Outpatient convalescent plasma therapy for high-risk patients with early COVID-19: a randomized placebo-controlled trial.

OBJECTIVES: The potential benefit of convalescent plasma (CP) therapy for coronavirus disease 2019 (COVID-19) is highest when administered early after symptom onset. Our objective was to determine the effectiveness of CP therapy in improving the disease course of COVID-19 among high-risk outpatients.
METHODS: A multicentre, double-blind randomized trial was conducted comparing 300 mL of CP with non-CP. Patients were ≥50 years, were symptomatic for <8 days, had confirmed RT-PCR or antigen test result for COVID-19 and had at least one risk factor for severe COVID-19. The primary endpoint was the highest score on a 5-point ordinal scale ranging from fully recovered (score = 1) or not (score = 2) on day 7, over hospital admission (score = 3), intensive care unit admission (score = 4) and death (score = 5) in the 28 days following randomization. Secondary endpoints were hospital admission, symptom duration and viral RNA excretion.
RESULTS: After the enrolment of 421 patients and the transfusion in 416 patients, recruitment was discontinued when the countrywide vaccination uptake in those aged >50 years was 80%. Patients had a median age of 60 years, symptoms for 5 days, and 207 of 416 patients received CP therapy. During the 28 day follow-up, 28 patients were hospitalized and two died. The OR for an improved disease severity score with CP was 0.86 (95% credible interval, 0.59-1.22). The OR was 0.58 (95% CI, 0.33-1.02) for patients with ≤5 days of symptoms. The hazard ratio for hospital admission was 0.61 (95% CI, 0.28-1.34). No difference was found in viral RNA excretion or in the duration of symptoms.
CONCLUSIONS: In patients with early COVID-19, CP therapy did not improve the 5-point disease severity score.

Introduction

Older patients and patients with medical comorbidities have an increased risk of hospitalization or death due to coronavirus disease 2019 (COVID-19). Over the last year, several antiviral therapies in the form of virus neutralizing monoclonal antibodies or as direct-acting antiviral drugs have shown to lower the risk of hospital admission when administered in the first days after disease onset.(1, 2, 3, 4, 5).

However, escape by new variants of concern (VOC) from neutralization by monoclonal antibodies has been observed for every single new VOC. Most recently, the Omicron BA.1 variant was shown to be completely resistant to casirivimab/imdevimab and much less susceptible to tixagevimab/cilgavimab(6), and while sotrovimab retained most of its activity against BA.1, this is no longer the case against BA.2.(7).

An alternative source of SARS-CoV-2 antibodies can be found in convalescent plasma (CP) from COVID-19 recovered and more recently also plasma from fully vaccinated people. CP has the advantage of being polyclonal and is less likely to be completely inactive against a new VOC. Just like monoclonal antibodies, the potential of CP as a treatment for COVID-19 lies in its use in high-risk patients early after symptom onset.(8) Here we report the results of the CoV-Early study: a randomized clinical trial designed to evaluate the effectiveness of CP in high-risk COVID-19 outpatients with less than 8 days of symptoms.

Methods

Study design

The CoV-Early study was a phase 3, multicentre, randomized, double-blind, placebo-controlled trial conducted in the Netherlands. The study was funded by ZonMW (grant-number 10430062010001) and supported by Sanquin (Dutch blood supply), which provided the plasma. A complete list of participating sites is available in our online protocol. The medical ethical review board of the Erasmus MC gave approval for the study (METC-2020-0682).

Randomization and intervention

All patients were randomized 1:1 CP or regular Plasma (non-CP) as placebo. Randomization was done using a web-based system (ALEA) and only a person from the transfusion laboratory was unblinded. Details about the CP and the control plasma as well as the neutralizing antibody assessment are provided in the online supplement.

Patients

Inclusion criteria were a nasopharyngeal PCR or a CE-antigen test that confirmed the SARS-CoV-2 infection, 7 or fewer days of symptoms at the time of screening (maximum of 8 on the day of transfusion) and not admitted to the hospital. In addition, patients needed to be at an increased risk for severe disease which we defined as one of the following: 1) 70 years or older, 2) 50 years or older with an additional comorbidity and 3) patients 18 years or older who are severely immunocompromised (criteria available in the full protocol).(9, 10) Patients were excluded when 1) their life expectancy was <28 days, 2) they were unable to provide informed consent, 3) COVID-19 symptoms were already improving, 4) they had a documented IgA deficiency or 5) they had a previous history of transfusion-related acute lung injury. All study participants provided written informed consent. Details on the recruitment and referral process are available in the online supplement.

