SARS-CoV-2 Reinfection Is a New Challenge for the Effectiveness of Global Vaccination Campaign: A Systematic Review of Cases Reported in Literature

Reinfection with SARS-CoV-2 seems to be a rare phenomenon. The objective of this study is to carry out a systematic search of literature on the SARS-CoV-2 reinfection in order to understand the success of the global vaccine campaigns. A systematic search was performed. Inclusion criteria included a positive RT-PCR test of more than 90 days after the initial test and the confirmed recovery or a positive RT-PCR test of more than 45 days after the initial test that is accompanied by compatible symptoms or epidemiological exposure, naturally after the confirmed recovery. Only 117 articles were included in the final review with 260 confirmed cases. The severity of the reinfection episode was more severe in 92/260 (35.3%) with death only in 14 cases. The observation that many reinfection cases were less severe than initial cases is interesting because it may suggest partial protection from disease. Another interesting line of data is the detection of different clades or lineages by genome sequencing between initial infection and reinfection in 52/260 cases (20%). The findings are useful and contribute towards the role of vaccination in response to the COVID-19 infections. Due to the reinfection cases with SARS-CoV-2, it is evident that the level of immunity is not 100% for all individuals. These data highlight how it is necessary to continue to observe all the prescriptions recently indicated in the literature in order to avoid new contagion for all people after healing from COVID-19 or becoming asymptomatic positive.


Introduction
The novel coronavirus (SARS-CoV-2) outbreak since December 2019 has continued to exhibit devastating consequences, and was declared as a pandemic by the World Health Organization in early 2020 [1][2][3]. To date, as of 17 October 2021, 240,421,359 infections have been confirmed, with 4,895,034 deaths [4]. In many countries, the vaccination campaign has started with the use of various vaccines recently put on the market and the total number of vaccine doses administered is 6,609,632,994. However, a new problem is emerging with regard to the evolution of the behavior of SARS-CoV-2: the possibility of reinfection of healed subjects after the first infection. On 25 August 2020, the first case of reinfection of SARS-CoV-2 was reported in international literature [5]. This event pointed out that infection by this virus does not uniformly confer protective immunity to all infected individuals [6]. Therefore, several critical questions are intriguing the researchers. Is SARS-CoV-2 reinfection a widespread phenomenon or is it limited to few subjects with immune deficits or specific comorbidities [6]? Can this phenomenon be due to a too weak, too short, or too narrow natural immune response to SARS-CoV-2, that is unable to protect Moderate evidence >2 nucleotide differences per month * in consensus between sequences that meet quality metrics above, ideally coupled with other evidence of actual infection (e.g., high viral titers in each sample or positive for subgenomic mRNA, and culture) Poor evidence but possible ≤2 nucleotide differences per month * in consensus between sequences that meet quality metrics above or >2 nucleotide differences per month * in consensus between sequences that do not meet quality metrics above, ideally coupled with other evidence of actual infection (e.g., high viral titers in each sample or positive for subgenomic mRNA, and culture)

Study Selection
Three independent investigators retrieved the studies that were the most relevant by titles and abstracts (ELM, LLM, MA). Subsequently, the full text of the retrieved papers was reviewed, and the most relevant papers were chosen according to the eligibility criteria. Then, we extracted the relevant data and organized them in tables. The original papers that were peer-reviewed and published in English and fulfilled the eligibility criteria were included in the final report, together with two works not reviewed at the time of preparation of this report [30,31].
The following inclusion criteria was used: a positive RT-PCR test carried out more than 90 days after the initial test in healed patients or a positive RT-PCR test carried out more than 45 days after the initial test that is accompanied by compatible symptoms or epidemiological exposure, after confirmed healing. This criteria corresponds to the CDC protocol designed to identify cases of real SARS-CoV-2 reinfection (Table 1) [32].
We considered the exclusion criteria for this study as follows: (1) papers conveying non-human studies including in vitro observations or articles focusing on animal experiments; (2) papers in which their full text were out of access; (3) any suspicious and duplicated results in the databases.

Data Extraction
After summarizing, we transferred the information of the authors, type of article (e.g., case reports), publication date, country of origin, age, gender, and clinical symptoms to a data extraction sheet. Three independent investigators collected this information and subsequently organized them in the tables. Finally, to ensure no duplications or overlap existed in the content, all the selected articles were cross-checked by other authors.

