Safety and Efficacy of Convalescent Plasma in COVID-19: An Overview of Systematic Reviews

Convalescent plasma (CP) from patients recovered from COVID-19 is one of the most studied anti-viral therapies against SARS-COV-2 infection. The aim of this study is to summarize the evidence from the available systematic reviews on the efficacy and safety of CP in COVID-19 through an overview of the published systematic reviews (SRs). A systematic literature search was conducted up to August 2021 in Embase, PubMed, Web of Science, Cochrane and Medrxiv databases to identify systematic reviews focusing on CP use in COVID-19. Two review authors independently evaluated reviews for inclusion, extracted data and assessed quality of evidence using AMSTAR (A Measurement Tool to Assess Reviews) and GRADE tools. The following outcomes were analyzed: mortality, viral clearance, clinical improvement, length of hospital stay, adverse reactions. In addition, where possible, subgroup analyses were performed according to study design (e.g., RCTs vs. non-RCTs), CP neutralizing antibody titer and timing of administration, and disease severity. The methodological quality of included studies was assessed using the checklist for systematic reviews AMSTAR-2 and the GRADE assessment. Overall, 29 SRs met the inclusion criteria based on 53 unique primary studies (17 RCT and 36 non-RCT). Limitations to the methodological quality of reviews most commonly related to absence of a protocol (11/29) and funding sources of primary studies (27/29). Of the 89 analyses on which GRADE judgements were made, effect estimates were judged to be of high/moderate certainty in four analyses, moderate in 38, low in 38, very low in nine. Despite the variability in the certainty of the evidence, mostly related to the risk of bias and inconsistency, the results of this umbrella review highlight a mortality reduction in CP over standard therapy when administered early and at high titer, without increased adverse reactions.


Introduction
The Coronavirus Disease 2019 (COVID-19) pandemic, caused by the Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2), is still a worldwide health crisis with devastating social and economic consequences. Despite the virus known for more than a year and a half, the management of COVID-19 remains challenging with still high mortality rates among severely affected patients [1,2]. A number of guidelines from experts have been released during the pandemic period suggesting several treatments for COVID-19 patients, including antiviral, hydroxychloroquine, steroid, anticoagulation and other supportive therapies [3,4]. However, recent evidence from large-scale studies failed to clarify the efficacy of most of the treatments proposed [5][6][7].
Convalescent plasma (CP), first introduced in 1890 by Emil von Behring to treat diphtheria and pertussis and then utilized in several other serious infectious diseases, including Ebola, severe acute respiratory syndrome (SARS), and Middle East respiratory syndrome (MERS), has been also proposed more recently as passive immunotherapy for treatment of SARS-CoV-2 infection [1]. Convalescent plasma is nowadays among the most MF). Both reviewers compared the articles identified. Studies were selected independently by two reviewers (MF and MC) with disagreements resolved through discussion and on the basis of the opinion of a third reviewer (FC). The two assessors also independently extracted quantitative and qualitative data from each selected study. Findings are presented in tabular format with supporting text (Table 1). Quantitative tabulation of results include: first author name and year of publication, the clinical condition under evaluation, principal characteristics of the study population, number of RCTs and non-RCTs included in the SR, intervention (CP versus control characteristics), the outcomes assessed, a quantitative synthesis (when available) of the estimates of interest (odds ratio (OR), risk ratio (RR), risk difference (RD), or mean difference (MD) with the 95% confidence intervals (CI), as reported in individual reviews), and the main results and conclusions of the SR. In addition, a three-color score was used for an immediate visual inspection of the CP-related effects with regard to the four main outcomes assessed: viral clearance, clinical improvement, mortality reduction and safety (green color: CP confers advantage over standard therapy or placebo; red color: CP does not confer advantage over standard therapy or placebo; yellow color: no clear advantage or disadvantage).

Assessment of Methodological Quality
We used A Measurement Tool to Assess Reviews (AMSTAR-2) critical appraisal checklist for SRs, a tool that evaluates both quantitative and qualitative reviews [8]. The tool is suitable for reviews including randomised and non-randomised studies. It includes 16 domains relating to the research question, review design, search strategy, study selection, data extraction, justification for excluded studies, description of included studies, risk of bias, sources of funding, meta-analysis, heterogeneity, publication bias, and conflicts of interest. Two review authors (MC, MF) independently assessed the quality of evidence in the included reviews and the methodological quality of the SRs. We resolved discrepancies through discussion or, if needed, through a third review author (FC). We did not exclude reviews based on AMSTAR-2 ratings, but considered the ratings in interpretation of our results.

Appraisal of the Quality of Evidence
The quality of evidence was appraised following the Grades of Recommendation, Assessment, Development, and Evaluation (GRADE) approach. Whenever available, the grading of the quality of evidence reported in the included reviews was considered to determine the quality of evidence. In a situation in which the grading of evidence was not reported by the authors of the study, the GRADE approach was applied in its five domains (risk of bias, indirectness, imprecision, inconsistency, and publication bias) based on the information available from the study [9].

