Convalescent Plasma for the Prevention and Treatment of COVID-19: A Systematic Review and Quantitative Analysis

Background: The COVID-19 pandemic, caused by a novel coronavirus termed SARS-CoV-2, has spread quickly worldwide. Convalescent plasma (CP) obtained from patients following recovery from COVID-19 infection and development of antibodies against the virus is an attractive option for either prophylactic or therapeutic treatment, since antibodies may have direct or indirect antiviral activities and immunotherapy has proven effective in principle and in many clinical reports. Objective: We seek to characterize the latest advances and evidence in the use of CP for COVID-19 through a systematic review and quantitative analysis, identify knowledge gaps in this setting, and offer recommendations and directives for future research. Methods: PubMed, Web of Science, and Embase were continuously searched for studies assessing the use of CP for COVID-19, including clinical studies, commentaries, reviews, guidelines or protocols, and in vitro testing of CP antibodies. The screening process and data extraction were performed according to PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. Quality appraisal of all clinical studies was conducted using a universal tool independent of study designs. A meta-analysis of case-control and randomized controlled trials (RCTs) was conducted using a random-effects model. Results: Substantial literature has been published covering various aspects of CP therapy for COVID-19. Of the references included in this review, a total of 243 eligible studies including 64 clinical studies, 79 commentary articles, 46 reviews, 19 guidance and protocols, and 35 in vitro testing of CP antibodies matched the criteria. Positive results have been mostly observed so far when using CP for the treatment of COVID-19. There were remarkable heterogeneities in the CP therapy with respect to patient demographics, donor antibody titers, and time and dose of CP administration. The studies assessing the safety of CP treatment reported low incidence of adverse events. Most clinical studies, in particular case reports and case series, had poor quality. Only 1 RCT was of high quality. Randomized and nonrandomized data were found in 2 and 11 studies, respectively, and were included for meta-analysis, suggesting that CP could reduce mortality and increase viral clearance. Despite promising pilot studies, the benefits of CP treatment can only be clearly established through carefully designed RCTs. Conclusions: There is developing support for CP therapy, particularly for patients who are critically ill or mechanically ventilated and resistant to antivirals and supportive care. These studies provide important lessons that should inform the planning of well-designed RCTs to generate more robust knowledge for the efficacy of CP in patients with COVID-19. Future research is necessary to fill the knowledge gap regarding prevention and treatment for patients with COVID-19 with CP while other therapeutics are being developed. (JMIR Public Health Surveill 2021;7(4):e25500) doi: 10.2196/25500 JMIR Public Health Surveill 2021 | vol. 7 | iss. 4 | e25500 | p. 1 https://publichealth.jmir.org/2021/4/e25500 (page number not for citation purposes) Peng et al JMIR PUBLIC HEALTH AND SURVEILLANCE