Procedures and outcomes

After consent, patients received CP or non-CP at one of the study sites and left the hospital one hour after the infusion. Patients were contacted to evaluate their illness severity on day 7, 14 and 28. This was assessed using a 5-point ordinal disease severity scale. A score of 1 indicates the patient is fully recovered within 7 days, 2 indicates continued symptoms at day 7, 3 indicates admitted to hospital but no invasive ventilation, 4 indicates admitted to hospital and invasive ventilation needed and 5 indicates death. Serum was collected at baseline to determine the antibody status preceding the transfusion.

Nasopharyngeal swab as well as serum was collected on day 1, 3, 7, 14 and 28 in a subgroup of patients able and willing to attend extra visits at 2 dedicated study sites. A SARS-CoV-2 PCR was performed on these swabs to determine the cyclic threshold (Ct)-value of the E-gene to evaluate the rate of viral decay for both study arms (Cobas® 6800 (Roche)). Antibody titers in Binding Antibody Units (BAU)/mL were determined on serum using an IgG anti-Spike SARS-CoV-2 test (Trimeric Spike antibody test, Liaison, Diasorin, Saluggia, Italy).

The primary outcome was the improvement on the 5-point ordinal disease severity scale. Improvement was determined using the highest score in 28 days. Secondary outcomes were hospital admission and symptom duration. Exploratory endpoints were the treatment modifying effect of age, clinical frailty using the clinical frailty score version 2.0, symptom duration before inclusion and the height of the antibody titers in CP on the disease severity scale.(11) Additional exploratory endpoints were the difference in probability of a positive PCR in 28 days after inclusion, the difference in Ct-value in 28 days and the difference in antibody levels in 28 days between CP and non-CP. We also explored the difference in length of hospital stay in the two groups. Details of the Bayesian statistical analysis are available in the online supplement.

Results

Recruitment

Between November 2020 and July 2021 (when the D614G and alpha B.1.1.7 variant were dominant) 3545 COVID-19 patients were contacted and screened. The reasons for a patient being excluded from participation are summarized in Figure 1 . A total of 421 outpatients were randomized, of whom 416 were transfused and included in the intention to treat (ITT) population and none were lost to follow-up. 207 were included in the CP group and 209 in the non-CP group. During study monitoring, the observation was made that 5 of the 416 patients had received the incorrect treatment (4 received non-CP instead of CP and 1 received CP instead of non-CP). These patients were included in the ITT analysis and analysed according to the study arm they had been randomized to. None of these 5 patients was hospitalized during follow-up. Based on the recommendation by the DSMB, the study was terminated on the 13th of July 2021 before the planned sample size of 690 was reached. More details are available in the online supplement.

Figure 1

CONSORT flowdiagram.

Baseline characteristics

Patients had a median age of 60 years (IQR 55–65 years), a median of 5 days of symptoms before inclusion (IQR 4-6 days), a median of 1 comorbidity (IQR 1–2) and a median frailty score of 2 (IQR 1-2) which was measured in 174 patients only. 93 (22.4%) patients were female. The median O2-saturation at baseline (right before transfusion) without supplementary oxygen was 97% (IQR 96%–98%). All but 30 participants (7.9%) were SARS-CoV-2 IgG antibody negative at baseline. 12 (2.9%) patients had been fully vaccinated and 21 (5.0%) patients had received one vaccination. The baseline characteristics of the patients in the CP and non-CP groups were comparable (Table 1 ). CP had a median neutralization level of 386 international units/mL (IQR 271-707 international units/mL).

