New review: The effect of sample site and collection procedure on identification of SARS‐CoV‐2 infection
How accurate are different types of sample collection for diagnosing COVID‐19 infection?
Key messages
• When used with RT‐PCR tests (a molecular test that detects genetic material in COVID‐19 using a technique called reverse transcription polymerase chain reaction), self‐collected gargle and deep throat saliva samples have a similar sensitivity compared to trained healthcare worker‐collected nasopharyngeal samples (taken from the back of the throat through the nose) in detecting COVID‐19.
• When used with RT‐PCR, samples collected from the nose, oropharynx (throat via the mouth), oral cavity, and other saliva collection methods are less sensitive for detecting COVID‐19 compared to healthcare worker‐collected nasopharyngeal samples.
• When used with rapid antigen tests (Ag‐RDTs; at‐home/self‐tests), samples collected from the nose have a similar sensitivity to healthcare worker‐collected nasopharyngeal samples in detecting COVID‐19.
Why is improving the diagnosis of COVID‐19 important?
Coronavirus disease (COVID‐19) is caused by infection with severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2). People with suspected COVID‐19 may decide to take a test to know whether they are infected, so that they can receive treatment, and follow recommended guidance to self‐isolate and inform close contacts. Not detecting COVID‐19 when it is present (a false negative result) risks spreading infection and results in missed opportunities for treatment.
Types of sample collection methods for diagnosing COVID‐19?
The type and quality of sample taken for confirmation of COVID‐19 affects the reliability of diagnosis. The most accurate type of sample to diagnose COVID‐19 is that taken by a trained healthcare worker from the back of the throat through the nose (a nasopharyngeal sample). This type of test detects genetic material in the virus using a technique called reverse transcription polymerase chain reaction (RT‐PCR). However, this sample is difficult to obtain correctly, causes discomfort and risks spreading infection if individuals cough or sneeze when the sample is taken. Alternative sample types, particularly those that can be self‐collected using rapid antigen tests (Ag‐RDTs; i.e. self‐tests), may reduce cost and discomfort and improve the safety of sampling. This may, in turn, improve access to and uptake of testing.
What did we want to find out?
We wanted to compare the sensitivity of different sample sites and collection methods in detecting COVID‐19 with molecular tests (RT‐PCR tests) or self‐based tests (Ag‐RDT tests).
What did we do?
We searched for studies that had compared the accuracy of nasopharyngeal samples to any alternative that could be used in patients outside of hospital, including nose (nasal) samples, throat samples taken through the mouth (oropharyngeal), gargle samples and saliva samples. We looked at the use of samples with either RT‐PCR or Ag‐RDTs. We also searched for studies that had compared different methods for taking samples, such as samples collected by a healthcare worker compared to those collected by individuals with no or minimal instructions.
What did we find?
The review included 106 studies with a total of 60,523 participants, of whom 11,045 had COVID‐19 infection. Fifty‐nine per cent of studies were conducted on adults and 79% on symptomatic or mixed symptomatic and asymptomatic participants. Sixty per cent of studies took place in Europe or the USA; just over half (55%) took place in dedicated COVID‐19 testing centres or in outpatient settings.
Main results
With RT‐PCR, on average:
‐ 100% of positive nasopharyngeal samples collected by healthcare workers would also test positive on self‐collected gargle samples or saliva samples (collected by coughing and then spitting (deep throat saliva));
‐ 88% of positive nasopharyngeal samples collected by healthcare workers would also test positive with self‐ or healthcare worker‐collected nose samples;
‐ 87% of positive nasopharyngeal samples collected by healthcare workers would also be detected by saliva self‐collected using spitting, 84% by saliva self‐collected using drooling and 79% by saliva self‐collected by sucking on a swab; and
‐ 83% of positive nasopharyngeal samples collected by healthcare workers would also be detected by self‐ or healthcare worker‐collected oropharyngeal samples.
With Ag‐RDTs, on average:
‐ 100% of positive nasopharyngeal samples collected by healthcare workers would also be detected by self‐collected or healthcare worker‐collected nose samples.
Summary results
The results of these studies indicate that in a group of 1000 people, of whom 230 (23%) have COVID‐19, then:
when used with PCR, compared to healthcare worker‐collected nasopharyngeal samples:
‐ no cases of COVID‐19 would be missed using self‐collected gargle samples (12 less to 5 more) or deep throat saliva samples (2 less to 48 more);
‐ 28 (16 to 39) fewer cases of COVID‐19 infection would be detected using healthcare worker‐ or self‐collected nose sample;
‐ 30 (18 to 41) fewer cases of COVID‐19 infection would be detected by saliva self‐collected using spitting, 37 (12 to 62) fewer by saliva collected by drooling and 48 (12 to 85) fewer by saliva collected by sucking on a swab;
‐ 39 (12 to 67) fewer cases of COVID‐19 infection would be detected by healthcare worker‐ or self‐collected oropharyngeal samples; and
when used with Ag‐RDTs, compared to healthcare worker‐collected nasopharyngeal samples:
no cases of COVID‐19 infection would be missed using healthcare worker‐ or self‐collected nose samples.
What are the limitations of the evidence?
It was often not clear whether included studies deliberately excluded inadequate samples or whether the results of the more accurate nasopharyngeal sample were known when alternative samples were interpreted. This may have resulted in alternative sample types appearing more accurate than they are in practice, decreasing the number of missed cases of COVID‐19 infection.
More than half of studies did not give information about how long participants had had symptoms at the time of sampling. This reduces our confidence in the comparison of different sample types.
Most studies evaluated self‐collected samples by adults with symptoms for use with RT‐PCR; therefore, the findings of this review may not be applicable to asymptomatic individuals or children. For studies conducted with Ag‐RDTs, it is unclear whether sensitivity estimates of nose samples are applicable to home use (self‐collected and self‐interpreted).
How up to date is this review?
The evidence is current to 22 February 2022.