Quality and Risk of Bias Assessment
As aforementioned, we applied the PRISMA checklist to ensure the quality and reliability of selected articles. Two independent researchers evaluated the consistency and quality of the articles and the risk of bias. In either case of discrepancy in viewpoints, a third independent researcher resolved the issue. The full text of selected articles was read, and the key findings were extracted.
Included studies underwent quality check and risk of bias assessment. This qualitative analysis was performed according Murad's quality checklist of case series and case report [33]. As reported, the scale consists of four parameters, to evaluate the (a) patient selection; (b) exposure ascertainment; (c) causality; (d) reporting. Each section contains one to four question to be addressed. As it is suggested we performed an overall judgement about methodological quality since questions 4, 5 and 6 are mostly relevant to cases of adverse drug events. Each requested field will be considered as adequate, inadequate or not evaluable. The table showing this tool for evaluating the methodological quality of case reports and case series, is reported in the original manuscript [33].

Results
In this study, 117 documents were identified using the systematic search strategy. After a primary review of 2201 retrieved articles, 379 duplicates were removed, and the title and abstract of the remaining 1822 resources were reviewed. After applying the selection

Results
In this study, 117 documents were identified using the systematic search stra After a primary review of 2201 retrieved articles, 379 duplicates were removed, an title and abstract of the remaining 1822 resources were reviewed. After applying th lection criteria, only 117 articles met the inclusion criteria and were included in the review ( Figure 1). Therefore, the cases confirmed according to these parameters wer ( Table 2).   Table 2. Cases of SARS-CoV2 reinfection in the international literature (all cases were again positive for SARS-CoV-2 after complete symptomatic recovery in addition to negative RT-PCR test for SARS-CoV-2, according to WHO recommendations [34]).

Demographic and Clinical Features of Reinfection Cases
Reinfection occurred across the world: 1 case from Austria, 1 from Bahrain, 5 from Bangladesh, 2 from Belgium, 31 from Brazil, 3 from China including 1 from Hong Kong, 2 from Colombia, 28 from the Czech Republic, 1 from Denmark, 2 from Ecuador, 10 from France, 2 from Gambia, 1 from Germany, 24 from India, 31 from Iran, 12 from Iraq, 1 from Israel, 5 from Italy, 1 from Japan, 1 from Lebanon, 1 from Libya, 4 from Mexico, 5 from Pakistan, 1 from Panama, 1 from Peru, 1 from Portugal, 6 from Qatar, 1 from South Korea, 1 from Switzerland, 8 from Saudi Arabia, 1 from South Africa, 9 from Spain, 1 from the Netherlands, 4 from Turkey, 9 from the United Kingdom, 42 from the United States of America (Figure 2).  Reinfection occurred across the world: 1 case from Austria, 1 from Bahrain, 5 from  Bangladesh, 2 from Belgium, 31 from Brazil, 3 from China including 1 from Hong Kong,  2 from Colombia, 28 from the Czech Republic, 1 from Denmark, 2 from Ecuador, 10 from  France, 2 from Gambia, 1 from Germany, 24 from India, 31 from Iran, 12 from Iraq, 1 from  Israel, 5 from Italy, 1 from Japan, 1 from Lebanon, 1 from Libya, 4 from Mexico, 5 from  Pakistan, 1 from Panama, 1 from Peru, 1 from Portugal, 6 from Qatar, 1 from South Korea, 1 from Switzerland, 8 from Saudi Arabia, 1 from South Africa, 9 from Spain, 1 from the Netherlands, 4 from Turkey, 9 from the United Kingdom, 42 from the United States of America (Figure 2).    The main risk groups were healthcare workers and patients with comorbidities. In total, 66/260 cases (2.3%) occurred among high risk groups, including healthcare workers (HCWs), doctors, students and nursing resident. A total of 91 cases (35%) occurred among patients with comorbidities, 48 in men and 38 in woman ( Figure 5). The evolution of the reinfection episode itself was more severe in 92/260 (35.3%) cases with the death only in 14/260 cases (5.3%), 7/260 male (2.65%) and 7/260 females (2.65%); 8 of these had a neoplastic immune system diseases, or transplant or other important comorbidities and 3 were over 80 years old ( Figure 6).  The main risk groups were healthcare workers and patients with comorbidities. In total, 66/260 cases (2.3%) occurred among high risk groups, including healthcare workers (HCWs), doctors, students and nursing resident. A total of 91 cases (35%) occurred among patients with comorbidities, 48 in men and 38 in woman ( Figure 5).  The main risk groups were healthcare workers and patients with comorbidities. In total, 66/260 cases (2.3%) occurred among high risk groups, including healthcare workers (HCWs), doctors, students and nursing resident. A total of 91 cases (35%) occurred among patients with comorbidities, 48 in men and 38 in woman ( Figure 5). The evolution of the reinfection episode itself was more severe in 92/260 (35.3%) cases with the death only in 14/260 cases (5.3%), 7/260 male (2.65%) and 7/260 females (2.65%); 8 of these had a neoplastic immune system diseases, or transplant or other important comorbidities and 3 were over 80 years old ( Figure 6).