Table 1

Baseline characteristics

Characteristic	Total (n = 416)	CPa (n = 207)	Non-CPa (n = 209)
Male sex – no. (%)	323 (77.6%)	164 (79.2%)	159 (76.1%)
Age - median (IQR)	60 (55 – 65)	59 (55 – 65)	60 (54 – 66)
O2 saturation - median (IQR)b	97 (96 – 98)	98 (96 – 98)	97 (96 – 98)
BMI – median (IQR)	27.4 (24.6 – 30.7)	27.4 (24.5 – 30.8)	27.5 (24.8 – 30.6)
Number of comorbidities - median (IQR)c	1 (1 – 2)	1 (1 – 2)	1 (1 – 2)
Obesity – no. (%)	46 (11.1%)	21 (10.2%)	25 (12.0%)
Cardiac and/or pulmonary disease – no.(%)	140 (33.7%)	72 (34.8%)	68 (32.5%)
Neurological disease – no.(%)	22 (5.3%)	15 (7.2%)	7 (3.3%)
Diabetes – no.(%)	29 (7.0%)	8 (3.9%)	21 (10.0%)
Chronic kidney disease – no.(%)	20 (4.8%)	11 (5.3%)	9 (4.3%)
Rheumatic disease – no.(%)	21 (5.0%)	11 (5.3%)	10 (4.8%)
Severe immunodeficiency – no.(%)	15 (3.6%)	5 (2.4%)	10 (4.8%)
Cancer – no.(%)	14 (3.4%)	7 (3.4%)	7 (3.3%)
Chronic liver disease – no.(%)	3 (0.7%)	1 (0.5%)	2 (1.0%)
HIV – no.(%)	1 (0.2%)	0 (0%)	0 (0.5%)
Days since first symptoms - median (IQR)	5 (4 – 6)	5 (4 – 6)	5 (4 – 7)
Positive antibody status at baseline – no. (%)‖	30 (7.9%)	17 (8.8%)	13 (7.1%)
Fully vaccinated at baseline – no. (%)	12 (2.9%)	6 (2.9%)	6 (2.9%)
Only one vaccination at baseline – no. (%)	21 (5.0%)	8 (3.9%)	13 (6.2%)

‖‖ The anti-S, Liaison IgG test (Diasorin) was available for 381 of the 416 patients at baseline.

Convalescent plasma.

Baseline oxygen saturation without supplementary oxygen.

Obesity, cardiac disease, lung disease, neurological disease, diabetes, chronic renal failure, cancer and/or liver disease. See the supplementary appendix for additional details of the comorbidities.

Primary endpoints

Table 2 shows the distribution of the 5-point disease severity scale. The estimated common OR for the highest disease status in the 28 days after randomization was 0.86 (95% credible interval 0.59–1.22) for patients treated with CP. Hospital admission occurred in 10 (4.8%) patients receiving CP versus 18 (8.6%) patients in the non-CP arm with an adjusted hazard ratio of 0.61 (95% confidence interval (CI) 0.28–1.34). Death occurred in 1 (0.5%) patient treated with CP and in 1 (0.5%) patient treated with placebo. No difference was found in the median duration of symptoms between CP and non-CP patients (13 days vs 12 days, p=0.99, Supplementary figure 1).

Table 2

Distribution of the outcome of the patients in the 28 days after inclusion across the 5-points disease severity scale.

Worst Disease Severity Score	Total	CPa	Non-CPa
	(n = 416)	(n = 207)	(n = 209)
Recovered – no. (%)b	112 (26.9%)	59 (28.5%)	53 (25.4%)
Continued symptoms – no. (%)c	274 (65.9%)	137 (66.2%)	137 (65.6%)
Admitted to hospital but no invasive ventilation – no. (%)	27 (6.5%)	10 (4.8%)	17 (8.1%)
Admitted to hospital and invasive ventilation – no. (%)	1 (0.2%)	0 (0%)	1 (0.5%)
Death – no. (%)	2 (0.5%)	1 (0.5%)	1 (0.5%)

Convalescent plasma.

Recovered with no symptoms within 7 days after inclusion.

Continued symptoms attributable to COVID-19 at day 7.