Introduction
SARS-CoV-2, the cause of COVID-19, was declared a pandemic in early 2020 by the World Health Organization [1,2]. This is the third coronavirus to emerge in the past two decades, causing multinational outbreaks and carrying substantial morbidity and mortality [3]. COVID-19 is characterized by a spectrum of symptoms, ranging from mild subclinical infection with self-limiting respiratory tract illness (dry cough, fever, fatigue, difficulty breathing) to severe progressive manifestations (acute respiratory distress, hypercoagulation, hyperinflammation, multi-organ dysfunction, death) in high-risk patients with known comorbidities (advanced age, diabetes, obesity, cardiopulmonary disease) [4,5]. Case-fatality rates range from 4% to 50%, with higher mortality observed in the most critically ill [6]. Growing evidence also suggests that some patients with COVID-19, including those with milder symptoms, will have a prolonged course of recovery including fatigue, cognitive impairment, and cardiopulmonary dysfunction [7]. As such, COVID-19 represents an overwhelming universal health crisis [8], and the burden of this disease continues to threaten lives and livelihoods worldwide [9]. As SARS-CoV-2 and its emerging new mutant strains (which may be associated with an increased efficiency of viral replication, transmission, and virulence in humans) continue to spread globally, international research efforts are being accelerated to identify effective preventive and therapeutic approaches to mitigate its impact [10][11][12].
The magnitude and urgency of this public health emergency has prompted global scientific collaborations to seek rapid solutions via repurposing of previously approved broad-spectrum antivirals (remdesivir, ritonavir, hydroxychloroquine, interferon) [13,14] and therapeutic doses of corticosteroids (dexamethasone, hydrocortisone, methylprednisolone) [15,16] for high-risk patients while fast-tracking development of vaccines and other novel therapeutics [17]. To that end, great advances in understanding the biology of this new coronavirus and the natural history of the disease have been achieved [18,19]. Moreover, the unprecedented development of multiple COVID-19 vaccines capable of eliciting immunological protection, in less than a year from identification of the causative agent, has been a remarkable success and remains the best hope for ending this pandemic [20].
Despite this incredible progress on COVID-19, many challenges remain post vaccine development including ongoing vaccine deployment, large-scale production and distribution of billions of vaccine doses [21], and uncertainty over the effectiveness of current vaccines against more transmissible new variants [22]. These factors, combined with public hesitation around vaccination, have casted doubt on the likelihood of achieving worldwide herd immunity in the near future [23]. Consequently, other therapeutic strategies to impair virus infection or to counteract further disease spread are still needed, at least until more effective drugs are available or vaccines are distributed and administered to everyone [24].
In the absence of definitive treatment against this new human pathogen, clinical management of hospitalized, severely ill patients remains mainly supportive care, including oxygen and mechanical ventilation, and is based largely on preclinical studies or previous experience with severe acute respiratory syndrome-related coronavirus (SARS-CoV) [25]. Thus, an effective evidence-based therapeutic intervention is urgently needed to reduce the morbidity, mortality, and length of in-hospital stay for patients with COVID-19.
Passive immunotherapy with convalescent plasma (CP), hyperimmune γ-globulin, or artificially produced monoclonal antibodies are beneficial for treatment or prophylaxis of several infections, and these approaches are under investigation as potential therapeutic modalities for the management and prevention of COVID-19 [26]. Passive immunotherapy with human convalescent blood products, in particular CP, is a promising strategy for the prevention and treatment of COVID-19 [27][28][29]. Although further research is needed to determine the utility of immunotherapy with CP or monoclonal antibodies for the treatment of patients who are symptomatic and potentially for use as postexposure prophylaxis, initial findings in limited clinical trials suggest these interventions are safe and can be effective, particularly when administered early in the course of treatment [29]. Experience suggests that CP therapy could be used as an empirical treatment modality to prevent further progression or promote early recovery in patients who are critically ill with COVID-19 [30,31]. CP has been used safely for decades to treat infectious diseases where no specific treatment is available [32,33]. In the late 19th and early 20th century, CP was given to treat a wide range of viral infections, including diphtheria, polio, measles, mumps, and Spanish influenza A (H1N1) [34][35][36]. Although no randomized trials were conducted, a retrospective meta-analysis of studies on the use of CP during the Spanish influenza flu pandemic showed a significant decrease in mortality in patients who received CP versus those given plasma from unexposed donors [37]. After World War II, plasma became a valuable pharmaceutical component, which used it for diverse products to successfully treat everything from bleeding disorders to immune deficiencies to hypovolemic shock [38]. Since then, CP has been used in outbreaks of Ebola and other coronavirus diseases including SARS-CoV and Middle East respiratory syndrome-related coronavirus (MERS-CoV) infection with varying efficacy [33]. CP was proven to be efficacious in patients with severe 2009 pandemic H1N1 flu, reducing respiratory tract viral load, serum cytokine responses, length of hospital stay, and patient mortality [39]. CP therapy involves transfusing whole or fractionated plasma, collected from patients that have recently recovered from SARS-CoV-2 infection, to confer passive humoral immunity in people who are infected or at risk of infection [29,40]. Furthermore, CP therapy has advantages over other proposed treatment: it requires low technology (and therefore it can be produced where required independent of pharmaceutical companies), it is low cost and has strong biological plausibility, and it has potential for rapid development and deployment (production is easily scalable as long as there are sufficient donors) [41][42][43]. Accordingly, on March 24, 2020, the Food and Drug Administration (FDA) approved the use of CP therapy as an emergency investigational new drug to treat patients with serious or immediately life-threatening COVID-19 infections [44]. Subsequently, on August 23, 2020, the FDA issued an Emergency Use Authorization (EUA) for CP for treating COVID-19 [45]. According to the FDA regulation, the plasma must be collected from recovered patients who can donate blood, have had no symptoms for 14 days, and have had negative results on COVID-19 tests. Both single-donor and pooled immuno-globulin products currently prioritize collection of convalescent donor plasma with high levels of neutralizing antibodies. Based on the preliminary data from clinical trials and considering the United States National Institute of Health and FDA recommendation, remdesivir and CP are the most promising potential for COVID-19 treatment [46]. CP for treating COVID-19 is accessible via the regulatory pathways (investigational new drug regulatory pathway). Another is expanded access, also called "compassionate use" emergency Investigational New Drug Application (an investigational medical product), to treat patients [47]. It should be noted that, currently, Regeneron's REGN-COV2 and Lilly's LY-CoV555, both of which are cocktail therapies comprising receptor binding domain (RBD)-reactive antibodies, have also been granted EUA for COVID-19 by the FDA [48,49].
On the other hand, systematic reviews have been conducted for current medications that have been used for the treatment of COVID-19. A comparative analysis of three treatment modalities for COVID-19, chloroquine and hydroxychloroquine, CP, and remdesivir, found that each modality had both favorable and unfavorable characteristics, but none showed clear evidence of benefit for early outpatient disease or prophylaxis; in particular, chloroquine or hydroxychloroquine is no longer a viable option [50], while CP therapy appeared to show clinical advantages for inpatient use [14]. Moreover, meta-analysis of the safety and efficacy of various interventions including the three treatments and dexamethasone or lopinavir-ritonavir showed that dexamethasone and remdesivir might be beneficial for patients with COVID-19, but the certainty of the evidence was low to very low, so more trials are needed [51].
Studies are currently underway to evaluate use of CP as treatment for patients with severe COVID-19 and to prevent infection (prophylaxis) in certain high-risk patients exposed to COVID-19. Currently, CP is being given to small numbers of hospitalized patients with severe or life-threatening COVID-19 illness [52]. Several case reports suggest treatment is helpful, but larger studies are still needed. Although there is a lot that is unknown, CP may work best for patients earlier in the disease course [53,54]. Therapy using CP may also be beneficial for prophylaxis against SARS-CoV-2 in individuals who are at high risk; there is considerable interest to leverage CP for frontline health care workers, first responders, other caregivers, and vulnerable individuals with underlying medical conditions [55,56]. This strategy has been previously used in SARS-CoV and MERS-CoV outbreaks [57]. Although the evidence for CP therapy remains inconclusive, preliminary trials for CP suggest that there may be some benefits, and there is growing consensus that CP is an important first-line immunotherapy for emerging viral infections when other specific treatments are not available [58]. Currently, several countries and health institutions are collecting CP for either empirical treatment or clinical trials [55,59]. However, research to date is at a high risk of bias, and randomized control trials are desperately needed to determine the efficacy and safety of this therapeutic option.
There are many ongoing trials and reviews, perspectives, commentaries, and guidelines published every day related to all aspects of COVID-19 CP, ranging from donor selection, plasma collection, testing, and storage to clinical use. In this paper, we sought to review all aspects of CP use for COVID-19, from detection of the level and activity of CP antibodies to appraisal of the quality and meta-analysis of original clinical studies of CP therapy, to characterize the knowledge gap and provide recommendations for future directions.

Data Extraction and Study Appraisal
All literature search results were screened independently by two reviewers. The commentaries in support of the use of CP for COVID-19 were considered positive, those suggesting improvements in CP treatment were categorized as neutral, and precautions against CP were determined to be negative. The review type was determined according to a typology of reviews by Grant and Booth [61]. The quality appraisal of included clinical studies was conducted using the Effective Public Health Practice Project (EPHPP) Quality Assessment Tool [62]. Specifically, each clinical study was evaluated for the following components: sample selection, study design, identification and treatment of confounders, blinding of outcome assessors and participants, reliability and validity of data collection methods, and withdrawals and dropouts. The overall rate of each study was determined by assessing the six component ratings. Those with no weak ratings and at least 4 strong ratings were rated strong. Those with less than 4 strong ratings and 1 weak rating were considered moderate. Those with 2 or more weak ratings were rated weak.

Analyses
Studies were analyzed separately according to their design (case report, case series, observational, or randomized trials). Clinical and methodological heterogeneities across the studies were assessed by examining the details of the patients, the baseline data, the interventions, and the outcomes to determine whether the studies were sufficiently similar.
For disease severity, severe COVID-19 is a clinical situation in which the patient has dyspnea, tachypnea (respiratory rate≥30 breaths/minute), blood oxygen saturation≤93% on room air, partial pressure of arterial oxygen to fraction of inspired oxygen ratio <300 PaO 2 /FiO 2 <300, or lung infiltrates >50% within 24-48 hours on chest x-ray [63]. Life-threatening disease is defined as respiratory failure, septic shock, or multiple organ dysfunction or failure [63].
Case and randomized controlled studies were combined in meta-analyses using Review Manager (Version 5.4, The Cochrane Collaboration). Data were pooled using an inverse variance method and analyzed using a random-effects model, as this approach accommodates clinical and statistical variations. Odds ratios (ORs) and 95% CIs were used as statistical measures for mortality, clinical improvement, and viral clearance as a dichotomous outcome. Mean and SD were the statistical measure used to describe length of hospital stay. In studies that reported data in medians and IQRs, mean and SD were estimated using the sample size in each study arm, the medians, and the first and third IQRs as demonstrated in the method published by Wan et al [64]. Heterogeneity was determined using the I 2 statistic and the chi-square test. High values of both tests (I 2 >40%, a significant chi-square value with P<.05) demonstrate high levels of inconsistency and heterogeneity.  The literature search was conducted on  June 19 and updated on October 22, 2020. The screening, full-text review, and extraction were managed online using Covidence. CP: convalscent plasma.  [194,195]