The evolution of the reinfection episode itself was more severe in 92/260 (35.3%) cases with the death only in 14/260 cases (5.3%), 7/260 male (2.65%) and 7/260 females (2.65%); 8 of these had a neoplastic immune system diseases, or transplant or other important comorbidities and 3 were over 80 years old ( Figure 6). Notably, reinfection occurred among patients whose initial infections were both asymptomatic/mild, 80% (207/260), and moderate/severe, 20% (53/260). The demonstration that moderate/severe initial infections do not necessarily provide enhanced protection against reinfection is important because patients with more severe infection have been found to have higher neutralizing antibody titers, which may be expected to confer protection. Additionally of note, the severity of the reinfection episode itself was less in 21/53 cases (40%). The observation that many reinfection cases were less severe than initial cases is interesting because it may suggest partial protection from disease [152] and argues against antibody-dependent immune enhancement, which can be seen with other viral pathogens. In the absence of routine surveillance, we would have expected a bias toward detection of symptomatic reinfection, underscoring the importance of prospective screening.
Another interesting datapoint is the detection of different clades or lineages detected by genome sequencing between initial infection and reinfection in 52/260 cases (20%). The current gold standard for identifying reinfection is detection of a distinct virus by genome sequencing. Detection of reinfection is most straightforward when viruses belong to a different clade or lineage, as this provides clear evidence of infection by a different virus [6]. Although reinfection is most apparent when viruses are different enough to distinguish by genome sequencing, it remains unclear whether these viral genomic differences play a causative role in reinfection. That is, does reinfection occur when viral genomic differences permit escape from an existing, but narrow, immune response to the initial infection? Answering this question will require detailed mapping of the relationship between virus substitutions and immune escape.

Quality and Risk of Bias Assessment
Briefly, only 14 studies fulfilled the quality checklist. "Selection-Does the patient(s) represent(s) the whole experience of the investigator (center) or is the selection method unclear to the extent that other patients with similar presentation may not have been reported?" checklist resulted unclear in most of the studies, because the patient selection method was unclear. In general, overall quality was satisfactory in all included studies.

Discussion
Since the first cases, a question has haunted all researchers: can a patient recovered from COVID-19 get sick again? The first confirmed case of reinfection occurred in a 33- Notably, reinfection occurred among patients whose initial infections were both asymptomatic/mild, 80% (207/260), and moderate/severe, 20% (53/260). The demonstration that moderate/severe initial infections do not necessarily provide enhanced protection against reinfection is important because patients with more severe infection have been found to have higher neutralizing antibody titers, which may be expected to confer protection. Additionally of note, the severity of the reinfection episode itself was less in 21/53 cases (40%). The observation that many reinfection cases were less severe than initial cases is interesting because it may suggest partial protection from disease [152] and argues against antibody-dependent immune enhancement, which can be seen with other viral pathogens. In the absence of routine surveillance, we would have expected a bias toward detection of symptomatic reinfection, underscoring the importance of prospective screening.
Another interesting datapoint is the detection of different clades or lineages detected by genome sequencing between initial infection and reinfection in 52/260 cases (20%). The current gold standard for identifying reinfection is detection of a distinct virus by genome sequencing. Detection of reinfection is most straightforward when viruses belong to a different clade or lineage, as this provides clear evidence of infection by a different virus [6]. Although reinfection is most apparent when viruses are different enough to distinguish by genome sequencing, it remains unclear whether these viral genomic differences play a causative role in reinfection. That is, does reinfection occur when viral genomic differences permit escape from an existing, but narrow, immune response to the initial infection? Answering this question will require detailed mapping of the relationship between virus substitutions and immune escape.

