Browse the corpus

Convalescent Plasma for Covid-19-Induced ARDS in Mechanically Ventilated Patients. BACKGROUND: Passive immunization with plasma collected from convalescent patients has been regularly used to treat coronavirus disease 2019 (Covid-19). Minimal data are available regarding the use of convalescent plasma in patients with Covid-19-induced acute respiratory distress syndrome (ARDS). METHODS: In this open-label trial, we randomly assigned adult patients with Covid-19-induced ARDS who had been receiving invasive mechanical ventilation for less than 5 days in a 1:1 ratio to receive either convalescent plasma with a neutralizing antibody titer of at least 1:320 or standard care alone. Randomization was stratified according to the time from tracheal intubation to inclusion. The primary outcome was death by day 28. RESULTS: A total of 475 patients underwent randomization from September 2020 through March 2022. Overall, 237 patients were assigned to receive convalescent plasma and 238 to receive standard care. Owing to a shortage of convalescent plasma, a neutralizing antibody titer of 1:160 was administered to 17.7% of the patients in the convalescent-plasma group. Glucocorticoids were administered to 466 patients (98.1%). At day 28, mortality was 35.4% in the convalescent-plasma group and 45.0% in the standard-care group (P = 0.03). In a prespecified analysis, this effect was observed mainly in patients who underwent randomization 48 hours or less after the initiation of invasive mechanical ventilation. Serious adverse events did not differ substantially between the two groups. CONCLUSIONS: The administration of plasma collected from convalescent donors with a neutralizing antibody titer of at least 1:160 to patients with Covid-19-induced ARDS within 5 days after the initiation of invasive mechanical ventilation significantly reduced mortality at day 28. This effect was mainly observed in patients who underwent randomization 48 hours or less after ventilation initiation. (Funded by the Belgian Health Care Knowledge Center; ClinicalTrials.gov number, NCT04558476.).

This randomized, two-group, open-label trial was conducted at 17 sites in Belgium and coordinated by the University Hospital of Liège. We encouraged inclusion soon after the initiation of invasive mechanical ventilation and stratified the randomization according to the delay between invasive mechanical ventilation and inclusion (≤48 hours vs. >48 to 120 hours). We assigned patients in a 1:1 ratio to receive either convalescent plasma with a neutralizing antibody titer against SARS-CoV-2 of at least 1:320 or standard care. The protocol, which was approved by the institutional review board at each center, was published previously13 and is available with the full text of this article at NEJM.org.

ts in a 1:1 ratio to receive either convalescent plasma with a neutralizing antibody titer against SARS-CoV-2 of at least 1:320 or standard care. The protocol, which was approved by the institutional review board at each center, was published previously13 and is available with the full text of this article at NEJM.org. Adult patients who had a score on the Clinical Frailty Scale of less than 6 (range, 1 to 9, with higher scores indicating greater frailty),14 who had been admitted to a participating ICU with a diagnosis of Covid-19–induced ARDS (corresponding to a score of 7, 8, or 9 on the World Health Organization [WHO] Clinical Progression Scale, with scores ranging from 0 to 10 and higher scores indicating more severe disease15), and who had received invasive mechanical ventilation for a maximum of 5 days were assessed for eligibility. ARDS was classified according to the Berlin definition.16 Covid-19–induced ARDS was defined as extended interstitial pneumonia on a computed tomographic scan or a chest radiograph within 10 days before inclusion and a positive result of a clinical SARS-CoV-2 nasopharyngeal polymerase-chain-reaction (PCR) test within 15 days before inclusion. Exclusion criteria were pregnancy, a previous episode of transfusion-related side effects, a medical decision to limit therapy, and participation in another Covid-19 trial.

