COVID-19 : Effectiveness of convalescent plasma (CP) transfusion to rescue severe patients

Since December 2019, a pneumonia associated with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), named as coronavirus disease 2019 (COVID-19) by World Health Organization (WHO), emerged in Wuhan, China (1–3).

The epidemic spread rapidly worldwide within 3 mo and was characterized as a pandemic by WHO on March 11, 2020. As of March 12, 2020, a total of 80,980 confirmed cases and 3,173 deaths had been reported in China. Meanwhile, a total of 44,377 confirmed cases and 1,446 deaths was reported in another 108 countries or regions.

Currently, there are no approved specific antiviral agents targeting the novel virus, while some drugs are still under investigation, including remdesivir and lopinavir/ritonavir (4, 5).

Although remdesivir was reported to possess potential antiviral effect in one COVID-19 patient from the United States, randomized controlled trials of this drug are ongoing to determine its safety and efficacy (6).

Moreover, the corticosteroid treatment for COVID-19 lung injury remains controversial, due to delayed clearance of viral infection and complications (7, 8). Since the effective vaccine and specific antiviral medicines are unavailable, it is an urgent need to look for an alternative strategy for COVID-19 treatment, especially among severe patients.

Convalescent plasma (CP) therapy, a classic adaptive immunotherapy, has been applied to the prevention and treatment of many infectious diseases for more than one century. Over the past two decades, CP therapy was successfully used in the treatment of SARS, MERS, and 2009 H1N1 pandemic with satisfactory efficacy and safety (9–12).

—————-

What is Convalescent Plasma?

Convalescent plasma is the liquid part of blood that is collected from patients who have recovered from the novel coronavirus disease, COVID-19, caused by the virus SARS-CoV-2. COVID-19 patients develop antibodies in the blood against the virus. Antibodies are proteins that might help fight the infection. Convalescent plasma is being investigated for the treatment of COVID-19 because there is no approved treatment for this disease and there is some information that suggests it might help some patients recover from COVID-19.

——————-

A meta-analysis from 32 studies of SARS coronavirus infection and severe influenza showed a statistically significant reduction in the pooled odds of mortality following CP therapy, compared with placebo or no therapy (odds ratio, 0.25; 95% confidence interval, 0.14–0.45) (13).

However, the CP therapy was unable to significantly improve the survival in the Ebola virus disease, probably due to the absence of data of neutralizing antibody titration for stratified analysis (14).

Since the virological and clinical characteristics share similarity among SARS, Middle East Respiratory Syndrome (MERS), and COVID-19 (15), CP therapy might be a promising treatment option for COVID-19 rescue (16).

Patients who have recovered from COVID-19 with a high neutralizing antibody titer may be a valuable donor source of CP. Nevertheless, the potential clinical benefit and risk of convalescent blood products in COVID-19 remains uncertain.

Hence, we performed this pilot study in three participating hospitals to explore the feasibility of CP treatment in 10 severe COVID-19 patients.

Results

Neutralizing Activity of CP against SARS-CoV-2.

The neutralizing activity against SARS-CoV-2 was evaluated by classical plaque reduction test using a recently isolated viral strain (1). Among the first batch of CP samples from 40 recovered COVID-19 patients, 39 showed high antibody titers of at least 1:160, whereas only one had an antibody titer of 1:32. This result laid the basis for our pilot clinical trial using CP in severe patients.

General Characteristics of Patients in the Trial.

From January 23, 2020, to February 19, 2020, 10 severe COVID-19 patients (six males and four females) were enrolled and received CP transfusion. The median age was 52.5 y (interquartile range [IQR], 45.0 y to 59.5 y) (Table 1).

None of the patients had direct exposure to Huanan Seafood Wholesale Market. The median time from onset of symptoms to hospital admission and CP transfusion was 6 d (IQR, 2.5 d to 8.5 d) and 16.5 d (IQR, 11.0 d to 19.3 d), respectively.

Three patients were affected by clustering infection. The most common symptoms at disease onset were fever (7 of 10 patients), cough (eight cases), and shortness of breath (eight cases), while less common symptoms included sputum production (five cases), chest pain (two cases), diarrhea (two cases), nausea and vomiting (two cases), headache (one case), and sore throat (one case). Four patients had underlying chronic diseases, including cardiovascular and/or cerebrovascular diseases and essential hypertension.

Nine patients received arbidol monotherapy or combination therapy with remdesivir (in one case not included in the current clinical trial), or ribavirin, or peramivir, while one patient received ribavirin monotherapy (Table 2).

Antibacterial or antifungal treatment was used when patients had coinfection. Six patients received intravenous (i.v.) methylprednisolone (20 mg every 24 h).

Table 1.

Clinical characteristics of patients receiving CP transfusion

Patient no.	Sex	Age, y	Clinical classification	Days of admission from symptom onset	Days of CP therapy from symptom onset	Clustering infection	Principal symptoms	Comorbidity
1	M	46	Severe	8	11	No	Fever, cough, sputum production, shortness of breath, chest pain	Hypertension
2	F	34	Severe	0	11	Yes	Cough, shortness of breath, chest pain, nausea and vomiting	None
3	M	42	Severe	8	19	Yes	Fever, cough, sputum production, shortness of breath, sore throat, diarrhea	Hypertension
4	F	55	Severe	10	19	No	Fever, cough, sputum production, shortness of breath	None
5	M	57	Severe	4	14	No	Fever, shortness of breath	None
6	F	78	Severe	8	17	Yes	Fever, cough, sputum production, shortness of breath, muscle ache	None
7	M	56	Severe	4	16	No	Fever, cough, sputum production, arthralgia	None
8	M	67	Severe	10	20	No	Fever, cough, headache, diarrhea, vomiting	Cardiovascular and cerebrovascular diseases
9	F	49	Severe	1	10	No	Cough, shortness of breath	None
10	M	50	Severe	3	20	No	Shortness of breath	Hypertension

• M, male; F, female.

Table 2.

Other treatments of ten patients receiving CP transfusion

	Drugs administered	Oxygen support
Patient no.	Antiviral treatment	Antibiotic or antifungal treatment	Corticosteroids treatment	Before CP therapy	After CP therapy
1	Arbidol 0.2 g q8h po.Ribavirin 0.5 g qd i.v.	Cefoperazone Sodium i.v.	None	High-flow nasal cannula, mechanical ventilation	Mechanical ventilation
2	Arbidol 0.2 g q8h po.	Cefoperazone Sodium i.v.	None	None	None
3	Arbidol 0.2 g q8h po.	Moxifloxacin i.v.	Methylprednisolone i.v.	High-flow nasal cannula, mechanical ventilation	High-flow nasal cannula
4	Ribavirin 0.5 g qd i.v.	Linezolid i.v.Imipenem-Sitastatin Sodium i.v.	Methylprednisolone i.v.	Mechanical ventilation	High-flow nasal cannula
5	Arbidol 0.2 g q8h po.	Moxifloxacin i.v.	Methylprednisolone i.v.	Low-flow nasal cannula	Low-flow nasal cannula
Remdesivir 0.2 g qd i.v.	Cefoperagone Sodium and Tazobactam Sodium i.v.
IFN-ɑ 500MIU qd inh.
6	Arbidol 0.2 g q8h po.	Cefoperazone Sodium i.v.Levofloxacin i.v.	Methylprednisolone i.v.	High-flow nasal cannula	High-flow nasal cannula
7	Arbidol 0.2 g q8h po.	Cefoperagone Sodium and Tazobactam Sodium i.v.	Methylprednisolone i.v.	High-flow nasal cannula	None
	Fluconazole i.v.
8	Arbidol 0.2 g q8h po.	None	None	None	None
Ribavirin 0.5 g qd i.v.
9	Arbidol 0.2 g q8h po.Oseltamivir 75 mg q12h po.	None	None	Low-flow nasal cannula	Low-flow nasal cannula (Intermittent)
Peramivir 0.3 g qd i.v.
10	Arbidol 0.2 g q8h po.IFN-ɑ 500 MIU qd inh.	Cefoperazone Sodium i.v.Caspofungin i.v.	Methylprednisolone i.v.	High-flow nasal cannula	High-flow nasal cannula

• po., per os; i.v., i.v. injection; inh., inhalation; q8h, every 8 h; qd, per day; q12h, every 12 h; MIU, million IU.

On computer-assisted tomography (CT), all patients presented bilateral ground-glass opacity and/or pulmonary parenchymal consolidation with predominantly subpleural and bronchovascular bundles distribution in the lungs. Seven patients had multiple lobe involvement, and four patients had interlobular septal thickening.

Effects of CP Transfusion.

Improvement of clinical symptoms.

All symptoms in the 10 patients, especially fever, cough, shortness of breath, and chest pain, disappeared or largely improved within 1 d to 3 d upon CP transfusion. Prior to CP treatment, three patients received mechanical ventilation, three received high-flow nasal cannula oxygenation, and two received conventional low-flow nasal cannula oxygenation.

After treatment with CP, two patients were weaned from mechanical ventilation to high-flow nasal cannula, and one patient discontinued high-flow nasal cannula. Besides, in one patient treated with conventional nasal cannula oxygenation, continuous oxygenation was shifted to intermittent oxygenation (Table 2).

Reduction of pulmonary lesions on chest CT examinations.