Quality and Risk of Bias Assessment
Briefly, only 14 studies fulfilled the quality checklist. "Selection-Does the patient(s) represent(s) the whole experience of the investigator (center) or is the selection method unclear to the extent that other patients with similar presentation may not have been reported?" checklist resulted unclear in most of the studies, because the patient selection method was unclear. In general, overall quality was satisfactory in all included studies.

Discussion
Since the first cases, a question has haunted all researchers: can a patient recovered from COVID-19 get sick again? The first confirmed case of reinfection occurred in a 33year-old Caucasian man of Hong Kong, that was admitted to the hospital for COVID-19 on 23 March 2020 [5]. After two negative tests by RT-PCR on days 21 and 22 he was discharged from the hospital and resumed his usual work [5]. Serological controls after the first infection showed that he did not produce virus neutralizing antibodies [139]. On 15 August 2020 after a 1-week trip in Spain, the patient returned to Hong Kong and was submitted to a collection of a deep throat saliva sample for RT-PCR as border surveillance and resulted positive [5]. The patient was asymptomatic until the new negative test. The viruses from the first and the second infection were phylogenetically distinct and the virus of first infection had a truncation in the 58AA open reading frame 8 gene, that could be responsible immune evasion [138]. However T cells and mucosal immunity might have played an important role in resolving the second infection, even if there was the absence of primary neutralizing antibodies [139].
In October 2020, Tillett et al. reported the first confirmed case of SARS-CoV-2 reinfection in the USA [27]. A 25-year-old man from Nevada, without known immune disorders, had PCR-confirmed SARS-CoV-2 infection in April, 2020 (cycle threshold (Ct) value 35·24; specimen A) [27]. He recovered in quarantine, testing negative by RT-PCR at two consecutive timepoints thereafter [27]. However, 48 days after the initial test, the patient tested positive again by RT-PCR (Ct value 35·31; specimen B) [27]. Viral genome sequencing showed that both specimens A and B belonged to clade 20C, a predominant clade seen in northern Nevada [27]. The genome sequences of isolates from the first infection (specimen A) and reinfection (specimen B) differed significantly, making the chance of the virus being from the same infection very small [27]. The particularity of this report is that SARS-CoV-2 reinfection resulted in worse disease than the first infection, requiring oxygen support and hospitalization [27]. The patient had positive antibodies after the reinfection, but whether he had pre-existing antibody after the first infection is unknown [27]. Both cases reported from Nevada and Hong Kong seem to confirm the possibility that the reinfections are due to a different variant of SARS-CoV-2.
The first important question to be answered is: are all cases reported in the literature as reinfection by SARS-COV-2 true reinfections?
A distinction must be made between true reinfection, relapsed infection, recurrence of positive (re-positive) nucleic acid detection [17,153], in fact one of the features of SARS-CoV-2 infection is prolonged virus shedding. Several studies reported persistent or recurrent elimination of viral RNA in nasopharyngeal samples starting from first contact with a positive subject [18][19][20]. Several explanations can exist in order to explain this phenomenon without it being a true reinfection. One possible explanation for testing positive after a previously negative result could be that the negative results after patient recovery were really false-negative results [154]. Literature reported that false-negative rates can be as high as 30% for SARS-CoV-2 PCR testing [155]. However, actually the KCDC (Korean Control Disease Center) determined recovery as two separate negative PCR results within 24 h [156]. In this way, patients positive after having two consecutive negative results would be positive for an increase in viral genetic material due to reinfection [156]. It is difficult to have two previous consecutive false-negative results [156]. Another possible explanation could be the contamination of the samples, but most testing centers are requiring testers to change personal protective equipment (e.g., gloves, gowns and masks) [156]. However, surely one of the main points to consider is the basis of PCR testing: the test is able to amplify nucleic acid in the sample, not fully active viral particles. The genetic material (RNA and DNA) left behind degrades over time [157]. Thus, positive PCR results after recovery may not necessarily signify reinfection, but rather the presence of leftover genetic material from previously active infection [156]. Therefore, a patient who retests positive for virus might not necessarily be experiencing a second, new SARS-CoV-2 infection [158]. True reinfection has criteria that must be considered, including isolation of the complete genome of the virus (and not just genomic fragments) in the second episode, identification of two different virus strains in two episodes of infection based on phylogenetic analysis; proof of virus infectivity in the second episode by virus isolation and evaluation of its cytopathic effect in cell culture; investigation of immune responses and their comparison in two