Methodological Quality
Of the included reviews, the majority (19/29; 65.5%) had ≥2 (from 2 to 12) unmet AMSTAR-2 methodological requirements, and nine (31.0%) had one unmet methodological requirement; one Cochrane review met all the methodologic requirements ( Table 2). Sixteen reviews (55.2%) had one or more methodological requirements partly met. Twenty-seven reviews (93.1%) did not report on the source of funding for the studies included in the review; 11 reviews (39.2%) did not register or publish a protocol. Eight reviews (27.6%) did not mention publication bias in methods and results, and failed to discuss the possible impact of publication bias on review findings. In five reviews, participants, interventions, comparators, and outcomes (PICO) were not clearly made explicit, and in 3 reviews design was not fully explained. In six reviews, the search strategy was not comprehensive. More than 85% of author teams perform study selection and screening in duplicate. Other unmet domains were related to the list of excluded reviews and reasons (seven reviews, 24.1%), assessment of risk of bias (three reviews), and sources of conflict of interest, including any funding that the authors receive for conducting the review (four reviews). Wardhani [35] Wenjing [36] Yuwono Soeroto [37] Zhang [38] Footnotes: four SRs were focused only on adult (>18 years) COVID-19 patients [11,20,31,33]. The main characteristics of the SRs included are summarized in Table 1.

Methodological Quality
Of the included reviews, the majority (19/29; 65.5%) had ≥2 (from 2 to 12) unmet AMSTAR-2 methodological requirements, and nine (31.0%) had one unmet methodological requirement; one Cochrane review met all the methodologic requirements (Table  2). Sixteen reviews (55.2%) had one or more methodological requirements partly met. Twenty-seven reviews (93.1%) did not report on the source of funding for the studies included in the review; 11 reviews (39.2%) did not register or publish a protocol. Eight reviews (27.6%) did not mention publication bias in methods and results, and failed to discuss the possible impact of publication bias on review findings. In five reviews, participants, interventions, comparators, and outcomes (PICO) were not clearly made explicit, and in 3 reviews design was not fully explained. In six reviews, the search strategy was not comprehensive. More than 85% of author teams perform study selection and screening in duplicate. Other unmet domains were related to the list of excluded reviews and reasons (seven reviews, 24.1%), assessment of risk of bias (three reviews), and sources of conflict of interest, including any funding that the authors receive for conducting the review (four reviews).

Footnotes:
Methodological requirement met, Methodological requirement partly met, Methodological requirement unmet. AMSTAR-2 domains: 1. Did the research questions and inclusion criteria for the review include the components of PICO? 2. Did the report of the review contain an explicit statement that the review methods were established prior to the conduct of the review and did the report justify any significant deviations from the protocol? 3. Did the review authors explain their selection of the study designs for inclusion in the review? 4. Did the review authors use a comprehensive literature search strategy? 5. Did the review authors perform study selection in duplicate? 6. Did the review authors perform data extraction in duplicate? 7. Did the review authors provide a list of excluded studies and justify the exclusions? 8. Did the review authors describe the included studies in adequate detail? 9. Did the review authors use a satisfactory technique for assessing the risk of bias (RoB) in individual studies that were included in the review? 10. Did the review authors report on the sources of funding for the studies included in the review? 11. If meta-analysis was performed did the review authors use appropriate methods for statistical combination of results? 12. If meta-analysis was performed, did the review authors assess the potential impact of RoB in individual studies on the results of the meta-analysis or other evidence synthesis? 13. Did the review authors account for RoB in individual studies when interpreting/ discussing the results of the review? 14. Did the review authors provide a satisfactory explanation for, and discussion of, any heterogeneity 15. If they performed quantitative synthesis did the review authors carry out an adequate investigation of publication bias (small study bias) and discuss its likely impact on the results of the review? 16. Did the review authors report any potential sources of conflict of interest, including any funding they received for conducting the review? Although AMSTAR-2 consists of 16 items, critical domains include items 2, 4, 7, 9, 11, 13, and 15.

Summary of the Effect of CP on the Main Outcomes
Methodological requirement met, four SRs were focused only on adult (>18 years) COVID-19 patients [11,20,31,33]. The main characteristics of the SRs included are summarized in Table 1.