According to chest CTs, all patients showed different degrees of absorption of pulmonary lesions after CP transfusion. Representative chest CT images of patient 9 and patient 10 are shown on Fig. 1. Patient 9, a 49-y-old female admitted 1 day postonset of illness (dpoi), showed the most obvious pulmonary image improvement.

At 10 dpoi, one dose of 200-mL transfusion of CP was given. The SARS-CoV-2 RNA converted to negative at 12 dpoi. Compared with the result at 7 dpoi, massive infiltration and ground-glass attenuation disappeared on CT image performed at 13 dpoi, accompanied by a much better pulmonary function. Patient 10, a 50-y-old male, was admitted 3 dpoi and was given a 200-mL transfusion of CP at 20 dpoi.

His chest CT presented massive infiltration and widespread ground-glass attenuation on admission and started to show a gradual absorption of lung lesions 5 d after CP transfusion. The SARS-CoV-2 RNA became negative at 25 dpoi.

Amelioration of routine laboratory criteria and pulmonary function.

Lymphocytopenia, an important index for prognosis in COVID-19 (2), tended to be improved after CP transfusion (median: 0.65 × 10⁹ per L vs. 0.76 × 10⁹ per L), 7 out of 10 patients showing an increase of lymphocyte counts (Fig. 2).

Concerning other laboratory tests, we observed a tendency of decrement of parameters indicative of inflammation and/or liver dysfunction as compared to the status before CP therapy.

These included C-reactive protein (CRP) (median: 55.98 mg/L vs. 18.13 mg/L), alanine aminotransferase (median: 42.00 U/L vs. 34.30 U/L), and aspartate aminotransferase (median: 38.10 U/L vs. 30.30 U/L) (Table 3).

The total bilirubin (median: 12.40 μmol/L vs. 13.98 μmol/L) remained unchanged, except for an obvious increment in patient 1 (Fig. 2). An increase of SaO₂ (median: 93.00% vs. 96.00%), a measurement constantly performed in most patients in our trial, was found, which could indicate recovering lung function.

This temporal relationship was notable despite the provision of maximal supportive care and antiviral agents.

Table 3.

Comparison of laboratory parameters before and after CP transfusion

Clinical factors	Before CP transfusion	After CP transfusion
CRP (mg/L, normal range 0 to 6)	55.98 (15.57 to 66.67)	18.13 (10.92 to 71.44)
Lymphocyte (109 per L, normal range 1.1 to 3.2)	0.65 (0.53 to 0.90)	0.76 (0.52 to 1.43)
Alanine aminotransferase (U/L, normal range 9 to 50)	42.00 (28.25 to 61.85)	34.30 (25.75 to 53.90)
Aspartate aminotransferase (U/L, normal range 15 to 40)	38.10 (28.50 to 44.00)	30.30 (17.30 to 38.10)
Total bilirubin (μmol/L, normal range 0 to 26)	12.40 (11.71 to 22.05)	13.98 (12.20 to 20.80)
SaO2 (%, normal range ≥ 95)	93.00 (89.00 to 96.50)	96.00 (95.00 to 96.50)

• SaO₂, oxyhemoglobin saturation.

Remarkably, patient 1, a 46-y-old male admitted 8 dpoi, had a very quick recovery, with much improved result of laboratory tests. He received antiviral drugs (arbidol and ribavirin) treatment and high-flow nasal cannula on admission. Mechanical ventilation was given at 10 dpoi for critical care support.

CP transfusion was performed at 11 dpoi. At 12 dpoi, the SARS-CoV-2 test turned to negative, with a sharp decrease of CRP from 65.04 mg/L to 23.57 mg/L and increment of SaO₂ from 86% to 90% (Fig. 3).

The mechanical ventilation was successfully weaned off 2 d after CP transfusion. At 15 dpoi, a steady elevation of lymphocyte count and a drop of aminopherase level were observed, indicating improvement of immunological and hepatic function.

Increase of neutralizing antibody titers and disappearance of SARS-CoV-2 RNA.

We determined neutralizing antibody titers before and after CP transfusion in all patients except one (patient 2) (Table 4). The neutralizing antibody titers of five patients increased and four patients remained at the same level after CP transfusion.

SARS-CoV-2 RNA, assayed by RT-PCR, was positive in seven patients and negative in three cases before CP transfusion.

Of note is that SARS-CoV-2 RNA was decreased to an undetectable level in three patients on day 2, three patients on day 3, and one patient on day 6 after CP therapy. These results are in support of a neutralizing effect of CP on serum SARS-CoV-2.

Table 4.

Comparison of serum neutralizing antibody titers and SARS-CoV-2 RNA load before and after CP therapy

Patient no.	CP transfusion date	Before CP transfusion	After CP transfusion
Date	Serum neutralizing antibody titers	Serum SARS-CoV-2 RNA load (Ct value)	Date	Serum neutralizing antibody titers	Serum SARS-CoV-2 RNA load (Ct value)
1	February 9	February 8	1:160	37.25	February 10	1:640	Negative
2	February 9	February 8	Unavailable	35.08	February 11	Unavailable	Negative
3	February 13	February 12	1:320	38.07	February 14	1:640	Negative
4	February 13	February 12	1:160	37.68	February 14	1:640	Negative
5	February 12	February 11	1:640	Negative	February 14	1:640	Negative
6	February 12	February 11	1:640	Negative	February 14	1:640	Negative
7	February 12	February 11	1:320	34.64	February 14	1:640	Negative
8	February 12	February 11	1:640	35.45	February 14	1:640	Negative
9	February 12	February 11	1:160	Negative	February 14	1:640	Negative
10	February 9	February 8	1:640	38.19	February 14	1:640	Negative

Outcome of patients treated with CP as compared to a recent historic control group.

A historic control group was formed by random selection of 10 patients from the cohort treated in the same hospitals and matched by age, gender, and severity of the diseases to the 10 cases in our trial.

Baseline characteristics of patients between CP treatment group and control group showed no significant differences, while clinical outcomes of these two groups were different: three cases discharged while seven cases in much improved status and ready for discharge in CP group, as compared to three deaths, six cases in stabilized status, and one case in improvement in the control group (P < 0.001; SI Appendix, Table S1).

Adverse Effects of CP Transfusions.

Patient 2 showed an evanescent facial red spot. No serious adverse reactions or safety events were recorded after CP transfusion.

Discussion

Our study explores the feasibility of CP therapy in COVID-19. All enrolled severe COVID-19 patients achieved primary and secondary outcomes.

One dose of 200-mL CP transfusion was well tolerated, while the clinical symptoms significantly improved with the increase of oxyhemoglobin saturation within 3 d, accompanied by rapid neutralization of viremia.

Severe pneumonia caused by human coronavirus was characterized by rapid viral replication, massive inflammatory cell infiltration, and elevated proinflammatory cytokines or even cytokine storm in alveoli of lungs, resulting in acute pulmonary injury and acute respiratory distress syndrome (ARDS) (17).

Recent studies on COVID-19 demonstrated that the lymphocyte counts in the peripheral blood were remarkably decreased and the levels of cytokines in the plasma from patients requiring intensive care unit (ICU) support, including IL-6, IL-10, TNF-ɑ, and granulocyte-macrophage colony-stimulating factor, were significantly higher than in those who did not require ICU conditions (2, 18).

CP, obtained from recovered COVID-19 patients who had established humoral immunity against the virus, contains a large quantity of neutralizing antibodies capable of neutralizing SARS-CoV-2 and eradicating the pathogen from blood circulation and pulmonary tissues (19).

In the present study, all investigated patients achieved serum SARS-CoV-2 RNA negativity after CP transfusion, accompanied by an increase of oxygen saturation and lymphocyte counts, and the improvement of liver function and CRP.

The results suggest that the inflammation and overreaction of the immune system were alleviated by antibodies contained in CP. The case fatality rates (CFRs) in the present study were 0% (0/10), which was comparable to the CFRs in SARS, which varied from 0% (0/10) to 12.5% (10/80) in four noncomparative studies using CP treatment (9, 20⇓–22).

Based on our preliminary results, CP therapy can be an easily accessible, promising, and safe rescue option for severe COVID-19 patients. It is, nevertheless, worth mentioning that the absorption of pulmonary lesions often lagged behind the improvement of clinical symptoms, as shown in patients 9 and 10 in this trial.

The first key factor associated with CP therapy is the neutralizing antibody titer. A small sample study in MERS-CoV infection showed that the neutralizing antibody titer should exceed 1:80 to achieve effective CP therapy (12).

To find eligible donors who have high levels of neutralizing antibody is a prerequisite. Cao et al. (23) showed that the level of specific neutralizing antibody to SARS-CoV decreased gradually 4 mo after the disease process, reaching undetectable levels in 25.6% (IgG) and 16.1% (neutralizing antibodies) of patients at 36 mo after disease status.

A study from the MERS-CoV−infected patients and the exposed healthcare workers showed that the prevalence of MERS-CoV IgG seroreactivity was very low (2.7%), and the antibodies titer decreased rapidly within 3 mo (24).

These studies suggested that the neutralizing antibodies represented short-lasting humoral immune response, and plasma from recently recovered patients should be more effective. In the present study, recently recovered COVID-19 patients, who were infected by SARS-CoV-2 with neutralizing antibody titer above 1:640 and recruited from local hospitals, should be considered as suitable donors.

