Skip to main content

Large-scale implementation of pooled RNA extraction and RT-PCR for SARS-CoV-2 detection

Our take —

Increased testing for SARS-CoV-2 will be limited by the availability of test kits. Pooling, where testing is done on combined batches of multiple samples, has the potential to increase throughput while maintaining sensitivity to detect SARS-CoV-2. In this study, authors evaluated two pooling methods for RT-PCR: two-stage Dorfman pooling and two-stage matrix pooling. Using Dorfman pooling, all positive samples were detected and pools with all negative samples were negative. The matrix pooling method also correctly identified all positive samples. Efficiency of these methods are dependent on the prevalence of SARS-CoV-2 in the tested population, and the authors provide a tool to help others to choose a pooling method and size. This study only tested small pool sizes, however, and the use of larger pools in large scale testing may need to be further validated.

Study design


Study population and setting

A single swab for combined deep nasal and oropharyngeal collection was obtained from symptomatic patients (hospital- or community-based) and asymptomatic populations being surveilled for SARS-CoV-2 (e.g., hospital workers and essential employees in non-healthcare settings). Two pooling strategies were compared: Dorfman pooling, where n samples are tested in a single pool and broken out for single sample testing if the pool is positive, and matrix pooling, where samples are ordered in a matrix and each row/column are pooled. In Dorfman pooling, samples from a positive pool are tested individually; in matrix pooling, individual samples are tested if there are two or more positives in both the columns and the rows. The authors tested the following: 186 samples individually, 23 pools of eight samples each with the with Dorfman pooling, and  three 5×5 arrangements for matrix pooling. Samples were pooled for both RNA extraction and PCR.

Summary of Main Findings

Dorfman pooling resulted in no loss of assay sensitivity and pools that were collectively positive in this method were found to have at least one positive individual sample. Matrix pooling resulted in all positive samples being accurately identified, with detection in both the row and column pools. Based on these results, an additional 2,168 samples were tested at the Hadassah Medical Centre in Israel via Dorfman pooling using 311 pooled reactions (14% of the kits that would have been used if the samples had all been run individually) to yield five positive samples.

Study Strengths

These studies utilized large sample sizes and compared the sensitivities of two different pooling strategies, resulting in less reagent use and testing time.


Samples that failed during the RNA extraction step will be missed resulting in false negatives. Small pools of eight samples were tested, and larger pool sizes will need to be validated. RT-PCR can also result in an “indeterminate” pool result, and the authors did not outline the downstream testing strategy for those results. This study only assessed a single PCR assay, and sensitivity of the pooling strategy likely varies by sensitivity of the PCR assay being used which may limit generalization to other assays.

Value added

Using pooling strategies reduces the amount of reagents/kits and time needed for individual testing and can potentially be used for large-scale testing. This study demonstrated two pooling strategies that increased sample throughput while maintaining sensitivity in detection of SARS-CoV-2.

This review was posted on: 5 November 2020