The Outcome of Hydroxychloroquine in Patients Treated for COVID-19: Systematic Review and Meta-Analysis

Background The pandemic of coronavirus disease 2019 (COVID‐19) caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) resulted in an unprecedented public health challenge worldwide. Despite urgent and extensive global efforts, the existing evidence is inconclusive regarding the medications used for the treatment of COVID-19. Purpose To generate an up-to-date evidence for the clinical safety and efficacy of hydroxychloroquine (HCQ) with or without azithromycin (AZ) among patients treated for COVID-19. Data Source. PubMed, Cochrane CENTRAL, LITCOVID, Web of Science, SCOPUS, BioRxiv, Embase, MedRxiv, and Wiley online library were searched from 2019/12/30 to 2020/05/23. Study Selection. Three investigators assessed the quality of the studies. Data Extraction. Data about study characteristics, effect estimates, and the quality of the studies were extracted by two independent reviewers and cross-checked by the third reviewer. Data Synthesis. The data of 6,782 (HCQ group, 3623; HCQ + AZ group, 1,020; control group, 2139) participants were included. HCQ was compared with standard care for virologic efficacy, disease progression, mortality, and adverse effects. HCQ was also compared with HCQ + AZ for QTc prolongation, admission to the intensive care unit, and mortality. The study found HCQ did not alter the rate of virologic cure (OR = 0.78; 95% CI: 0.39–1.56) and the risk of mortality (OR = 1.26; 95% CI: 0.66–2.39). The pooled prevalence for mortality was 5.8% (95% CI: 0.9%–10.8%). Moreover, HCQ did not impact disease progression (OR = 0.9; 95% CI: 0.36–2.29) but resulted in a higher risk of adverse effects (OR = 2.35; 95% CI: 1.15–4.8). HCQ was also compared against HCQ + AZ, and no difference was observed in QTc prolongation above 500 ms (OR = 1.11; 95% CI: 0.54–2.28), admission to the intensive care unit (OR = 0.92; 95% CI: 0.52–1.63), and mortality (OR = 0.88; 95% CI: 0.55–1.43). However, in the analysis of single-arm studies, about 11.2% (95% CI: 7.0%–15.5%) of patients have developed an absolute increase of QTc greater than 500 ms, and 4.1% (95% CI: 1.1%–7.1%) of patients discontinued their medication. Conclusion This meta-analysis and systematic review, which included a limited number of poorly designed studies of patients with COVID-19, revealed HCQ is intolerable, unsafe, and not efficacious. Similarly, HCQ + AZ combination was not different from HCQ alone in curbing mortality and ICU admission.