The median age of donors was lower than that of recipients (42.0 y vs. 52.5 y). Among the nine cases investigated, the neutralizing antibody titers of five patients increased to 1:640 within 2 d, while four patients kept the same level. The antibody titers in CP in COVID-19 seem thus higher than those used in the treatment of MERS patient (1:80) (12).

The second key factor associated with efficacy is the treatment time point. A better treatment outcome was observed among SARS patients who were given CP before 14 dpoi (58.3% vs. 15.6%; P < 0.01), highlighting the importance of timely rescue therapy (9).

The mean time from onset of illness to CP transfusion was 16.5 d. Consistent with previous research, all three patients receiving plasma transfusion given before 14 dpoi (patients 1, 2, and 9) in our study showed a rapid increase of lymphocyte counts and a decrease of CRP, with remarkable absorption of lung lesions in CT.

Notably, patients who received CP transfusion after 14 dpoi showed much less significant improvement, such as patient 10. However, the dynamics of the viremia of SARS-CoV-2 was unclear, so the optimal transfusion time point needs to be determined in the future.

In the present study, no severe adverse effects were observed. One of the risks of plasma transfusion is the transmission of the potential pathogen. Methylene blue photochemistry was applied in this study to inactivate the potential residual virus and to maintain the activity of neutralizing antibodies as much as possible, a method known to be much better than ultraviolet (UV) C light (25).

No specific virus was detected before transfusion. Transfusion-related acute lung injury was reported in an Ebola virus disease woman who received CP therapy (26). Although uncommon in the general population receiving plasma transfusion, this specific adverse reaction is worth noting, especially among critically ill patients experiencing significant pulmonary injury (27).

Another rare risk worth mentioning during CP therapy is antibody-dependent infection enhancement, occurring at subneutralizing concentrations, which could suppress innate antiviral systems and thus could allow logarithmic intracellular growth of the virus (28). The special infection enhancement also could be found in SARS-CoV infection in vitro (29). No such pulmonary injury and infection enhancement were observed in our patients, probably owing to high levels of neutralizing antibodies, timely transfusion, and appropriate plasma volume.

There were some limitations to the present study. First, except for CP transfusion, the patients received other standard care. All patients received antiviral treatment despite the uncertainty of the efficacy of drugs used. As a result, the possibility that these antiviral agents could contribute to the recovery of patients, or synergize with the therapeutic effect of CP, could not be ruled out.

Furthermore, some patients received glucocorticoid therapy, which might interfere with immune response and delay virus clearance. Second, the median time from onset of symptoms to CP transfusion was 16.5 d (IQR, 11.0 d to 19.3 d).

Although the kinetics of viremia during natural history remains unclear, the relationship between SARS-CoV-2 RNA reduction and CP therapy, as well as the optimal concentration of neutralizing antibodies and treatment schedule, should be further clarified. Third, the dynamic changes of cytokines during treatment were not investigated. Nevertheless, the preliminary results of this trial seem promising, justifying a randomized controlled clinical trial in a larger patient cohort.

In conclusion, this pilot study on CP therapy shows a potential therapeutic effect and low risk in the treatment of severe COVID-19 patients. One dose of CP with a high concentration of neutralizing antibodies can rapidly reduce the viral load and tends to improve clinical outcomes. The optimal dose and treatment time point, as well as the definite clinical benefits of CP therapy, need to be further investigated in randomized clinical studies.

Donors for Convalescent Plasma Transfusion.

Ten donor patients who recovered from COVID-19 were recruited from three participating hospitals. The recovery criteria were as follows: 1) normality of body temperature for more than 3 d, 2) resolution of respiratory tract symptoms, and 3) two consecutively negative results of sputum SARS-CoV-2 by RT-PCR assay (1-d sampling interval). The donor’s blood was collected after 3 wk postonset of illness and 4 d postdischarge. Written informed consent was obtained from each patient.

Plasma Preparation Procedure and Quality Control.

Apheresis was performed using a Baxter CS 300 cell separator (Baxter). Convalescence plasma for treatment was collected from 40 donors. The median age was 42.0 y (IQR, 32.5 y to 49 y). A 200- to 400-mL ABO-compatible plasma sample was harvested from each donor depending on age and body weight, and each sample was divided and stored as 200-mL aliquots at 4 °C without any detergent or heat treatment. The CP was then treated with methylene blue and light treatment for 30 min in the medical plasma virus inactivation cabinet (Shandong Zhongbaokang Medical Appliance Co., Ltd).

Serology Test and Real-Time RT-PCR Detection of SARS-CoV-2 and Other Pathogens.

The neutralizing activity of plasma was determined by plaque reduction neutralization test using SARS-CoV-2 virus in the high biosafe level (BSL-4) laboratory of Wuhan Institute of Virology, Chinese Academy of Sciences. Neutralization titer was defined as the highest serum dilution with 50% reduction in the number of plaques, as compared with the number of plaques in wells in the absence of novel coronavirus antibody as blank control. The neutralizing activity of the receptor-binding domain of antibody in the CP was detected by a sandwich enzyme-linked immunosorbent assay (ELISA).

SARS-CoV-2 IgG antibody titer was tested by ELISA. SARS-CoV-2 RNA was detected by RT-PCR assay, and the result was presented as cycle threshold (Ct) value (Shanghai BioGerm Medical Biotechnology Co., Ltd). Methylene blue residue was detected by the verified UV method. The serology screening for hepatitis B and C virus, HIV, and syphilis spirochete was negative. The protocols for neutralization assay, serological test, and real-time RT-PCR detection of SARS-CoV RNA are presented in SI Appendix.

Treatment.

All patients were admitted to the ICU and received antiviral therapy and other supportive care, while some patients received antibiotic treatment, antifungal treatment, glucocorticoid, and oxygen support at the appropriate situation. One dose of 200 mL of inactivated CP with neutralization activity of >1:640 was transfused into the patients within 4 h following the WHO blood transfusion protocol.

Data Collection.

Clinical information of all enrolled patients was retrieved from the hospital electronic history system, including the baseline demographic data, days of illness duration, presenting symptoms, different kinds of examination, and methods of treatment. Bacterial coinfection was identified by a positive culture from respiratory, urinary, or blood culture within 48 h of hospital admission.

Complications, including acute renal failure, acute coronary syndrome, myocarditis, ARDS, and nosocomial infection, were recorded. The applications of assisted mechanical ventilation, intranasal oxygen inhalation, and medication regimen were recorded. The SARS-CoV-2 RNA from the serum sample was monitored during treatment.

Outcome Measures and Definitions.

The clinical symptoms were recorded by attending physicians daily. The blood test and biochemical tests were carried out every 1 d to 2 d. SARS-CoV-2 RNA was detected every 2 d to 3 d. CT scan was repeated every 3 d to 5 d. The primary endpoint was the safety of CP transfusion.

The second endpoints were the improvement of clinical symptoms and laboratory and radiological parameters within 3 d after CP transfusion.

Clinical symptoms improvement was defined as temperature normalization, relief of dyspnea, and oxygen saturation normalization, and radiological improvement was defined as different degrees of absorption of lung lesions.

---

MEDLINE and EMBASE (initially CINAHL) electronic databases were searched from 1996 to present (April 15th 2020) using a mix of keywords such as COVID-19, COVID*, SARS and respective drug names, along with any relevant variants.

PubMed was searched daily during this period as a means to gain a rapid assessment of any emergent publications. Searches were conducted daily from March 15th to present to uncover any new evidence and evidence as considered from additional sources such as manuscript reference lists, clinical trials registers (such as the International Clinical Trial Registry Platform) and online trial portals that pre-publish studies not yet having completed the peer-review process.

For example, we have searched and will continue to search the largest clinical medicine preprint repository, medRxiv.org, on a daily basis.

The focus was any types of comparative effectiveness research (ideally RCTs studies) for all of the included therapeutic pharmacological interventions (adults and children) and this review was open to any study that could be informative, including case-series and observational designs.

Adults and children exposed to or with confirmed or suspected COVID-19 were and will be included. Trials that compare interventions head-to-head or against no intervention or placebo is the focus.

We have focused on comparative effectiveness studies that provide evidence on patient-important outcomes, but were open to all reported outcomes at this time6. No electronic database search restrictions were imposed.

If meta-analytical pooling was and is possible from retrieved evidence, this review would seek to do this to derive more precise estimates of effect and derive additional statistical power.

A risk of bias assessment was applied to RCTs as well as observational studies focusing on randomization, allocation concealment, blinding, attrition, or other relevant biases to the estimates of effect, as well as selection bias, residual confounding bias, statistical adjustment, matching (propensity score), stratification, or restriction, respectively [Chang Chen et al. Pre-publication. Favipiravir versus Arbidol for COVID-19: A Randomized Clinical Trial. medRxiv. url: https://www.medrxiv.org/content/10.1101/2020.03.17.20037432v1 (Accessed on March 22nd, 2020).]

The GRADE ‘outcome-centric’ method was applied to individual outcomes per study to derive a certainty/quality of evidence rating to establish how much confidence one could have in the estimates of effect. These are principally single studies and the approach was to consider the outcomes per study in a rapid manner to establish some sense of GRADE ‘lite’ rating per outcome and then to derive an overall rating.

