A case report was published in The Lancet Infectious Disease describing a genomic-confirmed SARS-CoV-2 reinfection. A 25-year old male in Nevada was first diagnosed with covid-19 on April 18, 2020. His symptoms included sore throat, cough, headache, nausea, and diarrhea. He was symptomatic until April 27th after which he remained asymptomatic until May 28th. Between April 27th and May 27th he had two negative covid-19 tests. On May 28th, he again had symptoms consistent with covid-19 including low oxygen levels, which had not occurred previously. In other words, it appears that the second infection was worse than the first.
The researchers used two different next-generation sequencing methods to confirm that the viral SARS-CoV-2 strains that infected the patient were genetically distinct. Although it is possible that the patient may have only been infected once and the SARS-CoV-2 virus merely mutated, the genetic discordance between the two samples was much greater than could reasonably be accounted for by such short-term evolutionary mutations.
This case implies that a prior covid-19 infection does not guarantee immunity. Although this is the only confirmed, peer-reviewed case in the literature, it is worrisome that this possibility exists. That said, it is likely that only a small fraction of patients are susceptible to re-infection, and that this will not be a common problem. The absence of many such cases so far is reassuring. Nor is this without precedent. There is a subset of people who can have chickenpox over and over, though most people are immune after one bout.