key: cord-1013316-4l3l1y0y authors: Wang, Hannah; Hogan, Catherine A; Verghese, Michelle; Solis, Daniel; Sibai, Mamdouh; Huang, ChunHong; Röltgen, Katharina; Stevens, Bryan A; Yamamoto, Fumiko; Sahoo, Malaya K; Zehnder, James; Boyd, Scott D; Pinsky, Benjamin A title: SARS-CoV-2 Nucleocapsid Plasma Antigen for Diagnosis and Monitoring of COVID-19 date: 2021-10-04 journal: Clin Chem DOI: 10.1093/clinchem/hvab216 sha: c87534e68a80ddcabee232a96c49f071106c201c doc_id: 1013316 cord_uid: 4l3l1y0y BACKGROUND: Detection of SARS-CoV-2 nucleocapsid antigen in blood has been described, but the diagnostic and prognostic role of antigenemia is not well understood. This study aimed to determine the frequency, duration, and concentration of nucleocapsid antigen in plasma and its association with COVID-19 severity. METHODS: We utilized an ultrasensitive electrochemiluminescence immunoassay targeting SARS-CoV-2 nucleocapsid antigen to evaluate 777 plasma samples from 104 individuals with COVID-19. We compared plasma antigen to respiratory nucleic acid amplification testing (NAAT) in 74 individuals with COVID-19 from samples collected ± 1 day of diagnostic respiratory NAAT, and in 52 SARS-CoV-2-negative individuals. We used Kruskal-Wallis tests, multivariable logistic regression, and mixed-effects modeling to evaluate whether plasma antigen concentration was associated with disease severity. RESULTS: Plasma antigen had 91.9% (95% CI 83.2-97.0%) clinical sensitivity and 94.2% (84.1-98.8%) clinical specificity. Antigen-negative plasma samples belonged to patients with later respiratory cycle thresholds (C(t)) when compared with antigen-positive plasma samples. Median plasma antigen concentration (log(10) fg/mL) was 5.4 (IQR 3.9-6.0) in outpatients, 6.0 (5.4-6.5) in inpatients, and 6.6 (6.1-7.2) in intensive care unit (ICU) patients. In models adjusted for age, sex, diabetes, and hypertension, plasma antigen concentration at diagnosis was associated with ICU admission (OR 2.8 [95% CI 1.2-6.2], P=.01), but not with non-ICU hospitalization. Rate of antigen decrease was not associated with disease severity. CONCLUSIONS: SARS-CoV-2 plasma nucleocapsid antigen exhibited comparable diagnostic performance to upper respiratory NAAT, especially among those with late respiratory C(t). In addition to currently available tools, antigenemia may facilitate patient triage to optimize intensive care utilization. 4 53 addition to currently available tools, antigenemia may facilitate patient triage to optimize 54 intensive care utilization. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 6 78 studies have demonstrated that the presence of SARS-CoV-2 ribonucleic acid (RNA) in plasma 79 at diagnosis is associated with the development of severe COVID-19 (9-12). 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 Table 3 ). For example, the adjusted risk of ICU admission increased from 16% to 226 39% with an increase from 4 log 10 fg/mL to 5 log 10 fg/mL plasma antigen, with a concomitant 227 decrease in non-ICU inpatient status and outpatient disposition of 40% to 36% and 44% to 26%, 228 respectively. Plasma antigen concentration, however, was not independently associated with 229 non-ICU inpatient status vs. outpatient status after multivariable regression. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 13 236 fg/mL) near the positivity threshold of 2.80 log 10 fg/mL. One individual was hospitalized for 237 cellulitis and was included as a negative control due to increased C-reactive protein. Another was 238 an ICU patient with myocardial infarction with increased C-reactive protein and procalcitonin of 239 unclear etiology. The third was an infant with rhinovirus bronchiolitis. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 14 258 To determine whether rate of plasma antigen decrease over time was associated with disease 259 severity, a linear mixed-effects model was constructed (Supplemental Fig. 5 ). From days 5-40 260 after symptom onset, plasma antigen concentration decreased linearly in relation to log 10 -261 transformed days after symptom onset (β = -7.9 [95% CI -9.2 to -6.5]) (Fig. 3) . Non-critical 262 inpatient vs. ICU status was not independently associated with change in antigen concentration 263 over time (β = 0.2 [95% CI -0.5 to 0.9]) (Supplemental Table 4 ). 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 281 Two prior studies have examined a single molecular array bead-based ELISA (SIMOA) for 282 SARS-CoV-2 nucleocapsid antigenemia, and have reported clinical sensitivities lower than that 283 observed in our study (64-74%), despite a reported limit of detection 10-fold lower than MSD S-284 PLEX (26,27). As such, this difference may be due to inclusion of a greater proportion of 285 patients with mild disease and/or plasma samples further out from diagnosis. In contrast, one 286 study reported a clinical sensitivity of 93% within 2 weeks of symptom onset for the COVID 287 Quantigene ELISA, which is more consistent with our observations (13). Interestingly, 288 nucleocapsid antigenemia had similar sensitivity for diagnosis of SARS-CoV-1 during the 289 original 2003 SARS outbreak (28, 29) . Similar to this other study (13), we also found that Laboratories utilizing this assay should consider using a local reference 300 population-based threshold rather than the assay's limit of detection, or the threshold reported in 301 this study. Of note, all false-positives in our study were near the threshold of positivity, 302 suggesting that orthogonal testing of borderline specimens (e.g., between 2.6-3.2 log 10 fg/mL) or 303 collection of a second specimen for additional testing may help increase clinical specificity. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 16 304 Further studies with larger numbers of more heterogeneous controls will be necessary to 305 establish whether this assay could be specific enough to be used for primary diagnosis. 