ObjectivesWe investigated machinelearningbased identification of presymptomatic COVID-19 and detection of infection-related changes in physiology using a wearable device.DesignInterim analysis of a prospective cohort study.Setting, participants and interventionsParticipants from a national cohort study in Liechtenstein were included. Nightly they wore the Ava-bracelet that measured respiratory rate (RR), heart rate (HR), HR variability (HRV), wrist-skin temperature (WST) and skin perfusion. SARS-CoV-2 infection was diagnosed by molecular and/or serological assays.ResultsA total of 1.5 million hours of physiological data were recorded from 1163 participants (mean age 44±5.5 years). COVID-19 was confirmed in 127 participants of which, 66 (52%) had worn their device from baseline to symptom onset (SO) and were included in this analysis. Multi-level modelling revealed significant changes in five (RR, HR, HRV, HRV ratio and WST) device-measured physiological parameters during the incubation, presymptomatic, symptomatic and recovery periods of COVID-19 compared with baseline. The training set represented an 8-day long instance extracted from day 10 to day 2 before SO. The training set consisted of 40 days measurements from 66 participants. Based on a random split, the test set included 30% of participants and 70% were selected for the training set. The developed long short-term memory (LSTM) based recurrent neural network (RNN) algorithm had a recall (sensitivity) of 0.73 in the training set and 0.68 in the testing set when detecting COVID-19 up to 2 days prior to SO.ConclusionWearable sensor technology can enable COVID-19 detection during the presymptomatic period. Our proposed RNN algorithm identified 68% of COVID-19 positive participants 2 days prior to SO and will be further trained and validated in a randomised, single-blinded, two-period, two-sequence crossover trial.Trial registration numberISRCTN51255782; Pre-results.

ObjectivesThe sensitivity of molecular and serological methods for COVID-19 testing in an epidemiological setting is not well described. The aim of the study was to determine the frequency of negative RT-PCR results at first clinical presentation as well as negative serological results after a follow-up of at least 3 weeks.MethodsAmong all patients seen for suspected COVID-19 in Liechtenstein (n=1921), we included initially RT-PCR positive index patients (n=85) as well as initially RT-PCR negative (n=66) for follow-up with SARS-CoV-2 antibody testing. Antibodies were detected with seven different commercially available immunoassays. Frequencies of negative RT-PCR and serology results in individuals with COVID-19 were determined and compared to those observed in a validation cohort of Swiss patients (n=211).ResultsAmong COVID-19 patients in Liechtenstein, false-negative RT-PCR at initial presentation was seen in 18% (12/66), whereas negative serology in COVID-19 patients was 4% (3/85). The validation cohort showed similar frequencies: 2/66 (3%) for negative serology, and 16/155 (10%) for false negative RT-PCR. COVID-19 patients with negative follow-up serology tended to have a longer disease duration (p=0.05) and more clinical symptoms than other patients with COVID-19 (p<0.05). The antibody titer from quantitative immunoassays was positively associated with the number of disease symptoms and disease duration (p<0.001).ConclusionsRT-PCR at initial presentation in patients with suspected COVID-19 can miss infected patients. Antibody titers of SARS-CoV-2 assays are linked to the number of disease symptoms and the duration of disease. One in 25 patients with RT-PCR-positive COVID-19 does not develop antibodies detectable with frequently employed and commercially available immunoassays.

Pan-immunoglobulin assays can simultaneously detect IgG, IgM and IgA directed against the receptor binding domain (RBD) of the S1 subunit of the spike protein (S) of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2 S1-RBD Ig). In this work, we aim to evaluate a quantitative SARS-CoV-2 S1-RBD Ig electrochemiluminescence immunoassay (ECLIA) regarding analytical, diagnostic, operational and clinical characteristics. Our work takes the form of a population-based study in the principality of Liechtenstein, including 125 cases with clinically well-described and laboratory confirmed SARS-CoV-2 infection and 1159 individuals without evidence of coronavirus disease 2019 (COVID-19). SARS-CoV-2 cases were tested for antibodies in sera taken with a median of 48 days (interquartile range, IQR, 43–52) and 139 days (IQR, 129–144) after symptom onset. Sera were also tested with other assays targeting antibodies against non-RBD-S1 and -S1/S2 epitopes. Sensitivity was 97.6% (95% confidence interval, CI, 93.2–99.1), whereas specificity was 99.8% (95% CI, 99.4–99.9). Antibody levels linearly decreased from hospitalized patients to symptomatic outpatients and SARS-CoV-2 infection without symptoms (p < 0.001). Among cases with SARS-CoV-2 infection, smokers had lower antibody levels than non-smokers (p = 0.04), and patients with fever had higher antibody levels than patients without fever (p = 0.001). Pan-SARS-CoV-2 S1-RBD Ig in SARS-CoV-2 infection cases significantly increased from first to second follow-up (p < 0.001). A substantial proportion of individuals without evidence of past SARS-CoV-2 infection displayed non-S1-RBD antibody reactivities (248/1159, i.e., 21.4%, 95% CI, 19.1–23.4). In conclusion, a quantitative SARS-CoV-2 S1-RBD Ig assay offers favorable and sustained assay characteristics allowing the determination of quantitative associations between clinical characteristics (e.g., disease severity, smoking or fever) and antibody levels. The assay could also help to identify individuals with antibodies of non-S1-RBD specificity with potential clinical cross-reactivity to SARS-CoV-2.