Methodological Quality
Of the included reviews, the majority (19/29; 65.5%) had ≥2 (from 2 to 12) unmet AMSTAR-2 methodological requirements, and nine (31.0%) had one unmet methodological requirement; one Cochrane review met all the methodologic requirements (Table  2). Sixteen reviews (55.2%) had one or more methodological requirements partly met. Twenty-seven reviews (93.1%) did not report on the source of funding for the studies included in the review; 11 reviews (39.2%) did not register or publish a protocol. Eight reviews (27.6%) did not mention publication bias in methods and results, and failed to discuss the possible impact of publication bias on review findings. In five reviews, participants, interventions, comparators, and outcomes (PICO) were not clearly made explicit, and in 3 reviews design was not fully explained. In six reviews, the search strategy was not comprehensive. More than 85% of author teams perform study selection and screening in duplicate. Other unmet domains were related to the list of excluded reviews and reasons (seven reviews, 24.1%), assessment of risk of bias (three reviews), and sources of conflict of interest, including any funding that the authors receive for conducting the review (four reviews).

Footnotes:
Methodological requirement met, Methodological requirement partly met, Methodological requirement unmet. AMSTAR-2 domains: 1. Did the research questions and inclusion criteria for the review include the components of PICO? 2. Did the report of the review contain an explicit statement that the review methods were established prior to the conduct of the review and did the report justify any significant deviations from the protocol? 3. Did the review authors explain their selection of the study designs for inclusion in the review? 4. Did the review authors use a comprehensive literature search strategy? 5. Did the review authors perform study selection in duplicate? 6. Did the review authors perform data extraction in duplicate? 7. Did the review authors provide a list of excluded studies and justify the exclusions? 8. Did the review authors describe the included studies in adequate detail? 9. Did the review authors use a satisfactory technique for assessing the risk of bias (RoB) in individual studies that were included in the review? 10. Did the review authors report on the sources of funding for the studies included in the review? 11. If meta-analysis was performed did the review authors use appropriate methods for statistical combination of results? 12. If meta-analysis was performed, did the review authors assess the potential impact of RoB in individual studies on the results of the meta-analysis or other evidence synthesis? 13. Did the review authors account for RoB in individual studies when interpreting/ discussing the results of the review? 14. Did the review authors provide a satisfactory explanation for, and discussion of, any heterogeneity 15. If they performed quantitative synthesis did the review authors carry out an adequate investigation of publication bias (small study bias) and discuss its likely impact on the results of the review? 16. Did the review authors report any potential sources of conflict of interest, including any funding they received for conducting the review? Although AMSTAR-2 consists of 16 items, critical domains include items 2, 4, 7, 9, 11, 13, and 15.

Summary of the Effect of CP on the Main Outcomes
Methodological requirement partly met, four SRs were focused only on adult (>18 years) COVID-19 patients [11,20,31,33]. The main characteristics of the SRs included are summarized in Table 1.

Methodological Quality
Of the included reviews, the majority (19/29; 65.5%) had ≥2 (from 2 to 12) unmet AMSTAR-2 methodological requirements, and nine (31.0%) had one unmet methodological requirement; one Cochrane review met all the methodologic requirements (Table  2). Sixteen reviews (55.2%) had one or more methodological requirements partly met. Twenty-seven reviews (93.1%) did not report on the source of funding for the studies included in the review; 11 reviews (39.2%) did not register or publish a protocol. Eight reviews (27.6%) did not mention publication bias in methods and results, and failed to discuss the possible impact of publication bias on review findings. In five reviews, participants, interventions, comparators, and outcomes (PICO) were not clearly made explicit, and in 3 reviews design was not fully explained. In six reviews, the search strategy was not comprehensive. More than 85% of author teams perform study selection and screening in duplicate. Other unmet domains were related to the list of excluded reviews and reasons (seven reviews, 24.1%), assessment of risk of bias (three reviews), and sources of conflict of interest, including any funding that the authors receive for conducting the review (four reviews).

ootnotes:
Methodological requirement met, Methodological requirement partly met, Methodological requireent unmet. AMSTAR-2 domains: 1. Did the research questions and inclusion criteria for the review include the comonents of PICO? 2. Did the report of the review contain an explicit statement that the review methods were established rior to the conduct of the review and did the report justify any significant deviations from the protocol? 3. Did the reiew authors explain their selection of the study designs for inclusion in the review? 4. Did the review authors use a omprehensive literature search strategy? 5. Did the review authors perform study selection in duplicate? 6. Did the reiew authors perform data extraction in duplicate? 7. Did the review authors provide a list of excluded studies and justify e exclusions? 8. Did the review authors describe the included studies in adequate detail? 9. Did the review authors use a tisfactory technique for assessing the risk of bias (RoB) in individual studies that were included in the review? 10. Did e review authors report on the sources of funding for the studies included in the review? 11. If meta-analysis was perrmed did the review authors use appropriate methods for statistical combination of results? 12. If meta-analysis was erformed, did the review authors assess the potential impact of RoB in individual studies on the results of the me--analysis or other evidence synthesis? 13. Did the review authors account for RoB in individual studies when interreting/ discussing the results of the review? 14. Did the review authors provide a satisfactory explanation for, and disussion of, any heterogeneity 15. If they performed quantitative synthesis did the review authors carry out an adequate vestigation of publication bias (small study bias) and discuss its likely impact on the results of the review? 16. Did the view authors report any potential sources of conflict of interest, including any funding they received for conducting the view? Although AMSTAR-2 consists of 16 items, critical domains include items 2, 4, 7, 9, 11, 13, and 15.