The overall rating is based on the lowest rating from among the critical/important patient outcomes. The reporting in these studies was very poor, scarce, and the general methodologies were very weak. This has been a rapid, albeit sub-optimal application of GRADE methods, while seeking to apply as much rigor to a flawed body of evidence emerging from the current reporting across COVID-19 research in general.[De Meyer et al. Lack of Antiviral Activity of Darunavir against SARS-CoV. url: https://doi.org/10.1101/2020.04.03.20052548doi (Accessed on April 8th 2020).].

Table 1: All COVID-19 in vitro lab and in vivo (clinical) human studies published from January 2020

Author; study design; year	Treatment arm vs comparator; sample size; age (mean/median); male %	Patient co- morbidities; additional medications reported besides the intervention/ control	Reported findings and author’s stated conclusion	Risk of bias (RoB)*; GRADE certainty of evidence rating**
Meplazumab (monoclonal antibody) There is no quality evidence to support a recommendation on its therapeutic use The effectiveness is being evaluated in various randomized clinical trials.
OBSERVATIONAL (clinical)
Bian1; observational; 2020	Add-on 10 mg meplazumab (n=17 patients) vs hospitalized patients in the same period as controls (n=11); 28; mean 56.1; 53.5%	32% hypertension, 10.7% cardiovascular disease, 10.7% diabetes: lopinavir/ritonavir, recombinant human interferon α-2b, glucocorticoid, and antibiotics.	Meplazumab treatment significantly improved the discharge (p=0.006) and case severity (p=0.021) in the critical and severe patients vs control; the time to being virus negative in treatment was reduced relative to the control group (median 3, 95% CI (1.5–4.5) vs. 13, (6.5– 19.5); p=0.014, HR=0.37, 95% CI (0.155–0.833)); suggested the need for further study in clinical trials as a potential therapeutic option in COVID-19.	High; Very low certainty1
Ivermectin There is no quality evidence to support a recommendation on its therapeutic use The effectiveness is being evaluated in various randomized clinical trials.
OBSERVATIONAL (in vitro)
Caly2; observational; 2020	One group: a single addition to Vero-hSLAM cells 2 hours post infection with SARS-CoV-2 isolate Australia/VIC01/2020 at a MOI of 0.1, followed by the addition of 5 µM ivermectin; NA	NA	Following a single addition to Vero-hSLAM cells 2 hours post infection, ivermectin at 24 hours contributed to a 93% reduction in viral RNA present in the supernatant of the samples treated with ivermectin compared to the vehicle DMSO. By 48 hours, there was an ~5000-fold reduction in viral RNA at 48 hours. Researchers concluded that ivermectin administration in vitro resulted in the effective loss of essentially all viral material by 48 hours, supporting further clinical study in COVID-19 patients.	High; Did not apply GRADE
OBSERVATIONAL (clinical)
Patel24; observational; 2020	Ivermectin (150 mcg/Kg once following initiation of mechanical ventilation) vs SoC (no ivermectin); 1,970; not reported; not reported	Not reported	A survival benefit was reported for ivermectin (mortality rate 18.6% vs 7.7%; HR 0.18, 95% CI (0.07-0.48), log rank (Mantel-Cox) p<0.001; length of hospital stay 10.9 +/- 6.1 days vs 15.7 +/- 8.1 days and ICU stay was 6.0 +/- 3.9 days vs 8.2 +/- 6.2 days, both p<0.001.	High; Very low certainty1
Siltuximab (monoclonal antibody) There is no quality evidence to support a recommendation on its therapeutic use The effectiveness is being evaluated in various randomized clinical trials.
OBSERVATIONAL (clinical)
Gritti3; observational; 2020	One group: patients received siltuximab at a median dose of 900 mg, ranging from 700	43% had hypertension, 23.8% diabetes, 19%	The results suggest a potential role of siltuximab in treating patients with ARDS secondary to SARS-CoV-2 infection.	High; Very low certainty1

	to 1,200 mg; received a second dose of siltuximab; 21; median 64.0 (IQR 48-75); 85.7%	cardiovascular disease, 4.7% malignancies, 4.7% chronic kidney disease, and 4.7% cerebrovascular disease; no other medication reported but siltuximab
Danoprevir (antiviral) There is no quality evidence to support a recommendation on its therapeutic use The effectiveness is being evaluated in various randomized clinical trials.
OBSERVATIONAL (clinical)
Chen4; observational; 2020	Treatment experienced (n=9) vs naïve patients (n=2), treatment naïve patients never received any antiviral therapies such as lopinavir/ritonavir and interferon nebulization before switching to danoprevir (all treated with danoprevir boosted by ritonavir in the presence or absence of interferon nebulization (the background therapy)); 11; median 44 (range 18-66); 36%	18% hypertension	After 4 to 12-day treatment with danoprevir boosted by ritonavir, all patients (n=11) discharged from the hospital based on normal body temperature for at least 3 days; there was substantial improvements in respiratory symptoms; the CT lung imaging revealed absorption and recovery of acute exudative lesions; there were 2 consecutive RT-PCR negative tests of SARS-CoV-2 nucleotide acid; researchers concluded that repurposing of danoprevir for COVID-19 should be considered within clinical trials.	High; Very low certainty1
Tocilizumab (monoclonal antibody) There is no quality evidence to support a recommendation on its therapeutic use The effectiveness is being evaluated in various randomized clinical trials.
OBSERVATIONAL (clinical)
Xu5; observational; 2020	All patients treated with tocilizumab; 21; mean 56.8 ± SD 16.5, ranged from 25 to 88 years; 85.7%	43% hypertension, 23.8% diabetes, 9.5% CHD, 4.8% COPD, 4.8% CKD, 4.8% bronchiectasis, 4.8% brain infarct, 4.8% auricular fibrillation; none reported	75.0% lowered oxygen intake and one patient required no oxygen therapy. CT scans showed lung lesion opacity was absorbed in 90.5%. The percentage of lymphocytes in peripheral blood returned to normal in 52.6% patients on the fifth day following treatment. Abnormally elevated C- reactive protein declined significantly in 84.2% of patients. No adverse reactions reported and 90.5% (n=19) discharged from hospital mean 13.5 days following the treatment with tocilizumab and the rest; 2 are undergoing good recovery; researchers concluded that tocilizumab should be considered within clinical trials for COVID-19.	High; Very low certainty1
Cellina34; case- series (1 patient); 2020	2 doses of tocilizumab (8 mg/kg), 12 hours apart, on day 7 and 8; 1 patient; 64; male	None reported; none reported	Patient without significant clinical history presented with syncope with normal vitals; ear temperature was 38 °C, oxygen saturation 99% on room air, chest X-Rays showed mild linear densities in the lower and middle left lung fields, laboratory investigations showed increased white blood cell count (10.900 per μL), elevated serum lactate level (250 U/L) and elevated reactive C protein (RCP) (89 mg/dL), other blood tests normal; COVID-19 detected in a throat swab sample by RT-PCR. Due to the worsening of the blood tests on the day 2, patient admitted; day 6, the patients developed dyspnea; decreased of oxygen saturation (90%) and further increase of CRP 336 mg/dL; white blood cell count was	Not applied; Not applied

			10.800 per μL; interleukin-6 was 80 ng/L; day 7, unenhanced chest CT showed the presence of diffused bilateral air space opacities, including ground glass opacities and consolidation; assisted ventilation started; patient administered 2 doses of tocilizumab (8 mg/kg), 12 hours apart, on day 7 and 8; day 9, CRP declined to 96 mg/dL and white blood cell count to 2.360 per μL; patient clinical condition gradually improved and ventilatory support was gradually stopped; day 14, repeat chest CT revealed mark improvement (size reduction of air cells opacities, density reduction of consolidations, some ground glass opacities, peripheral reticular opacities, reduction of pleural effusion and mediastinal lymphadenopathy).
Favipiravir (antiviral) There is no quality evidence to support a recommendation on its therapeutic use The effectiveness is being evaluated in various randomized clinical trials.
RCT (clinical)
Chang7; RCT (open-label); 2020	120 assigned to favipiravir group (116 assessed, routine treatment + 1600 mg on the first day twice a day, 600 mg from the second day to the end, twice a day) and 120 to arbidol group (120 assessed, 200 mg, 3 times a day to the end of the trial); 236; not reported clearly; 46.6%	27.9% hypertension, diabetes 11.4%, 95% COVID-19 pneumonia; none reported	Clinical recovery rate of day 7 between two groups, 61.2% favipiravir vs 5.7% arbidol (total patients), 71.4% vs 55.6% (moderate cases) respectively, 5.5% vs 0.0% (serious cases) respectively; patients with hypertension and/or diabetes 54.7% favipiravir vs 51.4% arbidol; adverse events 37/116 favipiravir vs 28/120 arbidol, note, 18 severe patients in the favipiravir group vs 9 severe patients in the arbidol group (imbalanced).	High; Very low certainty1
OBSERVATIONAL (clinical)
Cai6; observational; 2020	Oral FPV (Day 1: 1600 mg twice daily; days 2–14: 600 mg twice daily) plus interferon (IFN) α by aerosol inhalation in the FPV arm vs LPV/RTV (days 1–14: 400 mg/100 mg twice daily) plus IFN-α; 80 (n=35 FPV and n 45=in LPV/RTV); median 47 (35.75–61); 43.8%	None reported; no additional medications reported, standard care included oxygen inhalation, oral or intravenous rehydration, electrolyte correction, antipyretics, analgesics, and antiemetic drugs.	Viral clearance median time for FPV (Group A), was estimated to be 4 days (IQR: 2.5–9) and significantly shorter than the time for patients in control group (Group B), which was 11 d (IQR: 8–13) (P < 0.001). For chest CT changes, on Day 14 after treatment, the improvement rates of the chest CT in FPV significantly higher than those in the control arm (91.4% versus 62.2 %, 32/35 versus 28/45, P = 0.004). Adverse reactions in the FPV n=4 was four, significantly fewer than the 25 adverse reactions in the control arm (P < 0.001).	High; Very low certainty1
Darunavir (antiviral) There is no quality evidence to support a recommendation on its therapeutic use The effectiveness is being evaluated in various randomized clinical trials.
OBSERVATIONAL (in vitro)
De Meyer8; observational; 2020	Examined the in vitro antiviral activity of darunavir against a clinical isolate from a patient infected with SARS-CoV-2.	NA	Darunavir showed no activity against SARS-CoV-2 at clinically relevant concentrations (EC50 >100 μM). Remdesivir, used as a positive control, showed potent antiviral activity (EC50 = 0.38 μM).   Present findings do not support the use of darunavir for treatment of COVID-19.	Definitely high2 (risk of bias assessed for in vitro studies using OHAT tool); Very low certainty1
Nelfinavir (antiviral) There is no quality evidence to support a recommendation on its therapeutic use The effectiveness is being evaluated in various randomized clinical trials.

