Abstract
Background:
Coronavirus disease 2019 (COVID-19) has rapidly spread worldwide, but safe and effective treatment options remain unavailable. Numerous systematic reviews of varying qualities have tried to summarize the evidence on the available therapeutic interventions for COVID-19. This overview of reviews aims to provide a succinct summary of the findings of systematic reviews on different pharmacological and non-pharmacological therapeutic interventions for COVID-19.
Methods:
We searched PubMed, Embase, Google Scholar, Cochrane Database of Systematic Reviews, and WHO database of publications on COVID-19 from 1 December 2019 through to 11 June 2020 for peer-reviewed systematic review studies that reported on potential pharmacological or non-pharmacological therapies for COVID-19. Quality assessment was completed using A MeaSurement Tool to Assess systematic Reviews-2 (AMSTAR-2) measure.
Results:
Out of 816 non-duplicate studies, 45 were included in the overview. Antiviral and antibiotic agents, corticosteroids, and anti-malarial agents were the most common drug classes used to treat COVID-19; however, there was no direct or strong evidence to support their efficacy. Oxygen therapy and ventilatory support was the most common non-pharmacological supportive care. The quality of most of the included reviews was rated as low or critically low.
Conclusion:
This overview of reviews demonstrates that although some therapeutic interventions may be beneficial to specific subgroups of COVID-19 patients, the available data are insufficient to strongly recommend any particular treatment option to be used at a population level. Future systematic reviews on COVID-19 treatments should adhere to the recommended systematic review methodologies and ensure that promptness and comprehensiveness are balanced.
The reviews of this paper are available via the supplemental material section.
Introduction
Severe acute respiratory coronavirus 2 (SARS-CoV-2) was first detected in patients linked with Wuhan’s wet markets, but soon grew out of China and the coronavirus disease 2019 (COVID-19) was declared a pandemic on 11 March 2020. 1 Healthcare systems worldwide have been overwhelmed and continue to struggle with the soaring number of patients, limited supplies of personal protective equipment (PPE), and ventilators. Based on the existing evidence, the risk is relatively low for the general population, although elderly populations, immunocompromised people, and those with pre-existing medical conditions are at an increased risk of adverse outcomes.2,3 As of 5 October 2020, 35,330,119 COVID-19 patients have been detected, 1,038,958 of whom have died and 24,564,100 have recovered; 4 figures that are most likely underestimated due to the limited number of SARS-CoV-2 tests completed, inadequate confidence in the sensitivity and specificity of diagnostic testing, and various systems of case reporting and recovery definitions across the world. 5
The current evidence suggests that most COVID-19 patients may remain asymptomatic or experience mild clinical symptoms that require no specific treatments. Severe cases, however, suffer from a range of life-threatening conditions such as pneumonia, acute respiratory distress syndrome (ARDS), 6 and respiratory failure.3,7 To date, highly effective treatments for COVID-19 are unavailable; 8 however, numerous trials and studies have been conducted and are ongoing to develop and assess various pharmacological and non-pharmacological therapeutic options for the treatment and prevention of COVID-19.9,10 As of 5 October 2020, 3507 clinical trials have been registered on ClinicalTrials.gov 11 and the effectiveness of several classes of drugs such as antiviral, antibiotic, antimalarial agents, and corticosteroids are being investigated. 11
Given the urgency of the issue and the novelty of the disease, numerous systematic reviews of varying qualities and scope have tried to summarize the evidence on available therapeutic interventions for COVID-19; papers that have been published in several journals over a very short time. While these review papers are helpful, the overwhelming amount of information and the unprecedented number of COVID-19-related evidence syntheses of unclear quality, create challenges for policymakers and healthcare providers to sift through relevant and reliable research findings. This living overview of reviews aims to provide a succinct summary of the findings of peer-reviewed systematic reviews on different pharmacological and non-pharmacological therapeutic options for COVID-19, and assess the methodological and reporting quality of the existing systematic reviews.
Methods
The living overview of systematic reviews was guided by the methodological guidance on the conduct of overviews of reviews published by the Cochrane collaboration 12 and other recently developed guidelines.13,14 Inclusion criteria and analytical approaches were designed a priori and are documented in a protocol published in the Open Science Framework (https://osf.io/sby9u/).
