Favipiravir versus other antiviral or standard of care for COVID-19 treatment: a rapid systematic review and meta-analysis

Background The COVID-19 causing coronavirus is an enveloped RNA virus that utilizes an enzyme RNA dependent RNA polymerase for its replication. Favipiravir (FVP) triphosphate, a purine nucleoside analog, inhibits that enzyme. We have conducted this systematic review and meta-analysis on efficacy and safety of the drug FVP as a treatment for COVID-19. Methods Databases like Pubmed, Pubmed Central, Scopus, Embase, Google Scholar, preprint sites, and clinicaltirals.gov were searched. The studies with the standard of care (SOC) and FVP as a treatment drug were considered as the treatment group and the SOC with other antivirals and supportive care as the control group. Quantitative synthesis was done using RevMan 5.4. Clinical improvement, negative conversion of reverse transcription-polymerase chain reaction (RT-PCR), adverse effects, and oxygen requirements were studied. Results We identified a total of 1798 studies after searching the electronic databases. Nine in the qualitative studies and four studies in the quantitative synthesis met the criteria. There was a significant clinical improvement in the FVP group on the 14th day compared to the control group (RR 1.29, 1.08–1.54). Clinical deterioration rates were less likely in the FVP group though statistically not significant (OR 0.59, 95% CI 0.30–1.14) at the endpoint of study (7–15 days). The meta-analysis showed no significant differences between the two groups on viral clearance (day 14: RR 1.06, 95% CI 0.84–1.33), non-invasive ventilation or oxygen requirement (OR 0.76, 95% CI 0.42–1.39), and adverse effects (OR 0.69, 0.13–3.57). There are 31 randomized controlled trials (RCTs) registered in different parts of the world focusing FVP for COVID-19 treatment. Conclusion There is a significant clinical and radiological improvement following treatment with FVP in comparison to the standard of care with no significant differences on viral clearance, oxygen support requirement and side effect profiles.

around the globe is astronomical. As of 26 August 2020, the number of confirmed cases and deaths reported has reached 23,752,965 and 815,038 respectively [2]. This virus is getting transmitted mainly via respiratory tracts through droplets or respiratory secretions. The disease is characterized by asymptomatic to flu-like mild respiratory symptoms including shortness of breath (SOB) leading to pneumonia, acute respiratory distress syndrome (ARDS), and even multiple organ dysfunction in severe cases [3]. The coronavirus is an enveloped, non-segmented positive-sense RNA virus that utilizes an enzyme RNA dependent RNA polymerase (RdRp) for its replication which could be a potential target for the treatment development [4].
The road to discovering the effective prophylaxis and treatment is still an ongoing process. Numerous trials of medications of different categories have been conducted but none have succeeded to show promising results for effective treatment [5,6]. Some of the repurposed drugs like remdesivir are being utilized along with supportive care for the management of COVID-19 in different clinical settings.
Favipiravir (FVP) triphosphate, a purine nucleoside analog, competitively inhibits the enzyme RdRp. It has shown activity against influenza viruses, RNA viruses associated with viral hemorrhagic fever, and even against SARS-CoV-2 in vitro [7]. The evidence regarding FVP is relatively low as there have only been a handful of studies regarding its efficacy and safety among COVID-19 patients. We conducted this systematic review and metaanalysis to evaluate the efficacy and safety of the drug FVP as a treatment for COVID-19.

Objective
To determine the clinical improvement following the treatment with FVP in the cases of COVID-19, duration to attaining and percentage that attained negative conversion of RT-PCR following the treatment, adverse effects that were seen during the treatment, oxygen and mechanical ventilation requirements following the treatment.

Methods
We used PRISMA for the systematic review of available literature [8].

Criteria for considering studies for this review Types of studies
We included studies that were done to determine the safety and efficacy of FVP along with the standard of care (SOC) for COVID-19 diagnosed cases based on guidelines in comparison to the control group receiving standard of care alone. We only included the case series with more than 5 patients, randomized controlled trials, controlled clinical trials, prospective and retrospective studies where FVP was used in the management of COVID-19 patients in the qualitative analysis. Only the studies with both the treatment and the control groups were included in quantitative synthesis.

