Reports from South Korea today show doctors have detected 91 persons previously diagnosed with coronavirus (COVID-19) who tested negative now have retested positive for detectable virus. The tests used for viral clearance and detection are tests for viral RNA (JAMA explainer). Several possible explanations are possible; the previous negative test was a false negative result either because the test failed comletely or that levels of virus were below the sensitivity of the test, the virus has relapsed, or the patient has become reinfected with a variant of the coronavirus.
If this is a testing problem, it's not surprising. Various reports have suggested that some tests may have false negative rates of >50%. The performance characteristics of the S Korea test (Seegene's Allplex 2019-nCoV Assay) were better; manufacturer's data suggested a positive predictive value of 100% and a negative predictive value of 94% against a panel of specimens (Seegene report [PDF]). However, in real-life use, with variation is the quality of specimens, sample handling and transport issues, and lower prevalence of the virus in the test population, false negative cases will likely be greater than 6%.
Coronavirus, and all other viruses that employ RNA-directed RNA synthesis, have low fidelity replication. Spontaneous error rates in replication vary by viral family but can be on the order of 1 in 10000, meaning that for every million virions produced, 100 will have sequence differences. Some of these will be inactivating due to mutations in critical genes; others may alter external epitopes to which the immune response is directed. If virus replication in the host is supressed by the immune system, these mutations would allow relapse. Sequencing of various viral isolates now show a dizzing number of variants circulating (NextStrain database).
By the same token, a group of patients which cleared the initial infection could potentially become reinfected with a mutated strain which is not reocgnized by the immune system or, if the mutant is partially recognized, the reinfection could be attenuated in previously infected individuals, perhaps making many of these reinfections less symptomatic. For the naive patient, though, the infection by a mutated virus could be severe. The strain differences from country to country are interesting. Initial reports from China suggested that the so-called L strain was more aggressive than the intial S strain. At the time of these reports, these were the only 2 idenitified strains. In Italy, the predominant strain initially was the so-called L strain and death rates were about 10% while in S. Korea, the S strain initially predomnated and reported death rates are around 1%. Now, though, it is clear that there are many strains (see above NextStrain data) and it's not clear that any one is worse than the other. Differences in testing regimens, age of population, and aggressive public health measures seem likey to be more important to the difference between Italy and S Korea.
There's a lot unknown about the life cycle of COVID-19, the natural history of infection, and the immune response to the virus.
Stay safe: face mask, hand washing, and distancing.
Posted by Gordon, No Hair Blog, April 10, 2020
© nohair.net and the author