# Limitations of point-of-care testing for low SARS CoV-2 loads: Insights for future pandemics

## TL;DR

This study shows that rapid antigen tests for SARS-CoV-2 are less accurate when viral loads are low, suggesting they should not be used alone for diagnosis.

## Contribution

The study evaluates the performance of WHO-certified Ag-POCTs in low SARS-CoV-2 viral load samples, revealing significant limitations not previously highlighted.

## Key findings

- SD Biosensor detected 44.7% of positive samples, the highest among the three tested kits.
- Sensitivity of Ag-POCTs dropped to 32% for intermediate and 7% for low viral load samples.
- Ag-POCT results significantly deviated from manufacturer-reported performance in low viral load cases.

## Abstract

The emergence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has seen a surge in the development of diagnostic assays. However, the performance of antigen point-of-care tests (Ag-POCTs) on samples with low viral load has not been evaluated.

To evaluate the accuracy of three World Health Organization (WHO) certified Ag-POCTs in comparison to the reverse transcription polymerase chain reaction (RT-PCR) technique.

The study was conducted at Pietersburg Hospital Limpopo, South Africa between March 2020 and April 2023.

A total of 371 SARS-CoV-2 nasopharyngeal samples from the National Health Laboratory Service were tested using Ag-POCTs from Abbott Panbio, Roche RDT and SD Biosensor, following manufacturer instructions. All samples had RT-PCR results with Ct values between 13 and 45. Reverse transcription polymerase chain reaction results were compared and correlated with Ag-POCT results.

Of the 371 samples, the SD Biosensor Standard Q test kit detected the most positive isolates 166 (44.7%), followed by the Abbott Panbio. A total of 153 (41.2%) positives, while the Roche SD detected 134 (36.1%) samples. High viral load (Ct < 25) sensitivity and specificity exceeded 77%, while intermediate (Ct 25–35) and low viral load (Ct > 35) sensitivity and sensitivity dropped to 32% and 7%, respectively.

The performance rapid antigen tests was low on samples with low viral load with results markedly different from the manufacturer’s reported performance.

Rapid antigen tests should not be used alone for diagnosis, especially in samples with low viral load.