Introduction
e pandemic of coronavirus disease 2019 (COVID- 19) caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) resulted in an unprecedented public health challenge worldwide [1]. As of May 26, 2020, there were more than 5.3 million documented cases, and over 300,000 patients have succumbed to this disease globally [2]. e morbidity and mortality due to COVID-19 have found to increase with age and the presence of comorbid conditions such as diabetes, hypertension, coronary heart disease, or chronic obstructive lung disease [3].
With the rising death toll and a vaccine unlikely very soon, extensive global efforts are underway to develop safe and effective therapeutics against COVID-19 [4]. Among the efforts undergoing to treat the disease, repurposing of old medications is a compelling strategy for which their safety profile, pharmacokinetics, and potential drug interactions are well studied [5].
Initially, a combination of lopinavir and ritonavir was utilized as the first-line agent in Wuhan, China, the epicenter of the disease. However, a previous study [5] failed to show the beneficial clinical effects of this combination. Indeed, it has received considerable criticism from the scientific community [6].
Meanwhile, the aminoquinolines, chloroquine (CQ), and hydroxychloroquine (HCQ) have emerged as a potent inhibitor of SARS-CoV-2 in vitro, and some studies also demonstrated their clinical benefit among hospitalized COVID-19 patients [7][8][9]. On March 30, 2020, the United States Food and Drug Administration (FDA) granted emergency authorization that allowed the use of these drugs in hospitalized COVID-19-pneumonia [10].
To date, regardless of limited evidence, HCQ with or without azithromycin (AZ) is widely utilized in clinical settings to treat thousands of COVID-19 patients around the world [4]. e studies supporting the use of HCQ had suffered from methodological flaws including small sample size and ill quality of design creating difficulty in measuring the true clinical effects. e first study from France showed HCQ and AZ combination as an effective therapy for COVID- 19. In this open-label non-randomized clinical trial, a total of 20 patients were treated with HCQ at a dose of 200 mg three times daily for 10 days, and the data showed a significant reduction in viral carriage at day 6 post-inclusion compared to controls (70.0% clearance by day 6 vs. 12.5% clearance by day 6 in control groups). Interestingly, the addition of AZ to HCQ (n � 6) resulted in a 100% virological cure on day 6 postinclusion, compared with 57.1% virological cure in the HCQ alone arm (n � 14) and 12.5% virological cure in the control arm (n � 16) [8]. Similarly, Million et al. [11] showed HCQ/AZ combination to be safe with a lower death rate. On the contrary, a more recent study conducted by Molina and his colleagues [12] failed to show evidence of a strong antiviral activity or clinical benefit of HCQ in combination with AZ for the treatment of hospitalized patients with severe COVID-19. Notably, the patients included in this study belonged to the severe COVID-19 category and had significant comorbidities including solid and hematological cancers, HIV, and obesity.
A large observational study conducted in the USA reported that HCQ use among patients hospitalized with COVID-19 was not associated with either a greatly lowered or increased risk of intubation or death [4]. A very recent study by Mahévas et al. [13] and Rosenberg et al. [14] also did not show significant differences in terms of in-hospital mortality among patients receiving HCQ with/without AZ compared with standard care. Besides, a systematic review by Sarma and his colleagues [15] concluded that treatment with HCQ had benefits in terms of fewer cases showing radiological progression, time to body temperature normalization, and the number of cough days compared to standard treatment. However, no difference was seen in terms of virological cure, death, or clinical worsening of the disease. e safety of HCQ with/without AZ in COVID-19 patients, including cardiac arrest and QTc prolongation, was also investigated by several studies [14,16]. Interestingly, both drugs can potentially cause QTc prolongation, leading to life-threatening ventricular arrhythmias and torsade de pointes [17]. Critically ill admitted COVID-19 patients with multiorgan failure and metabolic derangements and those having other drugs that can increase the risk of QTc prolongation are at greater risk [18]. A study conducted in New York recorded higher rates of cardiac arrest among patients receiving a combination of HCQ and AZ [14]. Similarly, Ramireddy et al. [19] reported a significant number of patients with QTc-interval prolongation, and the highest QTc values were recorded in those treated with a combination of HCQ and AZ.
Despite numerous studies with small sample size, the efficacy and safety of HCQ in COVID-19 patients remained unclear. Given the inconclusiveness of the existing evidence and awaiting findings from large randomized controlled clinical trials to clear the controversy, we conducted a systematic review and meta-analysis in the interim to investigate the safety and effectiveness of HCQ in the clinical setup.

Objective
e objective of this review is to synthesize an evidence for the safety, efficacy, and tolerability of HCQ with or without AZ among patients treated for COVID-19.

Methods
is review was described by the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) framework.
e studies were identified from PubMed, Cochrane CENTRAL, LITCOVID, Web of Science, SCO-PUS, BioRxiv, Embase, MedRxiv, and Wiley online library. e search was conducted to include human studies published in English language from 2019/12/30 to 2020/05/23. e search terms included 2019-nCoV, 2019 novel coronavirus, COVID-19, coronavirus disease-2019, hydroxychloroquine, Plaquenil, and hydroxychloroquine sulphate. Details of the search strategy for some databases are annexed (Appendix A).
Study designs with single-group prospective/retrospective observational studies and controlled clinical trials were pooled using meta-proportion, while prospective/retrospective observational studies and controlled clinical trials comparing HCQ with or without AZ versus usual care or HCQ with HCQ plus AZ were pooled using RevMan version 5.3. Controlled clinical trials with serious risk of bias were not included in the pooled analysis, and their findings were narrated descriptively.

Inclusion and Exclusion Criteria.
Studies were considered if they included patients who received HCQ alone or in combination with other specific treatment modalities for COVID-19 infection. Both controlled clinical trials (CCTs) and observational studies with and without the comparator group were considered for inclusion. Data on at least one of the following outcomes had to be available for inclusion: virologic efficacy, mortality, disease progression, adverse effects, QTc prolongation, and drug discontinuation due to adverse effects (tolerability). Studies conducted among pediatric COVID-19 patients, case reports, preclinical studies, and studies that did not report outcomes with HCQ in COVID-19 were excluded.

Risk of Bias.
e risk of bias for comparative clinical trials [8,[20][21][22] was assessed using the Cochrane risk of bias tool for randomized controlled studies [23]. e study by Gautret et al. [8] was a non-randomized clinical trial and hence assessed using the ROBINS-I scale [24]. e results of the risk of bias for the studies included in the meta-analysis are found in Appendix B.
e Modified Newcastle-Ottawa Quality Assessment scale [25] was used for observational studies, and the full results are presented in Appendix B.

