Diagnostic Performance of Serological Assays in the Detection of SARS-CoV-2: A Review

Featured Application: This review provided a thorough analysis of the existing serological tests for the detection of SARS-CoV-2 infection: sensitivity, specificity, positive and negative predictive values of each assay were reported, percentages of IgM/IgG positive patients among cases and controls were shown. This paper may help clinicians choosing the most appropriate serological test for the diagnosis of COVID-19. Abstract: Introduction . The gold-standard method for diagnosis of the novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2 or COVID-19) foresees the examination of respiratory tract swabs by real-time reverse-transcription polymerase chain reaction (rRT-PCR). Another group of diagnostic tests, developed to overcome the limitations of RT-PCR, includes the serological assays, which have the purpose of detecting the antibody response to SARS-CoV-2 infection (IgM and IgG titers). The aim of this review was to establish the diagnostic capability of the existing serological tests in the detection of SARS-CoV-2 infection. Materials and Methods . Electronic research was conducted in PubMed, Scopus, Science Direct and Cochrane Library, and only 10 articles, testing 10 different types of serological assays, met the inclusion criteria and were consequently submitted to quality assessment and data extraction. Quantitative data about the sensitivity, specificity, positive/negative predictive value and IgM/IgG titer provided by each antibody test were reported in our review. Results . Almost all the serological tests used in the included items were recorded to ensure high sensitivity and specificity, identifying the presence of IgM and IgG antibodies against SARS-CoV-2 in patients with certain COVID-19 diagnosis (confirmed by RT-PCR) and in participants with suspected infection (SARS-CoV-2 clinical diagnosis and/or RT-PCR negative subjects). Conclusion. Serological tests may represent reliable diagnostic tools in the detection of SARS-CoV-2 infection, and they could be implemented complementary to real-time RT-PCR.


Introduction
The current global pandemic caused by the novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), whose initial outbreak was detected in December 2019 in Wuhan (China), represents a real threat to international health [1][2][3]. This new pathogen, which is an enveloped, non-segmented, positive sense RNA virus, belongs to the Coronaviridae family, as well as the severe acute respiratory syndrome coronavirus 1 (SARS-CoV-1) and Middle East respiratory syndrome coronavirus (MERS-CoV) [4,5]. Fever, cough expectoration, myalgia, fatigue, dyspnea and gastrointestinal symptoms are the most frequent clinical manifestations induced by SARS-CoV-2 [6][7][8]. In some cases, the progression of the illness is relatively asymptomatic, while, in other cases, pneumonia, acute respiratory distress syndrome (ARDS), sepsis or septic shock may occur [9,10]. The main SARS-CoV-2 transmission routes are represented by respiratory droplets (within a long distance of 2 m), generated by infected subjects during coughing and sneezing and by contact with contaminated surfaces [11][12][13][14]. No specific antiviral therapy against SARS-CoV-2 has been introduced yet: treatment protocols include broad-spectrum antiviral drugs (remdesivir, lopinavir, ritonavir, favipiravir), which should be administrated in the early stage of the infection, and antimalarial/autoimmune disease drugs (chloroquine, hydroxychloroquine) [15][16][17]. In order to limit the spread of the infection among the population, isolating the infected persons, adequate diagnostic strategies should be implemented. The most commonly used method to detect SARS-CoV-2 foresees the analysis of nasopharyngeal and throat swabs by real-time reverse-transcription polymerase chain reaction (rRT-PCR) test [18][19][20]. A complementary diagnostic procedure is represented by the chest computerized tomography (TC) tool, which may allow the detection of the virus in rRT-PCR falsenegative cases. Even if real-time RT-PCR plays a crucial role in COVID-19 diagnosis, it presents several limitations: it may not be able to identify the virus in the early stage of the infection, and it requires a long time to obtain the results, which may be influenced by external factors such as sampling operation method and the performance of detection kits and nucleic acid extraction from clinical material [21,22]. Another group of diagnostic tests used for SARS-CoV-2 identification is that of serological assays. Several different antibody tests have been developed, which differ depending on the targeted viral antigen (for example, nucleoprotein or spike protein); the most common SARS-CoV-2 biomarkers tested serological assays of IgM and IgG antibodies [23]. The purpose of this type of test may be the identification of PCR-negative cases and asymptomatic patients or the evaluation of vaccine response during clinical trials [24]. According to a study by Li-Xin Xie [25], in most cases IgM antibodies are present 3-5 days after the onset of symptoms, while IgG titer is higher in the recovery phase than in the acute one. Patel et al. [26] reported that seroconversion in infected patients may occur between 7 and 11 days after the exposure to the virus. Due to the delayed appearance of IgM and IgG antibodies, serological tests are unreliable in the acute phase of the infection. The Infectious Diseases Society of America stated that serological tests still remain clinically unverified and suggested that serology results alone should not be used for diagnostic decisions [27]. Although no clear evidence about the duration of immunity protection against SARS-CoV-2 has been recorded, the data from research by Bao et al. [28] demonstrate that primary SARS-CoV-2 infection in monkeys may generate an immunity response, which could be able to protect from subsequent exposures.
The aim of this research was to review literature in order to obtain an overview of existing serological tests for the detection of SARS-CoV-2 infection and to establish their reliability in the diagnosis of this new pathogen,