•
Wiwanitkit 2020 [196] This group of commentaries suggested that the risks associated with CP use (eg, adverse effects and blood-borne pathogen transmission) outweighed its benefits or other therapeutics for COVID-19.
Negative 46 different types of reviews (a total of 10 review types with unique features in terms of prescribed and explicit methodologies) on CP for treatment of virus infectious diseases (eg, SARS d , MERS e , EBOV f , and H1N1) and COVID-19 with safety and efficacy as main outcomes and recommendations. Some reviews also covered other aspects related to CP use, such as SARS-CoV-2 immunology, mechanism of action, CP donor selection, CP collection, pooling technologies, pathogen inactivation systems, banking of CP, timing and dose of CP treatment, patient selection, risk-benefit analysis, and list of ongoing registered clinical trials. Muruato et al 2020 [265] In vitro evaluation of CP potency for COVID-19 treatment could be measured by its binding capacity to the SARS-CoV-2 spike protein and neutralizing activity against pseudotyped and chimeric viruses and authentic SARS-CoV-2, which is useful to identify donors with high titers for CP for COVID-19 therapy. There were individual differences in the antibody level (neutralizing antibody titers <1:16 to >1:1024) and its changes over 12-60 days since onset of symptoms among representative convalescent patients. All clinical studies are therapeutic use of CP focusing on safety and efficacy, and they are further reviewed in the following section. The commentaries cover various aspects of CP, ranging from critiques of clinical studies [131,137,148,163,176,189] and literature review [145,221] to the stability of antibodies in CP [156,291], relevant news [180], and a response letter [164], while a majority focused on the safety and efficacy of CP. Most commentaries were in favor of CP therapy for COVID-19, recognizing the need for more high-quality evidence from large and well-designed clinical trials to show its efficacy, and other issues (eg, CP collection) still need to be addressed. Some commentaries proposed alternative or complementary CP-based approaches to COVID-19 that possess fewer risks [178,182] but may not be immediately available for clinical use. Only a few commentaries put more emphases on the potential risks over benefits of CP therapy [191][192][193][194]196].
In a particular correspondence, a metadata analysis of the efficacy of CP treatment based on 9 clinical studies (mostly case series) suggested that CP reduced viral loads (risk ratio 0. 13 [147]. In addition, the effects of CP on C-reactive protein levels and clinical improvement were not associated with the patient's age and the use of antivirals, antibiotics, and hydroxychloroquine. Several commentary papers and reviews advocated for the rationale of developing fast access to CP collection and treatment of patients with COVID-19 [34,47,59,148,199,229]. Among the reviews, most were descriptive overviews of existing literature and recommendations for clinical use and trial without any search strategies. Few were conducted following the PRISMA guidelines [222,224,225]. It is noteworthy that 1 systematic review and meta-analysis was on the safety and efficacy of CP therapy for other severe respiratory viral infections to provide indirect evidence for CP therapy for COVID-19 [206], and another 2 systematic reviews and meta-analyses were on completed and ongoing clinical studies on the safety and efficacy of CP or hyperimmune immunoglobulin transfusion in the treatment of COVID-19 [207,208]. One review and meta-analysis included 20 studies (1 RCT, 3 controlled nonrandomized studies of interventions, 16 noncontrolled nonrandomized studies of interventions) with 5443 participants [207]. The meta-analysis of 4 controlled studies (1 RCT and 3 controlled nonrandomized studies of interventions) with 339 patients could not support any effects of CP treatment on all-cause mortality at hospital discharge, time to death, or improvement of clinical symptoms at 7 days. The review also investigated the safety of CP based on 14 studies (5201 participants, with 5000 participants from 1 noncontrolled nonrandomized studies of intervention) and found very low-certainty evidence for safety. The review was recently updated, which included 19 studies with 36,081 patients treated by CP, and made the same conclusion [205]. The other review included 7 studies, including 2 RCTs and 5 cohort studies, with a total of 5444 patients [208]. The meta-analysis indicated that CP therapy reduced mortality and increased viral clearance and clinical improvement. It confirmed the safety of CP transfusion with very low incidence of serious adverse events. However, the risk of bias and quality assessment in both reviews indicated that the evidence for the efficacy and safety of CP therapy was of low quality, suggesting the need for a large well-designed RCT. In addition, a survey has been conducted for current registered clinical trials of CP therapy for COVID-19, including a description of their characteristics such as study design, patient populations, outcomes, eligibility criteria for CP donors, CP collection, antibody titer, and CP dose [218].
Protocols, programs, and standards have been developed to select donors and collect, process, characterize, store, distribute, and apply CP to patients in need [238,240,242,250], and to conduct clinical trials [239,246,247,254]. Regional and national programs for COVID-19 CP have been established [243,244] as well as a multi-criteria decision-making frame for both CP donor and receipt selection [241].
Some key findings and implications from the in vitro testing studies of CP antibodies should be considered: a variety of methods have been developed to measure CP antibody titers including gold standard neutralization assay using living SARS-Cov-2 [261,262]; enzyme-linked immunosorbent assay (ELISA) using the antigens derived from the virus, mostly in a microplate platform [257,258] and a few in lateral flow [266], microsphere [267], and microarray platforms [268]; and other methods (eg, polymerase chain reaction [PCR] tests) [272,273]. A number of studies showed a wide range of levels and neutralizing activities of anti-SARS-CoV-2 [264,267,289]. The neutralizing antibody levels declined within the first 3 months following diagnosis, suggesting a short optimum time window for the collection of CP with high neutralizing antibody titers [280]. A significant decrease was also observed in the antibody binding to the spike protein of SARS-CoV-2 and neutralizing capacity of plasma from convalescent donors at 6 and 10 weeks after symptoms onset [261]. The short duration of neutralizing antibody titers within months may have important implications for immunity and ongoing efforts to deploy CP for prevention and therapy of COVID-19 [165]. There is a significant correlation to various extents between ELISA-measured immunoglobulin (IgG) titer and neutralizing antibody titer [87,257,274,276,[278][279][280][283][284][285]288,289]. However, the ELISA-determined anti-SARS-CoV-2 IgG did not always inhibit the virus receptor binding [259]. Antibody binding to SARS-CoV-2 spike glycoprotein as measured by pseudovirus capture assay did not always translate into neutralization [261].
Highly sensitive and specific platforms for the detection of anti-SARS-CoV-2 antibodies are becoming increasingly important for evaluating potential CP donors and identifying individuals with seroconversion [292]. Various platforms demonstrate significant correlations with a SARS-CoV-2 plaque reduction neutralization assay, suggesting their use for screening of individuals who have recovered from SARS-CoV-2 infections. Notably, a novel multiplexed solid-phase chemiluminescence immunoassay has been developed and commercially available from Meso Scale Discovery for simultaneous detection of IgG binding to four SARS-CoV-2 antigens (trimeric spike, spike RBD, spike N terminal domain, and nucleocapsid antigen) and the quantification of antibody-induced angiotensin-converting enzyme 2 (ACE-2) and ACE-2-binding inhibition (pseudo-neutralization assay) [293].
In addition to neutralization and immune assays, biophysical and functional evaluation of CP showed that it may have diverse antiviral effects against SARS-CoV-2 beyond neutralization, namely, antibody-dependent cellular cytotoxicity, phagocytosis, and complement activation [290]. Moreover, CP could act not only on the viral infection but also on the antithrombin deficiency to reduce thromboembolic events [133].

