Skip to main content
NCBI home page
American College of Physicians - PMC COVID-19 Collection logoLink to American College of Physicians - PMC COVID-19 Collection
. 2020 Jul 6:M20-2889. doi: 10.7326/M20-2889

Clinical Validity of Serum Antibodies to SARS-CoV-2

A Case–Control Study

Giorgio Caturegli 1,2, Joshua Materi 1,2, Brittney M Howard 1,2, Patrizio Caturegli 1,2
PMCID: PMC7370852  PMID: 32628534

Serum antibodies develop upon infection with SARS-CoV-2 and are currently used to stratify donors of convalescent plasma and assess the prevalence of COVID-19. This study examines the use of serum antibodies to SARS-CoV-2 spike protein in the diagnosis of COVID-19.

Visual Abstract. Clinical Validity of COVID-19 Serum Antibodies.

Visual Abstract. Clinical Validity of COVID-19 Serum Antibodies  Serum antibodies develop upon infection with SARS-CoV-2 and are currently used to stratify donors of convalescent plasma and assess the prevalence of COVID-19. This study examines the use of serum antibodies to SARS-CoV-2 spike protein in the diagnosis of COVID-19.

Open in a new tab

Serum antibodies develop upon infection with SARS-CoV-2 and are currently used to stratify donors of convalescent plasma and assess the prevalence of COVID-19. This study examines the use of serum antibodies to SARS-CoV-2 spike protein in the diagnosis of COVID-19.

Abstract

Background:

The clinical utility of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) antibodies remains undefined.

Objective:

To determine the clinical validity and utility of SARS-CoV-2 antibodies.

Design:

Case–control study.

Setting:

First month of testing for coronavirus disease 2019 (COVID-19) by using a nucleic acid amplification test (NAAT) on nasopharyngeal swabs at the Johns Hopkins Hospital, Baltimore, Maryland (11 066 persons).

Participants:

Of the 11 066 tested persons, 115 (1%) were hospitalized adults investigated for COVID-19. Clinical record review was performed to classify them into a COVID-19 case group (n = 60) or a non–COVID-19 control group (n = 55). The laboratory control groups comprised 513 persons not tested by NAAT: 160 healthy laboratory employees, 101 persons positive for IgG antibodies against Epstein-Barr virus capsid antigen, 215 positive for thyroperoxidase antibody, and 37 positive for rheumatoid factor.

Measurements:

Serum IgG and IgA antibodies against SARS-CoV-2 spike protein were detected by using enzyme-linked immunosorbent assay.

Results:

Sensitivity and specificity of the SARS-CoV-2 IgG assay were 0.976 (95% CI, 0.928 to 0.995) and 0.988 (CI, 0.974 to 0.995), respectively, when performed 14 days or later after symptom onset, but sensitivity decreased at earlier time points. Immunoglobulin G developed rapidly and was sustained at high levels throughout follow-up (up to 58 days). Antibodies to SARS-CoV-2 predicted the odds of developing acute respiratory distress syndrome, which increased by 62% (CI, 48% to 81%; P < 0.001) for every 2-fold increase in IgG. Of 11 066 NAAT-tested patients, 457 were repeatedly NAAT-negative, and serum samples were obtained for 18 such patients: 6 COVID-19 case patients and 12 non–COVID-19 control patients. Antibodies were present in 5 of 6 case patients and none of the 12 control patients (P = 0.001).

Limitations:

The study was retrospective and performed at a single-center; the sample was small; follow-up was limited; and selection bias may have occurred.

Conclusion:

Antibodies to SARS-CoV-2 demonstrate infection when measured at least 14 days after symptom onset, associate with clinical severity, and provide valuable diagnostic support in patients who test negative by NAAT but remain clinically suspicious for COVID-19.

Primary Funding Source:

Clinical Immunology Laboratory, Department of Pathology, Johns Hopkins Hospital.

Serum antibodies are the component of the adaptive immune system used most frequently and to greatest effect by clinicians and epidemiologists. Antibodies have accompanied immunology since its inception as an academic discipline in the late 19th century (also enjoying numerous Nobel Prize recognitions), and are once more brought to center stage by the coronavirus 2019 (COVID-19) pandemic. First reported in Wuhan, China, in December 2019, severe acute respiratory syndrome-related coronavirus 2 (SARS-CoV-2) has infected 10 424 992 persons as of 30 June 2020 (1), causing severe disease in about 15% (2) and death in approximately 0.4% (3), due to diffuse alveolar damage featuring intra-alveolar edema and lymphoplasmacytic infiltrate (4).

SARS-CoV-2 is a single-stranded, positive-sense RNA, enveloped, helical virus that synthesizes 4 structural proteins: spike (S), nucleocapsid, matrix, and envelope (5). Spike is a trimeric protein that protrudes from the envelope, giving the virus its “crown” (6). Its S1 subunit mediates cell entry by binding to angiotensin-converting enzyme 2 after “priming” by transmembrane protease serine S2 (7). Given its size, location, and essential function, spike is predicted to be a key target of antibodies (8, 9).

Serologic testing for COVID-19 is considered at all levels of society for many purposes, from diagnosis and management of individual patients (10) to selection of convalescent patients as donors for antibody transfer to critically ill patients (11) and screening of blood or organ donors (12). Serology facilitates assessment of prevalence in at-risk communities (such as health care workers, homeless people, and assisted living residents, among others) and the general population—a prevalence which, as demonstrated in previous viral pandemics, is typically higher than expected (13–16).