Protocol and Registration
In order to provide a transparent and complete protocol for systematic reviews, the PRISMA statement [29] was followed for methods and inclusion criteria selection.

Clinical Question (PICO)
P: A population of participants with certain diagnosis of COVID-19 (obtained thorough realtime RT-PCR). I: Detection of IgM and IgG titers in patients with SARS-CoV-2 infection using serological assays. C: Comparison between IgM/IgG titer in SARS-CoV-2-positive patients and negative controls. O: Reliability of the existing serological assays for the diagnosis of COVID-19.

Inclusion and Exclusion Criteria
All the papers that aimed to evaluate the diagnostic capability of serological tests in the detection of SARS-CoV-2 infection and that met the following inclusion criteria were selected for our review: -Studies that reported serological assay quantitative data (percentages, rates) about the IgM and IgG titers of the analyzed blood samples. Reviews and case reports were not included in our study.

Search
In order to select items concerning the utilization of antibody tests to diagnose SARS-CoV-2infected patients, we conducted electronic research in the PubMed, Scopus, Science Direct and Cochrane Library databases, analyzing papers published by April 2020. No restrictions were imposed with regards to demographics or clinical characteristics of the included patients (age, gender, comorbidities), and both articles with and without negative control groups were considered. Only studies written in the English language were selected. We combined the following keywords with the Boolean term "AND": "serological test", "COVID-19", "antibodies" and "immune system".

Study Selection and Data Collection Process
Following the inclusion criteria, eligible studies for this review were selected by two researchers (Dorina Lauritano, Giulia Moreo), who independently examined title, abstract and full texts of each article found during the electronic search. The same researchers performed data extraction from the selected items: the number of enrolled patients/blood samples, patient and blood sample sources, diagnostic tool used to confirm the presence of SARS-CoV-2 infection in the included patients, type of serological assay tested, and quantitative data about the sensitivity, specificity, positive/negative predictive value and IgM/IgG titer provided by each antibody test. Only articles that used real-time RT-PCR as a diagnostic tool to confirm the positivity of SARS-CoV-2 were considered, and rates and percentages were used for the principal outcome measures. The flow chart used for this review is shown in Figure 1.

Quality Assessment
Quality assessment of the selected items was investigated using the Newcastle-Ottawa scale (NOS) [30], recording an high quality level for the included research: the studies' average quality was equal to 6.1, while the highest score was equal to 7 and the lowest one was 5 ( Table 1). The majority of studies included case and control groups, in which data were recorded using the same methodology. The presence of SARS-CoV-2 infection in the cases of all the selected articles was confirmed following the guidelines for diagnosis of COVID-19 [31]. Table 1. Quality evaluation of the included articles.

