ABSTRACT
The COVID-19 pandemic caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is an unprecedented event requiring frequent adaptation to changing clinical circumstances. Convalescent immune plasma (CIP) is a promising treatment that can be mobilized rapidly in a pandemic setting. We tested whether administration of SARS-CoV-2 CIP at hospital admission could reduce the rate of ICU transfer or 28-day mortality or alter levels of specific antibody responses before and after CIP infusion. In a single-arm phase II study, patients >18 years-old with respiratory symptoms with confirmed COVID-19 infection who were admitted to a non-ICU bed were administered two units of CIP within 72 h of admission. Levels of SARS-CoV-2 detected by PCR in the respiratory tract and circulating anti-SARS-CoV-2 antibody titers were sequentially measured before and after CIP transfusion. Twenty-nine patients were transfused high titer CIP and 48 contemporaneous comparable controls were identified. All classes of antibodies to the three SARS-CoV-2 target proteins were significantly increased at days 7 and 14 post-transfusion compared with baseline (P < 0.01). Anti-nucleocapsid IgA levels were reduced at day 28, suggesting that the initial rise may have been due to the contribution of CIP. The groups were well-balanced, without statistically significant differences in demographics or co-morbidities or use of remdesivir or dexamethasone. In participants transfused with CIP, the rate of ICU transfer was 13.8% compared to 27.1% for controls with a hazard ratio 0.506 (95% CI 0.165-1354), and 28-day mortality was 6.9% compared to 10.4% for controls, hazard ratio 0.640 (95% CI 0.124-3.298). IMPORTANCE Transfusion of high-titer CIP to non-critically ill patients early after admission with COVID-19 respiratory disease was associated with significantly increased anti-SARSCoV-2 specific antibodies (compared to baseline) and a non-significant reduction in Ku transfer and death (compared to controls). This prospective phase II trial provides a suggestion that the antiviral effects of CIP from early in the COVID-19 pandemic may delay progression to critical illness and death in specific patient populations. This study informs the optimal timing and potential population of use for CIP in COVID-19, particularly in settings without access to other interventions, or in planning for future coronavirus pandemics.
ABSTRACT
Background/Case Studies: COVID-19 convalescent plasma (CCP) is a promising therapeutic option, but efficacy remains unknown. In response to the emerging pandemic, multiple clinical trials assessing the efficacy and/or safety of CCP were registered in quick succession on ClinicalTrials.gov (CTG). However, concern is growing that several trials may not reach stated enrollment goals. We hypothesize that data harmonization among these trials may yield meaningful data that could not be obtained from underpowered individual trials. We have developed a multidimensional grid of relevant US clinical trials, with the goal of determining if data harmonization can be achievable by identifying common data standards. Study Design/Methods: All US trials registered on CTG as of 5/30/20 with CCP as treatment for COVID-19 patients were considered for inclusion. Studies were excluded if they were designed solely for compassionate use (expanded access), assessed the feasibility of CCP collection or delivery per se (no treatment endpoint), were exclusively in children, or were duplicates. Key parameters such as study design, patient population(s), and clinical and laboratory endpoints were extracted. Trials were grouped based on target patient populations and study design. Results/Findings: Between 3/22/20 and 5/30/20, 37 trials involving CCP were registered on CTG. 12 were excluded based on above criteria. The remaining 25 trials were organized based on patient populations and study design, yielding five distinct groups: (A) double blind randomized control trials (DBRCT) involving outpatients (N=4);(B) DBRCTs involving inpatients (N=6);(C) randomized open label trials involving inpatients (N=2);(D) single-arm open label trials involving inpatients (N=10);(E) single-arm open label trials involving both inpatients and outpatients (N=3). Among the 12 trials randomizing to two arms (groups A, B, C), the non-intervention arm varied (standard plasma, 5% albumin, LR, “standard care”). Among the 21 studies involving inpatients (groups B-E), both clinical severity of patients enrolled, and outcomes analyzed varied greatly. Conclusions: Meaningful groupings of CCP clinical trials based on common parameters were developed in the above framework. Overlaps as well as sources of heterogeneity among trials were identified. This framework allows the generation of preliminary recommendations to facilitate data harmonization across trials that may face accrual problems. Prospective pre-statistical harmonization of data across clinical trials by adoption of common data standards, including inclusion and exclusion criteria, study populations, intervention protocols, outcome measures and data dictionaries, can facilitate pooling of data and improve statistical power. Both trial-level and individual-patient-level meta-analyses may be employed as statistical strategies to analyze data.