A case of COVID-19 reinfection in the UK

Discussion

The authors believe COVID-19 is the most likely diagnosis for both episodes in Patient A. Differentials would include influenza and other viral illnesses. In the October episode, symptoms commenced 72 hours following significant contact with a confirmed COVID-19 positive case, who was coryzal and febrile at the time of contact. A COVID-19 NAAT also returned positive during the October episode.

The April episode commenced during the first peak of the pandemic, when the individual had been working on COVID-19 positive wards extensively. The April episode also demonstrated hallmark COVID-19 symptoms, including pronounced fevers and fatigue extending over weeks. While the April NAAT returned negative, nasopharyngeal NAAT testing is only evidenced to be 70% sensitive.² Antibody testing is evidenced to be approximately 98.7% specific.³ The combination of classic clinical symptoms, followed by a positive antibody test, leads the authors to believe a false positive antibody test is vanishingly unlikely in this case. Furthermore, both episodes occurred outside of typical flu seasons and Patient A had received flu vaccination for the preceding season.

The second infection episode was symptomatically distinct from the first, with predominance of coryzal symptoms absent in the April episode, reduced fatigue and a faster recovery. The May 2020 positive antibody test would indicate a degree of adaptive immune response following the April 2020 infection.⁴

In this case, the absence of an antibody test, negative or positive, immediately prior to the second infection in Patient A limits the analysis of this case with regards to the presence of residual immunity at the time of reinfection. It is likely that the antibodies pertaining to the initial immune response had either waned completely or provided partial protective immunity to the patient.

An antibody test in the weeks following October 2020 would be of equivocal value in understanding reinfection, since the second infection would likely prompt a discrete antibody response.

There is evidence in the literature that the COVID-19 immune response is variable and patient-specific with respect to the development of antibodies and to antibody persistence in serum over time. In considering the net protective effect of antibodies against a reinfection, the evidence is still inadequate and more research is warranted to clarify the interplay between the roles of adaptive and innate immunity.

The recent Icelandic humoral response study by Gudbjartsson et al⁵ concluded that antibody response was persistent within the 120-day timeframe used. However, there is evidence for modest decline in antibody titres after 120 days in the Icelandic data.

The recent paper by Iyer et al⁶ observed declining antibody titres over 90 days, with ‘median times to seroreversion of 71 and 49 days following symptom onset’.

The findings in this case of a reinfection after 178 days are more in keeping with the conclusions of Iyer et al. However, the Icelandic data do not extend to the longer timeframe of our case, and therefore do not necessarily contradict an antibody decline and subsequent reinfection after 120 days.

The close contact with a positive case preceding the October 2020 reinfection could support an argument for the role of large viral load exposure in reinfection. There is existing evidence relating viral load to severity.⁷

Unfortunately genomic analysis was not available in either the April or the October episode. Without genomic analysis it is difficult to speculate around discrete strains of COVID-19 in this case, but evidence does exist for a variety of strains.⁸

This report adds to a growing body of evidence of COVID-19 reinfection, as described in Table 1.^9–12