306 307 Similar to the two studies discussed above (13, 27), we observed that plasma antigen may also 308 have value in prognostication, though these studies did not attempt to adjust for possible 309 confounding demographic factors and comorbidities, or determine whether rate of antigenemia 310 decline was associated with disease severity. In our study, among diagnostic specimens (±1 d of 311 first positive NAAT), collected a median of 7 d (IQR 3-10) after symptom onset, plasma antigen 312 concentration was higher in patients requiring ICU admission, even after adjusting for age, sex, 313 and comorbidities. As such, a quantitative nucleocapsid antigenemia test at the time of diagnosis, 314 especially within the first week of symptom onset, may have potential value in triaging patients 315 for higher-level care, though its added value relative to other biomarkers such as RNAemia will 316 need to be evaluated (9,30). 318 Based on our mixed-effects model, we found no relationship between rate of antigen decrease 319 and disease severity. This suggests that nucleocapsid plasma antigen may serve as a biomarker of 320 tissue damage and vascular leakage, rather than as a driver of infection or disease response (31). 321 As such, longitudinal monitoring of plasma antigen concentrations is unlikely to be helpful in 322 predicting outcomes or response to therapy. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 18 350 Future prospective trials with larger numbers and standardized sample collection protocols will 351 be necessary to validate the diagnostic and prognostic value of this platform, especially in 352 outpatients and asymptomatic individuals. These trials should ideally evaluate other severity 353 variables and complications we were not powered to assess including oxygen needs, mechanical 354 ventilation, therapy, and time since symptom onset. Future studies may also compare RNAemia 355 with quantative antigenemia to determine whether these two potential biomarker assays might 356 offer adjunctive versus identical prognostic information. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 disease severity and outcome. Sci Immunol 2020;5:eabe0240. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 26 524 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 A subset of the specimens had been previously tested for presence of anti-nucleocapsid IgA using a laboratory-developed enzyme-linked immunosorbent assay (ELISA) 149 as previously described (14). In brief, 96-well Corning Costar high binding plates coated with recombinant SARS-CoV-2 nucleocapsid protein at a 151 concentration of 0.1 µg per well overnight at 4°C and incubated with plasma at a 1:100 dilution 152 for 1 hour at 37°C, with secondary detection by horseradish peroxidase conjugated goat anti-153 human IgG (-chain specific, catalog no. 62-8420, Thermo Fisher, 1:6,000 dilution), IgM (-154 chain specific The positivity thresholds were set as previously described 156 based on pre-pandemic samples: OD at 450 nm of ≥0.3 for IgG, ≥0.35 for IgM, and ≥0.1 for IgA Only plasma samples taken ±1 day of first positive diagnostic respiratory NAAT from unique 161 COVID-19 patients (n=74) along with specificity controls from SARS-CoV-2 respiratory 162 NAAT-negative unique patients (n=52) were included in the evaluation of plasma antigen 163 diagnostic performance. Clinical sensitivity and specificity were calculated using respiratory 164 NAAT as the gold standard, and were reported with exact All comparisons were 168 two-sided with Type I error set at .05. No correction for multiple comparisons was performed (IQR 39-72), and were comprised of 22% outpatients (n=23), 41% non-193 critical inpatients (n=42), and 38% ICU patients (n=39) Among the 68 antigen-positive individuals Analyzing the data by respiratory 207 NAAT platform and target did not reveal a stronger linear correlation between respiratory C t and 208 log-transformed plasma antigen values Individuals with antigen-negative plasma samples had a later median respiratory RT-qPCR C t 211 than did their antigen-positive counterparts One sample was from an outpatient who was already anti-N IgG positive Statistics with confidence No adjustments are needed for multiple comparisons Fitting linear mixed-effects models using 488 lme4 Quantification of SARS-CoV-2 antigen 490 levels in the blood of patients with COVID-19 Ultra-sensitive 492 serial profiling of SARS-CoV-2 antigens and antibodies in plasma to understand disease 493 progression in COVID-19 patients with severe disease Nucleocapsid protein as early 495 diagnostic marker for SARS A patient with asymptomatic severe 497 acute respiratory syndrome (SARS) and antigenemia from the 2003-2004 community 498 outbreak of SARS in Highly sensitive 500 quantification of plasma severe acute respiratory syndrome coronavirus 2 RNA sheds light 501 on its potential clinical value COVID-19: the vasculature unleashed 551 majority of the remaining variance, with each color representing a different individual (B). The Individual timelines of plasma SARS-CoV-2 nucleocapsid antigen concentration (top 555 panel, green diamond, left side Y-axis), respiratory swab reverse transcription quantitative 556 polymerase chain reaction (RT-qPCR) cycle threshold (C t ) values (top, orange + for positive RT-557 qPCR, orange x for negative RT-qPCR, right side inverted Y-axis), and anti-nucleocapsid 558 antibody concentrations (bottom panel The dashed horizontal line represents threshold of positivity 560 for both plasma antigen and respiratory RT-qPCR. Inpatient #2 and Intensive Care Unit Patient 561 (ICU) #9 are examples of the majority of antigen-positive individuals whose seroconversion 562 coincided with disappearance of plasma antigen (A-B). Inpatient #4 and #5, and ICU #2 all had 563 positive respiratory RT-qPCRs >1 month after earlier viral RNA clearance, without reappearance 564 of plasma antigen (C-E). Plots for all 83 individuals with >1