Clinical applications of COVID-19 serologic testing remain to be defined. A possible use is to complement the laboratory gold standard of COVID-19 diagnosis: reverse-transcriptase polymerase chain reaction assay, commonly referred to as “nucleic acid amplification test” (NAAT). These tests are predominantly performed on nasopharyngeal swabs, although samples from other anatomical sites, such as bronchoalveolar lavage, sputum, and endotracheal aspirate, are also tested. With increased use, NAAT begins to show limitations (17) arising from intermittent viral shedding (18), time since exposure (19), and nasopharyngeal swab technique (20). Cases where clinical suspicion remains high despite repeated negative NAAT results could especially benefit from serologic testing. Several recent studies have described the technical performance of antibody assays (8, 18, 20–27), but data on clinical sensitivity and specificity are scarce (15). We report the performance of a serum assay for SARS-CoV-2 spike protein, providing insights into antibody kinetics and clinical uses.

Methods

This study was approved by the institutional review board of the Johns Hopkins Hospital (IRB 00247645).

Study Design and Participants

We performed a case–control study designed to assess the clinical validity and utility of a serum antibody test for SARS-CoV2.

The study included 628 persons (876 serum samples) tested for SARS-CoV-2 antibodies: 115 hospitalized adults (363 serum samples) investigated for COVID-19 by using NAAT of nasopharyngeal swab, and 513 persons (513 serum samples) not tested by NAAT who provided a serum sample to the immunology laboratory (Figure 1). Clinical record review (without antibody testing) of all 558 patients who had more than 1 nasopharyngeal NAAT was also performed.

Figure 1. Study flow diagram.

Figure 1. Study flow diagram. The study included 628 participants tested for serum antibodies against spike protein. The study also included clinical record review of all 558 patients with repeated nucleic acid amplification testing of nasopharyngeal swabs (34 were tested for antibodies, whereas 524 were not). COVID-19 = coronavirus disease 2019; EBV = Epstein–Barr virus; NAAT = nucleic acid amplification test; SARS-CoV-2 = severe acute respiratory syndrome coronavirus 2.

Open in a new tab

The study included 628 participants tested for serum antibodies against spike protein. The study also included clinical record review of all 558 patients with repeated nucleic acid amplification testing of nasopharyngeal swabs (34 were tested for antibodies, whereas 524 were not). COVID-19 = coronavirus disease 2019; EBV = Epstein–Barr virus; NAAT = nucleic acid amplification test; SARS-CoV-2 = severe acute respiratory syndrome coronavirus 2.

Patients Investigated for COVID-19 and Tested by NAAT for SARS-CoV-2

First, we assembled a data set of NAATs on nasopharyngeal swabs performed by the Johns Hopkins microbiology laboratory during the first month of COVID-19 testing (11 March to 12 April 2020), and tallied a total of 11 699 tests. We then stratified the NAATs according to whether they were performed only once (n = 10 508) or more than once (1191 tests in 558 persons) in a given person, for a total of 11 066 persons. We then stratified those 11 066 persons by NAAT result and frequency, yielding 3 groups: NAAT single-positive (1453 [13%]), NAAT single-negative (9055 [82%]), and repeated NAATs (558 [5%]).

We next searched the clinical pathology laboratories for residual serum samples (samples retrieved from the laboratory after being ordered and tested for routine patient care) with an earliest collection date within the study period. A convenience sample featuring 115 hospitalized patients of the total 11 066 tested by NAAT was obtained (1%), representing the 3 groups indicated above. In particular, this sample included 207 serum samples for 38 single-positive patients, 43 serum samples for 43 single-negative patients, and 113 serum samples for 34 repeatedly tested patients (thus, a total of 363 serum samples from 115 hospitalized patients).

Persons With Laboratory Samples Tested for SARS-CoV-2 Antibodies, but Not by NAAT

The 513 persons in the laboratory control group were 160 healthy laboratory employees and 353 patients with a polyclonal activation of the antibody response (28). The latter subset was composed of 101 persons positive for IgG antibodies against Epstein–Barr virus (EBV) capsid antigen (the most prevalent viral antibody reported in our hospital population), 215 persons positive for thyroperoxidase antibody, and 37 persons positive for rheumatoid factor (humoral markers of the 2 most prevalent autoimmune diseases: Hashimoto thyroiditis and rheumatoid arthritis, respectively). Persons in the laboratory control group were further stratified by time of sample collection, distinguishing the study period from pre-2019 (thus preceding the emergence of COVID-19).

Patients Investigated for COVID-19 and Tested by NAAT, but Without Antibody Measurement

For the 558 patients with repeated NAAT, we reviewed clinical records, irrespective of whether serum was available for antibody assessment. This review aimed to identify the subset of patients for whom diagnostic application of SARS-CoV-2 serology could be most useful, considering that patients with repeated NAAT would be those for whom a clinical diagnosis of COVID-19 is most uncertain.

Detailed Clinical Record Review and Case Definition of the 115 Hospitalized Patients

Extensive review of electronic health records was performed, following each patient from admission to discharge. This review collected risk factors (travel history, sick contacts, occupation), presentation (symptom onset, fever, cough, sputum production, dyspnea, fatigue, chills, myalgia, headache, sore throat, rhinorrhea, nausea, emesis, diarrhea, anorexia, anosmia, dysgeusia), basic chemistry and hematologic analytes (complete blood cell count, complete metabolic panel), radiologic findings (chest plain radiography and computed tomography), comorbidities (hypertension, diabetes, chronic obstructive pulmonary disease, chronic kidney disease, congestive heart failure, cancer), smoking and alcohol history, body mass index, reason for repeated NAAT testing (as applicable), and complications (length of stay, intubation, acute respiratory distress syndrome [ARDS], acute kidney injury, shock).

Since NAAT results alone cannot incontrovertibly establish a COVID-19 diagnosis (29), we integrated the judgment of the clinical team with the collected variables to classify the 115 hospitalized patients as COVID-19 case patients (n = 60) or non–COVID-19 control patients (n = 55).