Definition of Controls Comparability Exposure Total
Cassaniti et al. 2020 Infantino et al. 2020 Jin et al. 2020 Lee et al. 2020 Li et al. 2020 Pan et al. 2020 Xiang et al. 2020

Study Selection and Characteristics
A total of 212 articles were identified after the electronic research in four different databases (PubMed, Scopus, Science Direct, Cochrane Library). Ten studies were excluded based on duplication removal and 202 items were assessed for eligibility. Overall, 192 articles were not included with reasons and only 10 articles were included in this review: 83 studies were excluded after examining the title, 47 after analyzing the abstract (absence of quantitative data about IgM and IgG titer), 41 after reading the full text (diagnosis confirmation not specified or absence of both IgM and IgG titer analysis), and 21 because of their study design (reviews or case reports). The main characteristics of each selected article are summarized in Table 2, which reports the number of enrolled patients and blood samples and their source, the diagnostic tool used to ascertain the positivity of SARS-CoV-2 (real-time RT-PCR), and the type of serological assay whose diagnostic capability was evaluated. By adding the participants of each study, a total of 1362 subjects were included in this review, of which 945 were cases (SARS-CoV-2-positive patients, whose diagnosis was confirmed by real-time RT-PCR), 310 were negative controls, and 107 were subjects with suspected infection (SARS-CoV-2 clinical diagnosis, real-time RT-PCR negative patients). Eight of the included articles were conducted in China, with the remainder in Italy. Data about IgM and IgG titer, sensitivity, specificity, and the positive and negative predictive value provided by each serological test were recorded in this review.

