Discussion

COVID-19 continues to spread among people and every day, new aspects of the disease are discovered. Furthermore, a new strain has recently been discovered in the United Kingdom with a more rapid spread rate (28). Despite the importance of identifying re-infected cases, there are still limited data on the rate of re-infection and the characteristics of the patients who test re-positive for SARS-CoV-2 for the second time.
By gathering existing literature data, our results showed that approximately 4.5% of patients might test positive again after disease resolution and negative PCR. However, in a single study, this number reached as high as 50% (29). Although many of the re-positive cases were discovered upon routine checkups in the first or second week after discharge, a few number of them tested re-positive after the 14-day mark, even as late as 35 days after first recovery (30).
More than 70% of patients show CT improvement upon re-positive PCR, compared with first discharge scans. Zhu et al. (31) reported that at the time of first discharge, at least 50% of the lung lesions were resolved in 54.3% and 29.4% of non-re-positive and re-positive patients, respectively (p = 0.029), and concluded that a slow rate of lung lesion absorption could lead to a higher chance of re-positive PCR.
Only three studies (26, 32, 33) performed contact tracing for their patients and checked whether they infected any other close contacts, which none of them did, suggesting an inactive disease, rather than a second infection. Although, out of eight re-positive patients in Wang et al.’s study, only two had contact with other people in their observation period (a total of seven close contacts), which is a small number, thus further investigations are required to determine the transmissibility of SARS-CoV-2 in re-positive cases. These results are consistent with Wolfel et al.’s findings, which stated that live virus could not be extracted from SARS-CoV-2 samples eight days after onset of symptoms, despite high levels of viral load (34).
It is a matter of debate whether the presence of anti-SARS-CoV-2 antibodies can immunize a patient from a second infection. Based on the results of our study, many patients had positive antibody profiles and yet, tested positive again. Non-re-positive patients had significantly higher rates of positive IgG titers, while IgM positivity was alike between re-positive and non-re-positive groups. As a result, we believe that a re-positive PCR after the first recovery is less likely in a patient with a positive IgG profile, whereas a positive IgM titer does not affect the likelihood of testing positive a second time. Nonetheless, there are some important limits to this analysis. Different studies had different methods for antibody measurement (e.g., time from disease resolution to sample collection) and reporting (e.g., definition of a “positive” antibody titer). Therefore, further studies are necessary to establish a correlation between antibody levels and the probability of a second positive nucleic acid test.
Although RT-PCR is the gold standard of SARS-CoV-2 detection (35), it is not completely sensitive; sensitivity of PCR tests could range from 98% down to 60.2% (36). Improper sampling is one of the reasons that may cause false negatives (37). False-negative results not only could account for underestimated rates of re-infection, but could also lead to a false assumption of recovery. As a result, re-positive PCR may not necessarily indicate a second infection, but rather due to the same virus that had not been cleared at all in the first place. Since samples for PCR could be obtained from nasopharyngeal swabs, oral swabs, sputum, or fecal swab, some experts suggest sampling from as many sites as possible to reduce the probability of false negatives (29). Nevertheless, it is assumed that PCR from sputum specimen has the highest sensitivity for SARS-CoV-2 detection, followed by nasal swab and oral swab (37, 38). On the other hand, although the presence of virus might be missed due to false negatives, not all positive tests indicate SARS-CoV-2 infection, as contamination or a cross-reaction with another organism’s genome could falsely turn a test positive (39). Therefore, a second positive PCR after testing negative might be due to false negative or false positive results of PCR and does not necessarily indicate a second infection. However, in the study by Li et al., patients had negative RT-PCR at least four times with one-day intervals between testings, which highly reduced the chance of false-negative result (37). Li et al. stated that there is a chance that the virus might have replicated and increased its viral load due to immunologic alterations throughout the disease course (37). Prolonged intermittent viral shedding is another explanation for negative PCR periods, as Wong et al. (26) indicate in their study, and has been previously described in SARS (40). Ye et al. also believe that second positive tests for SARS-CoV-2 in their study were subsequent to virus reactivation, rather than a second acquired infection (41). Immunosuppression due to corticosteroid administrations might also be a contributing factor (41). Some patients might not be able to fully clear the virus from their body and still test positive, even two to three months after infection (42). Also, remnants of SARS-CoV-2 have been observed in pulmonary tissue of a ready-for-discharge patient despite negative PCR (43), which might point to another source for virus re-emergence.