Databases and search strategy
Following the peer review of electronic search strategies guidelines, 15 we searched PubMed, Embase, Google Scholar, Cochrane Database of Systematic Reviews, and WHO database of publications on COVID-19 from 1 December 2019 through to 11 June 2020, for peer-reviewed review studies that synthesized and reported on potential pharmacological or non-pharmacological therapies for COVID-19. Search terms were combined using appropriate Boolean operators, and included subject heading terms and keywords relevant to COVID-19 and different treatment approaches. We also screened the reference lists of the included articles. Preprint papers that had not gone through peer review were excluded, given their potential for misinforming our conclusions. The live nature of our overview allows for relevant publications to be added to our body of evidence as they become visible in peer-reviewed journals. A detailed overview of the search strategy is provided in Supplemental Table 1.
Inclusion criteria
Systematic reviews, scoping reviews, and rapid reviews that summarized and synthesized primary studies about the effectiveness of pharmacological and non-pharmacological treatments for people with a confirmed diagnosis of COVID-19 using real-time reverse transcription polymerase chain reaction (RT-PCR), nasal/pharyngeal swabs, or other prespecified diagnostic approaches such as clinical or radiological findings were included in this overview. Studies on children or pregnant women were not excluded. Interventions could be of any duration, delivery, frequency, and intensity. Reviews were eligible if they provided an explicit description of inclusion criteria and included a systematic search of at least one electronic database. Letters, commentaries, expert opinion, theoretical and unstructured reviews as well as narrative reviews not meeting the above-mentioned eligibility criteria were excluded. Studies were not excluded based on language, and records published in languages other than English were assessed for eligibility using Google Translate (https://translate.google.com/).
Study selection
Two authors (MKH and MC) completed the title and abstract screening, independently. Citations that met our eligibility criteria or were unclear, were screened at full text by two independent reviewers (MKH and MC). Disagreements over the inclusion of studies were resolved through discussion or arbitration with the senior author (MK). Duplicate records were excluded.
Data extraction
A data extraction sheet was designed and piloted by two independent authors (MKH and MC). Data were extracted independently by the two authors (MKH and MC), and discrepancies were resolved through discussion or arbitration with the senior author (MK). Data were extracted on study type (e.g. systematic review, rapid review), publication date, information sources (e.g. MEDLINE, Embase), search date, number of included studies in each review, total sample size, age and sex of the participants, study design of included studies in the review, description of the pharmacological (e.g. antivirals, antibiotics, antimalarial) and non-pharmacological (e.g. oxygen therapy, invasive or non-invasive mechanical ventilation) treatments, and main findings.
Quality and overlap assessment of individual reviews
Two reviewers (MKH and MK) independently evaluated the methodological quality of included reviews using A MeaSurement Tool to Assess systematic Reviews-2 (AMSTAR-2). 16 This tool contains 16 items to appraise the methodological aspects of the systematic reviews that include randomized or non-randomized primary studies. The methodological quality for each review was rated as critically low (i.e. more than one critical flaw with or without non-critical weaknesses), low (i.e. one critical flaw with or without non-critical weaknesses), moderate (i.e. more than one non-critical weakness), and high (i.e. no or one non-critical weakness). Scores on the AMSTAR-2 tool range from 0 to 16 and higher scores correspond to higher quality of systematic reviews. 16 We also assessed the degree of overlap within the systematic reviews by calculating the recently developed metric of corrected covered area (CCA) based on the following recommended cut-offs (i.e. slight overlap: CCA of 0–5%; moderate overlap: 6–10%; high overlap: 11–15%; and very high overlap: >15%). 17
Data synthesis
Studies were summarized in a narrative fashion, and an overview of their methods and main findings were presented. Data synthesis was based on the pharmacological and non-pharmacological nature of the treatments.
Living overview of reviews
Given the rapidly increasing number of systematic reviews on COVID-19 treatment options, we will screen the online databases listed above on a bimonthly basis through December 2020 for new relevant evidence. If resources are available, we plan to update the overview of reviews when new peer-reviewed evidence that significantly alters the direction or strength of our original conclusions emerges.
Results
Literature search
The online database and reference search retrieved 992 studies, 176 of which were duplicate records and were excluded. A total of 686 records were excluded after the title and abstract screening. Of the remaining 130 articles that were assessed in the full-text screening phase, 45 were included in the overview (Figure 1). No additional systematic reviews were found through reference list screening.
Flowchart of screened and included studies in the overview of reviews on therapeutic interventions for COVID-19.