efore inclusion and a positive result of a clinical SARS-CoV-2 nasopharyngeal polymerase-chain-reaction (PCR) test within 15 days before inclusion. Exclusion criteria were pregnancy, a previous episode of transfusion-related side effects, a medical decision to limit therapy, and participation in another Covid-19 trial. Before requesting consent, the local investigator contacted a physician at the coordinating center, who checked the inclusion criteria. At the same time, the availability of convalescent plasma with an ABO group identical to the eligible patient’s group was checked on the Web site of the Belgian Red Cross. The capability of the patient to give consent was then assessed. If consent could not be obtained from the patient, it was asked of a relative. When patients became able to give consent, they were asked to do so. Donors were recruited by the Belgian Red Cross among adults who had been infected with SARS-CoV-2 and who had fully recovered between 28 days and 10 months earlier.13 Neutralizing antibody titers were determined with SARS-CoV-2 Nextstrain clade 20B (Wuhan-like, B.1.1), isolated from a Belgian patient, in 96-well plates containing confluent Vero E6 cells (ATCC CRL-1586)17 and were reported as 50% virus neutralization titer (NT50). The NT50 was the highest dilution of serum that neutralized the cytopathic effect in 50% of the cells (see the protocol).13,18 A neutralizing antibody titer of at least 1:320 was requested for the trial, and a titer of 1:160 was accepted in case of a shortage of plasma.

eported as 50% virus neutralization titer (NT50). The NT50 was the highest dilution of serum that neutralized the cytopathic effect in 50% of the cells (see the protocol).13,18 A neutralizing antibody titer of at least 1:320 was requested for the trial, and a titer of 1:160 was accepted in case of a shortage of plasma. Randomization was stratified according to the delay from the initiation of invasive mechanical ventilation and performed with the use of a Web-based program. Patients in the convalescent-plasma group were assigned to receive an infusion of 2 units (400 to 500 ml in total) of convalescent plasma of the identical ABO group within 24 hours after inclusion. No crossover was allowed between the convalescent-plasma group and the standard-care group. Investigators were aware of the trial-group assignments. The primary outcome was death by day 28 after randomization. Secondary outcomes included adverse events, inflammatory and anti–SARS-CoV-2 antibody responses, the Sequential Organ Failure Assessment (SOFA) score, the use of organ support, the length of hospital stay, and death by day 90 and 365.

Randomization was stratified according to the delay from the initiation of invasive mechanical ventilation and performed with the use of a Web-based program. Patients in the convalescent-plasma group were assigned to receive an infusion of 2 units (400 to 500 ml in total) of convalescent plasma of the identical ABO group within 24 hours after inclusion. No crossover was allowed between the convalescent-plasma group and the standard-care group. Investigators were aware of the trial-group assignments. The primary outcome was death by day 28 after randomization. Secondary outcomes included adverse events, inflammatory and anti–SARS-CoV-2 antibody responses, the Sequential Organ Failure Assessment (SOFA) score, the use of organ support, the length of hospital stay, and death by day 90 and 365. We anticipated a day 28 mortality of 40% in the standard-care group. We estimated that a relative difference of one third in day 28 mortality was realistic and relevant. With a two-sided alpha level of 0.05 and a beta level of 0.20, the number of patients to include for an absolute difference of −13.5 percentage points (convalescent plasma minus standard care) in day 28 mortality was 250 for each group. Because our estimate of the between-group difference was empirical, we performed sequential blinded interim analyses. For these analyses, two Lan–Demets spending functions were used to adjust the type I error. The analysis was performed on an intention-to-treat basis. For the primary analysis, there were no missing values. No imputation was made for secondary outcomes because the pattern of missing data was completely at random, and missing data were either very infrequent or not used for statistical analyses. The results are expressed as median (interquartile range) and as number (percent). The primary outcome was compared with the chi-square test and adjusted for the stratification factor with a binary logistic regression. The two-sided level of significance was 5%. We summarized secondary analyses with point estimates for differences between the two groups and 95% confidence intervals using the Hodges–Lehmann estimation of location shift and a Wald asymptotic confidence interval for quantitative and binary variables, respectively. The confidence intervals were not adjusted for multiplicity and should not be used to infer definitive treatment effects. The Fine–Gray model was used to analyze competing risks data. The secondary analyses should be considered exploratory. Statistical analyses were performed with the use of SAS software for Windows, version 9.4, and R software for Windows, version 4.0.2.