Summary of the Effect of CP on the Main Outcomes
Methodological requirement unmet. AMSTAR-2 domains: 1. Did the research questions and inclusion criteria for the review include the components of PICO? 2. Did the report of the review contain an explicit statement that the review methods were established prior to the conduct of the review and did the report justify any significant deviations from the protocol? 3. Did the review authors explain their selection of the study designs for inclusion in the review? 4. Did the review authors use a comprehensive literature search strategy? 5. Did the review authors perform study selection in duplicate? 6. Did the review authors perform data extraction in duplicate? 7. Did the review authors provide a list of excluded studies and justify the exclusions? 8. Did the review authors describe the included studies in adequate detail? 9. Did the review authors use a satisfactory technique for assessing the risk of bias (RoB) in individual studies that were included in the review? 10. Did the review authors report on the sources of funding for the studies included in the review? 11. If meta-analysis was performed did the review authors use appropriate methods for statistical combination of results? 12. If meta-analysis was performed, did the review authors assess the potential impact of RoB in individual studies on the results of the meta-analysis or other evidence synthesis? 13. Did the review authors account for RoB in individual studies when interpreting/ discussing the results of the review? 14. Did the review authors provide a satisfactory explanation for, and discussion of, any heterogeneity 15. If they performed quantitative synthesis did the review authors carry out an adequate investigation of publication bias (small study bias) and discuss its likely impact on the results of the review? 16. Did the review authors report any potential sources of conflict of interest, including any funding they received for conducting the review? Although AMSTAR-2 consists of 16 items, critical domains include items 2, 4, 7, 9, 11, 13, and 15. Overall mortality was the most common reported outcome. Great heterogeneity was ascertained in several SRs; thus, we performed subgroup analyses to control for sources of heterogeneity such as design of studies included in the review (e.g., RCTs and non-RCTs), titre of SARS-CoV-2 neutralizing antibodies (high or low), and time of administration (early or late). The results of our analyses are summarized in Table 3. Sixteen SRs reported the outcome mortality in RCTs and non-RCT. In 14 of these SRs the effect size favoured the CP arm compared to controls, while in two SRs it was unclear whether CP reduced mortality compared to controls; the quality of the evidence was very low in 2 SRs, low in four, moderate in nine, and from moderate to high in one. Eleven SRs analysed the outcome mortality in RCTs only, and all were consistent in concluding that CP did not reduce mortality compared to controls (moderate certainty of evidence in 10 SRs, from moderate to high certainty in one SR). The analysis of results from non-RCTs showed higher reduction in mortality in CP group compared to control in six out of seven SRs (low quality of evidence in four, very low in one, moderate in two).

Methodological Quality
Of the included reviews, the majority (19/29; 65.5%) had ≥2 (from 2 to 12) unmet AMSTAR-2 methodological requirements, and nine (31.0%) had one unmet methodological requirement; one Cochrane review met all the methodologic requirements ( Table  2). Sixteen reviews (55.2%) had one or more methodological requirements partly met. Twenty-seven reviews (93.1%) did not report on the source of funding for the studies included in the review; 11 reviews (39.2%) did not register or publish a protocol. Eight reviews (27.6%) did not mention publication bias in methods and results, and failed to discuss the possible impact of publication bias on review findings. In five reviews, participants, interventions, comparators, and outcomes (PICO) were not clearly made explicit, and in 3 reviews design was not fully explained. In six reviews, the search strategy was not comprehensive. More than 85% of author teams perform study selection and screening in duplicate. Other unmet domains were related to the list of excluded reviews and reasons (seven reviews, 24.1%), assessment of risk of bias (three reviews), and sources of conflict of interest, including any funding that the authors receive for conducting the review (four reviews).