Characteristics of the included reviews
Table 1 summarizes the overall characteristics of the 45 included reviews. All reviews were published between April and June 2020 and search dates were as early as 2 March 2020. 7 Reviews were published in 30 different journals, most of which had specific submission requirements for systematic reviews. The journal impact factor (JIF) of journals that published the systematic reviews ranged from 0 to 19.315 (median JIF = 3.689). All included reviews were published in English. The number of included primary studies in the reviews varied greatly from one 18 to 60 19 (median number of primary studies = 11). The majority (n = 30; 67%) of the reviews did not include a meta-analysis in their statistical analysis. Most reviews investigated pharmacological treatments for COVID-19.
Characteristics of the included studies in the overview of reviews.
Studies are listed on the chronological order of their publication dates.
CBM, Chinese Biomedical Database; CNKI, China National Knowledge Infrastructure; CENTRAL, Cochrane Central Register of Controlled Trials; ICTRP, WHO International Clinical Trials Registry Platform; JIF, Journal Impact Factor; ISRCTN, International Standard Randomized Controlled Trial Number; MA, Meta-Aanalysis; MERS, Middle East Respiratory Syndrome; NR, Not Rreported; nRCT, non-Randomized Controlled Trial; PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Anlyses; qRCT, quasi-Randomized Controlled Trial; RCTs, Randomized Controlled Trials; RR, Rapid Review; SARS: Severe Acute Respiratory Syndrome; SR, Systematic Review; SSRN, Social Science Research Network; WHO, World Health Organization.
A range of study designs was included across different systematic reviews; most of which included retrospective/prospective observational studies and case series/case reports. The target population was confirmed cases of COVID-19 patients and included children in three reviews43,46,47 and adult patients in the remaining studies. None of the included reviews reported on pregnant women.
Quality of included reviews
Supplemental Table 2 shows the details of AMSTAR-2 scores for each systematic review. The AMSTAR-2 score had a mean of 6.0 (range 0–13) and methodological drawbacks were frequent. According to AMSTAR-2, 29 systematic reviews were of critically low quality, nine were of low quality, and seven were of moderate quality. Seventeen reviews had pre-specified their clinical research question and inclusion criteria according to the components of population, intervention, comparators, and outcomes (PICO). Study selection and data extraction were performed in duplicate in 31 and 22 reviews, respectively. Only three reviews conducted a comprehensive literature search, and 13 reviews provided a list of excluded studies. Twenty-seven reviews described the characteristics of included studies in adequate detail and, 13 reviews assessed the risk of bias. In 15 reviews, appropriate meta-analysis methods were used for the statistical combination of results. Seven reviews provided sufficient explanation for any heterogeneity observed in the results, and eight studies carried out an adequate investigation of publication bias. Only in two reviews, the authors assessed the potential impact of risk of bias in individual studies on the results of the meta-analysis.24,28 Only three reviews accounted for the risk of bias when discussing the results.24,26,32 The value of the CCA was approximately 1.7, which corresponded to a slight overlap.
Therapeutic interventions for COVID-19
Details of the pharmacological and non-pharmacological therapeutic interventions and the main findings of each systematic review are summarized in Table 2. Most systematic reviews were conducted on observational studies without having suitable control groups for comparison. Moreover, systematic reviews reported on a range of different outcomes (e.g. viral clearance, viral load, improvement in radiology findings, time to a clinically meaningful response, mortality, intensive care unit [ICU] admission or duration, need for mechanical ventilation, disease progression, time to recovery, C-reactive protein level, white blood cell count, ARDS, hospitalization, fever reduction, cardiac failure) that made comparisons across the studies and drawing conclusions challenging.
Overview of therapeutic interventions for COVID-19 patients included in the study.
Studies are listed in the chronological order of their publication dates.
ACE, Angiotensin-Converting Enzyme; ARDS, Acute Respiratory Distress Syndrome; CPAP, Continuous Positive Airway Pressure; CQ, Chloroquine; CRRT, Continuous Renal Replacement Therapy; ECMO, Extracorporeal Membrane Oxygenation; HCQ, Hydroxychloroquine; HFNC, High Flow Nasal Cannula; HR, Hazard Ratio; ICU, Intensive Care Unit; IL, Interleukin; JAK2, Janus Kinase2; IMV, Invasive Mechanical Ventilation; IVIG, Intravenous Immunoglobulin; MD; Mean Difference, MERS-CoV, Middle East Respiratory Syndrome Coronavirus; NIV, Non-Invasive Ventilation; NR, Not Reported; OR, Odds Ratio; PCR, Polymerase Chain Reaction; RCT, Randomiized Controlledd Trials;g RD, Risk Difference; RNA, Ribonucleic Acid; RR, Risk Ratio; RT-PCR, Real-time Reverse Transcription Polymerase Chain Reaction; SARS-CoV, Severe Acute Respiratory Syndrome Coronavirus; WMD, Weighted Mean Difference.