Types of participants
The studies had patients with COVID-19 diagnosed as per guidelines who were enrolled either in FVP and SOC compared to standard of care alone in quantitative analysis.

Types of interventions
FVP along with the SOC was taken in the treatment arm and SOC alone in the control arm. SOC included other antivirals, respiratory support, antibiotics, immunomodulators, and herbal medicines.

Types of outcome measures
Our outcomes of interest were clinical improvements following the treatment with FVP in cases of COVID-19; negative seroconversion of RT-PCR; adverse effects that were seen during the treatment; oxygen and mechanical ventilation requirements.

Outcomes
The parameters for clinical improvements were symptomatic and radiological improvements (in CT scan), and clinical deterioration at 7 and 14 days after treatment between the treatment and control group. We also compared overall adverse effects that had occurred during the treatment and respiratory support requirements between the treatment and control groups. We also compared the time to negative RT-PCR and the percentage of negative RT-PCR at day 7 and 14 following treatment.

Search methods for identification of studies
Studies were independently screened by two reviewers (DBS and PB) using COVIDENCE and data were extracted for both quantitative and qualitative synthesis. The conflicts were resolved by taking the opinion of the third reviewer (NP). Assessment of biases and crosschecking of the selected studies were done by another reviewer (SK).

Electronic searches
We have included the electronic search strategy in Additional file 1.

Data collection and analysis
Databases like Pubmed, Pubmed central, Scopus, Embase, Google Scholar, bioRxiv, medRxiv, and clinicaltirals.gov were searched until 20th August, 2020. We decided to include the preprints because the studies on FVP are actively ongoing with very few papers published in academic journals. We extracted data for quantitative synthesis and analyzed it using RevMan 5.4.

Selection of studies
We included RCTs, controlled clinical trials, prospective and retrospective observational studies for all case series with more than 5 patients for our qualitative analysis in which FVP was used in the treatment of COVID-19 patients with sufficient details on outcomes. We included studies with the treatment groups in which patients received FVP and SOC in the treatment group and SOC alone in the control group for quantitative analysis. Studies lacking control groups were excluded in the quantitative analysis. We excluded studies where the outcomes of the patients receiving favipiravir were not properly defined. Case reports, reviews, protocols, in-vitro studies, and letters to editors were also excluded.

Data extraction and management
We evaluated the quality of the studies and included the outcome of interest in the quantitative synthesis.

Assessment of risk of bias in included studies
We used the Cochrane risk of bias (ROB) tool to analyze the risk of bias shown in Fig. 1 [9]. We used the NHLBI (National Heart, Lung, and Blood Institute) quality assessment tools (Additional file 2) to assess the risk of bias in observational studies and case series (Table 1) [10]. We used the RevMan 5.4 for the creation of risk-ofbias plots.

Assessment of heterogeneity
We assessed the heterogeneity using the I-squared (I 2 ) test. We used the Cochrane Handbook for Systematic Reviews of Interventions for interpretation of I 2 test done as follows based on "0-40%: might not be important; 30% to 60% may represent moderate heterogeneity; 50% to 90%: may represent substantial heterogeneity; 75% to 100%: considerable heterogeneity [16]. The importance of the observed value of I 2 depends on (1) magnitude and direction of effects and (2) strength of evidence for heterogeneity (e.g. P value from the chi-squared test, or a confidence interval for I 2 ). "

Assessment of reporting biases
We assessed the reporting biases through predetermined outcome reporting documentation.

Data synthesis
We did a statistical analysis using RevMan 5.4 software. We used Risk Ratio (RR)/ Odds Ratio (OR) for outcome estimation whenever appropriate with 95% Confidence Interval (CI). We used the fixed/randomeffects model as per the heterogeneities. We assessed the heterogeneity using the I 2 test. We analyzed the mean differences among the two groups for the duration of viral clearance using the median, sample size, and interquartile range whenever the means and standard deviations were not provided in the study [17].