Findings of Clinical Studies
As summarized in Table 2, there are considerable heterogeneities among the clinical studies in terms of the populations, the amount of CP received, and a variety of comparators. The CP therapy studies differed in the following aspects: patient demographics (eg, age, gender, and comorbidities), donors' selection (ie, age, gender, diagnosis of SARS-CoV-2 infection and of recovery, and anti-SARS-CoV-2 antibody titer required for plasma donation), plasma collection and biologic qualification (number, volume and frequency of donations, infectious disease markers, and pathogen inactivation), and treatment and disease characteristics (dose and timing of administration, stage of the disease at which to start CP treatment). The patient was transfused with 300 mL of CP on day 7 from onset of symptoms and another 300 mL of CP on day 12, and treated with antibiotics, low-molecular-weight heparin, hydroxychloroquine, and methylprednisolone Not reported A woman 29 years of age at 24 2/7 weeks of gestation Grisolia et al 2020 [71] Not reported The patient became afebrile and was tested negative for SARS-CoV-2 the following day after CP therapy, gradually improved and was weaned from the ventilator and discharged alive from the ICU on day 63 A total of 900 ml of CP was transfused at a slow infusion rate on day 31 after admission and treatment with a respirator, muscle relaxants, and antibiotics Obtained from two blood donors with one being diagnosed with highlevel anti-SARS-CoV-2 IgG antibody A previously healthy man in his 70s with severe COVID-19 admitted to ICU Hahn et al 2020 [72] Not reported The patient showed rapid improvement in symptoms and electrocardiogram findings, and was discharged 36 hours after the transfusion Im et al 2020 [74] Not reported The patient showed dramatic clinical and radiologic improvements and was discharged 2 weeks after admission with no infection of the newborns One unit of CP was transfused on the sixth day after hospital admission in addition to favipiravir and oxytocine Not reported A woman 26 years of age with a twin pregnancy at 36 weeks and 1 day gestation with confirmed COVID-19 Jafari et al 2020 [75] Not reported The patient's body temperature became normal and chest CT f was significantly better than at admission, and the patient was discharged on day 30 200  Hartman et al 2020 [63] No adverse effects except for a transient transfusion reaction (fever and hematuria) within 2 h of CP infusion in 1 patient 24 (63%) recovered and were discharged from the hospital, and 14 (37%) died. The survival patients received CP earlier in their course of disease (mean 15.3, SD 6.9 days) and hospital stay (mean 8.4, SD 6.8 days) compared to those who died with mean durations of 24.5 (SD 9.6) days and 16.6 (SD 9.5) days, respectively.
ABO-compatible CP was given in two consecutive 200-mL infusions (mean 18.7, SD 9.0) days following symptom onset. Another unit of CP was given to those with undetectable anti-SARS-CoV-2 antibodies.  [99] 1 patient experienced acute chest pain and dyspnea but improved over the following 12-24 h.
All showed clinical improvement and were discharged 9, 16, and 25 days after hospital admission with no evident infectious complications One or two units of CP were given on day 2, 4, or 7 after hospital admission, in addition to immunosuppressant/ antiviral/antibiotic Collected from donors at local and regional blood centers 3 kidney transplant recipients with COVID-19 treated with CP (1 female 65 years of age admitted to the general medicine service and a female aged 35 years and a male 36 years of age in the ICU) Naeem et al 2020 [100] No side effects Improvement in overall respiratory function and clinical condition over a period of 8 days, with 6 discharged and 2 died Olivares-Gazca et al 2020 [101] No adverse events except a fever during transfusion in 1 patient, resulting in infusion of only 100 mL All patients showed a decline in oxygen needs and ventilatory support with most effects seen in patients when CP was administered early in their disease course A single unit of 200 mL CP was given at an average time of 12 (range 4-41) days from illness; 3 patients received two units roughly 8 days apart in addition to other COVID-19 treatment and chemotherapy as required Collected from donors 18-56 days following full recovery from COVID-19 with anti-SARS-CoV-2 spike protein IgG titers 1:400-1:6400 as measured by ELISA 17 critically ill patients (mean age 56, range 24-81 years; 10 males) with COVID-19 and most patients had multiple medical comorbidities, including 6 with hematological malignancies Pal et al 2020 [102] No apparent transfusion-related adverse reactions 8 patients had de-escalating oxygenation support by day 7 post CP. 9 patients were discharged, 1 still hospitalized, and 3 patients died 3 months after the CP transfusion.
All patients received two ABO-compatible units of CP, for a total of 500 mL, at a median time of 8 days from symptom onset and additional therapies (hydroxychloroquine alone or in combination with azithromycin, steroids, anticoagulation, and immunosuppression) Tremblay et al 2020 [106] No adverse reactions Within 6 days after CP therapy, all patients became negative for two consecutive nucleic acid tests. Additionally, 4-9 days following the CP, 3 patients showed resolution of pulmonary lesion. 2 recovered and 3 died. 200 mL of crossmatching CP was transfused over 15 min initiated at a median of 37 (IQR 34-44) days from the onset of symptoms. In total, 3 patients received 400 mL and the other 2 received 1200 mL; all received antibiotics, antiviral, and anti-inflammatory agents.
Collected from the recently cured patients whose antibody titers were above 1:640 Xi et al 2020 [110] No adverse events A resolution of ground-glass opacities and consolidation in 5 out of 6 patients and an elimination of the virus in 2 in the following days of CP therapy One to three doses of ABO-compatible CP (200 mL/dose) at 6-31 days after admission. Each transfusion was administered over a 30-minute period.
Collected from patients at least 3 weeks following disease onset, two consecutive negative RT-PCR tests, and seropositive for anti-SARS-CoV-2 IgG and IgM 6 laboratory-confirmed critically ill COVID-19 patients (mean age 58, SD 16.4 years; 3 male) Ye et al 2020 [111] No adverse events The time from transfusion to negative RT-PCR test results ranged from 3 to 22 days. 3 were discharged from the hospital, and 1 remained in ICU up to the time of this writing One to eight doses of CP (200-2400 mL in total) 11-41 days after admission in addition to antiviral therapy Prepared from recovered patients without details 4 critically ill patients infected with SARS-CoV-2 (age: 31-73 years; 2 male) Zhang et al 2020 [112] No adverse events 6 of 8 patients showed an improvement in oxygen support status within 5 days from CP treatment, partial resolution of pulmonary lesions, and decreased viral load ABO-compatible and cross-matched CP were administered at one (3 patients)  ABO-compatible CP was transfused at approximately 4-13 mL/kg of recipient body weight and at approximately 10 mL for the first 15 minutes, which was then increased to approximately 100 mL per hour with close monitoring Collected based on routine plasma collection procedures via plasmapheresis from adults aged 18-55 years that were suitable for blood donation, initially diagnosed with COVID-19 but with 2 negative PCR results from nasopharyngeal swabs (at least 24 h apart) prior to hospital discharge, discharged for ≥2 weeks from the hospital, and had no persisting COVID-19 symptoms. CP S-RBD-specific IgG titer ≥1:640 correlating to serum neutralization titre of 1:80 103 patients (median age 70 years; 60 males, 58.3%) with severe and life-threatening COVID-19 randomized to receive CP in addition to standard treatment (n=52) or standard treatment (antiviral medications, antibacterial medications, steroids, human immunoglobulin, Chinese herbal medicines, and other medications) alone (control; n = 51) Li et al 2020 [126] a CP: convalescent plasma. b PCR: polymerase chain reaction.