Antiviral agents
Most reviews reported antivirals as a common treatment for patients. The most commonly used antiviral treatment modalities included HIV protease inhibitor (e.g. lopinavir20,23,25,28,30–34,39,42,44,46,48,53,56,58,60 and ritonavir),19,20,23,25,28,30,31,33,39,42,44,46,49,50,53,56,58,60 Influenza treatments (e.g. arbidol hydrochloride,20,23,28,30,31,33,34,44,46,49,53,56,58,60 and favipiravir),20,23,46,53,58 neuraminidase inhibitors (e.g. oseltamivir),19,23,31,33,44,46,53,56,60 nucleoside analogues (e.g. ribavirin),23,28,31,33,34,44,46,56,58 nucleotide analogues (e.g. remdesivir27,31,35,37,39,50,58 and ganciclovir),19,23,33,53,60 and biological response modifier-cytokines (e.g. interferon alfa).28,31,32,34,43,44,46,49,53,56,58,60 Study findings on the effectiveness of antiviral drugs in the treatment of COVID-19 were equivocal and varied greatly across different reviews. For example, the systematic review of case reports by Insiat Islam 58 showed that antiviral agents such as lopinavir/ritonavir and arbidol reduced SARS-CoV-2 viral load, and improved clinical symptoms. That study also reported that patients receiving favipiravir had a shorter SARS-CoV-2 clearance and improved chest imaging as compared with the lopinavir/ritonavir group. 58 However, in another review by Ford et al., lopinavir/ritonavir use was associated with a non-statistically significant shorter time to clinical improvement and lower mortality compared to the control group. 53 Moreover, Liu et al., 20 Xu et al., 39 and Chowdhury et al. 42 reported no meaningful benefit for lopinavir/ritonavir. In addition, Shi et al. reported the safety and effectiveness of antiviral agents for children with COVID-19 as unclear. 46 Overall, four reviews did not specify what antiviral agents were used in the treatment of COVID-19 patients7,43,48,54 and several studies did not report specific findings for antiviral agents used among the patients.
Antibiotic agents
Antibiotics to treat patients with COVID-19 were reported in 13 reviews.7,19,23,25,28,31–33,43,47,54,56,60 Macrolides (e.g. azithromycin),9,10,18,19,22,23,25,33,36,42,47,49,52,55,58,59 fluoroquinolone (e.g. moxifloxacin,19,23,28,31,33,41,47,60 levofloxacin),31,44,47 glycopeptides (e.g. tigecycline,47,57 vancomycin),19,31,44 cephalosporins (e.g. cefoperazone, 47 cefoselis, 31 cefdinir, 47 and cefepime),19,31,47 beta-lactams (e.g. imipenem sulbactam, 47 meropenem,19,31,47 piperacillin tazobactam), 31 and oxazolidinones (e.g. linezolid) 47 were the main classes of antibiotics used to treat COVID-19. Antibiotics were mainly used for addressing bacterial co-infections along with antivirals or antimalarial agents; however, their effectiveness was not directly investigated. Wang et al., who specifically evaluated the efficacy and safety of antibiotic agents in children with COVID-19, concluded that there was no direct evidence to support their efficacy. 47 Six reviews did not specify the antibiotic agents used in the treatment of infected patients.7,32,43,53,54,56
Chloroquine and hydroxychloroquine
A total of 23 systematic reviews evaluated the efficacy of chloroquine and hydroxychloroquine on treating COVID-19 patients,9,10,18,20,22,25,26,28,30,32,36,39,40,42,44,46,49,50,52,55,57–59 most of which concluded that these drugs were inefficient and unhelpful for COVID-19 patients. However, some in vitro data supported the ability of chloroquine and hydroxychloroquine in inhibiting the activity of SARS-CoV-2 and reported them as beneficial in both prophylactic and therapeutic interventions.10,58 Moreover, several systematic reviews9,18,20,22,25,26,36,42,49,52 that reported in favor of using this class of drugs, cited an open-label non-randomized controlled trial by Gautret et al., 61 and concluded that hydroxychloroquine treatment (600 mg/daily) was associated with SARS-CoV-2 viral load reduction/disappearance, and its effects were reinforced by azithromycin. Other studies also reported that using a 400 mg/daily dosage led to no clinical difference between hydroxychloroquine and the standard treatment group on day 7.9,10,20,22,42,52 Conversely, very few studies reported some clinically meaningful outcomes associated with the use of these drugs (e.g. a smaller number of cases showing radiological progression of lung disease,25,49,52 faster fever recovery and cough relief, 26 and virological clearance). 57
