RT-PCRs for SARS-CoV-2 are performed for various indications. First, asymptomatic individuals may be screened because they had previously been in close contact with somebody diagnosed with COVID-19. In this setting, it may be required to test several times to confirm or rule out an infection of the test subject. This approach is considered a public health measure, because it detects SARS-CoV-2 infection early in pre-symptomatic subjects and enables the breach of infection chains 13,34,35. Second, asymptomatic health-care workers and other staff working in hospitals or nursing homes, who are in contact with vulnerable groups of patients, can be screened for the safety of both co-workers and patients 36,37. Third, individuals with symptoms suggestive of COVID-19 such as fever, malaise, dry cough or loss of taste or smell ought to be tested for diagnostic purposes and initiation of subsequent specific health care measures.
The majority of diagnostic PCRs for SARS-CoV-2 are performed for screening of asymptomatic persons or in a diagnostic approach of patients with mild and unspecific symptoms and low to medium pre-test probability of actually carrying the disease. Therefore, PCR testing needs to be conducted on a large scale and with high sensitivity not only in a given country but worldwide. As a consequence, millions of swabs and PCR kits are needed globally during the ongoing pandemic, challenging resources and supply chains.
Confronted with limitations in resources and logistics, we investigated the utility of several kinds of swab sets in a well-defined cohort of subjects. We recruited symptomatic patients who were confirmed to have COVID-19 by RT-PCR and who were hospitalized for supportive care such as conventional oxygen therapy. These patients were re-swabbed for follow-up within five days after the initial positive RT-PCR. The focus of our study was the technical validation of alternative swabs systems used to collect oropharyngeal specimens and their sensitivity in identifying PCR-confirmed cases. However, one has to admit that PCR results may turn negative after a short period of time, at least in not severely affected subjects38.
As applies for most other sections of industry, biomedical companies which are involved in molecular diagnostics of COVID-19 and other diseases rely on central manufacturing and cross-border delivery to serve a global market. Their repertoire may extent from simple materials such as plastic swabs and tubes to highly sophisticated reagent kits and diagnostic platforms. In late March 2020, less than four weeks after the COVID-19 pandemic reached Austria, we started to experience severe shortages in materials and reagents in the diagnostic laboratories at the University Hospital of Innsbruck. The earliest and largest limitations affected pharyngeal swab sets because factories producing them for the global market are located in regions most heavily affected by COVID-19 such as Lombardy in Northern Italy. Therefore, we placed utmost priority on the active search for alternative swab sets. By April, the number of SARS-CoV-2 PCR tests requested rose to more than 400 per day at the University Hospital of Innsbruck. This meant a massive challenge for the local resources of a tertiary centre.
The earliest and simplest way to approach these limitations was to use dry plastic swabs which are compatible with PCR as recommended by some authorities. However, collecting an upper respiratory sample with a dry swab and eluting the viral RNA contained in the specimen with normal saline may reduce the diagnostic sensitivity of the subsequent RNA extraction and RT-PCR, consequently 39,40. This limitation has been known for other pathogenic viruses and was confirmed for SARS-CoV-2 in our study (Figure 1). Therefore, we aimed to improve the yield of viral RNA by mixing a VTM in-house from a small number of components that are more broadly available because many vendors deal with those reagents. However, we also purchased other specimen collection sets and included them in our comparative analysis. For instance, sets used in gynecology or urology which are suitable for the detection of Chlamydia, Mycoplasma and Ureaplasma spp. but also HPV, were evaluated to collect specimen from the posterior oropharyngeal wall. Furthermore, the inhouse VTM was used in combination with several swabs. However, when using one with a wooden stalk and a cotton tip, we failed to detect SARS-CoV-2 by RT-PCR. Concretely, we saw that the Raucotupf® swab resulted in high Ct values for the viral E gene and undetectable levels of the viral S gene as surrogates of poor sensitivity. Organic compounds present in wood and cotton could have inhibited PCR reactions. Specifically, tannic acid present in wood has been demonstrated to inhibit Taq polymerase 41.