icity and should not be used to infer definitive treatment effects. The Fine–Gray model was used to analyze competing risks data. The secondary analyses should be considered exploratory. Statistical analyses were performed with the use of SAS software for Windows, version 9.4, and R software for Windows, version 4.0.2. The statistical analysis plan13 included prespecified analyses of subgroups defined according to the median of C-reactive protein level, SOFA score, and time from admission to inclusion. Exploratory post hoc analyses addressed subgroups based on the periods determined by the predominant variant in Belgium during the trial and subgroups defined according to the median total antibody IgG and neutralizing antibody titers against SARS-CoV-2 at the time of inclusion.

Adult patients who had a score on the Clinical Frailty Scale of less than 6 (range, 1 to 9, with higher scores indicating greater frailty),14 who had been admitted to a participating ICU with a diagnosis of Covid-19–induced ARDS (corresponding to a score of 7, 8, or 9 on the World Health Organization [WHO] Clinical Progression Scale, with scores ranging from 0 to 10 and higher scores indicating more severe disease15), and who had received invasive mechanical ventilation for a maximum of 5 days were assessed for eligibility. ARDS was classified according to the Berlin definition.16 Covid-19–induced ARDS was defined as extended interstitial pneumonia on a computed tomographic scan or a chest radiograph within 10 days before inclusion and a positive result of a clinical SARS-CoV-2 nasopharyngeal polymerase-chain-reaction (PCR) test within 15 days before inclusion. Exclusion criteria were pregnancy, a previous episode of transfusion-related side effects, a medical decision to limit therapy, and participation in another Covid-19 trial. Before requesting consent, the local investigator contacted a physician at the coordinating center, who checked the inclusion criteria. At the same time, the availability of convalescent plasma with an ABO group identical to the eligible patient’s group was checked on the Web site of the Belgian Red Cross. The capability of the patient to give consent was then assessed. If consent could not be obtained from the patient, it was asked of a relative. When patients became able to give consent, they were asked to do so.

Donors were recruited by the Belgian Red Cross among adults who had been infected with SARS-CoV-2 and who had fully recovered between 28 days and 10 months earlier.13 Neutralizing antibody titers were determined with SARS-CoV-2 Nextstrain clade 20B (Wuhan-like, B.1.1), isolated from a Belgian patient, in 96-well plates containing confluent Vero E6 cells (ATCC CRL-1586)17 and were reported as 50% virus neutralization titer (NT50). The NT50 was the highest dilution of serum that neutralized the cytopathic effect in 50% of the cells (see the protocol).13,18 A neutralizing antibody titer of at least 1:320 was requested for the trial, and a titer of 1:160 was accepted in case of a shortage of plasma.

Randomization was stratified according to the delay from the initiation of invasive mechanical ventilation and performed with the use of a Web-based program. Patients in the convalescent-plasma group were assigned to receive an infusion of 2 units (400 to 500 ml in total) of convalescent plasma of the identical ABO group within 24 hours after inclusion. No crossover was allowed between the convalescent-plasma group and the standard-care group. Investigators were aware of the trial-group assignments.

The primary outcome was death by day 28 after randomization. Secondary outcomes included adverse events, inflammatory and anti–SARS-CoV-2 antibody responses, the Sequential Organ Failure Assessment (SOFA) score, the use of organ support, the length of hospital stay, and death by day 90 and 365.