Corticosteroids and glucocorticoids
Corticosteroids including methylprednisolone and dexamethasone were another common drug category used to treat COVID-19, mainly in addition to other classes of drugs such as antivirals and antibiotics.7,19,23,24,30–33,38–40,43,45,48,51,53,54,56–58 The evidence, however, was inconsistent about administering corticosteroids for COVID-19 patients. For example, Tobaiqy et al. 57 and Zhang et al. 33 recommended against using corticosteroids among COVID-19 patients due to increased risks of mortality. Lu et al. also reported higher rates of death (Relative risk [RR] = 2.0) among COVID-19 patients who were receiving corticosteroids in comparison with those who did not. 45 Conversely, Yang et al. 54 and Alijotas-Reig et al. 40 showed that patients with severe conditions were more likely to benefit from corticosteroid therapy. In addition, Ye et al. 24 reported that corticosteroids were only beneficial in reducing mortality (Hazard ratio [HR] = 0.41) among COVID-19 patients who also had ARDS. Some studies also reported no clinically meaningful benefits or harms for using corticosteroids among COVID-19 patients.38,39,51
Other pharmacological treatments
Ten reviews reported that immunoglobulin was used to treat some of the COVID-19 patients.7,19,23,31,32,40,43,53,56,57 These medications were used as co-interventions in COVID-19 treatment and their specific efficacy was not evaluated. Valk et al. 32 reported that the use of immunoglobulin was associated with improvement in clinical symptoms and a reduction in hospitalization period among COVID-19 patients. Moreover, Alijotas-Reig et al. 40 recommended a high intensity anti-inflammatory and immunomodulatory therapy for severe patients as a potential therapeutic option. In six studies, Interleukin-6 (IL-6) agents such as tocilizumab and sarilumab were used as supportive treatments for COVID-19.25,30,39,40,48,58 Some studies showed that tocilizumab could improve clinical status in severe to critically ill patients.39,40,48,58 The C-reactive protein level was significantly lower after the initiation of tocilizumab in patients. 30 In addition, five studies19,41,44,56,57 reported the usage of antifungal agents in the treatment of patients in some studies, mainly as an add-on therapy. Traditional Chinese medicine34,53,57,58 and herbal medicine23,28,31 were some of the other treatment co-interventions examined for COVID-19, yet their effectiveness was not fully assessed and few details were provided for these interventions.
Non-pharmacological interventions
Oxygen therapy and invasive or non-invasive mechanical ventilation were the two main supportive non-pharmacological interventions used in COVID-19 treatment approaches.7,19,25,29,31,32,43,44,46,50,53,54,56 As these interventions were mainly used as an add-on treatment in COVID-19 patients with severe or in critical conditions, reviews did not specifically assess or discuss their efficacy. As ARDS was one of the frequently reported complications among COVID-19 patients, the principles of treatment for these patients were focused on improving oxygenation and supporting the function of multiple organs. 33 However, included reviews did not specify the effectiveness of oxygen therapy and ventilation in COVID-19 patients. Moreover, two studies reported on the use of mesenchymal stromal cells but provided few details about its efficacy,21,39 five reported on the use of convalescent plasma,32,39,40,44,57 and three reported on the use of continuous renal replacement therapies (CRRT);7,19,54 none of which reported on the specific efficacy of these interventions.