Footnotes:
Methodological requirement met, Methodological requirement partly met, Methodological requirement unmet. AMSTAR-2 domains: 1. Did the research questions and inclusion criteria for the review include the components of PICO? 2. Did the report of the review contain an explicit statement that the review methods were established prior to the conduct of the review and did the report justify any significant deviations from the protocol? 3. Did the review authors explain their selection of the study designs for inclusion in the review? 4. Did the review authors use a comprehensive literature search strategy? 5. Did the review authors perform study selection in duplicate? 6. Did the review authors perform data extraction in duplicate? 7. Did the review authors provide a list of excluded studies and justify the exclusions? 8. Did the review authors describe the included studies in adequate detail? 9. Did the review authors use a satisfactory technique for assessing the risk of bias (RoB) in individual studies that were included in the review? 10. Did the review authors report on the sources of funding for the studies included in the review? 11. If meta-analysis was performed did the review authors use appropriate methods for statistical combination of results? 12. If meta-analysis was performed, did the review authors assess the potential impact of RoB in individual studies on the results of the meta-analysis or other evidence synthesis? 13. Did the review authors account for RoB in individual studies when interpreting/ discussing the results of the review? 14. Did the review authors provide a satisfactory explanation for, and discussion of, any heterogeneity 15. If they performed quantitative synthesis did the review authors carry out an adequate investigation of publication bias (small study bias) and discuss its likely impact on the results of the review? 16. Did the review authors report any potential sources of conflict of interest, including any funding they received for conducting the review? Although AMSTAR-2 consists of 16 items, critical domains include items 2, 4, 7, 9, 11, 13, and 15.

Summary of the Effect of CP on the Main Outcomes
Favouring CP,

Methodological Quality
Of the included reviews, the majority (19/29; 65.5%) had ≥2 (from 2 to 12) unmet AMSTAR-2 methodological requirements, and nine (31.0%) had one unmet methodological requirement; one Cochrane review met all the methodologic requirements ( Table  2). Sixteen reviews (55.2%) had one or more methodological requirements partly met. Twenty-seven reviews (93.1%) did not report on the source of funding for the studies included in the review; 11 reviews (39.2%) did not register or publish a protocol. Eight reviews (27.6%) did not mention publication bias in methods and results, and failed to discuss the possible impact of publication bias on review findings. In five reviews, participants, interventions, comparators, and outcomes (PICO) were not clearly made explicit, and in 3 reviews design was not fully explained. In six reviews, the search strategy was not comprehensive. More than 85% of author teams perform study selection and screening in duplicate. Other unmet domains were related to the list of excluded reviews and reasons (seven reviews, 24.1%), assessment of risk of bias (three reviews), and sources of conflict of interest, including any funding that the authors receive for conducting the review (four reviews). : Methodological requirement met, Methodological requirement partly met, Methodological requireet. AMSTAR-2 domains: 1. Did the research questions and inclusion criteria for the review include the comf PICO? 2. Did the report of the review contain an explicit statement that the review methods were established e conduct of the review and did the report justify any significant deviations from the protocol? 3. Did the reors explain their selection of the study designs for inclusion in the review? 4. Did the review authors use a nsive literature search strategy? 5. Did the review authors perform study selection in duplicate? 6. Did the reors perform data extraction in duplicate? 7. Did the review authors provide a list of excluded studies and justify ions? 8. Did the review authors describe the included studies in adequate detail? 9. Did the review authors use a ry technique for assessing the risk of bias (RoB) in individual studies that were included in the review? 10. Did authors report on the sources of funding for the studies included in the review? 11. If meta-analysis was perid the review authors use appropriate methods for statistical combination of results? 12. If meta-analysis was d, did the review authors assess the potential impact of RoB in individual studies on the results of the mes or other evidence synthesis? 13. Did the review authors account for RoB in individual studies when interiscussing the results of the review? 14. Did the review authors provide a satisfactory explanation for, and disf, any heterogeneity 15. If they performed quantitative synthesis did the review authors carry out an adequate ion of publication bias (small study bias) and discuss its likely impact on the results of the review? 16. Did the thors report any potential sources of conflict of interest, including any funding they received for conducting the lthough AMSTAR-2 consists of 16 items, critical domains include items 2, 4, 7, 9, 11, 13, and 15.

Summary of the Effect of CP on the Main Outcomes
Not clear effect of CP compared to controls. 1. Downgraded for risk of bias and heterogeneity. 2. Downgraded for risk of bias. 3. This judgment was based on the consistency of the results between RCTs and matched control studies and the corroborating evidence from dose-response studies and other uncontrolled case data. In aggregating data from all controlled studies, the meta-analyses provided precise estimates, did not demonstrate substantial heterogeneity, and demonstrated no evidence of publication bias. The inherent limitations of the included studies rendered the certainty of evidence judgment to be moderate to high. 4. Downgraded for risk of bias, heterogeneity and publication bias. 5. Downgraded for risk of bias, imprecision and heterogeneity. 6. Downgraded for imprecision. 7. Downgraded for risk of bias and imprecision. 8. Downgraded twice for risk of bias (confounding and publication bias). 9. Downgraded for imprecision and clinical heterogeneity.
Due to the clinical heterogeneity observed in many of the primary studies and statistical heterogeneity in many of the SRs, we performed subgroup analyses as specified in the protocol. High titre CP was more effective than lower titre in reducing mortality (moderate certainty of evidence in three SRs, low in one). In two SRs, it was unclear whether high titre CP reduces mortality compared to controls (low certainty of evidence). Subgroup analysis according to time to CP transfusion showed a reduction in mortality in early CP recipients compared to controls in three SRs (moderate certainty of evidence); by contrast a Cochrane review based on a single RCT (RECOVERY trial) [62] concluded that it is unclear whether early CP transfusion reduces mortality compared to controls (low quality of evidence), although transfusion after the first week of illness resulted in higher risk of mortality compared to early transfusion. We also performed subgroup analysis of mortality according to baseline severity of COVID-19, but there was heterogeneity in defining the clinical condition. In patients with moderate/severe infection, the effect of CP was unclear (three SRs, low-quality of evidence), while in patients with severe/critical infection, the results were more heterogeneous in the comparison, effect size and certainty of the evidence (Table 3).