Outcome Assessment and Statistical Analysis.
e virologic efficacy, mortality, disease progression, adverse effects, QTc prolongation, and drug discontinuation due to adverse effects (tolerability) were assessed. Virologic efficacy is defined as two negative results of SARS-CoV-2 in nasopharyngeal swab using RT-PCR assay with samples obtained 24 hours apart. Disease progression is defined as the need for admission to the intensive care unit, the need for mechanical ventilation, and hospital admission of previously mild cases. Adverse effect is defined as any adverse effect (side effect) reported in a study except QTc prolongation. QTc prolongation is defined as an increase of greater than 60 ms from baseline, and absolute QTc increases to greater than 500 ms.
Open Meta [Analyst] was used to analyse the proportions of mortality, QTc prolongation, and tolerability in single-arm studies [11,12,16,[26][27][28][29]. e pooled proportion of the outcomes was reported with its 95% confidence interval (CI). RevMan 5.3 was used to estimate the risk of virologic efficacy, mortality, disease progression, and adverse effects in studies that compared HCQ with usual care or HCQ with HCQ plus AZ. e odds ratios (ORs) and 95% CI were calculated to estimate the effect sizes. Meta-analysis using the Mantel Hazel method was conducted, and either the fixed-effect or random-effects model was applied. A fixed-effect model was used when the heterogeneity was low to moderate [23]; otherwise, the random-effects model was applied.

Results
e databases (9 databases) search produced 442 articles. After removing the duplicates and excluding 138 articles with thorough evaluation for inclusion using titles and abstracts, 56 full-text articles were assessed for eligibility. Furthermore, 30 full-text articles were subjected to critical appraisal, and 10 articles were dropped with reasons. Twenty full-text articles (5 controlled clinical trials with 288 patients and 15 observational studies with 6,742 patients) were included in the final analysis. Of these, 12 [4, 8, 13, 14, 20-22, 26, 30-32] were double-arm studies ( Figure 1). ese studies compared hydroxychloroquine (HCQ) either with usual/standard care or HCQ with HCQ and azithromycin (AZ). e details of the studies are described in Table 1. Two of the controlled clinical trials [20,22] were at the preprint stage.

Virologic Efficacy.
To estimate the risk of virologic cure, the data of 180 patients (90 HCQ and 90 non-HCQ groups) from two controlled clinical trials [20,21], with moderate risk of bias, were pooled. Although Gautret et al. [8] reported improved virologic cure rate among the HCQ group as compared to the non-HCQ group (16/20 versus 2/16), the finding was not included in the pooled data due to its serious risk of bias. Hence, the pooled result indicated that the virologic cure rate of the HCQ group was not statistically different from the non-HCQ/standard care group (OR � 0.78; 95% CI: 0.39-1.56). e test for the overall effect was Z � 0.69 (p � 0.49) (Tau 2 � 0.00; Chi 2 � 0.18, df � 1 (p � 0.67); I 2 � 0%) ( Figure 2).

Mortality.
e finding was generated from five observational studies [4,13,14,30,31] comprising the data of 2,864 COVID-19 patients (1,311 HCQ and 1,553 non-HCQ groups). e overall result indicated that treatment with HCQ did not result in improved survival (OR � 1.26; 95% CI: 0.66-2.39) as compared to the routine care. e test for the overall effect, Z � 0.71 (p � 0.48). However, the interpretation of this finding might be limited by the substantial heterogeneity; heterogeneity: Tau 2 � 0.42; Chi 2 � 24.91, df � 4 (p ≤ 0.001); and I 2 � 84% ( Figure 3). A controlled clinical trial [8] also reported one death out of 14 HCQ-exposed patients and no death out of 16 in the opposite arm. Of note, the study was removed from pooled analysis as it carries a serious risk of bias (Appendix A).
Likewise, the data of 1,487 COVID-19 patients from single-arm observational studies [12,13,16,17] were included to determine the pooled prevalence of mortality among patients treated with HCQ with or without AZ. e pooled prevalence was 5.8% (95% CI: 0.9%-10.8%) with considerable heterogeneity (I 2 � 92.28%, p < 0.001) ( Figure 4). e heterogeneity could be attributed to the age difference of the COVID-19 patients, as the studies with death reports had a median age of greater than 60 years.
In the three [11,16,27] of the observational studies, the cause of death was respiratory and multiorgan failure. ere was no death due to arrhythmogenic adverse effects.