Adverse drug events
A total of 13 systematic reviews reported adverse drug events (ADEs) for hydroxychloroquine/chloroquine phosphate.10,20,22,25,35,39,40,46,50,52,53,57,59 Two systematic reviews and meta-analyses showed that there was no statistically significant difference in the incidence of ADEs between hydroxychloroquine therapy and the control group (Odds ratio [OR] = 2.19, 95% confidence interval [CI] 0.59–8.18, 52 and RR = 1.65, 95% CI 0.50–5.50). 46 Studies found hydroxychloroquine to be safe with mild adverse reactions such as diarrhea, nausea, vomiting, rash, headache, lethargy, blurred vision, and transient abnormal liver functions.22,25,34,39,50,52,55,59 Moreover, hydroxychloroquine was not found to increase the occurrence of abnormal liver function test results, increased serum creatinine level, rash, headache, or anemia in comparison with the control group. 25 Some studies reported QT prolongation associated with hydroxychloroquine treatment.22,25,52,59 Hernandez et al. reported that more patients receiving hydroxychloroquine plus azithromycin had a QTc interval of 500 ms or greater [mean difference = 1.8%, 95% CI –14.9–18.5%]. Moreover, more patients had a QTc interval increase of ⩾ 60 ms compared to baseline (mean difference = 5.1%, 95% CI –7.6% to 17.8%) versus hydroxychloroquine alone. 25 Other electrocardiogram changes were first-degree atrioventricular block22,59 and left bundle-branch block. 59 One patient died in the hydroxychloroquine arm on day three despite a negative RT-PCR.52,59 Severe ADEs were not reported in patients treated with chloroquine,10,59 but those receiving a higher dose of chloroquine therapy experienced a slight increase in anemia and a large increase in serum creatinine level compared with those receiving a lower dose. 25 Moreover, 23% (22 of 95) of patients receiving chloroquine had a QTc interval greater than 500 ms. 25
ADEs were also reported in 60% of patients taking remdesivir,27,39 and the most common ADEs were rash, diarrhoea, hypotension, nausea, abnormal liver function, and renal impairment. Serious ADEs (e.g. acute kidney injury, septic shock, multi-organ failure, and acute respiratory failure) were noted in 18–23% of patients and 8–12% of patients discontinued remdesivir because of mild or serious ADEs.27,35,39 The review of Shi et al. reported no statistically significant difference in the incidence of ADEs (RR = 1.24, 95% CI 0.67–2.28) and serious ADEs (RR = 0.62, 95% CI 0.38–1.01) among patients receiving lopinavir/ritonavir and others groups. 46 The most common reported ADEs were gastrointestinal reactions, abnormal liver function, anemia, insomnia, bradycardia, or hypoxemia.20,39,46,50,53 Moreover, 13% of patients were unable to complete the full course of treatment due to anorexia, nausea, abdominal discomfort, or diarrhoea. 39 In the study by Tobaiqy et al., the patients who took up lopinavir/ritonavir also complained about psychiatric symptoms, gastrointestinal ADEs, skin eruptions, and hypokalemia. Abnormal liver function, raised serum uric acid, psychiatric symptom reactions, and gastrointestinal ADEs were also detected in the patients receiving favipiravir and arbidol. 57 Moreover, favipiravir caused diarrhoea, and two systematic reviews reported no ADEs for arbidol.20,39 Furthermore, no obvious ADEs were reported in children receiving antiviral therapies. 46
The combined therapy of herbal medicine with western medicine had minor ADEs including, nausea and vomiting, diarrhoea, liver damage, and reduced blood cell count. There was no statistically significant difference between the combination of herbal medicine and western medicine versus western medicine alone (risk difference = 0.06, 95% CI –0.04 to 0.15).28,34 Valk et al. reported moderate fever after the transfusion of convalescent plasma 32 and no other adverse reactions were reported.32,44 Moreover, ADEs were not reported for intravenous immunoglobulin (IVIG). 43 Mild adverse reactions were reported in patients treated with mesenchymal stromal cells (e.g. grade I allergic reaction, generalized skin rash, diarrhoea, transient desaturation, dyspnea, and hypotension). 21 Finally, hyperglycemia, hypernatremia, hypokalemia, coinfections (bacterial or fungal), multiple organ dysfunction syndrome, and ARDS were the most common ADEs associated with corticosteroid use and required routine monitoring.24,45,57
Discussion
This overview of reviews summarized the systematic reviews describing therapeutic interventions for COVID-19 patients. While the antiviral and antibiotic agents, corticosteroids, and antimalarial agents were the most common drug classes used to treat COVID-19, the systematic reviews showed no strong evidence to support their efficacy. Oxygen therapy and ventilatory support were also the most common supportive care in patients with respiratory distress and low oxygen saturation levels. The outcomes reported in the included review studies were either incomplete or unclear, and most did not report on ADEs of different treatment options which made interpretations and comparisons about the efficacy and safety of treatments challenging.
