seropositivity compared to PEH who does not engage in sex work (OR 2.4, 95% CI 1.12-4.70, p=0.02).

Symptoms and self-reported illness

Of the 628 PEH, 57 (9.1%) suspected previous infection with SARS-CoV-2 of whom 8 (14%) were seropositive. A total of 371 (59.1%) PEH reported being previous tested for SARS-CoV-2, of whom 11 (3.0%) were reportedly positive at the time. However, of those only 3 (27.3%) tested positive for SARS-CoV-2 antibodies in our study. Of the shelter workers, 154 (80.6 %) reported previous testing for SARS-CoV-2. No shelter workers reported having previously tested positive for SARS-CoV-2, but 21 (11%) suspected previous infection with SARS-CoV-2 of whom 5 (23.8%) were seropositive.

Seventeen of 43 (39.5%) seropositive PEH and 10 of the 12 (83.3%) seropositive shelter workers reported having had symptoms. Overall, a total of 441 (53.8%) participants in our study group reported having had one or more symptoms. Of those, 15 (3.3%) reported being hospitalized at the time of their symptoms. Among the PEH 303 (48.2%) reported experiencing one or more symptom, of whom 26 (8.6%) though it was attributable to SARS-CoV-2. Of the shelter workers 138 (72.3%) reported one or more symptom, and of those 20 (14.5%) thought the symptoms were attributable to SARS-CoV-2; significantly more than among the PEH (RR 1.7, 95%CI: 1.36-2.22, p<0.001). Similarly, shelter workers were more likely than PEH to have more than three symptoms (RR 1.7, 95%CI: 1.36-2.22, p<0.001). The most common symptoms reported in the combined cohort were fever (23.2%) and shivers (18.5%). However, there was no observed significant association between experiencing symptoms and seropositivity (RR 0.8, 95%CI: 0.50-1.38, p=0.49). Table 4 shows symptoms stratified by seropositivity. Symptoms in PEH and shelter workers, is illustrated in supplementary table 2.

Use of protective means against SARS-CoV-2 infection

Figure 3 shows use of protective means against SARS-CoV-2 in PEH compared to shelter workers. Only 25 (4%) of the PEH reported that they did not follow any of the recommended guidelines. Among the shelter workers only 3 (1.6%) reported not following the guidelines.

• Download figure
• Open in new tab

Figure 3: Percentage of PEH and shelter workers who follow the national SARS-CoV-2 measures and guidelines.

Blue: Shelter workers; Red: people experiencing homelessness (PEH). 95% CI is illustrated.

Strengths and Limitations

Our study has several limitations. First, the cross-sectional study designs do not allow determination of time of infection nor provide information on when participants became seropositive. Individuals who might already have tested positive by PCR at an earlier point in time, might have chosen not to participate in this study. Individuals anticipating a positive result might have chosen not to participate, fearful of the consequences such as isolation. If so, our results could be biased and the seroprevalence be underestimated. However, our apprehension is that the desire to participate and get tested was high (10.2% of an estimated 6 431 homeless people in Denmark). Of the 827 participants in our study group, 544 (67.0%) reported having previously been tested (nasopharyngeal swab and/or antibody test) and still participated. Further, questions on sociodemographic characteristics, physical health, use of drugs and alcohol, co-morbidities, symptom manifestations and the use of personal protective equipment against SARS-CoV-2 were self-reported, hence an information bias could have affected our results We compared our seropositivity findings to that of blood donors serving as a proxy to the background population, with some limitations to consider. First, blood donors are required to have a good general health and are ineligible to donate blood if they have ever engaged in sex work, have certain medical conditions, travel to certain international destinations and/or receive certain immunizations. Seropositivity could as a result potentially be lower in this group of the population compared to the background population. Although, blood donors due to their good health, are more represented in the labor market and so more at risk of SARS-CoV-2 infection. Another limitation is that blood donors and our study group were tested using different methods, POCT vs ELISA however both with high and similar sensitivity and specificity. Antibodies generated in response to SARS-CoV-2 exposure are generated in the weeks after the acute phase of the infection, thus it may not register in recently infected participants. However, recent studies have shown that SARS-CoV-2 IgM antibodies reach threshold to be detected 5-7 days after symptom onset (25). Thus, participants currently or recently infected with SARS-CoV-2 might not have been identified in this study. A strength of serology approach versus test by PCR is that it allows us to detect SARS-CoV-2 antibodies in those categorized as asymptomatic carriers and those with suspected SARS-CoV-2 despite negative PCR results. Furthermore, serology testing with a POCT is easy and can be performed as a self-test. It does not require a blood sample nor laboratory equipment, hence is less costly than a laboratory serology test and provide more rapid results.