Subgroup analysis and investigation of heterogeneity
In the case of heterogeneity, we tried the inverse variance, random-effect model. We then ran an analysis excluding non-randomized study to evaluate their impact on the overall result wherever appropriate. We presented Forest plots to visualize the degree of variation between studies.

Sensitivity analysis
For sensitivity analysis, we examined the effect of study based on their type (RCT and non-RCT) by excluding non-RCT studies when appropriate and re-running the analysis to find any differences.

Qualitative synthesis
We identified a total of 1798 studies after searching the electronic databases. After the removal of 462 duplicates, the title and abstracts of 1336 studies were screened. We        excluded 1284 studies after title and abstracts screening and 52 articles were assessed for full-text eligibility. A total of 43 articles were excluded for definite reasons. We included 9 studies in our qualitative study (Fig. 2). The summary of studies is discussed in Table 2.

Quantitative analysis
Four studies meet the criteria and are included in the quantitative synthesis. In the present meta-analysis, we have compared findings among randomized/nonrandomized controlled studies to extract outcome on viral clearances, improvements or deteriorations among FVP group in comparison to COVID-19 cases getting other antivirals or SOC, duration to viral clearance, the requirement of non-invasive mechanical ventilation/ oxygen support and adverse effects.

FVP versus other antivirals or SOC only; effectiveness
Among the treatment groups FVP in addition to SOC versus other antivirals or SOC we have compared the duration of viral clearance (negative RT-PCR) and radiological/ clinical improvement.

Viral clearance
The meta-analysis of risk ratios (RR) for FVP in addition to SOC effectiveness compared with other antivirals or SOC using random effect model among randomized and non-randomized studies showed that there were no significant differences between two groups (Day 7: RR 1.13, 95% CI 0.55 to 2.33; Day 14: RR 1.06, 95% CI 0.84 to 1.33). Also, there is no significant risk difference (RD) for viral clearance between two groups FVP in addition to SOC versus other antivirals or SOC (Day 7: RD 0.06, 95% CI − 0.34 to 0.45; Day 14: RD 0.03, 95% CI − 0.17 to 0.24) (Fig. 3). For heterogeneity, both subgroup assessments inverse variance method and excluding nonrandomized study by Cai et al. [18] showed no significant changes (Additional file 3/ Figs. 1 and 2).  (Fig. 4). Clinical improvement on the 7th and 14 th day among randomized controlled trials after excluding nonrandomized study by Cai et al. [18] showed slight improvement on favipiravir arm but statistically not significant (Additional file 3/ Fig. 3).

FVP versus other antivirals: clinical/CT deterioration
The meta-analysis on clinical deterioration rate at the end of study duration showed clinical deteriorations is less likely in the FVP treatment group than other antiviral agents though statistically not significant (OR 0.59, 95% CI 0.30 to 1.14; participants = 376; studies = 3; I 2 = 39%) (Fig. 5).

FVP group versus other antivirals or SOC group: Oxygen support or non-invasive ventilation
Meta-analysis on the oxygen support requirements and non-invasive mechanical ventilation among included randomized studies showed decreased odds of oxygen support among FVP group but it is not statistically significant (OR 0.76, 95% CI 0.42 to 1.39; participants = 255; studies = 2; I 2 = 0%) (Fig. 6).

Adverse effects
Meta-analysis comparing adverse effects between the treatment and the control groups showed lesser odds for adverse effect in the treatment arm but of no statistical significance (OR 0.69, 95% CI 0.13 to 3.57; participants = 376; studies = 3; I 2 = 88%) (Fig. 7). Overall adverse effects among randomized controlled trials after excluding non-randomized study by Cai et al. [18] showed slight increase in adverse effects among favipiravir arm but statistically not significant. This may be due to heterogeneity in treatments patients might be taking other than favipiravir or other standard treatment (Additional file 3/ Fig. 4).