OBSERVATIONAL (in vitro)
Yamamoto 9; observational; 2020	Assessed the 50% effective concentration (EC50), the 50% cytotoxic concentration (CC50), and the selectivity index (SI, CC50/EC50); C max-EC50 ratio (C max/EC50) and C trough- EC50 ratio (C trough/EC50) were also calculated to evaluate the safety and efficacy of the 9 antivirals (plus lopinavir, ritonavir, saquinavir, atazanavir, tipranavir, amprenavir, darunavir, and indinavir).	NA	Nelfinavir effectively obstructs replication of SARS-CoV- 2; the effective concentrations for 50% and 90% inhibition (EC50 and EC90) of nelfinavir was the lowest from among the 9 HIV-1 protease inhibitors.   Present in vitro findings are positive and support further clinical study of nelfinavir in COVID-19 patients. The methodology indicates a high risk of bias.	Definitely high2 (risk of bias assessed for in vitro studies using OHAT tool); Very low certainty1
Remdesivir (antiviral) There is no quality evidence to support a recommendation on its therapeutic use The effectiveness is being evaluated in various randomized clinical trials.
OBSERVATIONAL (clinical)
Holshue 10; case- series; 2020	1 COVID-19 patient (first in USA), aged 35 years, male, treated with remdesivir on compassionate use authorization	NA	Treatment with IV remdesivir began on the evening of day 7, and no adverse events were observed in association with the infusion. Vancomycin was discontinued on the evening of day 7, and cefepime was discontinued on the following day, after serial negative procalcitonin levels and negative nasal PCR testing for methicillin- resistant Staphylococcus aureus. On hospital day 8 (which was illness day 12), it was found that the patient’s clinical condition improved significantly, whereby the supplemental oxygen was discontinued, and his oxygen saturation values improved to 94 to 96% while he was breathing ambient air. Bilateral lower-lobe rales were no longer present. Appetite improved, and he was asymptomatic aside from intermittent dry cough and rhinorrhea. All symptoms resolved.	Not applied; Not applied
Grein, 11; case- series; 2020	Remdesivir, 53; median IQR 64 (48–71); 75	Hypertension 25%, diabetes 17%, hyperlipidemia 11%, asthma 11%; none reported	Researchers reported that at baseline, 30 patients (57%) were receiving mechanical ventilation and 4 (8%) were receiving ECMO. Based on a median follow-up of 18 days, 36 patients (68%) had an improvement in oxygen- support class, including 17 of 30 patients (57%) receiving mechanical ventilation who were extubated. A total of 25 patients (47%) were discharged, and 7 patients (13%) has died; mortality was 18% (6 of 34) among patients receiving invasive ventilation and 5% (1 of 19) among those not receiving invasive ventilation. Thirty-two patients incurred adverse events in follow-up.	High; Very low certainty1
Chloroquine/hydroxychloroquine There is no quality evidence to support a recommendation on its therapeutic use The effectiveness is being evaluated in various randomized clinical trials. Cardiovascular adverse events should be closely monitored
RCT (clinical)
Chen 12; RCT; 2020	Hydroxychloroquine (HCQ) 400 mg per day for 5 days vs control (conventional	None reported; nebulization with interferon alpha, and	Nucleic acid of throat swabs was negative in 13 (86.7%) HCQ cases and 14 (93.3%) cases in the control group (P>0.05), median duration from hospitalization to virus	High; Very low certainty1

	treatment); 30 (15:15); 48.5 mean; 70%	80.0% patients in the experimental group received Abidol vs 66.7% in control, 2 received lopinavir / ritonavir.	nucleic acid negative conservation was 4 (1-9) days in HCQ group, which is comparable to that in the control group [2 (1-4) days, median time for body temperature normalization in HCQ group was 1 (0-2) after hospitalization, which was also comparable to that in the control group 1(0-3), radiological progression was shown on CT images in 5 cases (33.3%) in the HCQ group and 7 cases (46.7%) in the control group. Researchers concluded that the standard dose of hydroxychloroquine sulfate does not show clinical effects in improving patient symptoms and accelerating virological suppression.
Chen13; RCT; 2020	5-day HCQ (n=31) (400 mg/d), control (n=31) received SoC; 62; 44.7 mean (SD 15.3); 46.8%	None reported; none reported	Body temperature recovery time and the cough remission time were significantly shortened in the HCQ treatment group (mean days and SD was 2.2 (0.4) in the HCQ groups vs 3.2 (1.3) in the control, p=0.0008. They also reported a greater proportion of patients with improved pneumonia (on chest CT) in the HCQ treatment group (80.6%, 25 of 31) relative to the control group (54.8%, 17 of 31). Four patients in the control group developed severe illness (none in the treatment group) and there were 2 mild adverse events in the HCQ group. The study group was generally younger, and the illness was mild on entry, suggestive that this was not an overly ill group to begin with and patients may have recovered on their own. No accounting of whether patients were taking any other medications prior to study entry or during the study.	High; Very low certainty1
Huang 14; RCT; 2020	Twice-daily oral of 500 mg Chloroquine (n=10) versus 400/100mg Lopinavir/Ritonavir (n=12) for 10 days; 22; 44.0 mean (36.5 to 57.5); 59.1%	None reported; none reported	Using RT-PCR, on day 13, all patients in the chloroquine group were negative, and 11 of 12 in the control group (lopinavir/ritonavir) were negative on day 14. Via lung CT on day 9, 6 patients in chloroquine group achieved lung clearance versus 3 in the comparison group. At day 14, the rate ratio based on CT imaging from the Chloroquine group was 2.21, 95% CI 0.81-6.62) relative to the control group. Five patients in the chloroquine group had adverse events versus no patients in the control group. This small RCT appeared to show better effectiveness of chloroquine over lopinavir/ritonavir in moderate to severely ill COVID-19 patients.	High; Very low certainty1
Silva Borba15; RCT; 2020	CQ (600mg CQ twice daily for 10 days or total dose 12g); or low dose CQ (450mg for 5 days, twice daily only on the first day, or total dose 2.7g); 81 (41 high doses vs 40 low dose); 51; 75	Hypertension 46.2%, diabetes 25.9%, alcoholism 26%, heart disease 9.2%, asthma 6.2%, CKD 7.5%, rheumatic disease 5.6%, liver disease 3.7%, TB 3.7%, HIV/AIDS 1.9%; corticosteroids 5.4%, ACE inhibitors 10.3%, oseltamivir 89.6%	There were 11 deaths (13.5%) in high dose and low dose users; the high dose CQ arm presented more QTc>500ms (25%), and a trend toward higher lethality (17%) than the lower dosage. Fatality rate was 13.5% (95%CI=6.9–23.0%), overlapping with the CI of historical data from similar patients not using CQ (95%CI=14.5-19.2%). In 14 patients with paired samples, respiratory secretion at day 4 was negative in only one patient; preliminary findings suggest that the higher CQ dosage (10-day regimen) should not be recommended for COVID-19 treatment because of its potential safety hazards	Low-moderate; Moderate certainty3
Tang16; RCT; 2020	HCQ (a loading dose of 1, 200 mg daily for three days followed by a maintained dose of 800 mg daily for the remaining days) vs SoC; 150; mean 46.1±14.7; 54.7%	Diabetes 14.0%, hypertension 6%, others 31%; 80 patients used other drugs after randomization (not clearly reported)	The overall 28-day negative conversion rate was not different between SOC plus HCQ and SOC group (85.4% versus 81.3%, p=0.34). Negative conversion rate at day 4, 7, 10, 14 or 21. A significant efficacy of HCQ on alleviating symptoms was observed (HR, 8.83, 95%CI, 1.09 to 71.3). There was a significantly greater reduction of CRP (6.98 in SOC plus HCQ versus 2.72 in SoC, milligram/liter, p=0.045) conferred by the addition of HCQ, which also led to more rapid recovery of lymphopenia, albeit no statistical significance. Adverse events found in 8.8% of SoC and 30% of HCQ recipients	High; Low certainty1

