Optimizing RT-PCR detection of SARS-CoV-2 for developing countries using pool testing

Mauricio J Farfan; Juan P Torres; Miguel ORyan; Mauricio Olivares; Pablo Gallardo; Carolina Salas

Our study demonstrates that pooling of 5 negative and/or 4 negative and one positive SARS-CoV-2 nasopharyngeal samples in the same RT-PCR run can effectively identify all negative samples and detect the positive sample. Furthermore, similar detection results were observed when comparing automated and manual extraction of the sample. Results of the sample without nucleic acid extraction, was unsatisfactory, with a significant increase in CT values, and thus for risk of a false negative result. . This study is complementary to ours as they used samples negative for other viruses, not including samples with known CT values. Another study from Israel, found that a single positive sample could be detected even in pools of extracted nucleic acid of up to 32 samples, with an estimated false negative rate of 10% 8 . Multi-sample pools can be a good alternative to increase testing throughput, using less reagents and offering faster results. This is relevant for underdeveloped or developing countries, where resources may be scarce. The possibility of increasing the number of samples for SARS-Cov-2 detection could significantly help countries with reduced resources, to obtain better outcomes for the COVID-19 pandemic. For post-pandemic screening of large populations, sample pooling also will represent an important alternative. Our study has the limitation of having performed only 31 pools on 63 nasopharyngeal samples (40 negatives and 23 positives), however, results were consistent and provide relevant information for the implementation of strategies that might allow optimizing the detection of SARS-CoV-2. We included 5 samples in each pool which seems adequate in our current situation with a near overall 10% positivity rate. In areas with lower positivity rates, especially in future post-pandemic testing, increasing sample numbers in the pool can be considered. Finally, we did not test the inclusion of more than one sample in each pool, however, we would not expect this to modify the observed results.