Patient Demographics
A total of 36,379 patients, with most patients (n=35,322) from a single study [97], have been treated with CP in all clinical studies included in this review. There is a patient heterogeneity across the clinical studies in terms of age (ranging from infant [81] and 6 [70] to 100 years [79]), gender, and different underlying diseases, in particular hypertension and diabetes [114,122,124,294]. Some case studies investigated CP therapy for COVID-19 in patients who were immune compromised or deficient [56,80,94,100,103,125].
A few studies reported the antibody titers of patients before CP transfusion, which varied from undetectable IgG RBD antibody levels (<1:50 serum dilution) to extremely high levels (1:25,600) [88]. Studies suggested that patients with low antibody levels may benefit more from CP therapy [88,125].

Donor Selection and CP Antibody Titer
Most individuals with previous laboratory-diagnosed SARS-CoV-2 infection developed measurable antibody responses and neutralizing antibodies. There is evidence for a significant decline in neutralizing antibody levels over time [280].
Studies suggest that the efficacy of CP depends on the antibody levels of the donor plasma and CP, with high antibody levels possibly conferring immediate immunity to recipients [122].
One key factor associated with CP therapy is the neutralizing antibody titer, and when the infused plasma has a high antibody titer, it may be of the greatest benefit [88,97,99,113]. Hence, it may be a prerequisite to find eligible donors who have high levels of neutralizing antibody.
Prior smaller studies have reported a variety of titer cut-offs [105,115]. The FDA has recommended that CP with a virus neutralizing antibody titer of ≥1:160 be used for therapeutic transfusion [295]. Recently, the FDA has updated its EUA to limit the authorization to the use of high titer CP for the treatment of hospitalized patients with COVID-19 early in the disease course and to those hospitalized who have impaired humoral immunity and cannot produce an adequate antibody response, and include additional tests to be used in the manufacture of COVID-19 CP [296]. Studies have reported the levels of CP antibody titer, ranging from no minimum neutralizing antibody titer level [96] to 1:640 [115], and an even wider range of RBD-specific IgG titer, from <1:160 to >1:6000 within the same study [88].
There was substantial heterogeneity in the antibody response among potential CP donors, but sex, age, and hospitalization emerged as factors that can be used to identify individuals with a high likelihood of having strong antiviral antibody levels [297]. In vitro testing of CP showed a tendency of higher neutralizing antibody titers from donors with increased disease severity, of advanced age, and of male sex; however, the clinical relevance of this difference needs to be investigated [109,270,276,277,283,284]. Moreover, pooling CP samples from many donors may prove more effective for increasing and standardizing anti-SARS-CoV-2 neutralizing antibody titers [19].
In addition, CP collection efforts should be organized around the temporal dynamics of the immune response to viral clearance and a rise in neutralizing antibody titer, with a recommended window for plasma collection beginning at 4 weeks after the resolution of symptoms and narrowing rapidly by 12 weeks [165].

Timing and Dose
One key factor associated with CP efficacy is the optimal treatment time point [115]. The phase of the disease that this treatment modality may be most beneficial is still a matter of some debate, with early versus intermediate-late stages of the cytokine storm reaction being associated with acute respiratory distress syndrome or other severe disease complications [298].
There was no therapeutic effect from CP treatment on severely or critically ill patients with COVID-19 more than 2 weeks after the onset of disease as reported by Liu et al [117]. However, CP therapy has been limited to patients with severe or critical COVID-19. The majority of patients were severe or critically ill with COVID-19, with only a few mild cases [79,90,109].
Similar to most viral illnesses, viremia in COVID-19 peaks in the first week of infection, and the primary immune response develops by days 10-14, which is followed by virus clearance. Therefore, transfusion of CP at the early stage of disease theoretically should be more effective [114,121,124]. CP appears to be of greater clinical benefit when administered early in the course of disease than delaying transfusion under the development of severe disease [63,108]; in principle, the course of disease does not exceed 3 weeks [67]. Studies have found that, regardless of COVID-19 severity at time of transfusion, patients that received CP earlier in their course of disease showed lower mortality, more rapid viral clearance, and shorter hospital stays [92,113].
Based on the current findings, CP treatment should be given to patients with COVID-19 at the right phase or severity of illness and at the right time point. It is known that most patients with mild COVID-19 can recover without treatment, and CP may be an improper therapy for those patients. For patients with end-stage COVID-19, treatment with CP may be unable to avert a poor outcome, as demonstrated by the current findings [108,124,294]. Therefore, CP treatment may be more beneficial if used in patients who are potentially critically ill with COVID-19 at an early stage of the disease. Thus, early recognition of patients with COVID-19 who are likely to become critically ill is important for timely treatment with CP [124]. This is in line with one of the first published RCTs of CP, in which Li and colleagues [126] found that clinical improvement was limited to those without life-threatening disease, with 91% improvement in the plasma group compared to 68% in the control arm [294]. A large multicenter study involving 35,322 patients found significant reductions in 7-and 30-day mortality with early use of CP containing high levels of SARS-CoV-2-specific IgG antibodies in a subset of patients [97].
Transfusion volume ranged from 2x50 mL [110] to 8x300 mL [112]. Total antibody dose could be calculated as the transfused volume of CP multiplied by SARS-CoV-2 neutralizing antibody titer. CP dose has also been recognized as a key characteristic that may influence CP-associated outcomes [187]. One study showed that patients transfused with 400 mL of CP tended to turn faster to viral clearance than those who received 200 mL of CP [113].