			with two serious adverse events in the HCQ group.
Barbosa28; quasi-	HCQ + supportive care vs	Not reported; not	HCQ administration was associated with increased need	High;
RCT; 2020	supportive care alone; 63 (32	reported	of escalation of respiratory support relative to those that	Low certainty1
(submitted to	HCQ vs 31 control); mean		did not receive HCQ at 5 days (p=0.013). The absolute
NEJM for peer	62.7 ± 15; 58.7%		lymphocyte change in the HCQ group was no different
review, abstract			than supportive care alone (p=0.41). HCQ use trended to
form)			ward worsening neutrophil-to-lymphocyte ratio relative to supportive care alone as well as a greater risk for
			intubation. There was no benefit of HCQ on mortality
			(4/31 HCQ and 1/32 in supportive care arm),
			lymphopenia, or neutrophil-to-lymphocyte ratio
			improvement.
OBSERVATIONAL (clinical)
Gautret 17; observational; 2020	HCQ 600 mg daily 6 d n=26 (AZ added depending on clinical presentation); 42; 26 HCQ, 16 control; 45.1 ± 22.0 (mean/SD); 41.7%	None reported; none reported	Researchers reported that 6 patients were asymptomatic, 22 had upper respiratory tract infection symptoms and eight had lower respiratory tract infection symptoms. Twenty cases were treated in this study and showed a significant reduction of the viral carriage at D 6-post inclusion compared to controls, and much lower average carrying duration than reported of untreated patients in the literature. Azithromycin (Z-Pak) added to hydroxychloroquine was significantly more efficient for virus elimination.	High; Very low certainty1
			Clinical follow-up and occurrence of side-effects were not discussed in the paper.
Gautret 18;	200 mg of HCQ three times	Cancer 6.3%,	Nasopharyngeal viral load tested by qPCR and negative	High;
observational; 2020	per day for ten days combined with AZ (500 mg on D1 followed	diabetes 11.2%, CAD 7.5%, hypertension 16.3%,	on day 8 was found in 93.7% of patients, not contagious (with a PCR Ct value<34) at day 10 was found in 98.7%, negative virus cultures on day 5 was found in 98.7%, and	Very low certainty1
	by 250 mg per day for the	chronic respiratory	length of stay in ICU (days) was a mean 4.6 days ± 2.1
	next four days); 80; 52.5	disease 10%, obesity	SD (n=65). Researchers reported that patients were
	median, 52.5%	5%; immune-	rapidly discharged from highly contagious wards with a
		suppressive	mean length of stay of five days. This study was judged to
		treatment 5%, non-	be at high risk of biased estimates due to it being a case-
		steroid anti-	series observational study with no control group. Based
		inflammatory	on reporting, the cohort appears to be younger and the
		treatment 2.5%	NEWS risk scoring system placed them all at very low
			risk of deteriorating, leaving one to speculate on if they
			would have recovered on their own. This group appears
			to be COVID-19 patients with mild illness. Patients may
			have recovered on their own.
Molina 19; observational’ 2020	HCQ 600 mg/d for 10 days and AZ 500 mg Day 1 and 250 mg days 2 to 5; 11; 58.7 mean, 64%	None reported; none reported	One patient, hydroxychloroquine and azithromycin were discontinued after 4 days because of a prolongation of the QT interval from 405 ms before treatment to 460 and 470 ms under the combination; They report that in the 10 living patients, repeated nasopharyngeal swabs were positive for COVID-19 RNA in 8 of the 10 patients (80%) at days 5 to 6 following treatment initiation. Researchers also questioned the one death and 3 ICU transfers14 that suggest a worsening clinical outcome. They conclude that there is “no evidence of a strong antiviral activity or clinical benefit of the combination of hydroxychloroquine and azithromycin for the treatment of our hospitalized patients with severe COVID-19”. This was a small consecutive series of patients followed to describe the response to the treatment, high risk of biased estimates.	High; Very low certainty1
Lane20;	Network cohort and self-	ARDS 58%, COPD	Data comprised 14 sources of claims data or electronic	High;
network cohort and case-series;	controlled case series study that involved 956,374 and	5%, depression 14.5%, diabetes	medical records from Germany, Japan, Netherlands, Spain, UK, and USA. Researchers found no excess risk of	Very low certainty1
2020	310,350 users of HCQ and	13.2%,	SAEs was when 30-day hydroxychloroquine and

	sulfasalazine, and 323,122	hyperlipidemia 30%,	sulfasalazine use were compared. However, when
and 351,956 users of HCQ-	pneumonia 5.7%,	azithromycin was added to hydroxychloroquine,
azithromycin and HCQ-	renal impairment	researchers reported an increased risk of 30-day
amoxicillin.	4.2%, UTI 14.2%	cardiovascular mortality HR 2.19 (95% CI 1.22-3.94), chest pain/angina HR 1.15 (95% CI 1.05-1.26), and heart
		failure HR 1.22 (95% CI 1.02-1.45)). The conclusion was
		that short-term hydroxychloroquine treatment was safe,
		but when azithromycin is added, it can induce heart
		failure and cardiovascular mortality, likely due to
		synergistic effects on QT length. Researchers urged
		caution in the use of this combination in COVID-19.
Chorin21;	HQC plus azithromycin; 84;	CAD 11%,	The QTc was prolonged maximally from baseline (days	High;
observational;	mean 63 +15; 74%	hypertension 65%,	3-4) and in 25 patients, the QTc increased more than	Very low certainty1
2020		CKD 7%, diabetes	40ms. They also found that in 9 patients (11%), the QTc
		20%, COPD 8%,	increased to >500 ms, indicative of a high-risk group for
		congestive heart	malignant arrhythmia and sudden cardiac death.
		failure 2%;
		Levofloxacin,
		Lopinavir/Ritonavir,
		or Tacrolimus 8%,
		Norepinephrine,
		Phenylephrine, or
		Vasopressin 13%,
		Amiodarone 7%
Mahévas22; observational; 2020	HCQ at a daily dose of 600 mg in the first 48 hours after hospitalisation vs no HCQ; 181; median 60 years (IQR 52 to 68 years); 71.1%   Note: in the HCQ group, 20% received concomitant azithromycin	Respiratory disease 11%, heart failure 3.3%, hypertension (cardiovascular illnesses) 51.9%, diabetes 8.3%, CKD 5%, immuno- depression 11.6%;	In terms of deaths or transfer to the ICU, 19% vs 21.6% occurred in the HCQ vs no HCQ groups respectively (RR 0.93 (0.48 to 1.81)), for day 7 mortality, 3.6% died in HCQ group vs 4.1% in the no-HCQ group (RR 0.61 (0.13 to 2.90)), occurrence of acute respiratory distress syndrome, 28.6% occurred in HCQ group vs 24.1% in no HCQ group (RR 1.15 (0.66 to 2.01)); in the 84 patients receiving HCQ within the first 48 hours, 8 (9.5%) experienced ECG modifications requiring HCQ discontinuation at a median of 4 days (3-9) after it began.	High; Very low certainty1
			Note: inverse probability of treatment weighting (IPTW) approach was used to closely approximate randomisation and try to balance the differences in baseline prognostic variables between treatment groups
Corticosteroids There is no quality evidence to support a recommendation on its therapeutic use The effectiveness is being evaluated in various randomized clinical trials.
OBSERVATIONAL (clinical)
Lu23;	Corticosteroid	Hypertension 45%,	28-day mortality rate was 39% (12 out of 31) in case	High;
observational; 2020	(methylprednisolone, dexamethasone, and hydrocortisone) vs no drug;	diabetes 17.7%, CVD 6.5%, COPD 1.5%; oseltamivir,	subjects and 16% (5 out of 31) in control subjects (P=0.09). Increased corticosteroids dosage was significantly associated with elevated mortality risk	Very low certainty1
	61 (31:31); 57.5 mean; 52%	arbidol,	(P=0.003) in matched cases after adjustment for
	lopinavir/ritonavir,	administration duration; every ten-milligram increase in
		ganciclovir,	hydrocortisone dosage was associated with additional 4%
		interferon-α	mortality risk (adjusted HR: 1.04, 95% CI: 1.01-1.07).
CONVALESCENT PLASMA (CP) There is no quality evidence to support a recommendation on its therapeutic use The effectiveness is being evaluated in various randomized clinical trials
OBSERVATIONAL (clinical)
Shen25; case-series; 2020	Convalescent plasma (CP) to all; 5; age range 36-73 years;	1 has hypertension and mitral	Following plasma transfusion, body temperature normalized within 3 days in 4 of 5 patients, the SOFA	High; Did not apply