Serum Antibody Testing

The 628 study participants contributed a total of 876 residual serum samples: 308 obtained longitudinally from 60 COVID-19 case patients, and 568 cross-sectionally from control persons (Figure 1). All serum samples were tested for the presence of IgG and IgA antibodies against the S1 domain of the SARS-CoV-2 spike protein by using a manual commercial enzyme-linked immunosorbent assay (ELISA) (Euroimmun), which received an Emergency Use Authorization by the U.S. Food and Drug Administration on 5 May 2020 (30). The assay yields a unitless ratio calculated by dividing the absorbance (optical density) of antibody reactivity in patient serum by that of an assay calibrator (a proprietary humanized monoclonal antibody to the SARS-CoV-2 spike protein). The assay was performed according to the manufacturer's recommendations, testing the serum samples at a 1:100 dilution and respecting specified incubation temperatures, with 2 main modifications. First, we extensively honed the assay to select an in-house threshold value of the ratio that best distinguished COVID-19 case patients from control persons, in addition to applying the cutoff suggested by the manufacturer (1.1 units for both IgG and IgA isotypes). We also serially diluted (1:1000 and 1:2000) samples that violated the Beer–Lambert law, which describes the linear relationship between absorbance and concentration (31), to obtain a quantitative value instead of a “greater than” estimate for samples with high antibody reactivity.

Statistical Analysis

Univariate logistic regression was used to select the antibody level that best classified clinical diagnosis into COVID-19 and non–COVID-19. As COVID-19 case patients for this analysis, we created 5 different scenarios based on the day the antibody was measured after recorded onset of symptoms. As non–COVID-19 control patients, we used all other control groups combined, among whom antibody levels did not differ. After each regression, we calculated the area under the receiver operating characteristic curve for the probability of having a COVID-19 diagnosis. We then used the closest-to-(0,1) method to select the antibody cutoff.

Multiple linear regression with generalized estimating equations (32) was used to analyze how antibodies in COVID-19 case patients evolved longitudinally after the onset of symptoms, considering that measures repeated over time in the same patient are likely to be correlated with one another. In addition to day after symptom onset, predictors featured in the final model included antibody isotype (IgG or IgA) and use of immunosuppressive drugs at admission. Kaplan–Meier nonparametric survival function was used to calculate the seroconversion rate of IgG and IgA after symptom onset in COVID-19 case patients.

Clinical characteristics of COVID-19 case patients and persons in the control group were compared by using nonparametric tests (Fisher exact χ2 test for low incidences and Wilcoxon rank-sum test) and were presented as proportions or medians with interquartile ranges. Because the antibody result was not normally distributed, results were transformed to natural logarithm for all analyses. Differences in mean antibody levels among groups were assessed by analysis of variance.

Data were analyzed by using Stata statistical software, release 16.1.

Role of the Funding Source

The study was funded internally by the Clinical Immunology Laboratory of the Department of Pathology, Johns Hopkins Hospital. The funder had no role in the design or conduct of the study, analysis of the data, or the decision to submit the manuscript for publication.

Results

The serum ELISA for SARS-CoV-2 antibodies performs well when used 14 days or more after symptom onset. Serum IgG antibodies against SARS-CoV-2 were significantly higher in COVID-19 case patients (median, 2.01 units [interquartile range, 0.16 to 44.33 units]) than in all persons in the control groups (median, 0.10 unit [interquartile range, 0.05 to 0.19 unit]; P < 0.001) (Figure 2, A). Time of assessment, however, was critical to the interpretation of antibody results: After time since symptom onset was grouped into 5 intervals (0 to 5 days, 6 to 9 days, 10 to 13 days, 14 to 25 days, and 26 to 59 days), median IgG increased from 0.06 units to 0.17, 1.11, 44.33, and 59.08 units, respectively (Figure 2, A). When measured at least 14 days after symptom onset, serum IgG to SARS-CoV-2 spike protein distinguished COVID-19 case patients from persons in the control group with a sensitivity of 0.976 (95% CI, 0.928 to 0.995) and a specificity of 0.988 (CI, 0.974 to 0.995), using the manufacturer's cutoff of 1.1 units (Table 1). Similar sensitivity (0.976 [CI, 0.928 to 0.995]) and specificity (0.989 [CI, 0.977 to 0.996]) were obtained by using the internally derived IgG cutoff (1.23 units). Serum IgA antibodies against SARS-CoV-2 spike protein measured at the same time point yielded similar sensitivity (0.984 [CI, 0.939 to 0.999]) but lower specificity (0.880 [CI, 0.851 to 0.905]) when the manufacturer cutoff (1.1 units) was used (Appendix Table 1). Sensitivity decreased to 0.951 (95% CI, 0.895 to 0.980) and specificity increased to 0.961 (95% CI, 0.942 to 0.975) when the internally derived IgA cutoff (2.26 units) was used (Appendix Table 1).

Figure 2. Biological characteristics of serum IgG antibodies against SARS-CoV-2 spike protein in the COVID-19 case group and the laboratory control groups.

Figure 2. Biological characteristics of serum IgG antibodies against SARS-CoV-2 spike protein in the COVID-19 case group and the laboratory control groups.A. Violin plot showing the distribution of IgG levels, indicating median and interquartile range. The horizontal dashed line represents the manufacturer cutoff (1.1 units). B. Overall relationship between IgG levels and day post symptom onset. The white circles represent a hypogammaglobulinemic patient receiving immunosuppressive therapy due to kidney transplant. C. Kaplan–Meier survival function of IgG and IgA seroconversion. D. Serum IgG and IgA antibody levels, stratified by the presence or absence of ARDS. ARDS = acute respiratory distress syndrome; COVID-19 = coronavirus disease 2019; EBV = Epstein–Barr virus; HCW = health care worker; RF = rheumatoid factor; SARS-CoV-2 = severe acute respiratory syndrome coronavirus 2; TPO = thyroperoxidase.