Safety
All studies that assessed adverse events have reported no or minimal adverse events [102,206]. Of major interest is one of the first large trials published so far-concerning the safety of 5000 recipients-that has identified only limited and nonunexpected transfusion complications [95]. The case series study focused on the safety of CP transfusion in COVID-19 reported that, out of 5000 patients, there were 7 transfusion-associated circulatory overload (TACO), 11 transfusion-related lung injury (TRALI), and 3 severe allergic reactions. However, the reported low incidence of serious adverse effects might be due to an extremely short time frame of observation (4 hours) [194]. The latest update of the study  involving 20,000 hospitalized adults with severe or  life-threatening COVID-19 further demonstrated low adverse  events because of the treatment, with 36 TACO, 21 TRALI, 21 severe allergic reactions, and 38 transfusion-related thromboembolic events [96]. Consistently, other studies reported no to minimal adverse events. Half of the case reports that assessed the safety of CP did not indicate any adverse events or complications related to its use. One case series study reported 5 serious adverse events in 4 out of 46 patients [118]. The controlled studies reported 15 adverse events out of 695 patients. Overall, among a total of 20,749 patients reported with safety data, the incidence of adverse events related to CP transfusion was less than 0.8%, comparable or even lower than the incidence of adverse events related to plasma transfusions in other clinical settings [299]. There has been no evidence so far of antibody-mediated enhancement of disease in patients with COVID-19 treated with CP despite the concern that this might be a possibility in the presence of reactive but nonneutralizing antibodies against SARS-CoV-2 [170].
Although it is not yet clear whether the SARS-CoV-2 virus is transmitted by blood [300], donor selection criteria in compliance with existing policies and routine procedures should be met and pathogens reduction by solvent-or detergent-based treatments or light-based methods (especially for noncovered or detected in screening tests) should be performed in each donated plasma product as a standard for any plasma production [157,230]. Ultraviolet light and riboflavin used in the pathogen reduction process could effectively reduce SARS-CoV-2 in plasma and blood products without decreasing the quality of the blood products [301]. More studies have shown that the pathogen reduction processes did not alter neutralizing antibodies [156,272].

Outcomes
These were measured by SARS-CoV-2 negative PCR tests, improvements of clinical symptoms assessed by respiratory distress and fever, computed tomography, time to death, length of hospital stay, and mortality at discharge.
All case reports showed either viral load decrease/clearance or different extents of improvements of clinical symptoms with no mortality. Preliminary evidence from case reports and case series is favorable, as significant clinical and biochemical improvement and hospital discharge have been reported.
COVID-19 severity and underlying diseases affected the outcome of CP treatment. A patient with lymphoma who underwent autologous stem cell transplantation showed persistent SARS-CoV-2 viral shedding for 74 days, even with the administration of CP [77]. On the other hand, 1 study reported that 2 patients with long-term positive viral infection for 8 weeks showed substantial improvement after treatment with CP and ritonavir-boosted danoprevir [107]. Similarly, another study showed that CP therapy could rapidly reduce viral loads in more than half of 27 patients with prolonged positivity of SARS-CoV-2 for a median of 44 days after symptom onset [109]. It should be noted that most of these patients had mild COVID-19 symptoms.
Studies demonstrated that CP could effectively improve the respiratory symptoms of severe patients and help them wean from oxygen support. However, patients in extremely critical or life-threatening conditions could not benefit from CP [63,122,124,294].
The case series reported a mortality rate of 24.4% in 35,666 patients, mainly from 1 study with 35,322 patients [97]. The case-control and randomized controlled studies included a total of 2289 patients in the control group and 695 patients in the CP group, and reported a total of 219 (9.6%) and 63 (9.1%) deaths in each group, respectively. The number of patients and the mortality rates varied remarkably among these studies, from 6 [124] to 1430 patients [122] and from 0% [115] to 93.3% [124], respectively. The mortality at discharge [114] or at 28-day posttransfusion [121,294] have been reported as a primary outcome. Some studies showed improved survival for the CP group compared to its control [115,117,122], more clinical improvements [115,117], and viral clearance [115,124]. The efficacy of CP on mortality, length of hospital stay, clinical improvement, and viral clearance was further analyzed by meta-analysis of controlled studies, as presented later.

Quality Assessment of Clinical Studies
As indicated in Table 3, 52 clinical studies showed overall weak quality, 9 had moderate quality, and 1 had strong quality. Patients often had underlying medical conditions (hypertension, diabetes). Case reports and series with limited number of patients were considered weak for selection of participants (high risk of selection bias). Some studies included only males with a total of 3 patients [117] or only pediatric patients with fewer than 4 children [70,88] and therefore were judged to be weak for sample selection. Studies that targeted a specific group (eg, older populations, median age >60 years) were rated with moderate selection bias [122,124,125,294], while studies that selected patients with a broad range of ages and balanced gender and comorbidities [114,121] were ranked as strong.  [126] With respect to the study design, case reports and series were considered to be weak; case-control studies and RCTs were determined to be moderate and strong, respectively. The confounders for case reports and series studies were ranked weak given the uncontrolled nature of these studies involving other therapeutic treatments and supportive care and the use of other treatment regimens, including antiviral medications along with CP transfusion. Two different analytical methods were used to control for confounding in 1 case series study [97] subsequently determined to be of moderate risk for confounders. This component was ranked to be strong for RCTs and moderate for case-control studies, except for 1 study by Duan et al [115] given the uncertain characteristics of participants selected into the intervention group and the use of a historical control group.
As CP treatment was not blinded to either outcome assessors or study patients in most studies, the blinding component was judged to be weak except for the RCT by Li et al [126], where the evaluation of clinical outcomes was performed by an investigator who was blind to the treatment.
If there was no detailed CP therapy in terms of CP collection, neutralizing antibody or anti-SARS-CoV-2 IgG titers, timing and dose of the treatment, and valid measures of clinical outcomes, the data collection methods of the study were deemed to be weak. Some case reports did not provide any information for CP donators, antibody titers, and adverse events [66,67].
There were no dropouts in the case reports and case series. One case series study where all patients were followed up for only 7 days [118] was ranked as moderate. In the RCT reported by Gharbharan et al [125], all 86 patients had been followed for at least 15 days after inclusion, and 75 and 32 patients for at least 30 and 60 days, respectively.
Both RCTs were terminated prematurely due to the concerns over the potential benefit of CP in the study population with high neutralizing antibody titers (≥1:160) at baseline [125] and the lack of patients with COVID-19 to reach the planned recruitment target of 200 patients [294], resulting in an underpowered study sample size. However, fewer studies were available to assess the effects of CP treatment on the length of hospital stay, clinical improvement, and viral clearance. We identified only 6 studies (1 RCT and 5 cohort studies) reporting the length of hospital stay, with a total of 366 and 1735 patients in the CP and control groups, respectively ( Figure 3). These studies had significant heterogeneity (P<.001; I 2 =95%) and, when combined, did not show any effects of CP treatment on the length of hospital stay (mean difference 0.84, 95% CI -3.35 to 5.02 days).