We anticipated a day 28 mortality of 40% in the standard-care group. We estimated that a relative difference of one third in day 28 mortality was realistic and relevant. With a two-sided alpha level of 0.05 and a beta level of 0.20, the number of patients to include for an absolute difference of −13.5 percentage points (convalescent plasma minus standard care) in day 28 mortality was 250 for each group. Because our estimate of the between-group difference was empirical, we performed sequential blinded interim analyses. For these analyses, two Lan–Demets spending functions were used to adjust the type I error. The analysis was performed on an intention-to-treat basis. For the primary analysis, there were no missing values. No imputation was made for secondary outcomes because the pattern of missing data was completely at random, and missing data were either very infrequent or not used for statistical analyses. The results are expressed as median (interquartile range) and as number (percent). The primary outcome was compared with the chi-square test and adjusted for the stratification factor with a binary logistic regression. The two-sided level of significance was 5%. We summarized secondary analyses with point estimates for differences between the two groups and 95% confidence intervals using the Hodges–Lehmann estimation of location shift and a Wald asymptotic confidence interval for quantitative and binary variables, respectively. The confidence intervals were not adjusted for multiplicity and should not be used to infer definitive treatment effects. The Fine–Gray model was used to analyze competing risks data. The secondary analyses should be considered exploratory. Statistical analyses were performed with the use of SAS software for Windows, version 9.4, and R software for Windows, version 4.0.2.

Between September 10, 2020, and March 9, 2022, a total of 1034 patients with Covid-19–induced ARDS were screened and 475 (45.9%) were included (Fig. S6 in the Supplementary Appendix, available at NEJM.org). We decided to end recruitment prematurely, on April 9, 2022, because no new case had been screened in 1 month, presumably owing to reduced virulence of the omicron BA.2 variant.19 At inclusion, 47 patients (9.9%) had mild ARDS, 274 (57.7%) had moderate ARDS, and 154 (32.4%) had severe ARDS16; 342 (72.0%) underwent randomization 48 hours or less after the initiation of invasive mechanical ventilation, and 133 (28.0%) underwent randomization more than 48 hours after the initiation of invasive mechanical ventilation. A total of 237 patients were assigned to receive convalescent plasma, and 238 patients were assigned to receive standard care. These numbers were 171 and 171, respectively, among patients who underwent randomization 48 hours or less after ventilation initiation and 66 and 67, respectively, among those who underwent randomization more than 48 hours after ventilation initiation. Relatives provided consent for all the patients. Four patients declined to continue participation after recovery; with the exception of data on vital status, subsequent data for these patients were not collected. No patient was lost to follow-up regarding vital status.

n more than 48 hours after ventilation initiation. Relatives provided consent for all the patients. Four patients declined to continue participation after recovery; with the exception of data on vital status, subsequent data for these patients were not collected. No patient was lost to follow-up regarding vital status. All the patients in the convalescent-plasma group received convalescent plasma except 1 who died before infusion, and none in the standard-care group received convalescent plasma. Adherence to the standard of care for invasive mechanical ventilation was similar in the two groups (Table S2). No patient received polyvalent immunoglobulins between inclusion and day 28. The inclusion rate followed the pandemic waves in Belgium (Fig. S1). We included 197, 163, 99, and 16 patients during periods when the ancestral virus and B.1.1.7 (alpha), B.1.617.2 (delta), and B.1.1.529 (omicron) variants, respectively, were predominant in Belgium.20 The data and safety monitoring board recommended that the trial be continued after each interim analysis (see the Supplementary Appendix). The baseline characteristics are shown in Table 1 and Table S1 and include some imbalance in hypertension and diabetes between the two groups. In the convalescent-plasma group, convalescent plasma with a neutralizing antibody titer of 1:160, 1:320, 1:640, and more than 1:640 was administered in 17.7%, 38.8%, 31.9%, and 11.6%, respectively, of the patients. Glucocorticoids were administered to 466 patients (98.1%).