	60%   Note: CP administered to all between 10 and 22 days after admission	insufficiency; antivirals (lopinavir/ ritonavir; interferon alfa-1b; favipiravir; arbidol; darunavir) and corticosteroid methylprednisolone	score decreased, and PAO2/FIO2 increased within 12 days (range, 172-276 before and 284-366 after). Viral loads also decreased and became negative within 12 days after the transfusion, and SARS-CoV-2–specific ELISA and neutralizing antibody titers increased following the transfusion (range, 40-60 before and 80-320 on day 7). ARDS resolved in 4 patients at 12 days after transfusion, and 3 patients were weaned from mechanical ventilation within 2 weeks of treatment. Of the 5 patients, 3 have been discharged from the hospital (length of stay: 53, 51, and 55 days), and 2 are in stable condition at 37 days after transfusion.	GRADE
Duan26; case- series; 2020	CP to all; 10; median age was 52.5 years (IQR, 45.0–59.5); 60%	Hypertension 30%, cardiovascular and cerebrovascular disease 10%; arbidol, ribavirin, remdesivir, Interferon-ɑ, oseltamivir, peramivir and corticosteroid methylprednisolone	Following transfusion, the level of neutralizing antibody quickly increased to 1:640 in five cases, and maintained at a high level (1:640) in remaining of cases. Researchers reported that the clinical symptoms were substantially improved. They also found an increase in oxyhemoglobin saturation within 3 days. Several parameters tended to improve as compared to pre-transfusion. Improved parameters included “increased lymphocyte counts and decreased C-reactive protein. Radiological examinations showed varying degrees of absorption of lung lesions within 7 days. The viral load was undetectable after transfusion in seven patients who had previous viremia”. No severe adverse effects.	High; Did not apply GRADE
Zhang27; case- series; 2020	CP to all; 4; 31, 55, 69, 73 years old and F, M, M, and pregnant F respectively	None reported; arbidol, lopinavir- ritonavir, ribavirin, interferon alpha inhalation, oseltamivir, albumin, zadaxin and immunoglobulin, antibacterial and antifungal drugs	Researchers reported no serious adverse reactions and all 4 patients recovered from COVID-19.	High; Did not apply GRADE
Pei29; case-series; 2020	CP to all three; 3; not reported; not reported	Not reported; not reported	There were 2 patients with negative conversions and 1 failure due to anaphylaxis shock (discontinued); 1st patient treated on 12th day admission, turned severe, 2nd treatment, then significantly improved (nucleic acid test became negative and symptoms improved) and met discharge criteria on 26th day, 2nd patient, treatment on 27th day, the nucleic acid test became negative 4 days later, 3rd patient was a 51-year old pregnant woman who suffered anaphylaxis shock and CP was discontinued).	High; Did not apply GRADE
Umifenovir (antiviral) There is no quality evidence to support a recommendation on its therapeutic use The effectiveness is being evaluated in various randomized clinical trials.
RCT (clinical)
Li30; RCT; 2020	Lopinavir/ritonavir (LPV/r) vs arbidol vs control; 44 (21, 16, 7 respectively); mean 49.4 years; 50%	Some type of underlying illnesses 34%; gamma globulin 11.3%, glucocorticoids 22.7%	The median time of positive-to-negative conversion of SARS-CoV-2 nucleic acid was 8.5 (IQR 3, 13) days in the LPV/r group, 7 (IQR 3, 10.5) days in the arbidol group and 4 (IQR 3, 10.5) days in the control group (p=0.751). Researchers reported that there were no statistical differences between the three groups in the rates of antipyresis, cough alleviation, improvement of chest CT or the deterioration rate of clinical status (all p > 0.05). Five (23.8%) patients in the LPV/r group experienced adverse events during the follow-up period versus none in the other groups.	High; Low certainty1

OBSERVATIONAL (clinical)
Chen31; observational; 2020	Favipiravir versus Arbidol open-label RCT; 236 (116 favipiravir, 120 arbidol); unclear; 46.6%	Hypertension 27.9%, 11.4% diabetes; moxifloxacin hydrochloride tablets, cephalosporins, antiviral drugs other than the experimental drugs, glucocorticoid and human serum albumin.	There was no significant difference in clinical recovery rate at day 7, whereby 71 (61%) recovered in the favipiravir arm and 62 (52%) in the arbidol group. In patients with hypertension and/or diabetes, 23 (54.76) recovered in the favipiravir arm and 18 (51.43) in the arbidol arm (no significant difference). There were no deaths in either arm and 1 respiratory failure in the favipiravir arm and 4 (3.33) in the arbidol arm. Researchers reported 37 adverse events in the favipiravir arm and 28 in the arbidol arm. The reporting in this study was very poor and the methodology was weak. This was described as a randomized study but it was not. No proper description of randomization, allocation concealment, or masking was provided.	High; Very low certainty1
Deng32; observational; 2020	Arbidol combined with LPV/r (n=16) vs LPV/r alone (n=17); 33; mean 44.5; 51.5%	Median number of comorbidities was 0 7 (range 0–2); corticosteroid therapy; a number of antibacterial therapy agents; vasopressors.	Researchers reported that COVID-19 was not detected for 12 of 16 patients’ nasopharyngeal specimens (75%) in the combination group after 7 days, relative to 6 of 17 (35%) in the monotherapy group (p < 0·05). “After 14 days, 15 (94%) of 16 and 9 (52·9%) of 17, respectively, SARS-CoV-2 could not be detected (p < 0·05)”. They reported that the chest CT scans were improving for 11 of 16 patients (69%) within the combination group following seven days relative to 5 of 17 (29%) in the monotherapy group (p < 0·05). The sample was very small (n=33) and this was a nonrandomized retrospective design which is a weak design.	High; Very low certainty1
Wang33; observational; 2020	Arbidol vs no arbidol; 67; median 42.0(35.0-62.0); 46%	Hypertension 13%, cardiovascular disease 12%, diabetes 10%, COPD 6%, malignancy 6%, asthma 3%, chronic hepatitis 1%; antivirals, antibiotics, antifungals, corticosteroids	Mortality rate was 7.5%. Patients were divided into the SpO2≥90% group (n=55) and the SpO2 < 90% n=14; all deaths occurred in SpO2 < 90%, median age of the  SpO2 <90% was 70.5, IQR 62-77, SpO2 <90% had more comorbidities (included the 5 that died) than SpO2≥90% group, 36% vs 7%, p=0.014, cardiovascular disease 36% vs 5%, p=0.07, diabetes 43% vs 2% p<0.001. SpO2 < 90% group had more fever and dyspnea; no persons died who were treated with arbidol (n=36 patients), and all 5 deaths occurred in the group that received no arbidol (n=31 patients). The study showed that elderly persons (older) with underlying medical conditions were at increased risk of death.	High; Very low certainty1
Lopinavir/ritonavir (LPV/r) protease inhibitor There is no quality evidence to support a recommendation on its therapeutic use The effectiveness is being evaluated in various randomized clinical trials.
RCT (clinical)
Li30; RCT; 2020	Lopinavir/ritonavir (LPV/r) vs arbidol vs control; 44 (21, 16, 7 respectively); mean 49.4 years; 50%	Some type of underlying illnesses 34%; gamma globulin 11.3%, glucocorticoids 22.7%	The median time of positive-to-negative conversion of SARS-CoV-2 nucleic acid was 8.5 (IQR 3, 13) days in the LPV/r group, 7 (IQR 3, 10.5) days in the arbidol group and 4 (IQR 3, 10.5) days in the control group (p=0.751). Researchers reported that there were no statistical differences between the three groups in the rates of antipyresis, cough alleviation, improvement of chest CT or the deterioration rate of clinical status (all p > 0.05). Five (23.8%) patients in the LPV/r group experienced adverse events during the follow-up period versus none in the other groups.	High; Low certainty1
Huang 14; RCT; 2020	Twice-daily oral of 500 mg Chloroquine (n=10) versus 400/100mg Lopinavir/Ritonavir (n=12)	None reported; none reported	Using RT-PCR, on day 13, all patients in the chloroquine group were negative, and 11 of 12 in the control group (lopinavir/ritonavir) were negative on day 14. Via lung CT on day 9, 6 patients in chloroquine group achieved	High; Very low certainty1

	for 10 days; 22; 44.0 mean (36.5 to 57.5); 59.1%		lung clearance versus 3 in the comparison group. At day 14, the rate ratio based on CT imaging from the Chloroquine group was 2.21, 95% CI 0.81-6.62) relative to the control group. Five patients in the chloroquine group had adverse events versus no patients in the control group. This small RCT appeared to show better effectiveness of chloroquine over lopinavir/ritonavir in moderate to severely ill COVID-19 patients.
Cao36; RCT; 2020	LPV/r (400 mg and 100 mg, respectively) twice a day for	Diabetes 11.6%, cerebrovascular	Time to clinical improvement — median no. of days (IQR) 16.0 (13.0 to 17.0) vs 16.0 (15.0 to 18.0);	High; Low certainty4
	14 days, in addition to	6.5%, cancer 3%;	Day 28 mortality — no. (%) n=19 (19.2) vs 25 (25.0)
	standard care vs standard	interferon on	intervention vs control respectively; Clinical
	care alone; 100 (99	enrollment 11.1%,	improvement — no. (%) day 28 n=78 (78.8) vs 70 (70.0);
	intervention 100 control);	vasopressors 22.1%,	ICU length of stay — median no. of days
	median 58 years IQR 49 to	glucocorticoid	(IQR) 6 (2 to 11) vs 11 (7 to 17); Hospital stay — median
	68 years; 60.3%	33.7%, antibiotic	no. of days (IQR) 14 (12 to 17) vs 16 (13 to 18)
		95%	The median interval time between symptom onset and randomization was 13 days (IQR, 11 to 16 days).
			Open-label
OBSERVATIONAL (clinical)
Ye35;	LPV/r vs plus adjuvant	Hypertension 17%,	Improvement in body temperature for both groups	High;
observational;	drugs only no LPV/r	diabetes 17%;	admission to the 10th day treatment; body temperature of	Very low certainty1
2020	(adjuvant drugs only); 47 (42	arbidol, moxifloxacin	intervention group declined faster than control, some
	treatment vs 5 control); aged		reductions in proportions of white blood cells,
	between 5 and 68, of which		lymphocytes and C-reactive protein in intervention vs
	9 were under 30 and 38 were		control, proportion with abnormal alanine
	over 30; 42%		aminotransferase and aspartate aminotransferase in intervention lower than control; reduced number of days
			testing negative in intervention group.
Deng32;	Arbidol combined with	Median number of	COVID-19 was not detected for 12 of 16 patients’	High;
observational; 2020	LPV/r (n=16) vs LPV/r alone (n=17); 33; mean 44.5; 51.5%	comorbidities was 0 7 (range 0–2); corticosteroid	nasopharyngeal specimens (75%) in the combination group arbidol plus LPV/r following 7 days, relative to 6 of 17 (35%) in the monotherapy group (p < 0·05). “After	Very low certainty1
		therapy; a number of	14 days, 15 (94%) of 16 and 9 (52·9%) of 17, respectively,
		antibacterial therapy	SARS-CoV-2 could not be detected (p < 0·05)”. They
		agents; vasopressors.	reported that the chest CT scans were improving for 11
			of 16 patients (69%) within the combination group
			following seven days relative to 5 of 17 (29%) in the
			monotherapy group (p < 0·05).
			The sample was very small (n=33) and this was a
			nonrandomized retrospective design which is a weak
			design.