Open in a new tab

A. Violin plot showing the distribution of IgG levels, indicating median and interquartile range. The horizontal dashed line represents the manufacturer cutoff (1.1 units). B. Overall relationship between IgG levels and day post symptom onset. The white circles represent a hypogammaglobulinemic patient receiving immunosuppressive therapy due to kidney transplant. C. Kaplan–Meier survival function of IgG and IgA seroconversion. D. Serum IgG and IgA antibody levels, stratified by the presence or absence of ARDS. ARDS = acute respiratory distress syndrome; COVID-19 = coronavirus disease 2019; EBV = Epstein–Barr virus; HCW = health care worker; RF = rheumatoid factor; SARS-CoV-2 = severe acute respiratory syndrome coronavirus 2; TPO = thyroperoxidase.

Table 1. Performance Characteristics of Serum IgG Against SARS-CoV-2 Spike Protein for the Diagnosis of COVID-19*.

Table 1. Performance Characteristics of Serum IgG Against SARS-CoV-2 Spike Protein for the Diagnosis of COVID-19*

Open in a new tab

Appendix Table 1. Performance Characteristics of Serum IgA Against SARS-CoV-2 Spike Protein for Diagnosis of COVID-19*.

Appendix Table 1. Performance Characteristics of Serum IgA Against SARS-CoV-2 Spike Protein for Diagnosis of COVID-19*

Open in a new tab

When measured before day 14, IgG and IgA antibodies yielded substantially lower sensitivity (Table 1 and Appendix Table 1). Positivity rates for the 2 isotypes in each control group are reported in Appendix Table 2).

Appendix Table 2. Positivity Rates of IgG and IgA Antibodies Against SARS-CoV-2 Spike Protein in the Control Groups*.

Appendix Table 2. Positivity Rates of IgG and IgA Antibodies Against SARS-CoV-2 Spike Protein in the Control Groups*

Open in a new tab

We found that SARS-CoV-2 antibodies rapidly reached high levels in COVID-19 case patients and remained elevated up to 2 months after symptom onset. From being undetectable in all patients until day 5 after symptom onset, IgG levels increased 1.78 units (95% CI, 1.62 to 1.97 units; P < 0.001) per day until 16 days after symptom onset (Figure 2, B). By that point, all case patients except 1 had elevated antibody results; the exception was a hypogammaglobulinemic patient receiving immunosuppressive therapy (tacrolimus and prednisone) for kidney transplant (Figure 2, B). Similarly, rapid kinetics was observed for the IgA isotype (data not shown). As the humoral response to SARS-CoV-2 matured, spike antibodies reached levels approximately 100-fold greater than those in the control group; these levels were sustained for the duration of follow-up, up to 59 days after symptom onset.

Immunoglobulin A seroconversion anticipates IgG by 2 days. Of 60 COVID-19 case patients, 49 had longitudinal serum collection, with a follow-up time ranging from 5 to 59 days after symptom onset (median, 23 days [interquartile range, 12 to 29 days]) (Appendix Figure, top). Of these, 36 patients had baseline (preseroconversion) serum collection. On time-to-event analysis, median time to seroconversion was 12 days after symptom onset for IgG, and all patients developed IgG by day 17 (Figure 2, C). Serum IgA levels correlated well with those of IgG: For every unit increase in the natural logarithm of IgG, the natural logarithm of IgA increased by 0.73 units (adjusted R 2 = 0.83; P < 0.001) (Appendix Figure, middle). On average, IgA was approximately 14 units higher than IgG (P < 0.001) and appeared earlier, with a median time to seroconversion of 10 days after symptom onset (Figure 2, C).

Appendix Figure. Additional study characteristics.

Appendix Figure. Additional study characteristics. Top. Distribution of follow-up time in 49 of the 60 COVID-19 patients for whom longitudinal serum samples were available. Middle. Linear relationship between serum IgG and IgA levels. Open circles indicate patients who exclusively converted IgA but not IgG before day 12 post symptom onset. Bottom. Distribution of time between first and last NAAT in 558 patients tested repeatedly. COVID-19 = coronavirus disease 2019; IQR = interquartile range; NAAT = nucleic acid amplification test; SARS-CoV-2 = severe acute respiratory syndrome coronavirus 2.

Open in a new tab

Top. Distribution of follow-up time in 49 of the 60 COVID-19 patients for whom longitudinal serum samples were available. Middle. Linear relationship between serum IgG and IgA levels. Open circles indicate patients who exclusively converted IgA but not IgG before day 12 post symptom onset. Bottom. Distribution of time between first and last NAAT in 558 patients tested repeatedly. COVID-19 = coronavirus disease 2019; IQR = interquartile range; NAAT = nucleic acid amplification test; SARS-CoV-2 = severe acute respiratory syndrome coronavirus 2.

We noted that SARS-CoV-2 antibody levels strongly predicted development of ARDS, after adjustment for G or A isotype and time since symptom onset. Both IgG and IgA levels were significantly elevated in COVID-19 case patients with ARDS than in less severe cases (P < 0.001) (Figure 2, D). The odds of ARDS increased by 62% (CI, 48% to 81%; P < 0.001) or 76% (CI, 54% to 111%; P < 0.001) for every 2-fold increase in IgG or IgA, respectively. In addition to a significantly higher incidence of intubation and ARDS, COVID-19 case patients were more frequently male and symptomatic (fever, cough, myalgia, fatigue, rhinorrhea, chills, anorexia, and headache), and stayed longer than non–COVID-19 control patients (Appendix Table 3).

Appendix Table 3. Clinical Characteristics of the COVID-19 Case Group and the Non–COVID-19 Control Group From the First Month of the Epidemic in Baltimore (March–April 2020)*.