episodes; epidemiologic data such as re-exposure history to COVID-19 patient in the second event and timing between episodes, with a longer time interval between two episodes favoring the reinfection hypothesis [17,159]. To date, positive retesting more than 83 days after the first positive test, along with other criteria, favors confirmation of reinfection, even if Turner et al. recently reported a patient with prolonged viral RNA shedding lasting 87 days after the initial positive clinical PCR test and 97 days after the onset of symptoms, probably due to the poor CD8+ T cell response during the first three months of his illness [160]. In addition to the abovementioned reasons, the disease clinical data are also useful in confirming the second episode, although the second episode may be asymptomatic [17]. A time interval where the patient is free of clinical signs between the two episodes is also necessary. In conclusion, only cases with clinical symptoms and RT-PCR positivity after negative tests following recovery from COVID-19 could be considered true SARS-CoV-2 reinfections. Recently Raveendran et al. suggested an interesting approach in order to individuate the reasons for a persistent RT-PCR positivity (Figure 7) [161]. According to this flow chart it is possible to individuate cases of persistent RT-PCR positivity due to reinfection or to presence of dead viral fragment or to persistent viral shedding. netic analysis; proof of virus infectivity in the second episode by virus isolation and evaluation of its cytopathic effect in cell culture; investigation of immune responses and their comparison in two episodes; epidemiologic data such as re-exposure history to COVID-19 patient in the second event and timing between episodes, with a longer time interval between two episodes favoring the reinfection hypothesis [17,159]. To date, positive retesting more than 83 days after the first positive test, along with other criteria, favors confirmation of reinfection, even if Turner et al. recently reported a patient with prolonged viral RNA shedding lasting 87 days after the initial positive clinical PCR test and 97 days after the onset of symptoms, probably due to the poor CD8+ T cell response during the first three months of his illness [160]. In addition to the abovementioned reasons, the disease clinical data are also useful in confirming the second episode, although the second episode may be asymptomatic [17]. A time interval where the patient is free of clinical signs between the two episodes is also necessary. In conclusion, only cases with clinical symptoms and RT-PCR positivity after negative tests following recovery from COVID-19 could be considered true SARS-CoV-2 reinfections. Recently Raveendran et al. suggested an interesting approach in order to individuate the reasons for a persistent RT-PCR positivity ( Figure 7) [161]. According to this flow chart it is possible to individuate cases of persistent RT-PCR positivity due to reinfection or to presence of dead viral fragment or to persistent viral shedding.  The second important question to be answered is: can SARS-CoV-2 re-infect a patient after recovery?
When any unwanted virus comes into contact with our body, also in the case of SARS-CoV-2 infection, most patients are able to develop specific antibodies neutralizing the spike proteins of this virus [5]. A recent study of Pilz et al. pointed out that the relatively low tentative reinfection rate (40 cases in 14,840 COVID-19 survivors of first wave-0.27%) ensures a good protection after natural infection for SARS-CoV-2 [162]. However there are three main mechanisms for reinfection: the immune response can be ineffective, strain-specific, or short-lived [156].
Monoclonal antibodies formed against the SARS-CoV-2 virus target the Spike (S) glycoprotein component, the receptor-binding domain of the virion [156]. SARS-CoV-2, however, has been shown to develop "escape mutants," or alterations, in the epitope of the S protein that contribute to host tropism and viral virulence [156]. Sui et al. reported that major variations exist in the S protein at positions 360, 479, and 487 [163]. They found that altering 1-2 amino acids at those positions led previously efficacious neutralizing antibodies to SARS-CoV-2 to a 20-50% reduction in binding capacity [163]. Theoretically, if SARS-CoV-2 is also able to form "escape mutants" in the S protein, IgG antibodies formed in patients may be less ineffective, though not completely, in neutralizing the virus [156]. This could mean that patients remain resistant to SARS-CoV-2 infection even after mutations, with antibody responses that are 50-80% efficacious [156].
Another possibility that could allow the reinfection of a patient is the duration of the body immune response [156]. Recent findings suggested that protective immunity does not occur in all infected individuals [164], supporting the possibility of reinfection [103], even if 93% of the infected produce neutralizing antibodies [165]. Their function is to prevent the virus from entering cells between 6 and 20 days after infection [166] with this mechanism: after the infection, B lymphocytes are activated and produce IgM, IgG and IgA antibodies. A subset of them (IgG and IgA) then manage to make the new viral particles harmless. The neutralizing antibodies, in turn, are accompanied by the activation of killer cells (T lymphocytes), specialized in recognizing and destroying the virus [167].