Results of Individual Studies
The data recorded in this review (Tables 3-6) referred to 10 different types of serological tests, which were performed in the included articles in order to establish whether the antibody response can be considered a reliable diagnostic tool for SARS-CoV-2 infection. The VivaDiag COVID-19 IgM/IgG Rapid Test lateral flow immunoassay (LFIA) was used in a sample of 110 patients in the study by Cassaniti et al. [32]: 19 out of 30 cases (63.3%), whose serum samples were taken at a median 7 days after the symptom onset, were positive for both IgM and IgG; five of them were weakly positive; all negative controls were recorded to be negative for both antibodies. The same study also included 50 patients at their first access to the emergency department, who were later tested for COVID-19 by RT-PCR, detecting 38 SARS-CoV-2-positive patients, among whom only seven (18.4%) showed positivity for both IgM and IgG. The sensitivity and specificity of the VivaDiag test were recorded to be 18.4% and 91.7% respectively, results that lead the authors not to recommend its use in the infection diagnostic process.  [38] 60 0/60 (0%) 3/60 (5%)  Two of the included items performed the serological test using an iFlash1800 [33] and iFlash3000 [34] fully automated chemiluminescence immunoassay (CLIA) analyzer from Shenzhen YHLO Biotechnology Co., Ltd (China), respectively.
The iFlash1800 CLIA analyzer guaranteed an overall sensitivity of 75% and a specificity of 100% for IgG and 92.2% for IgM. The collected blood samples analyzed thorough this test had a mean duration of 12 days (range 8-17 days) from the onset of symptoms.
The iFlash 3000 CLIA analyzer reported, among the cases, 41 IgM/IgG-positive (64.1%), 3 IgMpositive (16.7%) and 5 IgG-positive patients out of 64, with a sensitivity of 48.1% (IgM) and 88.9% (IgG) and a specificity equal to 100% (IgM) and 90.9% (IgG). The duration from first symptoms to the first serological test was, on average, 18 days: both IgM and IgG levels were not high during the first 5 days following symptom onset.
Lee et al. [35] used the ALLTEST 2019-nCoV IgM/IgG Rapid Test Cassette (Hangzhou ALLTEST Biotech Co., Ltd. Hangzhou, China) to identify SARS-CoV-2 IgM and IgG in 14 COVID-19-positive patients (six with symptoms and eight without symptoms or with mild symptoms). All the symptomatic subjects showed IgG positivity (6/6), while two of them were IgM-negative (2/6). Three of the asymptomatic patients had positive IgG, but none of them had positive IgM. This research reported the earliest detection of IgM and IgG, on day 5.
The sensitivity and specificity of the SARS-CoV-2 rapid IgM/IgG combined antibody test kit developed in the study by Li et al. [36] were 88.66% and 90.63%, respectively; 256 out of 397 cases and 1 out 128 negative controls were positive for both IgM and IgG. This serological test was performed from day 8 to day 33 after symptoms' appearance.
By using a colloidal gold-based immunochromatographic (ICG) strip, Pan et al. [37] analyzed 108 blood samples (starting from 7 days after symptom onset), of which 86 were taken from SARS-CoV-2-positive patients (cases) and 22 from subjects with suspected infection (with negative RT-PCR): 55.8% and 54.7% of SARS-CoV-2-positive blood samples showed positivity to IgM and IgG, respectively, while 36.4% and 59.1% of the blood samples taken from patients with suspected infection were positive for IgM and IgG, respectively.
In research by Xiang et al. [38] the antibodies against this new pathogen were found with an enzyme-linked immunosorbent assay (ELISA): of 66 positive participants, 51 were IgM-positive and 55 were IgG-positive; 21 patients with suspected infection out of 24 were IgM-positive and 17 IgGpositive; negative controls (60) were all negative for IgM and only three were positive for IgG. The detection of the antibodies was conducted from day 4 to day 28 after the beginning of the symptoms, recording an increase in both IgM and IgG rates: from 60% to 63.6% for IgM and from 40% to 90.9% for IgG.
The gold immunochromatography assay supplied by Innovita Co., LTd, China (CFDA approved) tested by Yongchen et al. [40] highlighted that among the coronavirus-2-positive group (21 persons) all the symptomatic patients (17 subjects) were recorded to be seropositive during the follow-up period. Zhang et al. [41] developed an in-house anti-SARS-CoV IgG/IgM ELISA kit (using a cross-reactive nucleocapsid protein from another SARS-related virus, Rp3, which is 92% identical to SARS-CoV-2 nucleocapsid protein), employing it to investigate the antibody response in 16 positive participants: this study demonstrated that IgM and IgG titers were low or undetectable on the day of first sampling (day 0), and that, on day 5, an increase in viral antibodies could be observed in nearly all patients: IgM and IgG positive rate increased from 50% (8/16) to 81% (13/16) and from 81% (13/16) to 100% (16/16).
The ELISA kit supplied by Beijing Wantai Biological Pharmacy Enterprise Co., Ltd., used by Zhao et al. [39] reported that 82.7% and 64.7% of 173 SARS-CoV-2-positive subjects were IgM-and IgG-positive, respectively, ensuring a sensitivity equal to 66.7% in the early phase of the illness.