rtension and diabetes between the two groups. In the convalescent-plasma group, convalescent plasma with a neutralizing antibody titer of 1:160, 1:320, 1:640, and more than 1:640 was administered in 17.7%, 38.8%, 31.9%, and 11.6%, respectively, of the patients. Glucocorticoids were administered to 466 patients (98.1%). At day 28, mortality was 35.4% (84 of 237 patients) in the convalescent-plasma group and 45.0% (107 of 238) in the standard-care group (P=0.03, before and after adjustment for the stratification factor). These values were 32.7% (56 of 171) and 46.8% (80 of 171), respectively, among patients who underwent randomization 48 hours or less after ventilation initiation and 42% (28 of 66) and 40% (27 of 67), respectively, among those who underwent randomization more than 48 hours after ventilation initiation (Table 2). The survival curves in the convalescent-plasma and standard-care groups separated at approximately day 17 (Figure 1), and the difference in restricted mean survival time (convalescent plasma minus standard care) at day 28 was 0.33 days (95% CI, −1.27 to 1.92). The neutralizing antibody titer of the infused convalescent plasma was not associated with mortality. The direction of the effect was opposite in Center 105 (Figure 2). In this center, the standard-care group had a shorter time from ICU admission to inclusion than the convalescent-plasma group (Fig. S8).

1.27 to 1.92). The neutralizing antibody titer of the infused convalescent plasma was not associated with mortality. The direction of the effect was opposite in Center 105 (Figure 2). In this center, the standard-care group had a shorter time from ICU admission to inclusion than the convalescent-plasma group (Fig. S8). The investigators reported 711 adverse events and classified 209 of these as serious (Table 3 and Table S7). Of the 209 serious adverse events, 184 (88.0%) were fatal. All 711 events were attributed to Covid-19 or complications of organ support, and none was directly attributed to convalescent plasma. The occurrence of secondary bacteremia and pneumonia, duration of organ support, and length of hospital stay were similar in the two groups. Mean neutralizing antibody and total IgG titers to SARS-CoV-2 increased in both groups after inclusion. In the prespecified subgroup analysis, a higher SOFA score at inclusion was associated with a greater effect. The effect appeared similar in the four periods defined a posteriori by the predominance of the successive variants20 and in the subgroups defined according to the median total antibody IgG and neutralizing antibody titers against SARS-CoV-2 at the time of inclusion.

At day 28, mortality was 35.4% (84 of 237 patients) in the convalescent-plasma group and 45.0% (107 of 238) in the standard-care group (P=0.03, before and after adjustment for the stratification factor). These values were 32.7% (56 of 171) and 46.8% (80 of 171), respectively, among patients who underwent randomization 48 hours or less after ventilation initiation and 42% (28 of 66) and 40% (27 of 67), respectively, among those who underwent randomization more than 48 hours after ventilation initiation (Table 2). The survival curves in the convalescent-plasma and standard-care groups separated at approximately day 17 (Figure 1), and the difference in restricted mean survival time (convalescent plasma minus standard care) at day 28 was 0.33 days (95% CI, −1.27 to 1.92). The neutralizing antibody titer of the infused convalescent plasma was not associated with mortality. The direction of the effect was opposite in Center 105 (Figure 2). In this center, the standard-care group had a shorter time from ICU admission to inclusion than the convalescent-plasma group (Fig. S8).

The investigators reported 711 adverse events and classified 209 of these as serious (Table 3 and Table S7). Of the 209 serious adverse events, 184 (88.0%) were fatal. All 711 events were attributed to Covid-19 or complications of organ support, and none was directly attributed to convalescent plasma. The occurrence of secondary bacteremia and pneumonia, duration of organ support, and length of hospital stay were similar in the two groups. Mean neutralizing antibody and total IgG titers to SARS-CoV-2 increased in both groups after inclusion. In the prespecified subgroup analysis, a higher SOFA score at inclusion was associated with a greater effect. The effect appeared similar in the four periods defined a posteriori by the predominance of the successive variants20 and in the subgroups defined according to the median total antibody IgG and neutralizing antibody titers against SARS-CoV-2 at the time of inclusion.

