Methods
COVID-19 infected patients (n = 117) were admitted to the ICU due to respiratory failure on average 10 (range 2–34 days) days after a positive COVID-19 PCR test. Eighty-nine patients were men (age median 60 years, range 25–86 years) and 28 women (age median 69 years, range 24–86 years).
The present investigation is a sub-study of a prospective observational study carried out at the ICU at Uppsala University hospital, a tertiary care hospital in Sweden at the time only treating COVID-19 patients. The study was approved by the National Ethical Review Agency (EPM; No. 2020 − 01623). Informed consent was obtained from the patient, or next of kin if the patient was unable to give consent. The Declaration of Helsinki and its subsequent revisions were followed. The protocol of the study was registered a priori (Clinical Trials ID: NCT04316884). STROBE guidelines were followed for reporting.
De-identified plasma collected pre-pandemic between 2016–2018 from apparently healthy persons (n = 76) were used as IgG antibody negative samples.
The AgPlus point-of-care technology was applied. The antigens used in the assay were the S1 and S2 COVID-19 subunits provided by The Native Antigen Company (Kidlington, UK). These proteins were biotinylated and immobilized onto streptavidin coated magnetic particles as the capture component of the assay. Diluted blood sample was added to a tube containing the dried streptavidin magnetic particles and allowed to incubate for three minutes. During the same incubation, dried 80 nm silver nanoparticle pre-coated with Mouse Anti Human IgG were rehydrated. Following a magnetic hold the sample supernatant was removed and replaced with the re-hydrate silver nanoparticles. The mouse Anti Human IgG binds to the Sars-COV-2 specific antibodies already bound to magnetic particles. Following another short incubation (3 minutes) and a magnetic hold, unbound silver particles were removed and the sample washed with a buffer solution. A further magnetic hold was performed and the wash buffer supernatant removed and replaced with an electrolyte containing read buffer. This sample was then added to an AgPlus electrode and the amount of silver present determined via anodic stripping voltammetry. The result was presented as nano Coulomb (nC). The total assay time was < 10 minutes.
Statistics
Non-parametric statistics was applied. Comparison between groups was calculated by Mann-Whitney U-test and correlations by Spearman rank. The statistics programme MedCalc® Statistical Software version 19.6 was used (MedCalc Software Ltd, Ostend, Belgium; https://www.medcalc.org; 2020). The determination of cut-off i.e., the limit between positive or negative IgG response was performed by Fleet Bioprocessing (Hartley Wintney,UK). A receiver operating characteristic curve was constructed. A receiver operating characteristic curve, or ROC curve, is a graphical plot that illustrates the diagnostic ability of a binary classifier system as its discrimination threshold is varied. The ROC curve is constructed by plotting the true positive rate against the corresponding false positive rate as the assay cut-off is varied.
The ROC curve data was used to determine the optimum assay cut-point. The ROC data was analysed in the following manner. Based upon the methodology of Liu(11), the CZ value was calculated as the product of the true positive rate and true negative rate for a given threshold value. In addition, the Youden function (J) was calculated by subtracting the false positive rate from the true positive rate for a given value of threshold. The optimum cut-off can then be set to maximise the CZ or J value, giving the optimum balance between assay sensitivity and specificity. Both methods yielded the same optimum cut-point for the assay.
Having determined the optimum cut-off, sample data were classified accordingly and the assay characteristics calculated as follows using a value of 0.95 for 1-α (ie. 95% confidence interval, CI).