Appendix Table 3. Clinical Characteristics of the COVID-19 Case Group and the Non–COVID-19 Control Group From the First Month of the Epidemic in Baltimore (March–April 2020)*

Open in a new tab

In patients who repeatedly test negative on nasopharyngeal NAAT but the index of clinical suspicion is still high, SARS-CoV-2 antibodies support COVID-19 diagnosis. Of 11 066 persons who underwent nasopharyngeal NAAT during the study period, 1517 (14%) tested positive and 9549 (86%) negative (Figure 3, top), a relatively high prevalence reflective of the referral nature of the Johns Hopkins Hospital. In 10 508 (95%) persons, NAAT was performed once, whereas it was repeated 2, 3, 4, or 5 times in 558 (5%) persons (Figure 3, top). The interval between first and last NAAT ranged from 1 to 28 days (median, 6 days [interquartile range, 2 to 12 days]), with 8% of tests repeated within 24 hours (Appendix Figure, bottom). Most repeated tests (457 of 558 [82%]) were consistently negative (Figure 3, middle), whereas the remaining ones were consistently positive (42 [8%]), negative followed by positive (36 [6%]), positive followed by negative (17 [3%]), or inconsistent (6 [1%]) (Figure 3, middle). Clinical record review demonstrated that the predominant reason for repeating NAAT was clinical suspicion of COVID-19 (325 of 558 [58%]), followed by administrative requirements, such as discharge placement requirement or patient request (87 [16%]); belonging to an at-risk population, such as health care employees or homeless (79 [14%]); or unknown (67 [12%]).

Figure 3. Integration of clinical data and molecular testing with serum antibodies to SARS-CoV-2 spike protein.

Figure 3. Integration of clinical data and molecular testing with serum antibodies to SARS-CoV-2 spike protein. Top. Classification of 11 066 patients undergoing nasopharyngeal NAAT during the first month of COVID-19 at the Johns Hopkins Hospital, by NAAT result and number per patient. Middle. Classification of NAAT results in 558 repeatedly tested patients. Bottom. Correlation between antibody and NAAT results, classified dichotomously as ever positive or negative, in 115 hospitalized patients. Closed circles indicate COVID-19 case patients; open circles indicate non-COVID-19 hospital control patients. The top left quadrant shows the 8 patients who tested positive by NAAT and negative by antibodies (the immunosuppressed transplant patient and 7 patients with serum collection between 2 and 9 day after symptom onset). The top right and bottom left quadrants show the correlation between NAAT and antibody results, which agreed in 102 patients (46 both positive and 56 both negative). In the lower left quadrant, the closed circle indicates the antibody- and NAAT-negative patient with COVID-19 who contributed a serum sample at day 10. The lower right quadrant indicates patients with false-negative results on NAAT who tested positive for SARS-CoV-2 antibodies. COVID-19 = coronavirus disease 2019; NAAT = nucleic acid amplification test; SARS-CoV-2 = severe acute respiratory syndrome coronavirus 2.

Open in a new tab

Top. Classification of 11 066 patients undergoing nasopharyngeal NAAT during the first month of COVID-19 at the Johns Hopkins Hospital, by NAAT result and number per patient. Middle. Classification of NAAT results in 558 repeatedly tested patients. Bottom. Correlation between antibody and NAAT results, classified dichotomously as ever positive or negative, in 115 hospitalized patients. Closed circles indicate COVID-19 case patients; open circles indicate non-COVID-19 hospital control patients. The top left quadrant shows the 8 patients who tested positive by NAAT and negative by antibodies (the immunosuppressed transplant patient and 7 patients with serum collection between 2 and 9 day after symptom onset). The top right and bottom left quadrants show the correlation between NAAT and antibody results, which agreed in 102 patients (46 both positive and 56 both negative). In the lower left quadrant, the closed circle indicates the antibody- and NAAT-negative patient with COVID-19 who contributed a serum sample at day 10. The lower right quadrant indicates patients with false-negative results on NAAT who tested positive for SARS-CoV-2 antibodies. COVID-19 = coronavirus disease 2019; NAAT = nucleic acid amplification test; SARS-CoV-2 = severe acute respiratory syndrome coronavirus 2.

Review of clinical records showed that 13 of 457 (3%) consistently NAAT-negative patients were clinically diagnosed with COVID-19 (Table 2). We obtained residual serum samples for 18 of these 457 patients: 12 were clinically classified as non–COVID-19 control patients and 6 as COVID-19 case patients. Antibodies for SARS were present in 5 of 6 (83%) COVID-19 case patients (Figure 3, bottom) and absent in all 12 non–COVID-19 control patients (Figure 3, bottom) (P = 0.001). The antibody- and NAAT-negative patient with COVID-19 (Figure 3, bottom) contributed only an early serum sample (day 10 after symptom onset), which was a premature time frame for adequate assessment of seroconversion. This patient was also the only of the 6 COVID-19 case patients who had NAAT of a sample from an additional anatomical site (bronchoalveolar lavage), which was positive.

Table 2. Diagnosis of COVID-19 in the 558 Patients Who Had Repeated NAAT of Nasopharyngeal Swabs.

Table 2. Diagnosis of COVID-19 in the 558 Patients Who Had Repeated NAAT of Nasopharyngeal Swabs

Open in a new tab

Correlating NAAT and antibody results (both dichotomously categorized as ever positive or negative) in all 115 study patients showed agreement in 102 patients, 46 both positive (Figure 3, bottom) and 56 both negative (Figure 3, bottom), for an overall concordance of 89%. Of 8 patients who tested positive by NAAT and negative by antibodies (Figure 3, bottom), 1 was the immunosuppressed transplant recipient described above, and the other 7 had only an early serum sample collection (between 2 and 9 days after symptom onset).

