Thread by @deeksj

Updated @cochranecollab review of Rapid Tests for Covid-19 is here

Rapid, point‐of‐care antigen and molecular‐based tests for diagnosis of SARS‐CoV‐2 infection - @jacdinnes @deeksj - 2021 | Cochrane Library

cochranelibrary.com/cdsr/doi/10.10…

1/12

Update included electronic searches to end of Sept, and other resource up to mid Nov 2020. Next update is already underway. Many thanks to the great crowd of people involved in putting this together.

2/12

Included both lateral flow antigen tests – data on 16 tests (of 92 with regulatory approval) in 48 studies (n=20,168)

And rapid molecular tests – data on 5 tests (of 43 with regulatory approval) in 30 studies (n=3,919)

3/12

Lateral flow antigen tests showed substantial variation in sensitivity.

Higher sensitivity in:

symptomatic (72%) than asymptomatic (58%);

first week of symptoms (78%) vs later (51%).

differences between brands

Specificity consistent – average of 99.6 (99.0, 99.8)

4/12

For testing people with SYMPTOMS
for tests with 5 of more studies

Sensitivity of SD Biosensor was 80% which meets the WHO acceptable use criteria.

Abbott Panbio was lower (74%),

Innova (56%)

Coris Bioconcept (34%)

Other tests have data in review.

5/12

Less data on testing people who were ASYMPTOMATIC

Only Abbott Panbio and SD Biosensor evaluated in >200 individuals.

Abbott Panbio – sensitivity in symptomatic was 72% vs 58% in asymptomatic

SD Biosensor – sensitivity in symptomatic was 79% vs 61% in asymptomatic.

6/12

Data on Innova in asymptomatics was published after search and will be included in review update

Liverpool showed sensitivity of 40% from 70 cases
University of Birmingham showed sensitivity of 3% from 8 cases

Not convincing or a lot of data for testing millions of people

7/12

No evaluations were found for:

Testing symptom free children

The accuracy of test strategies

Where infectiousness is the target condition rather than infection (there is no reference standard that can
be used for this).

8/12

Most studies of Molecular tests were undertaken in laboratories not at point of care

Results showed Xpert Xpress had higher sensitivity (99%) than ID NOW (79%). Only 1 study of DNA Nudge and 2 of Samba II, both promising

9/12

Implication 1

Some lateral flow tests have strong evidence that they can be used as the first test in people within the first 5 days of symptoms.

Their use could radically improve test and trace with early start to contact tracing

-ve results should be verified with PCR

10/12

Implication 2

Few studies have evaluated lateral flow tests for screening people without symptoms.

They miss many cases - not suitable for test-to-enable or test-to-release

No studies for screening children, repeated test strategies or detecting infectiousness.

11/12

Implication 3

Rapid molecular tests are promising but evaluation in settings where they are intended to be used is required.

12/12