Given the lack of an effective antiviral treatment for COVID-19, drugs previously developed to treat other viral infections are being tested. The most widely used antiviral agents are lopinavir, ritonavir, oseltamivir, arbidol, remdesivir, and cytokines. The timing, duration, and dose of antiviral agents varied greatly across the studies, and most patients were provided with other interventions that may have contributed to outcomes such as recovery and death. Overall, although there are early findings about the potential promise in the safety and efficacy of combined interferon beta-1b, lopinavir–ritonavir, and ribavirin, the evidence supporting their efficacy and safety has been rather inconsistent and uncertain.20,39,42,46,49,53,62 The evidence on remdesivir has also been mixed and requires further higher quality data for better assessment of its efficacy and safety. While remdesivir use was not associated with statistically significant clinical benefits and only led to a numerical reduction in time to clinical improvement among 237 patients enrolled in a randomized, double-blind, multicentre clinical trial in China, 63 there is some evidence that supports a modest improvement in clinical outcomes among COVID-19 patients who have received remdesivir.27,35,37 Nonetheless, as outlined in several systematic reviews in our study, the certainty of the evidence for the efficacy and safety of antiviral agents in COVID-19 treatment remains low. On the other hand, a high proportion of patients with COVID-19 were treated with antibiotics, despite the lack of aetiological evidence. This is concerning as an emerging body of evidence suggests that secondary bacterial or fungal infections such as Acinetobacter baumannii, Klebsiella pneumonia, and Aspergillus flavus may occur in COVID-19 patients.47,64 For example, a brief report describing the characteristics of 3200 COVID-19-related deaths from Italy reported superinfection in 8.5% of patients. 65
Despite the initial excitement about the potential significant effectiveness of chloroquine and hydroxychloroquine for treating COVID-19,52,61,66 the findings of our overview are in line with more recent high-quality studies67–69 such as the Randomized Evaluation of COVid-19 thERapY (RECOVERY) trial 70 that suggest little to no benefits in using these drugs for COVID-19 treatment. Moreover, in a large multinational study of over 96,000 patients with confirmed COVID-19 in 671 hospitals around the world, hydroxychloroquine or chloroquine, used alone or with a macrolide, were indeed harmful to in-hospital outcomes for COVID-19.69,71 Findings from another well-designed recent randomized controlled trial also indicated that in comparison with standard of care, hydroxychloroquine was not associated with higher rates of negative conversion and led to a higher probability of experiencing ADEs among its recipients. 67 Overall, the harms of using this group of drugs seem to overweigh their benefits for COVID-19 patients and the US Food and Drug Administration’s decision to caution against the use of these drugs seems well supported by the existing evidence.72,73
Corticosteroids were also assessed as largely unhelpful in most reviews. Patients with severe conditions were more likely to require corticosteroids and they were associated with increased mortality in patients with coronavirus pneumonia. 54 Moreover, recent studies have recommended against using corticosteroids for the routine treatment of COVID-19 patients due to increased risks of steroid-induced osteonecrosis of the femoral head (ONFH).74,75 Corticosteroids may still be a viable option for patients in critical conditions or septic shock, and prescribing them with bisphosphonates and vitamin E might help address some of the complications such as mortality and mechanical ventilation.76–78 Immunoglobulin, gamma globulin, traditional Chinese and herbal medicine were other treatment options for treating COVID-19; however, they were relatively uncommon at the time of writing this review and data about their efficacy are limited. 79
Previous studies have shown that an increased level of cytokines, IL-6 in severe patients in particular, may be attributed to cytokine release syndrome.80,81 As IL-6 plays an important role in the cytokine storm,80,81 it may serve as a possible treatment approach in severe patients. Based on six systematic reviews that assessed tocilizumab, this IL-6 blockade agent could be beneficial in the treatment of severely ill COVID-19 patients. These findings are in line with recent cohort studies that have suggested that the IL-6 blockade does not impair the specific antibody response against SARS-CoV-2 and may help reduce the risk of death among severely ill COVID-19 patients;82,83 nonetheless, they warrant further investigation in sufficiently controlled trials.
This overview identified oxygen therapy and ventilation as the two main non-pharmacological interventions used for COVID-19 patients. As the disease progresses, greater amounts of oxygen are needed. In patients with COVID-19, the decision to use high-flow oxygen via a nasal cannula or the initiation of non-invasive ventilation is controversial and there are no data describing whether these modalities were successful at avoiding intubation. 84 A systematic review on severe and critically severe COVID-19 patients suggested that the principles of treatment for these patients should be lung-protective and focus on improving oxygenation. 85 CRRT and convalescent plasma therapy were other therapeutics assessed by some reviews but their effects on COVID-19 patients were not fully evaluated. The result of a retrospective cohort study showed the patients in the CRRT group had better survival rates than those in the non-CRRT group. 86 Furthermore, case series showed that convalescent plasma therapy was well tolerated and could potentially improve the clinical outcomes through neutralizing viremia in severe COVID-19 cases.8,87,88 However, the optimal dose, time point, and the clinical benefits of this intervention need further investigations in larger and well-controlled trials.