Discussion
This review aimed to provide an overview of the existing serological tests, assessing their capacity to detect the presence of IgM and IgG antibodies against SARS-CoV-2 in blood samples taken from patients. During SARS-CoV-2 infection, an overactivation of natural immunity cells, such as macrophages and monocytes, has been recorded, which consequently causes the diminution of lymphocytes and the increase of neutrophils, interleukin-6 (IL-6) and reactive protein C (PCR) [42]. With regard to the adaptive immune response, the activation of B lymphocytes of the humoral immunity (which occurs approximately after one week from infection) leads to the production of specific antibodies against SARS-CoV-2: IgM and IgG [43]. According to the literature, IgM antibodies constitute the first line of defense during a viral infection and can indicate the acute phase of the disease; IgG antibodies represent long-term immunity and immunological memory, and their presence highlights that exposure to the pathogen has occurred several days before [43,44]. The study by Demey et al. [45] tested four different immunochromatographic assays, demonstrating that the antibodies' detection time was, on average, 8-10 days after the symptoms' onset, while the case report of Thevarajan et al. [46] recorded a progressive increment of anti-SARS-CoV-2 IgM and IgG from day 7 until day 20 from the beginning of the illness. Two of the articles selected for this review reported that the IgM-positive rate tended to increase at first but then declined, while those of IgG were higher than of IgM at all times [34], and that IgM and IgG positive rates increased from 50% to 81% and from 81% to 100%, respectively [41]. Pan et al. [37] divided the infection into three stages: early (1-7 days from the onset), intermediate (8-14 days) and late (more than 15 days), establishing that the positive rate of IgM was raised from 11.1% (early stage) to 78.6% (intermediate stage) and 74.2% (late stage), whereas that of IgG was 3.6% in the early stages and 57.1% and 96.8% in the intermediate and late stages, respectively. Infantino et al. [33] was able to detect IgM and IgG from day 10 to day 30 and from day 20 onwards after COVID-19 infection, respectively. On the basis of the results obtained by Lee et al. [35], the persistence of positive real-time RT-PCR seemed to be shorter in symptomatic patients, who developed IgM antibodies. The median seroconversion time detected by Zhao et al. [39] was day 12 and day 14 for IgM and IgG respectively. Among the five patients with severe symptoms analyzed by Yongchen et al. [40], the antibody response was individuated within week 2, and three out of five of these subjects developed IgG response prior to viral clearance, indicating that high levels of this new pathogen viral load may provide an early antibody response [47,48]. Almost all the included items agreed on the fact that serological tests could be effective and reliable diagnostic tools for SARS-CoV-2 infection identification, since they are able to provide high sensitivity and specificity, and that their utilization should be complementary to the execution of real-time RT-PCR. Because of its several limitations, rRT-PCR could also report negative results in infected individuals: very early or late collection of swabs, poor quality of the specimen containing insufficient material quantity, or wrong technical procedures. Furthermore, it takes a long time to generate RT-PCR test results, it requires first-rate certified laboratory facilities with ad-hoc educated staff, and it may provide different results depending on the sampling site (oropharyngeal or nasopharyngeal swabs) [49]. Antibody assays could provide a faster, less expensive and simpler (no laboratory training need) method to diagnose COVID-19 [38]. As Li et al. [36] stated, serological tests may be used to screen the possible asymptomatic carriers [50], knowing that the majority of them develop anti-SARS-CoV-2 antibodies and, since these assays are able to individuate IgM and IgG simultaneously, they could be used for early diagnosis (detecting IgM) and for monitoring during the therapy [37]. The narrative review by Cheng et al. [51] highlighted that serological tests' negative results should not be a reason to exclude SARS-CoV-2 infection, considering that the patient may have been recently exposed to the pathogen. Moreover, present or past infections due to other coronaviruses could lead to crossreactivity of antibody to non-SARS-CoV-2 coronavirus proteins. According to the same authors, this type of assay could be useful for epidemiological studies, vaccine studies and risk assessment of health care workers. In contrast, Cassaniti et al. [32] also performed the serological assay on 30 healthy volunteers, 10 of whom had been infected in the past with OC43, 229E, HKU1 and NL63 coronavirus, reporting no cross-reactivity with antibodies against these pathogens. Infantino et al. [33] confirmed the absence of cross reaction with other coronaviruses but demonstrated that cytomegalovirus (CMV) infections and some rheumatic diseases could interfere with the test.

Conclusion
The global pandemic caused by the novel SARS-CoV-2 is threatening international health and screening of the population on a large scale has become imperative. Besides the use of real-time RT-PCR to individuate the presence of COVID-19, the detection of specific antibodies in response to this pathogen thorough serological assays may represent a reliable diagnostic protocol. Serological assays seem to be able to overcome the nucleic acid test's limitations, ensuring the diagnosis in asymptomatic patients and in false-negative RT-PCR case.

Conflicts of Interest:
The authors declare no conflict of interest.