Meta-analyses
Similarly, 4 studies (2 RCTs and 2 cohort studies) assessed the clinical improvement with the number of patients in both CP and control groups. As depicted in Figure 4, a larger portion of the patients in the CP group showed improved clinical status compared to that in the control, but the difference was not statistically significant (OR 1.54, 95% CI 0.79-3.01; I 2 =43%).
Based on the 3 studies (1 RCT and 2 cohort studies) with a total of 63 and 65 patients in the CP and control groups, respectively, we found that the use of CP increased the viral clearance significantly (OR 26.21, 95% CI 4.36-157.66; I 2 =43%) as shown in Figure 5.    . Efficacy of CP treatment on viral clearance in COVID-19 patients. Data from 3 controlled clinical trials were pooled using an inverse variance method and analyzed using a random-effects model. Odds ratios and 95% CIs were used as statistical measures for viral clearance as a dichotomous outcome. CP: convalescent plasma.
Except for the high heterogeneity among the studies on assessing the length of hospital stay (I 2 =0.98; P<.001), the heterogeneity among the studies assessing the clinical improvement and viral clearance was mild (I 2 =43%, P=. 16 and I 2 =38%, P=.20, respectively). Furthermore, since the included studies on the efficacy of CP treatment for mortality are homogenous (I 2 =0; P=.99), the overall effect on the mortality from the meta-analysis seems to be conclusive.

Mechanisms of Action
The biological basis for efficacy of CP entails the transfer of specific antiviral immunoglobulins (antibodies) and other bioactive substances in the plasma of patients in the convalescent phase of COVID-19 infection [233,302]. In theory, administration of CP containing high levels of polyclonal neutralizing antibodies (comprised mainly of IgG, with smaller amounts of IgM, IgA) can confer immediate pathogen-specific protection by inhibiting viral infection in a susceptible person [303]. However, findings suggest considerable variation in antibody titers and the duration of protective anti-SARS-CoV-2 IgG and IgM immunity observed in recovered CP donors [304,305]. A recent population-based study of humoral immune responses to SARS-CoV-2 demonstrated that >90% of people who recovered from COVID-19 were seropositive on virus-specific pan-immunoglobulin assays by day 25, and hospitalized patients seroconverted more frequently than nonhospitalized people. Furthermore, anti-SARS-CoV-2 antibody titers remained stable in recovered patients for the next 2 months, suggesting a durable immunoglobulin response [306]. Aside from CP, pooled human immunoglobulins may also be prepared from plasma as a concentrated antibody-containing solution to be administered as intravenous, subcutaneous, or intramuscular immunoglobulin. These pooled plasma-derived immunoglobulin products benefit from the polyclonal response of each individual donor and from the interindividual variability in such responses [307]. In addition, purified, high-titer hyperimmune immunoglobulin formulations can be obtained from vaccinated or convalescing donors, which have known levels of plasma-derived neutralizing antibodies that may prove valuable against COVID-19 [33,207,225].
Although not fully elucidated, the protective mechanisms of CP are based on direct and indirect antiviral activities, including antibody neutralization of viral infectivity [233,307]. In the case of SARS-CoV-2 pathogenesis, the viral spike glycoprotein is critical to the dissemination and pathogenesis of the virus [308]. The spike protein mediates binding of SARS-CoV-2 to host cell ACE-2 surface receptors, thereby acting as the first step in cellular entry and infection. Several lines of evidence from studies of SARS-CoV and CoV-2 show that infected hosts produce neutralizing antibodies directed against the RBD of the homotrimeric spike protein and can block infection by preventing viral entry and subsequent replication [309]. Other beneficial immune effects of CP are thought to include enhanced antibody-dependent cellular cytotoxicity, complement activation, and phagocytosis, along with restoration of the vascular endothelial glycocalyx [34,200]. Moreover, a majority of convalescent patients display robust antiviral SARS-CoV-2-specific T cell responses, with enhanced in vivo priming and expansion of CD8+ cytotoxic T cells and a higher frequency of CD4+ memory T cells in those who recovered from severe COVID-19, which may provide long-term antiviral protection even if antibodies wane [310]. Therefore, T cells could help to control SARS-CoV-2 infection and serve as correlates of protective antiviral immunity [311].
As new strains of SARS-CoV-2 with several dominant mutations in the spike protein have been identified recently, crucial questions associated with the possible reinfection of recovered patients and the efficiencies of vaccines designed based on early epidemic strains have arisen [22]. Recent findings show that sera collected from convalescent COVID-19 patients in early 2020 vaccinated with RBD-based vaccines efficiently neutralize viral variants of D614G and B.1.1.7 but weakly neutralize those of 501Y.V2, suggesting a warning to recovered patients and developed vaccines [312]. These results show that, as mutations accumulate in the RBD, spike proteins may acquire an antigenic shift that enable SARS-CoV-2 variants with loss-of-neutralization potency in vitro against emerging variants and eventually resist the current vaccines. Therefore, intensive monitoring of virus mutations and timely adjustments to the spike sequences of designed vaccines and updated antibody cocktail therapies, targeting highly conserved regions, are required to control the viral pandemic [313].

