Rapid test for COVID-19 not so perfect

Panbio, a rapid antigen test for SARS-CoV-2, has good specificity but limited sensitivity, as shown in a study. As such, all negative tests have to be confirmed with RT-PCR, especially in the presence of related symptoms, in settings with high community infection rate, and in cases of direct contact with confirmed COVID-19 patients.

Positivity in the rapid test seemed to depend on the days since the onset of symptoms, with an initial sensitivity of 69.1 percent during the first week of symptoms and decreasing considerably with a longer duration of symptoms. [Int J Infect Dis 2021;doi:10.1016/j.ijid.2021.10.027]

Aside from the days since onset of symptoms, viral load (estimated with the cycle threshold (Ct) of RT-PCR) also showed a strong association with positivity in the rapid test. Specifically, Panbio had a sensitivity of 82 percent (95 percent confidence interval [CI], 76–87) in the presence of a Ct ≤25.

The investigators stressed that the test was performed >5 days after the onset of symptoms among symptomatic individuals, which meant that the positivity rate would decrease significantly due to an expected lower viral load. “Nevertheless, even in this group in which sensitivity is considerably lower, a positive test would be highly informative.”

In the study, a total of 1,060 participants (mean age 47 years, 47 percent with a self-reported comorbidity) recruited from eight hospitals in Mexico underwent both rapid antigen and RT-PCR tests. Of these, 915 patients had any respiratory symptoms (mean duration of symptoms, 6 days); the remaining 145, who were usually contacts of a positive relative or coworker, were asymptomatic.

Considering all the test-pairs analysed, 44.5 percent were positive in the RT-PCR test and 25 percent were positive in the rapid test. The overall sensitivity of the Panbio test was 54.2 percent (95 percent CI, 51–57), which the investigators noted was much lower than that reported by the manufacturer in the insert. The corresponding positive likelihood ratio was 35.7, negative likelihood ratio was 0.46, and receiver-operating characteristics (ROC) curve area was 0.77.

The presence of symptoms predicted positivity in the rapid test, with a sensitivity of 58.1 percent in the symptomatic group as opposed to only 26.3 percent among individuals who were asymptomatic but had a positive RT-PCR result.

Meanwhile, the specificity of the rapid test was consistently above 97.8 percent.

A recent Cochrane review, which included 78 studies (20 pre-prints) and 24,087 samples (about one third positive for SARS-CoV-2), confirmed that the sensitivity of antigen tests varied according to the presence or absence of symptoms (72 percent vs 58 percent), time from symptom onset (first vs second week of onset: 78 percent vs 51 percent, respectively), and viral load (Ct ≤25 vs >25: 94.5 percent vs 40.7 percent). [Cochrane Database Syst Rev 2021; 3:CD013705]

The pooled sensitivity of the Panbio test (11 studies) was 75.1 percent among symptomatic patients during the first week of symptoms, but dropped to 58 percent among asymptomatic individuals. Specificity was high at 99.5 percent. [Cochrane Database Syst Rev 2021; 3:CD013705]

“It is noteworthy that, with symptoms of longer duration, the cost-effectiveness of applying point-of-care rapid antigen tests would drop progressively. Thus, given that specificity is high, the greatest clinical advantage of utilizing a rapid antigen test would present when the result is positive, in which case, a confirmatory RT-PCR would not be needed,” according to the investigators.

On the other hand, in the case of a negative rapid antigen test result, confirmation with RT-PCR will be required, they continued.  

“In any case, according to our observations, if a point-of-care rapid antigen test were employed, 69.1 percent of the RT-PCR tests would be avoided during the first week of symptoms, which represents a considerable saving. [This] represents an enormous advantage in settings where decision-making is needed, and where the lack of infrastructure and high costs render it difficult to implement molecular testing,” the investigators pointed out.