In this trial involving patients admitted to the ICU with Covid-19–induced ARDS, administration of convalescent plasma with neutralizing antibody titers of at least 1:160 against SARS-CoV-2 early after the initiation of invasive mechanical ventilation reduced mortality at day 28. This effect was observed mainly in the patients who underwent randomization 48 hours or less after the initiation of invasive mechanical ventilation. The survival curves in both the overall population and the patients who underwent randomization 48 hours or less after ventilation initiation separated near day 17. Among the secondary outcomes, the results for inflammation, vasopressor support, and the number of adverse events tended to be better in the convalescent-plasma group. In prespecified subgroup analyses, we observed a greater effect in patients with a higher illness severity at inclusion.

itiation separated near day 17. Among the secondary outcomes, the results for inflammation, vasopressor support, and the number of adverse events tended to be better in the convalescent-plasma group. In prespecified subgroup analyses, we observed a greater effect in patients with a higher illness severity at inclusion. Our trial differs from previous trials of convalescent plasma in Covid-19 in three significant ways. First, we used convalescent plasma selected for higher neutralizing antibody titers as determined by a virus neutralization test. Plasma with higher titers is recommended by most experts and the Food and Drug Administration21,22 because low titers have been incriminated in several treatment failures.23,24 We used convalescent plasma with a neutralizing antibody titer of 1:320 in 82.3% of the patients and a titer of 1:160 in the remaining 17.7%. We used this lower titer owing to a shortage of donors in the early phase of the pandemic, and we accepted it because it still exceeded the titer that has been recommended.21,25 In our trial, convalescent plasma was assessed for NT50 in two academic centers, which ensured that a similar technique was used to assess all plasma.21

is lower titer owing to a shortage of donors in the early phase of the pandemic, and we accepted it because it still exceeded the titer that has been recommended.21,25 In our trial, convalescent plasma was assessed for NT50 in two academic centers, which ensured that a similar technique was used to assess all plasma.21 Second, our trial focused on patients with ARDS, corresponding to a score of 7, 8, or 9 on the WHO Clinical Progression Scale.15 Such a high severity was chosen because no antiviral treatment was available and the virus is often present for a prolonged period in most of these patients.7,8 In our trial, as in most Belgian centers after the first wave of the pandemic,26 invasive mechanical ventilation was initiated on the basis of statements by experts,27 which favored the previous use of high-flow nasal oxygen in an attempt to prevent tracheal intubation and invasive mechanical ventilation. By contrast, in the other trials addressing the efficacy of convalescent plasma, patients receiving invasive mechanical ventilation were only subgroups of the included patients,28,29 mortality was addressed as a secondary outcome,24 or invasive mechanical ventilation was received by a low number of patients,30 which led to inconclusive results.

trials addressing the efficacy of convalescent plasma, patients receiving invasive mechanical ventilation were only subgroups of the included patients,28,29 mortality was addressed as a secondary outcome,24 or invasive mechanical ventilation was received by a low number of patients,30 which led to inconclusive results. Third, to ensure homogeneity of the cohort, we chose to include patients as early as possible after the initiation of invasive mechanical ventilation. Because several tertiary centers expected to receive patients transferred from other institutions, we included patients until the fifth day of invasive mechanical ventilation and stratified randomization to assess the treatment effect in the first 48 hours of invasive mechanical ventilation with the lowest risk of bias.

ecause several tertiary centers expected to receive patients transferred from other institutions, we included patients until the fifth day of invasive mechanical ventilation and stratified randomization to assess the treatment effect in the first 48 hours of invasive mechanical ventilation with the lowest risk of bias. Mortality in the standard-care group was 45.0%, a finding consistent with those of previous studies. This high mortality may be due to the age of the population and the direct nature of the pulmonary injury.3 Our interpretation of the benefit with respect to mortality is that convalescent plasma is directed against the cause of Covid-19–induced ARDS in patients receiving invasive mechanical ventilation. Convalescent plasma may reduce the quantity of virus alive in the lungs of patients who continue to shed the virus.7,8 It may counteract the delayed kinetics of neutralizing antibody production in these patients,6 a phenomenon that was observed in a placebo-controlled trial.31 A decrease in viral load might reduce inflammation, as we observed in our convalescent-plasma group. The median time between viral infection and ARDS occurrence is 12 days.32 This delay is close to the one we observed between the administration of convalescent plasma and the separation in survival curves between the two groups. This finding is consistent with our hypothesis that neutralizing antibodies act at the beginning of the cascade between infection and inflammation.