Main Findings
This systematic review summarizes a variety of evidence on the use of CP for treatment of COVID-19. Though the focus of this review was to identify and assess the quality of clinical studies reporting CP treatment for COVID-19, the broad search strategy identified a large number of studies related to various aspects of CP use, highlighting substantial research in this field.
The data on this topic is being rapidly generated and reported. Most are commentary and review articles and protocol or guidance descriptions on the theme of CP treatments for COVID-19. The main findings according to each group of articles dealing with COVID-19 CP were: Our meta-analysis of controlled studies showed significant reduction in mortality by CP therapy in comparison to controls. Similar meta-analysis of the efficacy of CP therapy on different types of infectious disease found a 44% reduction in the mortality of patients with COVID-19 [208]; a 25% reduction in other severe acute respiratory infections [33]; and a 32% reduction in SARS-CoV infection, severe influenza, and Ebola infection [209]. In contrast, the meta-analysis from 4 RCTs on CP treatment for influenza infection (n=572 patients) showed no convincing effects on deaths [206]. Another recent systematic review of 1 RCT and 3 controlled nonrandomized studies of CP therapy in patients with COVID-19 reported a potential reduction in mortality, time to death, and improvement of clinical symptoms but was unable to provide any opinion regarding the efficacy of CP treatment for COVID-19 due to paucity in quantitative synthesis [207].
Our meta-analysis showed no effect of CP on the length of hospital stay (mean difference), which is consistent with another meta-analysis of 3 RCTs for the effect of CP on the length of hospitalization in other severe respiratory viral infections, as reported by Devasenapathy et al [206]. Other systematic reviewers reported mixed results of both reduced length of hospital stay and no effects on the length of hospitalization in SARS-CoV infection, severe influenza, and Ebola infection [209], suggesting that the effectiveness of CP in reducing hospital length of stay might be dependent on early administration of the therapy, and its use as prophylaxis is more likely to be beneficial than treating severe disease [33]. However, the optimal timing and dosage of CP therapy remains to be defined.
The insignificant effect of CP on the improvements of clinical COVID-19 symptoms is comparable to another systematic review and meta-analysis of 5 studies with a total of 259 patients with COVID-19, showing more clinically improved patients treated with CP than no CP treatment but was not statistically significant (OR 2.06, 95% CI 0.8-4.9; I 2 =44%) [208]. In contrast, the meta-analysis of 9 controlled and uncontrolled studies showed improved clinical status of patients with COVID-19 when compared to baseline (ROM 0.53, 95% CI 0.36-0.79; P<.01; n=149) [147].
The significant increase in the viral clearance is also consistent with the other meta-analysis of 2 studies with a total of 144 patients, suggesting that the use of CP helps in viral clearance significantly [208], and with the meta-analysis of 9 controlled and uncontrolled studies showing reduced viral loads [147].
Various tools have been developed for quality assessment involving slightly different components and ranking criteria [314]. We used the EPHPP tool as it can be used for all types of clinical studies. This is a generic tool used to evaluate a variety of intervention study designs such as RCTs, before-and-after, and case-control studies [62]. A study has shown differences in quality assessment for RCTs between the EPHPP and the Cochrane Collaboration Risk of Bias tool [315].
Overall, clinical studies and systematic reviews have confirmed that CP caused few or no serious adverse events with low-quality evidence.
Consistent with other reviews [207,208], our quality appraisal showed that the present studies on the efficacy of CP are generally of low quality, although there are certain agreements and discrepancies between our assessment and others on the overall quality of case and randomized controlled studies on the use of CP for COVID-19, as different assessment tools have been used. Only 1 high-quality (low risk of bias in the underlying study results) RCT by Li et al [126] was identified in our assessment using the EPHPP tool, which is in agreement with the assessment in the systematic review by Sarkar et al [208], but was rated to be unclear in another systematic review by Piechotta et al [207], even though both reviews used the same Cochrane risk-of-bias tool (RoB 2.0) for the RCT.
The overall quality of the case-control studies in our assessment lies in between the risk of bias assessed by other two systematic reviews conducted by Piechotta et al [207] and Sarkar et al [208]. Specifically, the study by Duan et al [115] was considered weak in our quality assessment but was critical as assessed by Piechotta et al [207] and moderate risk of bias by Sarkar et al [208] in their reviews. The case-control study reported by Liu et al [117] was of moderate quality in our assessment but was critical and had a low risk of bias as assessed by Piechotta et al [207] and Sarkar et al [208], respectively, using the same Risk of Bias in Non-randomized Studies-of Interventions. The case-control study reported by Zeng et al [124] was moderate in our assessment, agreeing with the assessment in the systematic review by Sarkar et al [208], but was rated to be a critical risk of bias in the systematic review by Piechotta et al [207]. In addition to controlled and randomized studies, EPHPP could be used to assess the quality of case reports and series studies [62]. The overall quality of all case reports and series were weak based on our assessment.
Considering the promising evidence from existing clinical data, there is a clear need for RCTs on large patient numbers to evaluate the efficacy of CP therapy. Apart from sample size and the noncomparative, nonrandomized study design, numerous limitations hamper the interpretation of the aforementioned studies, such as the superimposition of effects mediated by other antiviral treatments, antibiotics, and glucocorticoids administered concomitantly with CP. As a whole, these studies indicate that patients receiving transfusions earlier than 14 days post infection may benefit from CP treatment [228,230].

Limitations
There are 2 systematic reviews and meta-analysis to appraise the literature on CP therapy for patients with COVID-19. However, this review covers the latest literature as of the date of our manuscript submission and provides insights about various aspects for the subject on the use of CP for COVID-19 that needs further investigation. The primary limitation of this review is that most data identified are nonrandomized (only 2 out of 64 clinical studies were randomized, with only 1 being of high quality), and therefore, confounding is highly inevitable. Furthermore, study populations, interventions, and measured outcomes have important clinical and methodological heterogeneity, which reflects an overall low to moderate quality of evidence identified by the appropriate quality assessment tool.
Publication bias may be another potential limitation given that the majority of early clinical studies on COVID-19 lacked original data, and those that did were rushed and did not include the appropriate measures to reduce bias [316]. Among the 243 papers included in this review, 32.5% (n=79) were commentaries, 18.9% (n=46) were reviews, and 7.8% (n=19) were protocols that did not contain any new data. We then evaluated the quality of the original clinical studies using the validated tool and found that more than 80% (52/64) were at risk of bias, mainly because of few participants, unrepresentative patient selection, poor study design, no control of confounders, and no blinding.

Future Directions
We summarized various aspects of the evidence on the use of CP in patients with COVID-19. However, important gaps in knowledge remain. Notably, the following areas require further investigation.
Well-designed prospective observational studies, preferentially RCTs, with well-defined characteristics for both CP donors and recipients are warranted to answer questions concerning the effects on mortality or other important clinical outcomes such as improvement in symptoms and respiratory status. The placebo or control should include standard-of-care or normal fresh frozen plasma. The plasma exchange has shown therapeutic effects for severe COVID-19 acute respiratory distress syndrome with multiple organ failure [317].
In vitro testing showed variable or diverse neutralizing antibody titers among individual donors, suggesting that an adequate pooling strategy of plasma units from different donors could reduce the variability of neutralizing antibody titers of CP and compensate deviations of individual antibody titers [289]. Clinical studies on the safety and efficacy of pooled CP should be conducted.
The COVID-19 pandemic has substantially reduced the national ability to provide blood products for medical care in an emergency [318], which further highlights the need to secure a stockpile of blood products with a long shelf life (eg, freeze-dried plasma) to be self-sufficient in a national crisis. Current CP protocols specify that, once thawed, CP may be stored for up to 5 days at 4 °C, similar to that of fresh frozen plasma. A recent study has demonstrated long-term stability of anti-SARS-CoV-2 spike antibodies in donor CP for 42 days when stored under refrigerated conditions [291]. There will be a need to stockpile freeze-dried CP for future waves of the pandemic for several years. Additionally, global concern over the potential for future waves of infection to occur before effective vaccines or drug therapies are available has many looking at other biological sources for large-scale production of neutralizing SARS-CoV-2 antibodies. Taking this into consideration, we are developing COVID-19 convalescent freeze-dried plasma. As this is a pooled plasma product of 10 donors, we also hypothesize that convalescent freeze-dried plasma will have higher anti-SARS-CoV-2 neutralizing antibody titers and activity than single donor CP. As well, this product may be administered in a hypertonic solution for those patients who cannot tolerate large volume CP transfusions.

Conclusions
There is still limited evidence but accumulating interest in CP treatment for COVID-19. The theoretical reasons for the likely efficacy of passive immunization, the urgent need felt by clinicians worldwide for effective treatment options for COVID-19, and the promising results offered mainly by retrospective clinical studies must be balanced against the lack of efficacy in the RCTs of CP and hyperimmune globulin therapy in severe influenza and COVID-19.
CP may be of greatest benefit for patients who are early in their illness and have not yet generated endogenous antibodies, and when the infused CP has a high antibody titer. Recurring observations suggested that treatment with CP within 4-5 days of symptom onset might be more effective than later treatment.
Our systematic review and analysis emphasizes the low quality of clinical studies. These studies could provide important lessons that should inform the planning of adequately powered and properly designed RCTs to evaluate the promise of CP therapy for patients with COVID-19.
Future research is necessary to fill the obvious knowledge gaps regarding CP treatment for patients with COVID-19. In brief, we offered recommendations around the need for a large-scale properly designed RCT, the potential prophylactic use of CP, selection criteria for both CP donors and recipients, development of antibodies with higher potency than CP, and freeze-dried CP as a long-term strategy against the pandemic.