Another option to circumvent backorders in plastic swabs is to produce them in a 3D printer as recently validated in other laboratories 42,43. Besides, manufacturers from other branches of industry, which are familiar with the production of plastic ware, may start to produce customized tubes. On the one hand, these ideas do not resolve the limitation of sensitivities of dry swabs. On the other hand, tubes for molecular biology applications may require specifications and purities not all companies can provide. This may be especially relevant for RNases that can contaminate surfaces including the outside of swabs or the inside of tubes or being present in biological samples 44. However, in our validation, we did not see any effect of the addition of an RNase inhibitor that would prevent the negative effects of possible contamination of collection tubes. Thus, when tubes are produced under clean room conditions and handled with gloves during sampling and laboratory work-up, RNases may not be of concern for detecting SARS-CoV-2 by RT-PCR.
Companies that produce complete swab sets with VTM typically do not report their detailed composition. While the components used are often disclosed upon request, the exact quantities present in the mixture are usually not provided. Therefore, users have to rely on recommendations and recipes published by health authorities and institutions such as the CDC. Importantly, the in-house system produced according to CDC recommendations by the Innsbruck Hospital Pharmacy was also low-priced. Concretely, the costs for its constituents were approximately €0.45 per unit, not including the plastic swab. Therefore, the in-house system was highly efficient not only from a diagnostic but also from an economic point of view.
Our study has several limitations. First, the study was monocentric by design, precluding general recommendations. Second, we relied on the Realstar® RT-PCR system provided by Altona® which is in RUO status and detects the E gene and the S gene along with an internal extraction and amplification control. However, when we validated its diagnostic performance in comparison with the Roche® Cobas SARS-CoV-2 test or an in-house RT-PCR for SARSCoV-2 using the primers and probes published by the CDC, we found no significant differences between these methods, suggesting that our results may not depend on a specific target sequence or type of polymerase. Third, we only enrolled hospitalized patients with moderate symptoms. Fourth, our study exclusively deals with oropharyngeal swabs, a common specimen taken for SARS-CoV-2 PCR 45 because we had observed in that advance serial sampling at the posterior oropharyngeal wall is technically less challenging for healthcare professionals, physically less demanding for test subjects and thus more reproducible than in the nasopharyngeal cavity. Fifth, due to a limited number of eligible patients, the sample size of our study is rather small and further studies which have a prospective and multicentric design or recruit out-patients as well may be required for more general conclusions.
One of the most striking findings of our study is that none of the swabs systems used could confirm the previous RT-PCR diagnosis of COVID-19 with 100% accuracy. In fact, the inhouse system reached the highest percentage of bona fide true positive results, concretely, 81.3%. While these numbers may first appear counterintuitively low, they support the idea that the timing and mode of specimen collection, transport and storage conditions as well as the sensitivity of the subsequent RT-PCR are critical for an accurate diagnosis. Therefore, when the clinical suspicion for infection is high, repeated testing may be required which will consume additional resources, though 46,47. Furthermore, many reports suggest that some patients with COVID-19 can display intermittently or continuously negative RT-PCR results in upper respiratory specimens 48,49. Unfortunately, the type of swab sets used is not reported in many studies rendering comparisons and general conclusions unnecessarily difficult.
In summary, our study emphasizes that materials and methods have to be thoroughly validated in diagnostic laboratories. This is especially relevant when alternative systems are adapted 50. We could also demonstrate that, by employing published protocols, the cooperation of the diagnostic laboratories and the Hospital Pharmacy at a tertiary center was able to circumvent limitations in the regional supply of swab sets from the global market. Our results argue for stringent standardization of specimen collection not only for diagnostic SARS-CoV-2 PCR but also for intervention studies on COVID-19 many of which include RT-PCR results as secondary end points.