Seroconversion of IgM and IgG antibodies occurs the first week after onset of symptoms, seroconversion rates rise until the fourth week and decline thereafter, by the seventh week IgM antibodies are not detected in most cases, even if some reports showed IgM antibodies to persist for up to 8 months post-COVID-19 [168], whereas IgG antibodies persist longer for a period of time yet unknown [169]. Immunoglobulins alone are not truly sufficient to confer long-term immunity to coronavirus [156]. CD4+ T-cells and memory CD8+ T-cells with their products, such as effector cytokines and IFN-γ, are important in providing protection from coronavirus [170]. In fact, when the infection is over, in the following weeks or months, the antibodies drop: the virus is no longer there, they are no longer needed. However, the memory cells remain in the body, ready to intervene in case of need. All the studies so far show that a long-lasting immune response occurs. A very recent study carried out in collaboration between the Policlinico San Matteo in Pavia and the Karolinska Institute in Stockholm quantifies this "time" more precisely: memory cells persist for at least 6-8 months after infection [171]. Considering that the disease erupted just under a year ago, this is the maximum observation time possible to date, but it could be much longer [171]. Previous studies showed that virus-specific memory CD8+ T-cells were found to persist for up to 6 years after a SARS associated coronavirus infection, but memory B-cells and accompanying antibodies were undetectable at that time [172]. However Vetter et al. hypothesized that reinfection can be due to a loss of protection elicited after the first episode for a progressive reduction of protective antibody titers [144,173].
We can conclude that antibody formation and longevity of immunity in a subject could be dependent by the strain of virus, its severity and age of subject [174].
Khoshkam et al. tried to classify the recovered and immunized subjects in four categories: (1) Infected cases with very mild symptoms or asymptomatic without any humoral immune response or elicited memory. (2) Infected cases with mild to moderate symptoms with low humoral immunity and low cellular immunity.
(3) Infected cases with moderate or severe symptoms with highly activated humoral immunity and elicited memory. (4) Infected cases with moderate or severe symptoms with highly activated humoral immunity and low cellular immunity [175].
They hypothesized that reinfection may happen in groups 1 and 2, which may also develop the severe disease in the future due to the absence or low levels of acquired immunity [175]. Individuals in group 3 are more protective against further exposures and they may show long-term immunity since they develop increased elicited memory in defense of SARS-CoV-2 [175]. The last group may show rapid response against reinfection; they may not be safe for longer periods because of the non-imprinted memory of immunity [175].
The question to be solved is whether these antibodies can neutralize each SARS-CoV-2 clade and guarantee immunity to subsequent contact. Reinfection from SARS-CoV-2 with a genetically distinct strain of SARS-CoV-2 is, in theory, possible in patients immediately after recovery from COVID-19. SARS-CoV-2 infection may not confer immunity against a different SARS-CoV-2 strain, so more research is needed. SARS-CoV-2, even if it is a virus similar to that of the flu, seems to have a more stable genome and the response that the immune system generates is towards several fragments of the viral proteins and not just one. In fact, the mutations observed so far (and, perhaps, also the new English variant, at least until proven otherwise) are not associated with a change in the severity of the disease.
The new variants are accumulating mutations in different spike domains, such as the alpha variant or B.1.1.7 lineage (also known as 501Y.V1 or VOC202012/01), the beta variant or B.1.351 lineage (501Y.V2), the gamma variant or P.1 lineage (501Y.V3) and the delta variant or B.1.617.2 lineage [176]. All these variants have cumulated at least nine non-synonymous mutations/deletions throughout the Spike coding region. For example, the case reported by Harrington et al. showed that anti-SARS-CoV-2 antibodies were still present shortly before onset of reinfection, with no evidence of antibody waning [82]. This may raise some concerns about immune evasion by the alpha variant, which is a concern with the high number of spike region mutations seen. However, the study has a bias: there were no assays for SARS-CoV-2 antibodies recognizing spike antigen in the second reinfection, while the tested antibodies recognized "N" antigen, so it is difficult to point out an evident role of antibodies in the reinfection. The 501Y.V2 variant, or beta variant, is characterized by eight mutations in the spike protein-coding sequences that can improve its ability to transmission [151]. The case reported by Zucman et al. showed that beta variant can be more aggressive than non-VOC SARS-CoV-2 [151]. The last, the delta variant, is characterized by P681R and L452R mutations that can help the delta variant spread. For all these reasons it is necessary to investigate urgently the possibility of these new variants to escape the vaccine action. The immune responses generated by mRNA and adenoviral vector-based vaccines are restricted to the Spike glycoprotein, so new variants with big antigenic drift could reduce their efficiency and determine a growing number of reinfections.