Being infected a second time after full recovery usually requires contact with another infected patient. However, most of the re-positive patients in our gathered studies tested positive in their post-discharge observation or quarantine period, and therefore a transmitted infection is unlikely. Nevertheless, an analysis of 8922 cases in South Korea (30) after the termination of their quarantine period (which requires two negative PCR tests with at least 24 hours in between) revealed that 292 (3.3%) patients tested positive again. The study challenges the re-activation or second infection argument, stating that re-positive tests are mainly due to technical limitations and the inability of current testing methods to differ viral nucleic acid remnants of dead viruses from an active virus. The grounds on which Kang makes these claims are acuity of SARS-CoV-2 and the fact that it is not a chronic infection, high chances of immunity after the first infection based on previous data from other coronaviruses, absence of documented disease transmission from re-positive patients to other individuals, and negative in-vitro viral activity tests of re-positive samples (30).
Taking clinical features into account could help in identifying a new or re-activated infection from false positives or previous false negatives. Approximately one-third of the patients express symptoms upon re-positive. The symptoms alone do not indicate a second infection, as they have to be compared with the previous episode of infection. Worsened symptoms, signs, and health status is more likely to be a result of re-infection. Wang et al. (32) retrospectively studied 131 discharged COVID-19 patients for four weeks, of which 8 tested re-positive during this period. However, only one was deemed a recurrent infection, as the positive PCR test was accompanied by deteriorated CT as well as fever. Qiao et al. also ruled out false-positive tests resulting from viral nucleic residues, due to the presence of symptoms and imaging findings (27).
Furthermore, a similar rate of IgM positivity in re-positive and non-re-positive cases increases the chance of a testing error and makes reinfection unlikely (42). Additionally, inflammatory markers in Yuan et al.’s (44) study were within the normal range in most re-positive patients, further reducing the probability of active disease.
Chinese guidelines require two negative PCR tests with a 24-hour interval before discharging a patient. Considering the probability of a re-positive test even after two negative tests, some experts believe that the currently practiced guidelines are not practical enough for recognizing clearance of the virus from the body (25). Increasing testing intervals and utilizing other diagnostic procedures such as antibody testing could assist in achieving this goal (25). Hao et al. demonstrated that when the number of negative PCRs increases from two to three, chances of a future re-positive test reduces by approximately 70% (16.4% vs 4.8%, respectively).
Most of the above-mentioned studies do not believe that re-positives in their population are second acquired infections, but rather consequences of testing inaccuracies or re-activation of the previous disease. The main factors contributing to these views are the few numbers of days between initial negative/discharge and re-positive PCR, no transmission to close contacts in re-positive patients, and testing positive without an identifiable infection source. On the other hand, numerous case reports and case series have been published, describing patients who have tested positive many days after their first infection, making false test results and virus re-emergence unlikely. In a recent case report by Tillett et al. (45), a 25-year-old man was infected by a genetically different strain of SARS-CoV-2 almost two months after his first episode of infection. Additionally, the case also had major symptoms such as self-reported fever and dyspnea. This is the first documented and confirmed re-infection of a patient with SARS-CoV-2 in North America, which also raises doubts regarding immunity after an episode of COVID-19.
In another case presented by Wang (46), PCR of the reported case turned positive three more times, after initial discharge. The number of days in between discharges and the next positive PCR were 13, 37, and 25. The patient did not infect his close contacts, and the study does not report any symptoms upon re-positives. Moreover, ground-glass opacities were present on the patients’ chest CT until the day before his 3rd discharge from the hospital. Total antibody titers were positive for SARS-CoV-2 since his second admission, further strengthening the idea that a positive antibody titer does not prevent a second positive PCR test, however, it might lead to a clinically suppressed relapse as the patient had no symptoms.
Lafaie et al. (47) provide another evidence for the role of antibodies in COVID-19 re-positive cases, where they report an 84-year-old female who presented with hypothermia and respiratory symptoms 41 days after her first positive test. She was tested positive again 4 days after the beginning of her second episode of symptoms, with relatively low cycle thresholds (17.5 and 18.1), and a worsened chest CT. Neutralizing antibodies were negative until one week after symptom onset, after which they became weakly positive. The patient, unfortunately, died 13 days after symptom onset.
On the other hand, in a case series (48) of 11 patients with re-positive PCR at least 21 days after initial symptom presentation, all cases experienced symptoms (for a median of 10 days) after a symptom-free interval, and all had COVID-19 matching CT scans. Nevertheless, six out of nine patients who underwent serologic antibody testing had positive or weakly positive antibody levels, raising doubt about the role of antibodies in disease outcome.
Multiple other case series and case reports are also available in which re-positive patients present worsened symptoms and/or chest imaging, suggesting a relapsing disease or re-infection (49-57). Some of the patients experienced relapse symptoms after drug cessation (58, 59).