Discussion

We report SARS-CoV-2 antibody characteristics in an early cohort of patients with COVID-19 from the Johns Hopkins Hospital, highlighting sensitivity and specificity, seroconversion kinetics, correlation with disease progression, and diagnostic potential in NAAT-negative cases. Besides epidemiologic and therapeutic applications, the study shows the potential contribution of serology to COVID-19 diagnosis, which currently relies on integrating symptom surveillance, radiographic findings, and NAAT results.

Antibodies to SARS-CoV-2 begin to appear approximately 10 days after symptom onset and rapidly increase to a high and sustained plateau, in keeping with IgG and IgA seroconversion reported in other infectious diseases and with SARS-CoV-2 incubation (33–35). When chronology is taken into account by measuring at least 14 days after symptom onset, serologic testing has excellent specificity and sensitivity. Conversely, premature testing can result in false-negative results and probably explains the suboptimal assay performance reported elsewhere (24). A wide analytical measuring range (0.01 to 300 units) coupled with proper serum dilution allows assay results to be reported as continuous numbers rather than titers, enabling precise monitoring of humoral kinetics and accurate stratification of convalescent plasma donors. Selecting the appropriate time to first measure antibodies is challenging because nonspecific and mild symptoms, altered mental status, language barriers, and other patient or physician factors can obfuscate the clinical history.

Patients with a clinical diagnosis of COVID-19 who do not have antibodies can be considered to have false-negative results, which probably arise from inaccurate estimation of infection time but also antigenic variation in spike protein or immunosuppression (19, 27, 36). In our cohort of 60 hospitalized patients with COVID-19, in fact, the 8 false-negative results were in patients with immunosuppression or follow-up not beyond day 10. Conversely, false-positive results are seen those who have antibodies in the absence of clinically diagnosed COVID-19. In our cohort, 5 of 513 (1%) persons in the laboratory control groups had false-positive results, but their antibody level was low and remained stable over time, in sharp contrast to the humoral kinetics observed in COVID-19 case patients. Two of these persons with false-positive results were hospital employees with serum collected during the study period and could thus represent asymptomatic infection, especially prevalent in younger healthy individuals (37). The other 3 results predated COVID-19 and could represent cross-reactivity with other human betacoronaviruses (38) or conserved viral antigens. For instance, EBV induces a robust polyclonal antibody response and predisposes to autoimmunity, possibly via molecular mimicry (39). Of 101 EBV-positive persons in our study, however, only 2 (2%) had SARS-CoV-2 antibodies of low titer, a proportion not significantly different from that seen in the healthy laboratory employees and non–COVID-19 patients.

In addition to clinical implications of antibody kinetics and levels, our study revealed a strong relationship between antibodies and development of ARDS in COVID-19. This novel finding provides further insights on the contribution of the immune system to COVID-19 pathogenesis, as demonstrated in cytokine storm (35, 40, 41).

An important application of serology is to support diagnosis in patients who test negative by NAAT but remain at high clinical suspicion of COVID-19. Since its inception just a few months ago, NAAT has been rapidly scaled up, but like any new test, is starting to show its limitations, possibly arising from intermittent viral shedding and nasopharyngeal swab technique (18, 20). Recent reports estimate false-negative results to be as frequent as 20%, depending on timing of measurement (29). Considering the risks of disease progression and further contagion, identification of false-negative results is imperative, especially given the prevalence of the pandemic. Until the virus becomes endemic, antibodies can play a useful role in the diagnosis of acute illness.

The observation that 5% of patients underwent NAATs repeatedly (as many as 5 times) and frequently (median interval of 6 days between first and last test, with 8% repeated within 24 hours) suggested diagnostic uncertainty. Review of clinical record for all repeatedly tested patients indeed confirmed that clinical suspicion of COVID-19 was the most common reason for repetition. Applying antibody determination to the NAAT false-negative subset revealed the remarkable utility of serology: SARS-CoV-2 infection was demonstrated in all but 1 patient, for whom serum collection was limited to 10 days after symptom onset.

Our study has limitations, which include a retrospective design, a relatively small number of cases from a single center, convenience sampling of hospitalized patients, a relatively brief duration of follow-up, imprecision in what represents the true time zero for antibody response, and incomplete isotype characterization owing to lack of IgM assay. Convenience sampling of hospitalized patients (Berkson bias) limits conclusions about pretest characteristics of case patients and the control group. Extended follow-up is always desirable but is currently precluded by the recent emergence of COVID-19. Nevertheless, every patient who seroconverted maintained antibodies for the duration of follow-up. An accurate determination of time of infection can rarely be evinced on clinical grounds, but its estimation by time of symptom onset revealed itself to be precise and clinically applicable. As recently reported, IgM appears and wanes rapidly, thus limiting diagnostic utility and appropriate characterization of convalescent plasma donors (42).

In conclusion, we report the clinical utility of SARS-CoV-2 antibodies by describing their kinetics, association with disease severity, and utility in diagnosing COVID-19 in patients with false-negative results on NAAT.

Footnotes

This article was published at Annals.org on 6 July 2020.

* Mr. Caturegli and Mr. Materi contributed equally.

