Reinfection is rare for individuals previously infected with SARS-CoV-2. The CDC advises that individuals who have been infected are typically protected from reinfection for at least 3 months. However, this period is almost certainly a lower bound on the true duration of typical, protective immunity. Recent work [here] shows that immunity is likely durable for 6 months and more. At present, the duration of immunity is unknown for Covid-19. Other coronaviruses have immunity of variable duration, with some coronaviruses that cause the common cold having relatively short immunity whereas other coronaviruses, such as the virus that caused the 2009 SARS outbreak, resulting in immunity for many years after infection.

At present, reinfection with Covid-19 appears to be rare. Early reports showed that individuals who had apparently recovered from Covid-19 had tested positive for the virus, which raised concerns that it was possible to be re-infected. However, evidence suggests that even if the virus can be detected after recovery from illness, such positive PCR signals generally do not imply that individuals are shedding infectious virus that can cause disease in others. Recent studies on the immune response to coronavirus infection have shown that most people who have recovered from Covid-19 have `neutralizing' antibodies that prevent the virus from entering human cells. The presence of antibodies and adaptive immune responses make it likely that people are protected from repeated infections. Note that the emergence of more transmissible variants has raised new questions with respect to the potential reinfection of previously infected individuals. We continue to monitor reports as part of efforts to leverage immune status as a means to reduce new infections for all.

Analytical features of serological tests are available here. Although early serological tests had questionable analytical features, there has been significant improvement. Based on the best-performing currently approved assays and assuming a prevalence of 5%, scientists estimate that >90% of individuals who test positive for antibodies for Covid-19 are truly immune (the positive predictive value, PPV) and >99% of individuals who test negative for antibodies for Covid-19 are truly susceptible (negative predictive value, NPV) at the time of testing. If seroprevalence is higher, test results would be more accurate overall.

The FDA's website provides updated information on the accuracy of currently approved serological tests.

Recent work from the groups of Matt Collins (Emory University) and Christopher Heaney (Johns Hopkins University) have led to the development of a saliva-based test for antibodies. The study ‘COVID-19 Serology at Population Scale: SARS-CoV-2-Specific Antibody Responses in Saliva’ was published in the Journal of Clinical Microbiology in December 2020. These saliva-based tests present new opportunities to evaluate the presence of antibodies at population scales without using phlebotomy.