Most systematic reviews scored poorly on the AMSTAR-2 tool; 29 were assessed as critically low, nine as low quality, and seven as moderate quality. Most of the included studies in the reviews were observational, which was indeed inevitable as the findings of large clinical trials are not fully available. We also noted that the findings of the studies were not likely to be biased due to overlap. Nonetheless, these limitations have significant implications for policy and clinical decision-making. It is important to ensure that the urgency for summarizing information on the effective treatment options for COVID-19 is not used as a justification for low-quality evidence syntheses. Indeed, promptness and comprehensiveness must be balanced to ensure decision-makers and clinicians are not provided with misleading information. Moreover, reporting outcomes about the safety of treatment options for COVID-19 is of utmost importance and should not be overlooked.
We acknowledge four main limitations of our overview of reviews. First, the actual effects of different treatments is still unclear, and our understanding continues to improve; however, our findings are important and informative for the decisions that need to be made based on the existing evidence. Second, evaluating the efficacy of therapeutic interventions was not a primary objective in most of the included systematic reviews; therefore, sufficient information and details were not provided in some studies. Third, the heterogeneity of the studied population, unclear definition of treatment and outcome variables as well as comparison groups of reviews limited our statistical analysis and we could not conduct a meta-analysis or network meta-analysis. We are hoping to update our live overview as more high-quality data and statistically comparable outcomes become available. Finally, the methodological quality of reviews was very concerning and the evidence on treatments was mostly from a few countries, China in particular. It is important to revisit what we know as treatment data from other countries are released. Despite these limitations, this overview followed a rigorous methodological approach and contributes to the literature by providing a comprehensive summary of all available evidence about COVID-19 treatment options as well as assessing the confidence in the findings of the existing systematic reviews.
Conclusion
The number of reviews on pharmacological and non-pharmacological therapies for COVID-19 is rapidly increasing. However, many reviews still show significant methodological flaws limiting definite conclusions about the efficacy and safety of therapeutic interventions. In particular, very little is known about the cost-effectiveness of existing treatments, particularly in low and middle-income countries. More high-quality, evidence-based clinical trials with proper design and adequate sample size would help reach more reliable results about the use of specific therapeutic interventions in the treatment of COVID-19 in the short and long term. Overall, this review demonstrates that although some therapeutic interventions (e.g. corticosteroids, remdesivir) may be helpful to certain subgroups of COVID-19 patients, the available data are insufficient to strongly recommend any particular treatment to be used at a population level.
Supplemental Material
sj-pdf-1-tar-10.1177_1753466620976021 – Supplemental material for Therapeutic interventions for COVID-19: a living overview of reviews
Supplemental material, sj-pdf-1-tar-10.1177_1753466620976021 for Therapeutic interventions for COVID-19: a living overview of reviews by Malahat Khalili, Maryam Chegeni, Sara Javadi, Mehrdad Farokhnia, Hamid Sharifi and Mohammad Karamouzian in Therapeutic Advances in Respiratory Disease
Supplemental Material
sj-pdf-2-tar-10.1177_1753466620976021 – Supplemental material for Therapeutic interventions for COVID-19: a living overview of reviews
Supplemental material, sj-pdf-2-tar-10.1177_1753466620976021 for Therapeutic interventions for COVID-19: a living overview of reviews by Malahat Khalili, Maryam Chegeni, Sara Javadi, Mehrdad Farokhnia, Hamid Sharifi and Mohammad Karamouzian in Therapeutic Advances in Respiratory Disease
Supplemental Material
sj-pdf-3-tar-10.1177_1753466620976021 – Supplemental material for Therapeutic interventions for COVID-19: a living overview of reviews
Supplemental material, sj-pdf-3-tar-10.1177_1753466620976021 for Therapeutic interventions for COVID-19: a living overview of reviews by Malahat Khalili, Maryam Chegeni, Sara Javadi, Mehrdad Farokhnia, Hamid Sharifi and Mohammad Karamouzian in Therapeutic Advances in Respiratory Disease
Footnotes
Author contribution(s)
Conflict of interest
The author(s) declare that there is no conflict of interest.
Ethics
Given the secondary nature of this overview and no interaction with humans, ethics approval was not required.
Funding
Mohammad Karamouzian is a member of Pierre Elliot Trudeau Foundation’s COVID-19 impact committee, and is supported by the Vanier Canada Graduate Scholarship as well as the Pierre Elliott Trudeau Foundation Doctoral Scholarship.
Supplemental material
The reviews of this paper are available via the supplemental material section.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