Another possibility that could allow the reinfection of a patient is the reactivation of dormant virus which is commonly seen in immunosuppressed patients with some viruses, such as Epstein Barr, cytomegalovirus and herpes groups [90], but it is necessary to sequence viral genome for differential diagnosis between viral reactivation or reinfection with a different strain.
For all these reasons, it is important to identify cases of reinfection to understand if the "immunological memory" affects the symptoms during a second infection, a crucial fact, in particular, to predict the effectiveness of the vaccination campaign. If in the second time the symptoms are generally reduced, as in the Hong Kong [5], Belgium and the Netherlands [29] patients, this suggests that the immune system is responding as it should. However, if symptoms are consistently more severe during a second COVID-19 attack, as in the case of the Nevada [27] or Ecuador [28] subjects, it may be that the immune system makes matters worse. The mechanisms that could account for a more severe secondary infection can only be speculated. First, a very high dose of virus might have led to the second instance of infection and induced more severe disease [177]. Second, it is possible that reinfection was caused by a more virulent variant of the virus, or more virulent in this patient's context [27]. Third, a mechanism of antibody-dependent enhancement might be the cause, a means by which specific Fc-bearing immune cells become infected with virus by binding to specific antibodies [27]. In fact, the clinical course of some severe COVID-19 cases has been worsened by abnormal immune responses that damage healthy tissue. Patients who experienced that problem during a first infection may have immune cells that are induced to respond disproportionately the second time too. Sometimes antibodies produced in response to SARS-CoV-2 can facilitate the virus during a second infection rather than fight it [178][179][180][181][182][183][184]. The phenomenon [185][186][187][188][189] is rare, but researchers have found worrying signs of it while trying to develop vaccines against the coronaviruses responsible for severe acute respiratory syndrome and Middle East respiratory syndrome [190] and against SARS-CoV-2 [191][192][193][194].
As researchers accumulate more examples of reinfection, the situation should become clearer. Depending on the criteria used, rates of reinfection can vary widely [195]. There are some reports about retrospective observational study such as that of Pilz et al. that reported 40 cases of tentative reinfection in Austria, but these data are limited by the lack of detailed clinical characteristics [162]. For this reason, in November 2020 the Centers for Disease Control and Prevention pointed out the following criteria to define reinfection with SARS-CoV-2: detection of SARS-CoV-2 RNA (with Ct values < 33 if detected by RT-PCR) >90 days after the first detection of viral RNA whether or not symptoms were present and paired respiratory specimens from each episode that belong to different clades of virus or have genomes with >2 nucleotide differences per month [32]. Cases in which detection of SARS-CoV-2 RNA is present >45 days to 89 days apart are considered reinfections if the second symptomatic episode had no obvious alternate explanation for the COVID-19-like symptoms or there was close contact with a person known to have laboratory diagnosed COVID-19 and paired specimens are available with the Ct values and sequence diversity noted above.
However, the ability to re-infect does not mean that a SARS-CoV-2 vaccine cannot be effective. Some vaccines, for example, require a "booster" dose to maintain protection. Learning more about reinfection could help researchers in developing truly effective vaccines by showing them which immune responses are important for maintaining immunity. For example, researchers may find that people become vulnerable to reinfection after antibodies drop below a certain level, and so they can modify vaccination strategies accordingly using a booster dose to maintain that level of antibodies. At a time when health authorities are grappling with the dizzying logistical difficulties of vaccinating the world population against SARS-CoV-2, the need for a booster injection is a necessity that complicates the management of the vaccination campaign, but it does not make long-term immunity from SARS-CoV-2 impossible. However, some researchers fear that vaccines will only reduce symptoms during a second infection, rather than prevent it altogether. While giving some advantages, this possibility could turn vaccinated individuals into asymptomatic carriers of SARS-CoV-2, putting vulnerable populations at risk. The elderly, for example, are among the most affected by COVID-19, but they tend not to respond well to vaccines. For all these reasons, it would be interesting to see data on how much virus SARS-CoV-2 reinfected individuals spread.