Clinical trials
Focusing on the safety and efficacy of FVP for COVID-19 treatment along with different parameters, there are 31 RCTs registered in different parts of the world as of 25 August 2020 (Additional file 4) [22]. Five of such trials have recently been completed from Egypt, Iran, and Turkey. Among the registered RCTs, 14 trials are recruiting participants, 6 trials have not yet started recruiting, and 4 trials are active but not recruiting any participants. One of the trials has been withdrawn thus not been included in this calculation. According to the location provided in 31 trials, a maximum number of trials are regulated by Turkey.

Discussion
Our meta-analysis was focused on the assessment of the clinical outcome and adverse effects following therapy with FVP because it has emerged as one of the treatments repurposed for COVID-19. Although some Fig. 3 Forest plot for risk ratios and risk differences regarding FVP in addition to SOC effectiveness for viral clearance compared with other antivirals or SOC promise has been shown by remdesivir and plasma therapy, the lack of highly efficacious and safe treatment for COVID-19 remains one of the biggest conundrums of the twenty-first century. Our study found that patients had a significant improvement in FVP groups on both the 7 th and 14 th day of treatment (Day 7: RR 1.25, 95% CI 1.01 to 1.53; Day 14: RR 1.29, 95% CI 1.08 to 1.54). The clinical deterioration is less likely in the FVP treatment groups than other antiviral agents (OR 0.59, 95% CI 0.30 to 1.14) following treatments though of no statistical significance. There were no significant differences between the two groups in terms of viral clearance (Day 7: RR       1.33). There were lesser odds for adverse effect in the treatment group but of no statistical significance (OR 0.69, 95% CI 0.13 to 3.57). In general, there were tolerable minor side effects like nausea, vomiting, diarrhea and an increase in transaminases and no serious lifethreatening complications following the FVP treatment. The possible side effects can however not be credited to favipiravir alone because the patients in treatment groups were receiving other drugs in 3 trials except the one done by Ivashchenko et al. [21]. As this is the first meta-analysis comparing the clinical outcome and adverse effects among patients receiving FVP compared to standard of care, we could not compare our findings with other meta-analyses.
Although good promise has been shown by FVP, additional randomized double-blind clinical trials are needed to give a definite opinion about the rationale of the drug. We could only include four studies for our quantitative analysis and one of the studies among them was nonrandomized. The sample size was small in our studies which could decrease the power of our study. The duration of treatment and dosages were different among various studies in qualitative analysis. Two of the RCTs that were included for our analysis had a varied duration of treatment as well. Lack of randomization may have led to selection bias in the non-randomized studies. Blinding was not applied to source studies leading to biases. Selective reporting may have been a problem in Chen's study [19] because of the limited observation time frame. It is important to determine the appropriate dose and duration of treatment with FVP because low dose therapy is found to be a bad prognostic factor for clinical improvement and widespread variations in treatment duration among studies and lack of effective plasma concentrations of drug in critically ill patients [13,14]. Due to the early evidence of potential benefits shown by this drug in clinical improvement as well as imaging improvement, it is necessary to conduct large-scale prospective, double-blind randomized controlled trials or wait for the result of ongoing studies to come. This will embolden the evidences led by our study and eliminate biases so that definitive advice for treatment can be given in the coming days.

Conclusion
Our study concludes that patients had clinical and radiological improvements following the treatment with FVP in comparison to that of the standard of care though no significant differences on viral clearance, oxygen support requirement and side effect profile. The results of ongoing clinical trials should be obtained to give any definite judgment on whether the treatment with FVP is the best option among antiviral treatments for COVID-19 or not. Till then, our meta-analysis supports judicial use of FVP in clinical settings.

Learn more biomedcentral.com/submissions
Ready to submit your research Ready to submit your research ? Choose BMC and benefit from: ? Choose BMC and benefit from: