Implementation of Practical Surface SARS-CoV-2 Surveillance in School Settings

Over the past 2 years, the COVID-19 pandemic has claimed the lives of over 5 million people worldwide (1). Due to high density occupancy of indoor spaces for prolonged periods of time, schools are often of concern for transmission, leading to widespread school closings to combat pandemic spread when cases rise. However, K-12 schools are important resources for communities, which, besides education and childcare, often provide food, authoritative and trusted information, and a sense of belonging and security (2, 3). Therefore, alternative approaches that keep children in school are highly desirable. Performing pediatric clinical testing, such as SARS-CoV-2 detection via RT-qPCR, is expensive, difficult from a consent perspective, and increasingly politicized. Wastewater testing, although highly effective even at the level of individual cases and buildings (4), can only identify SARS-CoV-2 among the subset of individuals that defecate at school, and often cannot provide spatial resolution at finer levels. A complementary method that we have deployed in Safer At School Early Alert (SASEA) (5) is surface sampling, which allows detection of SARS-CoV-2 from surfaces within a classroom. This is a screening method, not a diagnostic; however, even without high sensitivity, surface sampling provides visibility into environments where individuals will not consent to testing and where cases are not picked up through wastewater. In particular, surface sampling can localize cases within a single classroom (5).

In previous work, we showed SARS-CoV-2 persists on a range of school-relevant surfaces using our SASEA workflow, based on collecting swabs into a 0.5% sodium dodecyl sulfate (SDS) wt/vol solution (Acros Organics, 230420025), performing nucleic acid extraction on the Kingfisher Flex liquid-handling robot (Thermo Scientific), and performing RT-qPCR using the QuantStudio 7 (5). We implemented this protocol in the research phase of SASEA (3, 6). However, this protocol requires specialized reagents and equipment that is not generally available to public health laboratories, and we needed to test its generality using workflows already operational in the San Diego County Public Health Laboratory (SDPHL), which employs Viral Transport Medium (VTM) (NEST Scientific USA, 202016) for sample collection, and the PerkinElmer (PE) workflow for nucleic acid extraction and RT-qPCR (Table S1) (7). Adapting surface sampling to this widely used clinical workflow would enable its application to an entirely new sampling modality and allow surface sampling to be incorporated into a wide range of programs in schools, prisons, nursing homes, and other high-risk, indoor settings.

To assess whether conclusions drawn from our established Thermo pipeline (6) could be generalized to the more widespread PerkinElmer pipeline, we first compared the performance of both methods using contrived samples. Briefly, the Thermo pipeline collects surface swabs into Matrix tubes containing 0.5% wt/vol SDS in water, extracts nucleic acids using the Omega MagBind Viral RNA/DNA kit (SKU: M6246-03) using the KingFisher Flex platform, and detects SARS-CoV-2 presence through a miniaturized 3 μL-reaction version of the TaqPath COVID-19 Combo kit (ThermoFisher Scientific, A47814) on a QuantStudio 7 Pro qPCR machine (Thermo Fisher Scientific). The PerkinElmer pipeline follows the Emergency Use Authorization of sample collection into Viral Transport Media, requires a heat-inactivation step (65C for 15 min), extracts nucleic acids using the chemagic 360 platform and reagents, and detects SARS-CoV-2 presence using the standard 15 μL-reaction protocol of the PerkinElmer New Coronavirus Nucleic Acid Detection kit on an Analytik Jena qTOWER³ 84 G real-time PCR system.

To manufacture the contrived samples, we deposited 10 μL of a heat-inactivated SARS-CoV-2 dilution series (strain WA-1, SA-WA1/2020) on laminated cards, making triplicate cards for each of the three concentrations used (206, 1024, 4096 genomic equivalents (GEs)/μL). Nine replicates were swabbed with 0.5% SDS and processed by the Thermo pipeline while the other nine were swabbed with VTM and processed by the PerkinElmer pipeline. We found that the two platforms yielded highly correlated results (Pearson correlation, r = 0.810, P = 2.87 × 10⁻⁶) but the SDS/Thermo pipeline was more sensitive by ~4 Cq units on average (Fig. 1).

To test whether these conclusions extended to a real-world setting, we collected duplicate biological replicates of 30 samples from isolation housing in which known COVID-19 patients, confirmed by positive anterior nares RT-qPCR, were housed (IRB-approved research under HRPP UCSD protocol 200477). The distribution of sampled surfaces is given in Table 1, and their spatial localization in one of the apartments is given in Fig. 2 as an example. In a crossover protocol to separate the effects of the swabbing/transport medium and RNA extraction from the effects of the qPCR assay, we ran each protocol through RNA extraction, exchanged RNA between the UCSD lab and the SDPHL, then subjected the resulting RNA to RT-qPCR on the other platform (Table S2). This created four sets of samples summarized in Table S3.

We found that each laboratory performed best when sample extraction and RT-qPCR processing occurred in the same facility, presumably because of RNA degradation during transit (Fig. 3A and B). A Kruskal-Wallis H test confirmed that the mean Cq’s differ significantly across the pipeline combinations (H = 26.9, P = 6.22 × 10⁻⁶) (Fig. 3C). Pairwise Mann-Whitney U tests between the groups showed that the Cq ranks are significantly different between groups processed at the different PCR facilities, but there is no significant difference for groups within PCR facility. The SDPHL RT-qPCR assay performed better on the samples processed with the PE pipeline than on the Thermo/PE paired samples (mean Cq difference 0.88, U = 0.693, P > 0.1), whereas the UCSD RT-qPCR assay performed better with the samples processed with the Thermo pipeline than with the paired PE/Thermo samples (mean Cq difference 1.2, U = 1.48, P > 0.1) (Fig. 3C). However, the Thermo workflow provided an advantage in sensitivity of 5.1 Cq units on average over the PE workflow yielding a P-value of 5.86 × 10⁻³ after correcting for multiple comparisons (FDR-Benjamin/Hochberg) (U = 3.72). None of the assays were perfect: the Thermo pipeline detected 6 samples that the PE pipeline counted as negative, and the PE pipeline detected 2 samples that the Thermo pipeline counted as negative (Table S2). To calculate the sensitivity of each pipeline, we assumed that if at least one of the pipeline combinations (Thermo, Thermo/PE, PE, PE/Thermo) detected the presence of virus, then that sample could be considered a true positive resulting in a true positivity rate of 91% for the Thermo pipeline and 73% for the PE pipeline. However, we were unable to make any assumptions about the true negative rate and so are unable to calculate the specificity of each pipeline.

We note that although the PE assay is less sensitive, this level of accuracy is sufficient for projects such as SASEA where the goal is to screen environments for further resource allocation for COVID-19 mitigation efforts rather than to perform a diagnostic test. In conclusion, although the optimized Thermo protocol we developed for SASEA offers considerable sensitivity advantages, the PE assay is still pragmatically useful for classroom SARS-CoV-2 surveillance and can leverage large existing investments in infrastructure and expertise.

Surface sampling with two transport media (VTM and SDS) and subsequent nucleic extraction and SARS-CoV-2 readout with two RT-qPCR pipelines (Thermo and PerkinElmer) are compared in a factorial design.

This research was supported by NIH grant (K01MH112436) to RFM, the County of San Diego Health and Human Services Agency (Contract 563236), and the Career Award for Medical Scientists from the Burroughs Wellcome Fund to A.F.C. We thank Marisol Chacon, Sydney C. Morgan, Alhakam Nouri, Ashley Plascencia, Christopher A. Ruiz, and Lizbeth Franco Vargas for their support with environmental SARS-CoV-2 detection as part of the EXCITE Lab.

The following reagent was deposited by the Centers for Disease Control and Prevention and obtained through BEI Resources, NIAID, NIH: SARS-Related Coronavirus 2, Isolate USA-WA1/2020, NR-52281.

For a companion article on this topic, see https://doi.org/10.1128/mSystems.00109-22.