While the World Health Organization (WHO) currently recommends nasopharyngeal swabs for COVID-19 testing, there are several different sample and swab types that can be used for diagnostic tests. They include:
Nasal swab (inside the nose)
Nasopharyngeal swab (deep inside the nose, reaching the back of the throat)
Oropharyngeal swab (throat, through the mouth)
Even in people with confirmed COVID-19 infections, the virus is not found equally in these sites, calling into question which is the most accurate.
Which Sample Is Best for COVID-19 Testing?
While it's still too early for a definitive answer to which sample type allows for the most accurate test results, a May 19 pre-print of a meta-analysis of 11 studies found that sputum testing was the most effective, identifying 71% of positive cases.1 Because each of the 757 patients involved in the analysis had a confirmed COVID-19 diagnosis, this means sputum sample testing still missed 29% of cases.
What Is Sputum?
Sputum, or phlegm, is the mucousy substance secreted by cells in the lower airways of the respiratory tract. You can obtain a sputum sample by forcefully coughing into a container.
The meta-analysis showed nasopharyngeal swabs were the second most accurate with a 54% positivity rate. Oropharyngeal swabs were the least accurate with a positivity rate of 43%.
A May 26 study, which was not included in that meta-analysis, found that nasal swabs were almost as good at detecting the virus as nasopharyngeal swabs.2
What This Means For You
It's hard to get large numbers of people to take a diagnostic test that requires a painful sample like a nasopharyngeal swab. Sputum samples—which can be collected by coughing and spitting—are painless and easy to provide. Knowing that research shows sputum sample COVID-19 tests are among the most accurate is further encouragement to consider that option if you need to get tested.
A regulatory analysis published by scientists at the Rutgers Clinical Genomics Laboratory evaluated the effectiveness of their COVID-19 diagnostic test using various samples and swab types. Using 30 samples confirmed to be positive for COVID-19, scientists found 100% of nasopharyngeal swabs confirmed these positive results. They also found that self-collected saliva samples completely agreed with nasopharyngeal test results.3 At least one other study has also found that saliva tests give similar results to nasopharyngeal swabs.4
How Accurate Are COVID-19 Tests?
The accuracy of a COVID-19 test will depend, at least a little, on the specific test being used. In the United States, most direct testing for the virus uses a laboratory technique called rt-PCR, which, in theory, can detect even small amounts of the virus in a sample. However, the sensitivity and specificity will vary both by the specific test used and the type of swab administered.5
Sensitivity vs. Specificity
Sensitivity is the percentage of people who are infected who actually show a positive test result.
Specificity is the percentage of people who are not infected who actually show a negative test result.
Small, early studies of individuals eventually diagnosed with COVID-19 found that 11%6 to 30%7 of them initially incorrectly tested negative, even when they displayed symptoms.
Fortunately, the tests that are currently available in the U.S. should perform better. The NxTAG CoV Extended Panel Assay, which received emergency use authorization from the FDA in March, for example, shows low instances of false positives and false negatives, demonstrating 97.8% sensitivity and 100% specificity.8 This test uses nasopharyngeal swab samples.
A Word From Verywell
Not all COVID-19 swab tests are the same. Research suggests that nasopharyngeal swabs are better than throat swabs. Sputum tests may be even better. Still, if your local testing center is only offering throat swabs, don't walk away. Some information is better than none.
Because no test is perfect, take any negative test results with a grain of salt. If you're feeling sick, isolate yourself as much as possible to avoid infecting others in case your results are wrong.