Swab pooling: A new method for large-scale RT-qPCR screening of SARS-CoV-2 avoiding sample dilution

Abstract

To minimize sample dilution effect on SARS-CoV-2 pool testing, we assessed analytical and diagnostic performance of a new methodology, namely swab pooling. In this method, swabs are pooled at the time of collection, as opposed to pooling of equal volumes from individually collected samples. Paired analysis of pooled and individual samples from 613 patients revealed 94 positive individuals. Having individual testing as reference, no false-positives or false-negatives were observed for swab pooling. In additional 18,922 patients screened with swab pooling (1,344 pools), mean Cq differences between individual and pool samples ranged from 0.1 (Cr.I. -0.98 to 1.17) to 2.09 (Cr.I. 1.24 to 2.94). Overall, 19,535 asymptomatic patients were screened using 4,400 RT-qPCR assays. This corresponds to an increase of 4.4 times in laboratory capacity and a reduction of 77% in required tests. Therefore, swab pooling represents a major alternative for reliable and large-scale screening of SARS-CoV-2 in low prevalence populations.

Fig 1. Molecular screening for SARS-CoV-2 through swab pooling method.

1) From each individual, two swabs are collected. One nasopharynx swab is collected from one nostril and stored in an individual 3 mL tube. Then, another swab is collected through the other nostril and stored in a 5 mL tube containing up to 15 other individuals (pool tube). 2) In the laboratory, the RNA from pooled swabs is extracted, and SARS-CoV-2 detection is performed using RT-qPCR. If a given tested pool presents a positive result, all corresponding individual samples are then processed to identify infected patients.

Fig 2. Swab pooling and dilution effects.

(A) In a laboratory controlled experiment, 16 positive samples were inoculated in negative individual and pool samples maintaining the dilution factor of 1.67 between individual (3 mL) and swab pooling (5 mL) samples. Mean Cq differences between individual and pool samples estimated for E and RdRp genes along with 95% credibility intervals are shown in the top-left corner of the graphs. (B) Expected Cq variations (ΔCq) for sample pooling methods with 10, 16 or 32 samples compared to swab pooling in different amplification efficiencies. Expected ΔCq was calculated using Efficiency^slope = dilution factor [–31]. In swab pooling, the number of samples are not related to the dilution factor.

Fig 3. Paired analysis of pool and individual tests.

(A) Results for samples analyzed in parallel as pools and as individual tests. (B) Total number of positive samples for each positive pool. (C) Correlation between the number of positive samples within a given pool and the corresponding pool Cq (R: Pearson’s corr. coefficient; ⍴: Spearman’s rank corr. coefficient). Points were colored by the total number of individuals within the pool (pool size). (D) Cycle quantification values for positive pools and corresponding positive individual samples.

Fig 4. Large-scale molecular screening of SARS-CoV-2 with swab pooling.

(A) Number of positive pools related to the original pool size and the number of positive samples within the pool. (B) Correlation between Cq values from pools and their respective individual samples stratified by the number of positive samples within the pool (1, 2, and 3 or more positive individual samples). Point color represents the pool size (from less than 10 to 16 individuals). Cq values for the three marker genes tested were included. Correlation coefficients are presented in the figure (R: Pearson’s corr. coefficient; ⍴: Spearman’s rank corr. coefficient). (C) Cq values from individual tests and corresponding pools were assigned to each patient. A hierarchical model with patient-specific intercepts was used to estimate mean variations between individual and pool samples across varying genetic markers. Estimates for mean Cq differences along with 95% credible intervals are shown in the top-left corner of each graph.