Notes and considerations:

*ratings are high vs moderate-low vs low RoB; note, high risk for RCTs would be for serious flaws in randomization, allocation concealment, blinding, severe data loss, baseline imbalances etc. and for observational non-randomized studies (single or two-arm), there could be no adjustment for confounders, no masking, stratification etc.

**ratings are high, moderate, low, very low certainty (GRADE); note using GRADE, RCTs start as high certainty/quality evidence, observational studies start as low certainty/quality; for imprecision, the focus is on sample size, number of reported events, width of confidence intervals (if reported); note also that the use of GRADE in this application for RCTs and observational studies focuses mainly on risk of bias and imprecision given we are dealing with single studies and domains of consistency (heterogeneity), indirectness, and publication bias are not ideally applicable. We would consider the magnitude of effect, dose-response, and plausible residual confounding for observational designs.

1risk of bias and imprecision (small sample sizes, small event numbers), downgrade one level each

2risk of bias for in vitro studies uses OHAT risk of bias tool/NTP

url: file:///C:/Users/Owner/Downloads/SESSION_2_02_ROONEY%20(2).PDF whereby questions such as i) was administered dose or exposure level adequately randomized ii) was allocation to study groups adequately concealed and iii) can we be confident in the exposure characterization, were answered. Rating are definitely high, probably high, probably low, definitely low.

3imprecision downgrade one level due to small sample size and events.

4risk of bias downgrade due to open-label and imprecision due to small sample size and events; down-grade of two levels For more information on full systematic reviews from Methodological experts: Paul Elias Alexander, Ludovic Reveiz [email protected]

---

References

1. P. Zhou et al – ., A pneumonia outbreak associated with a new coronavirus of probable bat origin. Nature 579, 270–273 (2020).
2. N. Chen et al – , Epidemiological and clinical characteristics of 99 cases of 2019 novel coronavirus pneumonia in Wuhan, China: A descriptive study. Lancet 395, 507–513 (2020).
3. World Health Organization – , Coronavirus disease (COVID-19) Pandemic. https://www.who.int/emergencies/diseases/novel-coronavirus-2019. Accessed 11 March 2020.
4. H. Lu – , Drug treatment options for the 2019-new coronavirus (2019-nCoV). Biosci. Trends 14, 69–71 (2020).
5. M. Wang et al – ., Remdesivir and chloroquine effectively inhibit the recently emerged novel coronavirus (2019-nCoV) in vitro. Cell Res. 30, 269–271 (2020).
6. M. L. Holshue et al – ; Washington State 2019-nCoV Case Investigation Team, First case of 2019 novel coronavirus in the United States. N. Engl. J. Med. 382, 929–936 (2020).
7. C. D. Russell, J. E. Millar, J. K. Baillie , Clinical evidence does not support corticosteroid treatment for 2019-nCoV lung injury. Lancet 395, 473–475 (2020).
8. L. Shang, J. Zhao, Y. Hu, R. Du, B. Cao , On the use of corticosteroids for 2019-nCoV pneumonia. Lancet 395, 683–684 (2020).
9. Y. Cheng et al ., Use of convalescent plasma therapy in SARS patients in Hong Kong. Eur. J. Clin. Microbiol. Infect. Dis. 24, 44–46 (2005).
10. B. Zhou, N. Zhong, Y. Guan , Treatment with convalescent plasma for influenza A (H5N1) infection. N. Engl. J. Med. 357, 1450–1451 (2007).
11. F. Hung et al ., Convalescent plasma treatment reduced mortality in patients with severe pandemic influenza A (H1N1) 2009 virus infection. Clin. Infect. Dis. 52, 447–456 (2011).
12. J. H. Ko et al ., Challenges of convalescent plasma infusion therapy in Middle East respiratory coronavirus infection: A single centre experience. Antivir. Ther. 23, 617–622 (2018).
13. J. Mair-Jenkins et al .; Convalescent Plasma Study Group, The effectiveness of convalescent plasma and hyperimmune immunoglobulin for the treatment of severe acute respiratory infections of viral etiology: A systematic review and exploratory meta-analysis. J. Infect. Dis. 211, 80–90 (2015).
14. J. van Griensven et al .; Ebola-Tx Consortium, Evaluation of convalescent plasma for Ebola virus disease in Guinea. N. Engl. J. Med. 374, 33–42 (2016).
15. P. I. Lee, P. R. Hsueh , Emerging threats from zoonotic coronaviruses-from SARS and MERS to 2019-nCoV. J. Microbiol. Immunol. Infect., in press.
16. L. Chen, J. Xiong, L. Bao, Y. Shi, Convalescent plasma as a potential therapy for COVID-19. Lancet Infect. Dis. 20, 398–400 (2020).
17. R. Channappanavar, S. Perlman, Pathogenic human coronavirus infections: Causes and consequences of cytokine storm and immunopathology. Semin. Immunopathol. 39, 529–539 (2017).
18. C. Huang et al., Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China. Lancet 395, 497–506 (2020).
19. G. Marano et al ., Convalescent plasma: New evidence for an old therapeutic tool? Blood Transfus. 14, 152–157 (2016).
20. V. W. Wong, D. Dai, A. K. Wu, J. J. Sung, Treatment of severe acute respiratory syndrome with convalescent plasma. Hong Kong Med. J. 9, 199–201 (2003).
21. K. M. Yeh et al., Experience of using convalescent plasma for severe acute respiratory syndrome among healthcare workers in a Taiwan hospital. J. Antimicrob. Chemother. 56, 919–922 (2005).
22. L. K. Kong, B. P. Zhou, Successful treatment of avian influenza with convalescent plasma. Hong Kong Med. J. 12, 489 (2006).
23. W. C. Cao, W. Liu, P. H. Zhang, F. Zhang, J. H. Richardus, Disappearance of antibodies to SARS-associated coronavirus after recovery. N. Engl. J. Med. 357, 1162–1163 (2007).
24. Y. M. Arabi et al ., Feasibility of using convalescent plasma immunotherapy for MERS-CoV infection, Saudi Arabia. Emerg. Infect. Dis. 22, 1554–1561 (2016).
25. M. Eickmann et al ., Inactivation of Ebola virus and Middle East respiratory syndrome coronavirus in platelet concentrates and plasma by ultraviolet C light and methylene blue plus visible light, respectively. Transfusion 58, 2202–2207 (2018).
26. M. Mora-Rillo et al .; La Paz-Carlos III University Hospital Isolation Unit, Acute respiratory distress syndrome after convalescent plasma use: Treatment of a patient with Ebola virus disease contracted in Madrid, Spain. Lancet Respir. Med. 3, 554–562 (2015).
27. B. Benson, M. Moss, C. C. Silliman, Transfusion-related acute lung injury (TRALI): A clinical review with emphasis on the critically ill. Br. J. Haematol. 147, 431–443 (2009).
28. S. B. Halstead, Dengue antibody-dependent enhancement: Knowns and unknowns. Microbiol. Spectr. 2, AID-022-2014 (2014).
29. S. F. Wang et al ., Antibody-dependent SARS coronavirus infection is mediated by antibodies against spike proteins. Biochem. Biophys. Res. Commun. 451, 208–214 (2014).
30. World Health Organization , Clinical management of severe acute respiratory infection when Novel coronavirus (nCoV) infection is suspected: Interim guidance. https://www.who.int/publications-detail/clinical-management-of-severe-acute-respiratory-infection-when-novel-coronavirus-(ncov)-infection-is-suspected. Accessed 13 March 2020.
31. National Health Commission of China , Guideline for diagnosis and treatment for novel coronavirus pneumonia (fifth edition). http://www.nhc.gov.cn/xcs/zhengcwj/202002/3b09b894ac9b4204a79db5b8912d4440.shtml. Accessed 4 February 2020.