Challenges and issues of SARS-CoV-2 pool testing – Authors' reply

We appreciate the comments on our letter,¹ in which we described a strategy to identify asymptomatic people infected with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in large populations of uninfected people when testing capacity is low and laboratory infrastructure is overwhelmed. We proposed pool testing to screen for individuals who might spread SARS-CoV-2 without showing any symptoms. He and colleagues² reported temporal patterns of viral shedding and inferred from their data that viral load peaks 0·7 days before symptom onset and estimated that 44% of SARS-CoV-2 infections occur during the pre-symptomatic stage of the index case. Studies indicate a steady decline of viral loads and infectiousness after symptom onset.2, 3

We recommended restricting pool testing to asymptomatic people, excluding patients with severe acute respiratory illness and high-risk contacts, and for when testing capacity is limited.¹ For symptomatic individuals we use individual testing so not to miss patients with low or borderline SARS-CoV-2 viral loads and to obtain diagnostic results as quickly as possible.

We proposed that pool sizes should “accommodate different infection scenarios”, which should “be optimised according to infrastructure constraints”, and suggested that “sub-pools can further optimise resource use when infection prevalence is low.”¹ We investigated 164 pools in total and mainly used pool sizes of five (81 pools) or ten (45 pools), as further illustrated in the appendix. Beyond that, we explored the pooling of 30 samples with sub-pools of ten samples (five pools; appendix). During the period of our study (March 13–21, 2020), the positivity rate was below 2% in samples examined by pool testing. We agree with Jens Eberhardt and colleagues that pool sizes should be smaller than 30 when prevalence is higher than 2%.

We agree with Baijayantimala Mishra and colleagues that the swab sampling procedure is of paramount importance. The testing for host nucleic acid might help to confirm accurate sampling and improve the accuracy of a negative test result. Yet, as SARS-CoV-2 is released in mucous or saliva,³ the absence of cellular control signals does not necessarily prove that mucous or saliva is missing and that swab taking needs to be redone.

Jaehyeon Lee and colleagues highlighted another point in our letter. In our study, we used the Copan Liquid Amies Elution Swab Collection and Preservation System for sampling and pooled media before RNA extraction. We performed nucleic acid extraction from 400 μL of our single or pooled samples. From theoretical considerations one would expect dilution effects and a resulting increase of Ct values in pools. We observed a puzzling increase of detection sensitivity in pools containing single samples with high Ct values (appendix). This increase did not occur randomly but was reproducible, systematic, and significantly associated with higher Ct value samples for both E-gene and S-gene RT-PCRs (appendix). To explain our observation, we hypothesised that samples with higher Ct values might have gained detection efficiency through an RNA carrier effect in pools from the other negative samples with potentially higher cellular RNA content. We are currently addressing this interesting question in a further study, as well as whether different swab collection systems affect this phenomenon. The concern raised by Lee and colleagues that this phenomenon might cause false-positive results is not supported by our data obtained with now more than 3900 pools assessed in our institution since March, 2020.

In a broader context, several distinct steps contribute to accurate test results. Major contributing factors are adequate sample collection, quality of swabs, transport media, efficient nucleic acid extraction from a sufficient amount of material, and a highly sensitive detection method. All these steps need to be optimised and validated within the laboratory to obtain optimal pool testing efficiency and accuracy.

Owing to highly diverse laboratory settings, it might be difficult to harmonise worldwide pool testing protocols for SARS-CoV-2. However, we would be grateful if national authorities could guide SARS-CoV-2 pool testing procedures as has been done for blood donor pool testing in Germany⁴ and recently been announced by the US Food and Drug Administration for SARS-CoV-2.⁵