(b) Outcome "Viral Clearance"
The outcome viral clearance (rate of patients with negative reverse transcription polymerase chain reaction (RT-PCR) test for SARS-CoV-2 after a positive test at baseline) was reported in 6 SRs after 3 days, and in two SRs after 7-14 days (Table 4). On day 3, four out of six SRs reported an increase in viral clearance in CP recipients compared to controls (from low to moderate certainty of evidence), while in two SRs it was unclear whether CP reduced the viral clearance compared to controls (from low to moderate certainty of evidence). On day 14, CP showed significantly higher viral clearance rate compared to standard supportive therapy in one SR (low certainty of evidence), while in a Cochrane review based on two RCTs it was unclear whether CP increases viral clearance compared to controls (low certainty of evidence). It was not possible to perform subgroup analyses for these outcomes due to the relatively low number of primary studies and SRs available.

(c) Outcome "Clinical Improvement"
Clinical improvement was reported in nine SRs. Six SRs concluded that it is unclear whether CP increases rate of clinical improvement compared to controls (in three SRs moderate certainty of evidence, in three low certainty of evidence). Three SRs showed that CP increase rate of clinical improvement compared to controls (low certainty of evidence in two SRs, moderate certainty of evidence in one SR) (see Table 4).

(d) Outcome "Length of Hospital Stay"
Length of hospital stay was reported in six SRs. All concluded that it is unclear whether CP decreases length of hospital stay compared to controls (low quality of certainty in four SRs, moderate certainty of evidence in two SRs) ( Table 4).

(e) Outcome "Adverse Events"
Serious adverse events to CP transfusion were reported in five SRs and overall adverse events in two SRs. All SRs concluded that the frequency of adverse reactions was similar in CP and control groups (low quality of certainty in four SRs, moderate certainty of evidence in one SR) ( Table 5).

GRADE Assessment
Of the 89 analyses on which GRADE judgements were made, effect estimates were judged to be of high/moderate certainty in four analyses, moderate in 38, low in 38, and very low in nine. For the outcome mortality, the judgment was very low in nine analyses, low in 19, moderate in 19 and moderate/high in four.    AMSTAR-2 methodological requirements, and nine (31.0%) had one unmet methodological requirement; one Cochrane review met all the methodologic requirements ( Table  2). Sixteen reviews (55.2%) had one or more methodological requirements partly met. Twenty-seven reviews (93.1%) did not report on the source of funding for the studies included in the review; 11 reviews (39.2%) did not register or publish a protocol. Eight reviews (27.6%) did not mention publication bias in methods and results, and failed to discuss the possible impact of publication bias on review findings. In five reviews, participants, interventions, comparators, and outcomes (PICO) were not clearly made explicit, and in 3 reviews design was not fully explained. In six reviews, the search strategy was not comprehensive. More than 85% of author teams perform study selection and screening in duplicate. Other unmet domains were related to the list of excluded reviews and reasons (seven reviews, 24.1%), assessment of risk of bias (three reviews), and sources of conflict of interest, including any funding that the authors receive for conducting the review (four reviews).
Footnotes: Methodological requirement met, Methodological requirement partly met, Methodological requirement unmet. AMSTAR-2 domains: 1. Did the research questions and inclusion criteria for the review include the components of PICO? 2. Did the report of the review contain an explicit statement that the review methods were established prior to the conduct of the review and did the report justify any significant deviations from the protocol? 3. Did the review authors explain their selection of the study designs for inclusion in the review? 4. Did the review authors use a comprehensive literature search strategy? 5. Did the review authors perform study selection in duplicate? 6. Did the review authors perform data extraction in duplicate? 7. Did the review authors provide a list of excluded studies and justify the exclusions? 8. Did the review authors describe the included studies in adequate detail? 9. Did the review authors use a satisfactory technique for assessing the risk of bias (RoB) in individual studies that were included in the review? 10. Did the review authors report on the sources of funding for the studies included in the review? 11. If meta-analysis was performed did the review authors use appropriate methods for statistical combination of results? 12. If meta-analysis was performed, did the review authors assess the potential impact of RoB in individual studies on the results of the meta-analysis or other evidence synthesis? 13. Did the review authors account for RoB in individual studies when interpreting/ discussing the results of the review? 14. Did the review authors provide a satisfactory explanation for, and discussion of, any heterogeneity 15. If they performed quantitative synthesis did the review authors carry out an adequate investigation of publication bias (small study bias) and discuss its likely impact on the results of the review? 16. Did the review authors report any potential sources of conflict of interest, including any funding they received for conducting the review? Although AMSTAR-2 consists of 16 items, critical domains include items 2, 4, 7, 9, 11, 13, and 15.

