Validation of a combined ELISA to detect IgG, IgA and IgM antibody responses to SARS-CoV-2 in mild or moderate non-hospitalised patients

Abstract

Background: Frequently SARS-CoV-2 results in mild or moderate disease with potentially lower concentrations of antibodies compared to those that are hospitalised. Here, we validated an ELISA using SARS-CoV-2 trimeric spike glycoprotein, with targeted detection of IgG, IgA and IgM (IgGAM) using serum and dried blood spots (DBS) from adults with mild or moderate disease.

Methods: Targeting the SARS-CoV-2 trimeric spike, a combined anti-IgG, IgA and IgM serology ELISA assay was developed using 62 PCR-confirmed non-hospitalised, mild or moderate COVID-19 samples, ≥14 days post symptom onset and 624 COVID-19 negative samples. The assay was validated using 73 PCR-confirmed non-hospitalised, mild or moderate COVID-19 samples, ≥14 days post symptom onset and 359 COVID-19 negative serum samples with an additional 81 DBSs. The assay was further validated in 226 PCR-confirmed non-hospitalised, mild or moderate COVID-19 samples, ≥14 days post symptom onset and 426 COVID-19 negative clinical samples.

Results: A sensitivity and specificity of 98.6% (95% CI, 92.6-100.0), 98.3% (95% CI, 96.4-99.4), respectively, was observed following validation of the SARS-CoV-2 ELISA. No cross-reactivities with endemic coronaviruses or other human viruses were observed, and no change in results were recorded for interfering substances. The assay was stable at temperature extremes and components were stable for 15 days once opened. A matrix comparison showed DBS to correlate with serum results. Clinical validation of the assay reported a sensitivity of 94.7% (95% CI, 90.9-97.2%) and a specificity of 98.4% (95% CI, 96.6-99.3%).

Conclusions: The human anti-IgGAM SARS-CoV-2 ELISA provides accurate and sensitive detection of SARS-CoV-2 antibodies in non-hospitalised adults with mild or moderate disease. The use of dried blood spots makes the assay accessible to the wider community.

Keywords: COVID-19; Dried blood spot; Low prevalence; Mild disease; Serological; Trimeric spike protein.

Fig. 1

The distribution of responses and sensitivity and specificity of the human anti-IgGAM SARS-CoV-2 ELISA. The cumulative standard normal distribution of optical densities of (A) COVID-19 negative samples (n = 624, blue) and PCR-confirmed non-hospitalised, mild or moderate COVID-19 samples, ≥14 days post symptom onset (n = 62, red). (B) ROC curve of sensitivity (98.5%, 95% CI 91.4–99.7%) and specificity (97.6%, 95% CI 96.1–98.5%).

Fig. 2

IgGAM response in COVID-19 negative (n = 359) and PCR-confirmed non-hospitalised, mild or moderate COVID-19 samples, ≥14 days post symptom onset (n = 73). A sensitivity and specificity of 98.6% (95% CI 92.6–100%) and 98.3% (95% CI 96.4–99.4%), respectively, was reported using an assay cut-off of 1.0.

Fig. 3

Precision of the human anti-IgGAM SARS CoV-2 ELISA. Precision was measured using clinical and lab standards guidelines EP12-A2 in 40 replicates and assessed at clinical cut-offs C5, C50 and C95. The mean and coefficient of variation (%) was calculated for each clinical cut-off.

Fig. 4

Stability and cross reactivity of antibody positive or antibody negative serum samples on the anti-IgGAM SARS CoV-2 ELISA. The assay was assessed for (A) interfering substances, bilirubin, haemoglobin and triglyceride (n = 5 replicates per test) in antibody positive or antibody negative serum samples; and (B) evaluated for cross reactivity with human viruses and other endemic coronaviruses (n = 39). Additionally, the kit components were tested for (C) stability at extreme temperatures (-20 °C, 4 °C and 37 °C) (n = 3) and (D) stability once opened (n = 4).

Fig. 5

Comparison of the anti-IgGAM SARS-CoV-2 ELISA response in DBS and serum samples. The response was measured in 81 DBS and serum matched samples. A comparison between the two collection methods were assessed using (A) Passing-Bablok analysis (r = 0.959, y = 0.16 + 0.91×) and (B) Bland-Altman analysis (central line demonstrates mean difference and broken lines shows the limits of agreement, −0.76 to 0.86).

Fig. 6

Clinical study. The cumulative standard normal distribution of (A) COVID-19 negative samples (n = 426, blue) and PCR confirmed, mild or moderate COVID-19 positive samples, ≥14 days post symptom onset samples (n = 226, red). (B) ROC curve of sensitivity (94.7%, 95% CI 90.9–97.2%) and specificity (98.4%, 95% CI 96.6–99.3%).