Subgroup and laboratory analyses

No significant treatment modifying effect of age, clinical frailty, symptom duration or height of antibody titers in CP were found, as all these interaction terms had a p-value >0.1 (Supplementary figure 2 and 3, Supplementary table 1). Supplementary table 2 shows the distribution of the 5-point disease severity scale in non-CP patients across these subgroups. 85 patients were followed with nasopharyngeal SARS-CoV-2 PCR testing over time. No difference in the probability of having SARS-CoV-2 detected by PCR over time was observed between CP and placebo (p=0.51, Supplementary figure 4). Similarly, there was no difference in evolution of Ct-values over time between the 2 groups (p=0.35, Supplementary figure 5). The antibody levels of CP rose faster after transfusion when compared to placebo (p<0.0001, Figure 2 ), but this effect was short-lived. Supplementary figure 6 and 7 show the individual antibody titers and Ct-values. The median duration of hospital stay was 6 days (IQR 4.5 – 9 days) in the patients treated with CP and was 4 days (IQR 3.5 – 10.5 days) in the control group (p = 0.56, post hoc analysis)

Figure 2

Antibody titers over time between CP and non-CP patients using a mixed effects model. The line represents fitted log(antibodies in BAU/ml + 1) and the dotted line represents 95% confidence intervals. 85 patients had a day 1 measurement, 82 patients had a day 3 measurement, 79 patients had a day 7 measurement, 83 patients had a day 14 measurement and 82 patients had a day 28 measurement.

Discussion

In this randomized trial we evaluated whether 300mL of CP improved the outcome of COVID-19 in outpatients at risk for severe disease. The outcome was evaluated using a 5-point ordinal disease severity scale which included speed of recovery, hospital admission, ICU admission and survival during a 28-day follow-up period. Overall, CP did not significantly prevent a more severe disease course. The OR (0.58) was numerically lower in the subgroup of patients with ≤5 days of symptoms. The duration of symptoms and the time to become PCR negative was not reduced. Hospital admission was decreased from 8.6% to 4.8% but the 95%CI of the OR was wide (0.28-1.34).

This is the fifth clinical trial on early CP therapy for COVID-19. Some of the previous studies recruited different populations which makes the comparison across studies difficult. Libster et al. observed a beneficial effect of CP but focused on patients aged 75 or older with ≤72 hours of symptoms.(12) The small sample size and the specific patient population precludes definite conclusions. The second trial by Korley et al. recruited patients presenting with COVID-19 at the emergency room.(13) These patients were therefore probably in a later inflammatory stage of the disease at a time when the benefit of CP could be lower. Indeed, a large number of the hospital admissions in this trial occurred on the day of study enrolment and consequently diluted any potential treatment effect. In a third randomized trial by Alemany et al. CP did not reduce the risk of hospital admission in patients aged 50 or older.(14) Similar to our study, this trial was terminated early due to the rapid SARS-CoV-2 vaccination uptake in Spain and was therefore underpowered. In the most recent and largest trial on CP for outpatients with COVID-19, Sullivan et al. observed a significant reduction in hospital admission (from 6.3% to 2.9%) with CP.(15) The benefit of CP was limited to the patients with 5 or fewer days of symptoms at the time of transfusion. The non-significant reduction in hospital admission from 8.6 to 4.8% that we observed is in agreement with these findings.

Our study has several strengths. First, we found that more than 90% of patients were SARS-CoV-2 antibody negative at baseline. This demonstrates that we were able to identify patients with early disease and in whom clinical benefit from CP could be expected. Second, as the CP donors were selected with a whole SARS-CoV-2 PRNT50, we were certain that the CP we used contained functional antibodies. This is also demonstrated by the significant difference in antibody levels during follow-up between CP and non-CP.

Our study also has its limitations. Despite that patients were selected based on their age and having at least one comorbidity, the hospital admission rate in the control arm was 8.6% rather than the 20% we anticipated in the sample size calculation. This illustrates that even in an unvaccinated population and while using data provided by the RIVM on risk factors for hospital admission, identifying patients at risk for a poor outcome is challenging.

Secondly, the study was discontinued after 421 rather than the planned 690 inclusions. This decision was made after a meeting with the DSMB at the time when the vaccination uptake in people 50 years or older had reached 80% and monoclonal antibodies had become available as outpatient treatment in the Netherlands. This made further recruitment futile since the vaccination rate lowered the hospital admission rate far below our anticipated hospitalization rate. Also, randomizing high-risk patients to a placebo when monoclonal antibodies are available was ethically not acceptable. In conclusion, our study was therefore insufficiently powered to detect the relatively large effect size we anticipated reflected by the wide credible and confidence intervals. A meta-analysis of individual patient data from all available studies on CP for outpatients with early COVID-19 is therefore needed and will help to get a more precise estimate of the actual effect size. The ongoing Cochrane meta-analysis of CP for COVID-19 will be looking at studies on outpatients with COVID-19 in their next update to fill this knowledge gap.(16).