nce is 12 days.32 This delay is close to the one we observed between the administration of convalescent plasma and the separation in survival curves between the two groups. This finding is consistent with our hypothesis that neutralizing antibodies act at the beginning of the cascade between infection and inflammation. We believe our trial has a good potential for generalizability to most patients with Covid-19–induced ARDS because almost 50% of such patients were included in the centers during their participation in the trial. Our results should be confirmed in patients with ARDS with different Covid-19 variants than those we studied. Confirmation could be accomplished by continuing collection campaigns and delivering the most recently collected plasma available.

patients were included in the centers during their participation in the trial. Our results should be confirmed in patients with ARDS with different Covid-19 variants than those we studied. Confirmation could be accomplished by continuing collection campaigns and delivering the most recently collected plasma available. Our trial has limitations. First, it was not blinded. We chose this design intentionally, because the use of a placebo would have added an extra fluid volume of 500 ml, which is against the recommended conservative approach when treating ARDS.33 To reduce the risk of bias due to the lack of blinding, we chose a nonsubjective primary outcome, death by day 28, that was not likely to be influenced by the knowledge of the trial-group assignment. The open design may have resulted in subtle differences in clinical decisions after randomization. Such potential differences may have been in either direction, depending on the confidence of each investigator regarding convalescent plasma. In our trial, if one group had been treated with a lower intensity, we assume the survival curves would have diverged in the days immediately after the randomization process, which was not the case. In addition, the results were not driven by a particular center. Second, the convalescent plasma was obtained between April 2020 and May 2021, when the ancestral virus and then its alpha variant were predominant in Belgium, and their neutralizing antibodies might have been less active against subsequent variants. However, a difference in treatment effect was not apparent between the periods when the delta or omicron BA.1 variants were predominant in Belgium. Third, the trial ended prematurely because of the absence of new inclusion, presumably owing to the reduced virulence of the omicron BA.2 variant. We ultimately included 475 patients, representing 95% of the planned inclusions, and it is unlikely that the results would have been modified by including 25 patients more. The reduced virulence of new variants may limit the immediate effect of our results, but future SARS-CoV-2 variants with increased virulence may occur. Fourth, we have not standardized our in-house neutralizing antibody titers against an international standard. The quantitative results generated by this method are known to vary among laboratories because, as a reference standard, it is mostly used as a research tool.

-CoV-2 variants with increased virulence may occur. Fourth, we have not standardized our in-house neutralizing antibody titers against an international standard. The quantitative results generated by this method are known to vary among laboratories because, as a reference standard, it is mostly used as a research tool. However, because the convalescent plasma that we selected for passive immunization had a higher NT50 than that measured in 85% of all the convalescent plasma we tested in a previous large-scale study,34 we believe that, using this qualitative selection threshold, other investigators should approach the potency of our own convalescent-plasma preparations. Fifth, data were missing for a substantial number of quantitative nasopharyngeal PCR and antibody measurements, and the interpretation of the corresponding secondary results should be cautious. The administration of plasma collected from convalescent donors and documented to have neutralizing antibody titers of at least 1:160 to patients with Covid-19–induced ARDS within 5 days after the initiation of invasive mechanical ventilation significantly reduced mortality at day 28. In a prespecified analysis, this effect was mainly observed in the patient group that underwent randomization 48 hours or less after the initiation of invasive mechanical ventilation.