References

  • 1. Dong E, Du H, Gardner L. An interactive web-based dashboard to track COVID-19 in real time [Letter]. Lancet Infect Dis. 2020;20:533-534. [PMID: 32087114] doi:10.1016/S1473-3099(20)30120-1 [DOI] [PMC free article] [PubMed]
  • 2. Huang C, Wang Y, Li X, et al. Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China. Lancet. 2020;395:497-506. [PMID: 31986264] doi:10.1016/S0140-6736(20)30183-5 [DOI] [PMC free article] [PubMed]
  • 3. National Center for Health Statistics, Centers for Disease Control and Prevention. COVID-19 pandemic planning scenarios. 20 May 2020. Accessed at www.cdc.gov/coronavirus/2019-ncov/hcp/planning-scenarios.html on 1 June 2020.
  • 4. Schaller T, Hirschbühl K, Burkhardt K, et al. Postmortem examination of patients with COVID-19. JAMA. 2020. [PMID: 32437497] doi:10.1001/jama.2020.8907 [DOI] [PMC free article] [PubMed]
  • 5. Wu F, Zhao S, Yu B, et al. A new coronavirus associated with human respiratory disease in China. Nature. 2020;579:265-269. [PMID: 32015508] doi:10.1038/s41586-020-2008-3 [DOI] [PMC free article] [PubMed]
  • 6. Walls AC, Park YJ, Tortorici MA, et al. Structure, function, and antigenicity of the SARS-CoV-2 spike glycoprotein. Cell. 2020;181:281-292. [PMID: 32155444] doi:10.1016/j.cell.2020.02.058 [DOI] [PMC free article] [PubMed]
  • 7. Curfman G. Renin-angiotensin-aldosterone inhibitors and susceptibility to and severity of COVID-19. JAMA. 2020. [PMID: 32558905] doi:10.1001/jama.2020.11401 [DOI] [PubMed]
  • 8. Okba NMA, Müller MA, Li W, et al. Severe acute respiratory syndrome coronavirus 2-specific antibody responses in coronavirus disease patients. Emerg Infect Dis. 2020;26:1478-1488. [PMID: 32267220] doi:10.3201/eid2607.200841 [DOI] [PMC free article] [PubMed]
  • 9. Sisk JM, Frieman MB, Machamer CE. Coronavirus S protein-induced fusion is blocked prior to hemifusion by Abl kinase inhibitors. J Gen Virol. 2018;99:619-630. [PMID: 29557770] doi:10.1099/jgv.0.001047 [DOI] [PMC free article] [PubMed]
  • 10. Cheng MP, Papenburg J, Desjardins M, et al. Diagnostic testing for severe acute respiratory syndrome–related coronavirus 2. A narrative review. Ann Intern Med. 2020;172:726-734. [PMID: 32282894] doi:10.7326/M20-1301 [DOI] [PMC free article] [PubMed]
  • 11. Shen C, Wang Z, Zhao F, et al. Treatment of 5 critically ill patients with COVID-19 with convalescent plasma. JAMA. 2020. [PMID: 32219428] doi:10.1001/jama.2020.4783 [DOI] [PMC free article] [PubMed]
  • 12. Cai X, Ren M, Chen F, et al. Blood transfusion during the COVID-19 outbreak [Editorial]. Blood Transfus. 2020;18:79-82. [PMID: 32267830] doi:10.2450/2020.0076-20 [DOI] [PMC free article] [PubMed]
  • 13. Martinez DA, Hinson JS, Klein EY, et al. SARS-CoV-2 positivity rate for Latinos in the Baltimore-Washington, DC region. JAMA. 2020. [PMID: 32556212] doi:10.1001/jama.2020.11374 [DOI] [PMC free article] [PubMed]
  • 14. Steensels D, Oris E, Coninx L, et al. Hospital-wide SARS-CoV-2 antibody screening in 3056 staff in a tertiary center in Belgium. JAMA. 2020. [PMID: 32539107] doi:10.1001/jama.2020.11160 [DOI] [PMC free article] [PubMed]
  • 15. Bendavid E, Mulaney B, Sood N, et al. COVID-19 antibody seroprevalence in Santa Clara County, California. medRxiv. Preprint posted online 30 April 2020. doi:10.1101/2020.04.14.20062463 [DOI] [PMC free article] [PubMed]
  • 16. Kelen GD, Fritz S, Qaqish B, et al. Unrecognized human immunodeficiency virus infection in emergency department patients. N Engl J Med. 1988;318:1645-1650. [PMID: 3163774] [DOI] [PubMed]
  • 17. Woloshin S, Patel N, Kesselheim AS. False negative tests for SARS-CoV-2 infection—challenges and implications. N Engl J Med. 2020. [PMID: 32502334] doi:10.1056/NEJMp2015897 [DOI] [PubMed]
  • 18. Lou B, Li T, Zheng S, et al. Serology characteristics of SARS-CoV-2 infection since the exposure and post symptoms onset. medRxiv. Preprint posted online 27 March 2020. doi:10.1101/2020.03.23.20041707 [DOI] [PMC free article] [PubMed]
  • 19. Kucirka L, Lauer S, Laeyendecker O, et al. Variation in false negative rate of RT-PCR based SARS-CoV-2 tests by time since exposure. medRxiv. Preprint posted online 10 April 2020. doi:10.1101/2020.04.07.20051474 [DOI] [PMC free article] [PubMed]
  • 20. Amanat F, Stadlbauer D, Strohmeier S, et al. A serological assay to detect SARS-CoV-2 seroconversion in humans. medRxiv. Preprint posted online 16 April 2020. [PMID: 32511441] doi:10.1101/2020.03.17.20037713 [DOI] [PMC free article] [PubMed]
  • 21. Guo L, Ren L, Yang S, et al. Profiling early humoral response to diagnose novel coronavirus disease (COVID-19). Clin Infect Dis. 2020. [PMID: 32198501] doi:10.1093/cid/ciaa310 [DOI] [PMC free article] [PubMed]
