Evaluating SARS-CoV-2 Saliva and Dried Blood Spot Surveillance Strategies in a Congregate Population

The optimal approach to COVID-19 surveillance in congregate populations remains unclear. Our study at the US Naval Academy in Annapolis, Maryland, USA, assessed the concordance of antibody prevalence in longitudinally collected dried blood spots and saliva in a setting of frequent PCR-based testing. Our findings highlight the utility of salivary-based surveillance.

Collection Kit (Neoteryx, https://www.neoteryx. com) and tested them for SARS-CoV-2 reactive IgG by using an in-house multiplex microsphere-based immunoassay. The antigenic targets were a prefusionstabilized SARS-CoV-2 spike glycoprotein ectodomain trimer and a nucleocapsid protein; we detected antigen-specific IgG levels by using a Bio-Plex 200 HTF multiplexing systems (Bio-Rad, https://www. bio-rad.com) and reported results as median fluorescence intensity (MFI).
We collected saliva samples by using an Oracol S14 collection device (Malvern Medical Developments, https://www.malmed.co.uk) and tested them as previously described (11). We tested samples for IgG binding to any of 7 SARS-CoV-2 antigen components (2 SARS-CoV-2 nucleocapsid proteins, 3 receptor-binding domain [RBD] proteins, and 2 spike proteins) by using a multiplex immunoassay. After background subtraction, we classified samples positive for RBD and nucleocapsid IgG as indicative of prior infection, whereas we classified samples positive for only RBD IgG as indicative of SARS-CoV-2 vaccination.
As part of routine clinical care, USNA's Brigade Medical Clinic collected nasopharyngeal swab specimens from all returning midshipmen in August and throughout the school year when they visited the clinic with symptoms of respiratory illness. In addition, each week we randomly selected 15% of the asymptomatic midshipmen population for reverse transcription PCR (RT-PCR) screening; we also tested 100% of in-season varsity athletes each week. We excluded from weekly testing all participants who had confirmed positive SARS-CoV-2 infection during the preceding 90 days. We tested nasopharyngeal swab samples by using SARS-CoV-2 RT-PCR and made results accessible through electronic medical records.
We compared seroconversion rates with cumulative frequencies of molecularly confirmed infections. We calculated correlation coefficients for spike IgG and nucleocapsid IgG MFI in saliva and DBS. We used the Cohen kappa coefficient (κ) to measure concordance of saliva with DBS nucleocapsid IgG and spike IgG positivity and to measure concordance of PCR tests with seroconversions.
This study was approved by the Uniformed Services University Institutional Review Board under protocol IDCRP-129. All participants provided written informed consent.

showed evidence of SARS-CoV-2 infection based on spike IgG values in DBS.
Among the participants who were serologically negative for SARS-CoV-2 at enrollment, 18 seroconversions were detected in saliva, 19 were detected in DBS, and 19 were detected by PCR by the end of follow-up (Table 1); however, at V4 (postvaccination), additional cases were detected by DBS and saliva that were missed by PCR testing. One participant had a positive PCR result before a serologic result; the PCR test was conducted in August 2020, and the participant had no record of seroconversion through the end of the study.
Spike IgG MFI in saliva and DBS were significantly correlated at all 3 timepoints ( Figure 1); high spike IgG values at V4 were consistent with the participants receiving vaccinations in March-April 2021. Nucleocapsid IgG MFI in saliva and DBS also were significantly correlated at all 3 timepoints (Figure 2).

Conclusions
This study, conducted among a population of midshipmen at USNA in the first year of the COVID-19 pandemic, employed blood and saliva collection at multiple visits to evaluate the validity of salivary antibody surveillance. We observed concordance between DBS and saliva for the detection of spike and nucleocapsid IgG, and both biospecimen types were similar to RT-PCR for detection of cases. We noted that all vaccinees mounted a spike IgG response in DBS by V4, consistent with the known immunogenicity of these vaccines, but only 49.1% vaccinees had detectable nucleocapsid IgG at V4, indicating a substantive SARS-CoV-2 infection attack rate in the first half of 2021.
This assessment of SARS-CoV-2 detection in a congregate setting can help inform approaches for detection of SARS-CoV-2 in populations before and after vaccination. Prior evidence shows that PCR testing is an efficient method of infection control in congregate communities if administered regularly but that asymptomatic cases may still be undetected (12). These findings may apply to surveillance for other respiratory infections, such as influenza. A limitation to this study was the inability to directly match RT-PCR testing with blood and saliva collection, small sample size with paired samples, and loss to follow-up after the end of the academic year.
In summary, this assessment supports using saliva testing as a less invasive, more feasible surveillance method for monitoring changes in disease prevalence and susceptibility in large populations. Future directions include validation of alternative antibody targets, in both serum and saliva, which can discriminate antibody prevalence in the context of preexisting vaccination and postinfection hybrid immunity.