Summary of the Effect of CP on the Main Outcomes
Favouring CP, AMSTAR-2 methodological requirements, and nine (31.0%) had one unmet methodological requirement; one Cochrane review met all the methodologic requirements ( Table  2). Sixteen reviews (55.2%) had one or more methodological requirements partly met. Twenty-seven reviews (93.1%) did not report on the source of funding for the studies included in the review; 11 reviews (39.2%) did not register or publish a protocol. Eight reviews (27.6%) did not mention publication bias in methods and results, and failed to discuss the possible impact of publication bias on review findings. In five reviews, participants, interventions, comparators, and outcomes (PICO) were not clearly made explicit, and in 3 reviews design was not fully explained. In six reviews, the search strategy was not comprehensive. More than 85% of author teams perform study selection and screening in duplicate. Other unmet domains were related to the list of excluded reviews and reasons (seven reviews, 24.1%), assessment of risk of bias (three reviews), and sources of conflict of interest, including any funding that the authors receive for conducting the review (four reviews). : Methodological requirement met, Methodological requirement partly met, Methodological requireet. AMSTAR-2 domains: 1. Did the research questions and inclusion criteria for the review include the comf PICO? 2. Did the report of the review contain an explicit statement that the review methods were established e conduct of the review and did the report justify any significant deviations from the protocol? 3. Did the reors explain their selection of the study designs for inclusion in the review? 4. Did the review authors use a nsive literature search strategy? 5. Did the review authors perform study selection in duplicate? 6. Did the reors perform data extraction in duplicate? 7. Did the review authors provide a list of excluded studies and justify ions? 8. Did the review authors describe the included studies in adequate detail? 9. Did the review authors use a y technique for assessing the risk of bias (RoB) in individual studies that were included in the review? 10. Did authors report on the sources of funding for the studies included in the review? 11. If meta-analysis was perd the review authors use appropriate methods for statistical combination of results? 12. If meta-analysis was , did the review authors assess the potential impact of RoB in individual studies on the results of the meor other evidence synthesis? 13. Did the review authors account for RoB in individual studies when interiscussing the results of the review? 14. Did the review authors provide a satisfactory explanation for, and dis-, any heterogeneity 15. If they performed quantitative synthesis did the review authors carry out an adequate ion of publication bias (small study bias) and discuss its likely impact on the results of the review? 16. Did the thors report any potential sources of conflict of interest, including any funding they received for conducting the lthough AMSTAR-2 consists of 16 items, critical domains include items 2, 4, 7, 9, 11, 13, and 15.   SRs analyzed COVID-19 patients with no limitation of age, gender or ethnicity, while four SRs were focused only on adult (>18 years) COVID-19 patients [11,20,31,33]. The main characteristics of the SRs included are summarized in Table 1.

Methodological Quality
Of the included reviews, the majority (19/29; 65.5%) had ≥2 (from 2 to 12) unmet AMSTAR-2 methodological requirements, and nine (31.0%) had one unmet methodological requirement; one Cochrane review met all the methodologic requirements ( Table  2). Sixteen reviews (55.2%) had one or more methodological requirements partly met. Twenty-seven reviews (93.1%) did not report on the source of funding for the studies included in the review; 11 reviews (39.2%) did not register or publish a protocol. Eight reviews (27.6%) did not mention publication bias in methods and results, and failed to discuss the possible impact of publication bias on review findings. In five reviews, participants, interventions, comparators, and outcomes (PICO) were not clearly made explicit, and in 3 reviews design was not fully explained. In six reviews, the search strategy was not comprehensive. More than 85% of author teams perform study selection and screening in duplicate. Other unmet domains were related to the list of excluded reviews and reasons (seven reviews, 24.1%), assessment of risk of bias (three reviews), and sources of conflict of interest, including any funding that the authors receive for conducting the review (four reviews).

