Bayesian Prediction of Severe Outcomes in the LabMarCS: Laboratory Markers of COVID-19 Severity - Bristol Cohort

ObjectivesTo develop cross-validated prediction models for severe outcomes in COVID-19 using blood biomarker and demographic data; Demonstrate best practices for clinical data curation and statistical modelling decisions, with an emphasis on Bayesian methods. DesignRetrospective observational cohort study. SettingMulticentre across National Health Service (NHS) trusts in Southwest region, England, UK. ParticipantsHospitalised adult patients with a positive SARS-CoV 2 by PCR during the first wave (March - October 2020). 843 COVID-19 patients (mean age 71, 45% female, 32% died or needed ICU stay) split into training (n=590) and validation groups (n=253) along with observations on demographics, co-infections, and 30 laboratory blood biomarkers. Primary outcome measuresICU admission or death within 28-days of admission to hospital for COVID-19 or a positive PCR result if already admitted. ResultsPredictive regression models were fit to predict primary outcomes using demographic data and initial results from biomarker tests collected within 3 days of admission or testing positive if already admitted. Using all variables, a standard logistic regression yielded an internal validation median AUC of 0.7 (95% Interval [0.64,0.81]), and an external validation AUC of 0.67 [0.61, 0.71], a Bayesian logistic regression using a horseshoe prior yielded an internal validation median AUC of 0.78 [0.71, 0.85], and an external validation median AUC of 0.70 [0.68, 0.71]. Variable selection performed using Bayesian predictive projection determined a four variable model using Age, Urea, Prothrombin time and Neutrophil-Lymphocyte ratio, with a median AUC of 0.74 [0.67, 0.82], and external validation AUC of 0.70 [0.69, 0.71]. ConclusionsOur study reiterates the predictive value of previously identified biomarkers for COVID-19 severity assessment. Given the small data set, the full and reduced models have decent performance, but would require improved external validation for clinical application. The study highlights a variety of challenges present in complex medical data sets while maintaining best statistical practices with an emphasis on showcasing recent Bayesian methods.

Chronic SARS-CoV-2 infection and viral evolution in a hypogammaglobulinaemic individual.

There is widespread interest in the capacity for SARS-CoV-2 evolution in the face of selective pressures from host immunity, either naturally acquired post-exposure or from vaccine acquired immunity. Allied to this is the potential for long perm persistent infections within immune compromised individuals to allow a broader range of viral evolution in the face of sub-optimal immune driven selective pressure. Here we report on an immunocompromised individual who is hypogammaglobulinaemic and was persistently infected with SARS-CoV-2 for over 290 days, the longest persistent infection recorded in the literature to date. During this time, nine samples of viral nucleic acid were obtained and analysed by next-generation sequencing. Initially only a single mutation (L179I) was detected in the spike protein relative to the prototypic SARS-CoV-2 Wuhan-Hu-1 isolate, with no further changes identified at day 58. However, by day 155 the spike protein had acquired a further four amino acid changes, namely S255F, S477N, H655Y and D1620A and a two amino acid deletion ({Delta}H69/{Delta}V70). Infectious virus was cultured from a nasopharyngeal sample taken on day 155 and next-generation sequencing confirmed that the mutations in the virus mirrored those identified by sequencing of the corresponding swab sample. The isolated virus was susceptible to remdesivir in vitro, however a 17-day course of remdesivir started on day 213 had no effect on the viral RT-PCR cycle threshold (Ct) value. On day 265 the patient was treated with the combination of casirivimab and imdevimab. The patient experienced progressive resolution of all symptoms over the next 8 weeks and by day 311 the virus was no longer detectable by RT-PCR. The {Delta}H69/{Delta}V70 deletion in the N-terminus of the spike protein which arose in our patient is also present in the B.1.1.7 variant of concern and has been associated with viral escape mutagenesis after treatment of another immunocompromised patient with convalescent plasma. Our data confirms the significance of this deletion in immunocompromised patients but illustrates it can arise independently of passive antibody transfer, suggesting the deletion may be an enabling mutation that compensates for distant changes in the spike protein that arise under selective pressure.

Kinetics and performance of the Abbott Architect SARS-CoV-2 IgG antibody assay

ObjectivesTo assess the performance (sensitivity and specificity) of the Abbott Architect SARS-CoV-2 IgG antibody assay across three clinical settings. MethodsAntibody testing was performed on three clinical cohorts of COVID-19 disease: hospitalised patients with PCR confirmation, hospitalized patients with a clinical diagnosis but negative PCR, and symptomatic healthcare workers (HCWs). Pre-pandemic respiratory infection sera were tested as negative controls. The sensitivity of the assay was calculated at different time points (<5 days, 5-9 days, 10-14 days, 15-19 days, >20 days, >42 days), and compared between cohorts. ResultsPerformance of the Abbot Architect SARS-CoV-2 assay varied significantly between cohorts. For PCR confirmed hospitalised patients (n = 114), early sensitivity was low: <5 days: 44.4% (95%CI: 18.9%-73.3%), 5-9 days: 32.6% (95%CI, 20.5%-47.5%), 10-14 days: 65.2% (95% CI 44.9%-81.2%), 15-20 days: 66.7% (95% CI: 39.1%-86.2%) but by day 20, sensitivity was 100% (95%CI, 86.2-100%). In contrast, 17 out of 114 symptomatic healthcare workers tested at >20 days had negative results, generating a sensitivity of 85.1% (95%CI, 77.4% - 90.5%). All pre-pandemic sera were negative, a specificity of 100%. Seroconversion rates were similar for PCR positive and PCR negative hospitalised cases. ConclusionsThe sensitivity of the Abbot Architect SARS-CoV-2 IgG assay increases over time, with sensitivity not peaking until 20 days post symptoms. Performance varied markedly by setting, with sensitivity significantly worse in symptomatic healthcare workers than in the hospitalised cohort. Clinicians, policymakers, and patients should be aware of the reduced sensitivity in this setting.