Thirdly, we acknowledge that the much lower hospital admission rates observed with the currently circulating Omicron VOC in an almost fully vaccinated population of which many have also recovered from COVID-19 will make the magnitude of any potential benefit from CP smaller.

Fourthly, a selection of the perfect placebo can be complicated since each potential placebo has advantages and disadvantages (regulatory as well as practical). First, the obvious advantage of plasma as a control intervention is its identical physical characteristics and therefore perfect masking of the patient. Furthermore, by using plasma as the control intervention, the effect of the SARS-CoV-2 neutralizing antibodies on top of any other theoretical effect that plasma may have on COVID-19 can be studied. However, much higher Ig doses are typically used when this immune modulating effect is pursued so we considered this effect unlikely. Furthermore, any additional impact of plasma on the disease course would then be present in both study arms.

Finally, it should be emphasized that at the time the study was designed, we were uncertain about the minimum dose of antibodies required to result in a clinically relevant antiviral effect. In fact, this remains an important knowledge gap and based on the currently available evidence, it is unlikely that the dose of 250-300mL from a single donor with a minimum neutralization titer of 1:160 that we used in this trial is the optimal dose.(17) Fortunately, the collection of plasma with substantially higher virus neutralizing antibody titers has become less challenging. We recently tested BA.2 virus neutralization properties of plasma from 103 donors that had been fully vaccinated and boostered with a mRNA vaccine. Twenty donors (19.4%) had a PRNT50 titer of 1:10.000 or higher against Omicron BA.2. Consequently, treatment with CP containing much higher doses of virus neutralizing antibodies than used in the current study has become possible. With the BA.1, BA.2 and now also BA.4/5 variants escaping neutralization by many of the currently available monoclonal antibodies, the interest in CP as a treatment for COVID-19 will likely increase again. Because the protection from vaccination against Omicron VOC is often suboptimal in the immunocompromised patient, the morbidity of COVID-19 in these patients continues to be substantial despite the lower pathogenicity of the Omicron VOC.(18, 19, 20) As a result, CP could become a valuable part of the anti-COVID-19 armamentarium for selected patients. In our opinion, the focus of future studies on the clinical efficacy of CP should therefore be on the immunocompromised patient. Indeed, it is very unfortunate that 2 years into the pandemic, data from large randomized trials on the value of CP in immunocompromised patients are still lacking completely.(21).

In conclusion, in high risk COVID-19 outpatients, treatment with CP in the first week after symptom onset did not improve the overall disease severity. The reduced hospital admission rate that we observed is in line with data from Sullivan et al. and should be the subject of an individual patient data meta-analysis of the currently available trials on CP for outpatients with early COVID-19.

Conflicts of interest

We declare no conflicts of interest.

Funding

This study was made possible by a research grant from ZONMW, the Netherlands (10430062010001). Sanquin Blood Supply provided convalescent plasma free of charge for study sites in the Netherlands.

Access to data

BJAR, AG and GP have full access to the data and BJAR is the guarantor for the data.

Contribution

All authors contributed to the final approval of the version to be submitted.

Arvind Gharbharan, Carlijn Jordans, Lisa Zwaginga, Peter van Wijngaarden, Jan den Hollander, Faiz Karim, Elena van Leeuwen-Segarceanu, Robert Soetekouw, Jolanda Lammers, Douwe Postma, Linda Kampschreur, Geert Groeneveld, Hannelore Götz, Susanne Bogers: contributed to the acquisition of the data.

Arvind Gharbharan, Grigorios Papageorgiou, Nan van Geloven, Bart Rijnders: contributed to the analysis and interpretation of the data.

Arvind Gharbharan, Carlijn Jordans, Grigorios Papageorgiou, Nan van Geloven, Francis Swaneveld, C. Ellen van der Schoot, Bart Haagmans, Marion Koopmans, Corine Geurtsvankessel, Jaap Jan Zwaginga, Casper Rokx, Bart Rijnders: contributed to the conceptualization and intellectual input of the protocol and manuscript.