  • 22. Zhao J, Yuan Q, Wang H, et al. Antibody responses to SARS-CoV-2 in patients of novel coronavirus disease 2019. Clin Infect Dis. 2020. [PMID: 32221519] doi:10.1093/cid/ciaa344 [DOI] [PMC free article] [PubMed]
  • 23. Du Z, Zhu F, Guo F, et al. Detection of antibodies against SARS-CoV-2 in patients with COVID-19 [Letter]. J Med Virol. 2020. [PMID: 32243608] doi:10.1002/jmv.25820 [DOI] [PMC free article] [PubMed]
  • 24. Lassaunière R, Frische A, Harboe ZB, et al. Evaluation of nine commercial SARS-CoV-2 immunoassays. medRxiv. Preprint posted online 10 April 2020. doi:10.1101/2020.04.09.20056325
  • 25. Adams ER, Aisnworth M, Anand R, et al. Evaluation of antibody testing for SARS-Cov-2 using ELISA and lateral flow immunoassays. medRxiv. Preprint posted online 20 April 2020. doi:10.1101/2020.04.15.20066407
  • 26. Whitman JD, Hiatt J, Mowery CT, et al. Test performance evaluation of SARS-CoV-2 serological assays. medRxiv. Preprint posted online 17 May 2020. [PMID: 32511497] doi:10.1101/2020.04.25.20074856
  • 27. Infantino M, Damiani A, Gobbi FL, et al. Serological assays for SARS-CoV-2 infectious disease: benefits, limitations and perspectives. Isr Med Assoc J. 2020;22:203-210. [PMID: 32286019] [PubMed]
  • 28. van den Heuvel H, van der Meer-Prins EMW, van Miert PPMC, et al. Infection with a virus generates a polyclonal immune response with broad alloreactive potential. Hum Immunol. 2019;80:97-102. [PMID: 30773170] doi:10.1016/j.humimm.2018.10.014 [DOI] [PubMed]
  • 29. Kucirka LM, Lauer SA, Laeyendecker O, et al. Variation in false-negative rate of reverse transcriptase polymerase chain reaction-based SARS-CoV-2 tests by time since exposure. Ann Intern Med. 2020. [PMID: 32422057] doi:10.7326/M20-1495 [DOI] [PMC free article] [PubMed]
  • 30. Center for Drug Evaluation and Research, U.S. Food and Drug Administration. EUROIMMUN Anti-SARS-CoV-2 ELISA (IgG) EUA letter of authorization. 5 May 2020. Accessed at www.fda.gov/media/137606/download on 28 May 2020.
  • 31. Mayerhöfer TG, Pipa AV, Popp J. Beer's law—why integrated absorbance depends linearly on concentration. Chemphyschem. 2019;20:2748-2753. [PMID: 31544999] doi:10.1002/cphc.201900787 [DOI] [PMC free article] [PubMed]
  • 32. Diggle PJ, Liang KY, Zeger SL. Analysis of Longitudinal Data. Clarendon Pr; 1995.
  • 33. Lauer SA, Grantz KH, Bi Q, et al. The incubation period of coronavirus disease 2019 (COVID-19) from publicly reported confirmed cases: estimation and application. Ann Intern Med. 2020;172:577-582. [PMID: 32150748] doi:10.7326/M20-0504 [DOI] [PMC free article] [PubMed]
  • 34. Loeffelholz MJ, Tang YW. Laboratory diagnosis of emerging human coronavirus infections—the state of the art. Emerg Microbes Infect. 2020;9:747-756. [PMID: 32196430] doi:10.1080/22221751.2020.1745095 [DOI] [PMC free article] [PubMed]
  • 35. Vardhana SA, Wolchok JD. The many faces of the anti-COVID immune response. J Exp Med. 2020;217. [PMID: 32353870] doi:10.1084/jem.20200678 [DOI] [PMC free article] [PubMed]
  • 36. Woo PC, Lau SK, Wong BH, et al. Longitudinal profile of immunoglobulin G (IgG), IgM, and IgA antibodies against the severe acute respiratory syndrome (SARS) coronavirus nucleocapsid protein in patients with pneumonia due to the SARS coronavirus. Clin Diagn Lab Immunol. 2004;11:665-668. [PMID: 15242938] [DOI] [PMC free article] [PubMed]
  • 37. Nishiura H, Kobayashi T, Miyama T, et al. Estimation of the asymptomatic ratio of novel coronavirus infections (COVID-19) [Letter]. Int J Infect Dis. 2020;94:154-155. [PMID: 32179137] doi:10.1016/j.ijid.2020.03.020 [DOI] [PMC free article] [PubMed]
  • 38. Huang AT, Garcia-Carreras B, Hitchings MDT, et al. A systematic review of antibody mediated immunity to coronaviruses: antibody kinetics, correlates of protection, and association of antibody responses with severity of disease. medRxiv. Preprint posted online 17 April 2020. [PMID: 32511434] doi:10.1101/2020.04.14.20065771 [DOI] [PMC free article] [PubMed]
  • 39. McClain MT, Heinlen LD, Dennis GJ, et al. Early events in lupus humoral autoimmunity suggest initiation through molecular mimicry. Nat Med. 2005;11:85-9. [PMID: 15619631] [DOI] [PubMed]
  • 40. Del Valle DM, Kim-Schulze S, Hsin-Hui H, et al. An inflammatory cytokine signature helps predict COVID-19 severity and death. medRxiv. Preprint posted online 30 May 2020. [PMID: 32511562] doi:10.1101/2020.05.28.20115758
  • 41. Kadkhoda K. COVID-19: an immunopathological view. mSphere. 2020;5. [PMID: 32321823] doi:10.1128/mSphere.00344-20 [DOI] [PMC free article] [PubMed]
  • 42. Long QX, Liu BZ, Deng HJ, et al. Antibody responses to SARS-CoV-2 in patients with COVID-19. Nat Med. 2020;26:845-848. [PMID: 32350462] doi:10.1038/s41591-020-0897-1 [DOI] [PubMed]

Articles from Annals of Internal Medicine are provided here courtesy of American College of Physicians

RESOURCES