Discussion
The outbreak of the COVID-19 pandemic has greatly accelerated the clinical trial research evaluating the safety and efficacy of CP as emergency therapy. According to the study by Muller-Olling and colleagues [63], CP is the second most frequent investigational medicinal product evaluated in COVID-19-related clinical trials and increasing interest in this form of immunotherapy is documented by the fact that more than 140 clinical trials specifically evaluating CP in COVID-19 have been registered to date worldwide [63]. Studies on CP as treatment for COVID-19 have qualitatively evolved during the COVID-19 pandemic period in response to the advances in the knowledge of this disease and to the results of the published clinical trials [64]. They can be generally classified into three generations: the first-generation clinical trials, performed at the beginning of the first pandemic wave, utilized CP with high-titer anti-SARS-CoV-2 neutralizing antibodies usually in hospitalized patients with severe, advanced COVID-19 [65,66]. However, it soon became evident that, since CP was working by blocking the viral replication, the earlier it was used the more effective it was in preventing the disease progression [1]. Thus, second generation clinical trials were focused on the use of high-titer CP early in the course of COVID-19 (within 3 days from symptom onset or hospitalization) [67]. Finally, to optimize the possible beneficial effect of plasma, the more recent third generation trials are evaluating the CP infusion in particular populations of patients at high risk of development of severe or critical COVID-19, such as those with impaired humoral immunity, onco-hematological disorders or other severe cardiovascular or respiratory co-morbidities [68][69][70]. Moreover, some of these trials encompass the pre-transfusion evaluation of recipients' anti-SARS-CoV-2 antibody levels in order to capture those patients with lack or insufficient antibody response and therefore most likely to benefit from CP passive immunotherapy [1]. Based on the above, it is evident that the clinical trials conducted during the 18-month COVID-19 pandemic are widely heterogeneous in terms of study design and CP administration schedule, disease and patients characteristics. In this extremely uncertain and changing context, typical of emergency situations such as those of the COVID-19 pandemic, it is evident that even the systematic reviews and meta-analyses have produced heterogeneous results. This overview of reviews includes data from twenty-nine systematic reviews, based on more than 600 overlapping reports and 53 individual primary studies (43 controlled trials, including 17 RCTs and 26 non-RCTs, and 10 uncontrolled trials (single arm studies). We believe that makes this the largest review to date within this subject area, and hope this will make it particularly helpful to decision makers. The results of RCTs are not always consistent with the results of observational studies, and differences in estimated magnitude of treatment effect are very common, often resulting in overestimation of treatment effects in observational studies [71]. Interpretation of the results obtained from both RCTs and observational studies, as well as from systematic reviews including both types of study design, can help understand the efficacy/effectiveness and safety of a therapeutic option [72]. For this reason we performed, where available, subgroup analyses of the effect size obtained in the overall comparison, in RCTs and in observational studies. For the outcome most commonly reported, overall mortality, it was possible to perform subgroup analysis of SRs according to study design, antibodies titre and time of transfusion. While the majority of SRs reporting this outcome in non-RCTs and in non-RCTs + RCTs showed a reduction in mortality in CP recipients compared to controls, when the analysis was limited to RCTs it was unclear whether CP reduced mortality compared to controls. It was also clear that most of the included studies (both RCTs and non-RCTs) were at risk of bias and showed significant clinical, methodological and statistical heterogeneity. Overall the certainty of the evidence was from low to moderate in the majority of the SRs. Subgroup analyses according to neutralizing antibody titres and time of CP transfusion showed a reduction in mortality in the majority of SRs when high titres of antibody and early transfusions were administered (from low to moderate certainty of the evidence). The other secondary outcomes (i.e., viral clearance, clinical improvement and length of hospital stay) were addressed by only a minority of SRs with a high level of uncertainty, so that no definitive conclusions can be drawn. However, CP seemed to be effective in increasing viral clearance as compared with standard therapy, particularly within the first three days from CP transfusion (from low to moderate certainty of the evidence). Additionally, CP did not increase the risk of adverse events between intervention and control groups, confirming the safety of this procedure [22]. Although the SRs tried to address, at least in part, the heterogeneity of the results (on the basis of neutralizing titre, time of CP infusion and study design) it was almost impossible to evaluate in this overview the clinical heterogeneity of primary studies, related to different disease conditions at baseline, concomitant therapies, patients characteristics. Limitations to the methodological quality of reviews most commonly related to absence of a protocol (11/29) and funding sources of primary studies (27/29).
In conclusion, despite these limitations and based on the analysis of the main outcome mortality, this overview of systematic reviews supports the safety and efficacy of the clinical use of CP over standard therapy when administered at high titer and early during the course of COVID-19. Further pooled qualitative and quantitative analyses from a new systematic review based on individual patients data rather than on aggregate data (that are often insufficient for a thorough analysis) or from adequately powered third generation clinical trials (i.e., assessing the early use of high titer CP administered in populations of patients with inadequate antiviral response and at increased risk of developing severe COVID-19) are needed to pinpoint exactly where and when CP can give the greatest clinical benefit in COVID-19.