The real problem to be solved is, therefore, the duration of immunity conferred by a COVID-19 episode. 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 [146]. In considering the protective effect of antibodies against a reinfection, the evidence is still inadequate, and more research is necessary in order to clarify the interplay between the roles of adaptive and innate immunity. A recent study of Gudbjartsson et al. reported that Icelandic humoral response to SARS-CoV-2 infection was persistent within the 120-day timeframe used with a modest decline in antibody titers after 120 days [196]. Iyer et al. observed declining antibody titers over 90 days, with "median times to sero-reversion of 71 and 49 days following symptom onset" [197].
The genetic analysis of all the new cases reported as reinfection would help in understanding if the reinfection would be due to a new infection by a different SARS-CoV-2 or a reinfection by the same virus for a decline of immune response, but unfortunately genomic analysis is not available for some of these cases.

Conclusions
All these findings are useful and contribute towards the role of vaccination in response to the COVID-19 infections. Collected data show a wide range of situations: spanning a broad distribution of ages, risk groups, baseline health status and reinfection severity compared to the initial infection. Reinfection occurred as early as 45 days or >300 days after the initial infection. Common explanations for reinfection can be either waning SARS-CoV-2 antibodies or the presence of viral escape mutations [198]. While several cases of SARS-CoV-2 reinfection did involve infection with a different clade, it is noteworthy that mutations were identified throughout the genomes and the frequency of mutations within the S gene was not elevated relative to the rest of the genome [199]. In addition, individuals with more severe reinfections did not have significantly greater frequency of S gene mutations [199]. Finally, the presence of rare mutations was uncommon in the re-infecting virus, which largely mirrored the contemporaneously circulating variants in the region of infection, as reported by Choudhary et al. [199]. Concerning the problem of recognizing reinfection and persistent infection, two factors generally differentiated them. First, reinfections have so far been largely described in immunocompetent individuals while the majority of persistent COVID cases have been in immunosuppressed patients [199]. Secondly, phylogenetic analysis can generally differentiate between reinfection and persistent infection, especially in cases where persistent infection allowed the longitudinal collection of >2 sequences [199]. Due to the reinfection cases with SARS-CoV-2, it is evident that the level of immunity is not 100% for all individuals. Reinfection with SARS-CoV-2 is a possibility in both vaccinated and unvaccinated individuals, because vaccines to the virus may not translate to total immunity [199]. Recently breakthrough infections were reported following mRNA vaccination in healthy subjects [200,201], despite evidence of effective immune response among the breakthrough subjects [202]. Another study reported that eight symptomatic SARS-CoV-2 infections occurred in fully vaccinated healthcare workers (incidence rate 4.7 per 100,000 person-days adjusted) [203]. This type of challenge was also observed during the process of vaccine preparation for influenza [204]. Even though several vaccines are ready, the presence of more than 80 genotypical variants of the virus, possibility of reinfection, and short duration of seropositivity for neutralizing antibodies raise the concern that vaccination may not result in an effective and long-term immunity against SARS-CoV-2. Furthermore, immunoglobulin levels may not correlate with viral shedding and risk of transmissibility of SARS-CoV-2 [205]. Additionally, the short duration of immunity against the virus may not allow for increasing homogeneity of affected populations in a nonspecific time frame. These factors raise concerns that eliminating the COVID-19 pandemic may not be as feasible as once assumed and that we must rely more on prevention of transmission until more aspects of the virus and its pathogenicity are discovered. A recent study suggested that among persons with previous SARS-CoV-2 infection, full vaccination provides additional protection against reinfection [206]. In fact, among previously infected Kentucky residents, those who were not vaccinated were more than twice as likely to be reinfected compared with those with full vaccination [206]. Data from literature are comforting: out of hundreds of millions of people infected with the virus and then cured, only a few are reported cases of confirmed reinfection [199]. Despite the appearance of different variants of the virus, vaccines seem to help us for the near future. However, the presence of immunosuppressed or transplanted subjects requires us to continue to observe the precautionary rules useful to prevent the spread of the virus. In fact, it is imperative that all individuals, whether previously diagnosed with COVID-19 or not should take identical precautions to avoid reinfection with SARS-CoV-2 till the time when community immunity had been achieved [207]. All eligible persons should be offered vaccination, including those with previous SARS-CoV-2 infection, to reduce their risk for future infection [206].
This report highlights how it is necessary to continue to observe all the prescriptions recently indicated in the literature [208][209][210] in order to avoid new contagion for all patients after healed from COVID-19 or asymptomatic positive, since the infection does not ensure complete immunity in 100% of cases.