Disease Progression.
e outcome coded as "disease progression" included the need for admission to the intensive care unit, the need for mechanical ventilation, and hospital admission of previously mild cases. e data of 3,003 COVID-19 patients (1,699 HCQ and 1,304 non-HCQ group) extracted from one controlled clinical trial [22] and five observational studies [4,13,14,31,32] were pooled. e overall random effect analysis indicated HCQ therapy did not appear to halt disease progression (OR � 0.9; 95% CI: 0.36-2.29). e test result for the overall effect, Z � 0.21 is finding was also replicated by subgroup analysis of observational studies (OR � 1.15; 95% CI: 0.45-2.95). However, in subgroup analysis, the finding from a single controlled clinical trial (OR � 0.17; 95% CI: 0.04-0.86) was in favor of the HCQ group ( Figure 5).
ough the study had no risk of bias, the small sample size (31 in each arm) used made the interpretation of the finding extremely tricky. Several studies also indicated HCQ may not improve the rate of disease progression in COVID-19 patients. In a controlled clinical trial [8], which was not included in the pooled analysis, 3 patients were progressed to severe disease in the HCQ group (3/14 versus 0/14). Moreover, in three noncomparative studies [7,17,21,22], out of 1,192 COVID-19 patients treated with HCQ with or without AZ, 42 patients were transferred to ICU and intubated.

QTc Prolongation.
QTc prolongation was reported in two ways in most of the studies. e cut-off points, an increase in greater than 60 ms from baseline, and absolute QTc increase to greater than 500 ms were used as a threshold to discontinue medications responsible or suspected to cause QTc prolongation. ese cut-off points were described by recent guidelines and FDA [18,35]. An absolute QTc prolongation greater than 500 ms was reported in two observation studies [26,31], which compared HCQ alone with HCQ plus AZ. A study by Rosenberg et al. [14] reported that after ECG screening, 80 patients (12.6%) had a QTc prolongation in the combination group, whereas 39 patients (16.7%) developed QTc prolongation in the HCQ group.
In the current review, the data of 291 patients from two observational studies [26,31] were pooled to estimate the   risk of QTc prolongation in patients exposed to HCQ versus HCQ plus AZ. e result indicated the risk of QTc prolongation above 500 ms due to HCQ was statistically not different from those receiving HCQ plus AZ (OR � 1.11; 95% CI: 0.54-2.28) with the test for the overall effect of Z � 0.27 (p � 0.79); heterogeneity: Chi 2 � 0.00, df � 1 (p � 0.97); I 2 � 0% (Figure 7).
However, the findings from noncomparative studies appeared much concerning. Seven single-arm studies reported data on QT prolongation after HCQ ± AZ exposure. Six studies (1,657 patients) were reported a baseline increase in QTc by more than 60 ms [11,12,16,[26][27][28] after HCQ ± AZ exposure. e result of the analysis showed about 13.0% (95% CI: 3.8%-22%) of patients had an increase of QTc by more than 60 ms from the baseline. Considerable heterogeneity was present between studies (Q � 104.16, I 2 � 95.2%, p < 0.001) (Figure 8). Several studies also raised concerns regarding QTc prolongation following HCQ exposure with or without other medications. In four studies [16,[26][27][28], concomitant use of other QT-prolonging medications was reported (   of therapy. Similarly, from five single-arm studies [16,[26][27][28]36], which comprised of 607 COVID-19 patients treated with HCQ ± AZ, about 11.2% (95% CI: 7.0%-15.5%) of patients had developed an absolute increase of QTc greater than 500 ms ( Figure 9). However, a study by Million et al. [11] reported there were no patients who had an increased QTc greater than 500 ms. A meta-regression was conducted to assess the effect of baseline use of other drugs suspected to cause QTc prolongation to above 500 ms. Indeed, the result did not show any significant effect (coefficients (Q) � −0.0; 95% CI: −0.002-0.000; p � 0.79).
In turn, the pooled prevalence of drug discontinuation due to the increased QTc prolongation to greater than 500 ms in three studies [26][27][28] and to greater than 60 ms from baseline in one study [12] among COVID-19 patients treated with HCQ with or without AZ was 4.1% (95% CI: 1.1-7.1). e heterogeneity among the included studies was also acceptable enough (I 2 � 52.19, p � 0.099) ( Figure 10). In one study [11], three patients had discontinued the treatment due to adverse events other than QTc prolongation.

Discussion
e global community is in the state of urgency to mitigate the health and economic crisis instigated by COVID-19. Chloroquine (CQ) and its derivative hydroxychloroquine (HCQ) have traditionally been used for the treatment of malaria and certain autoimmune diseases. e drugs have possible activity against SARS-CoV-1 and SARS-CoV-2 in vitro and in clinical practice [36]. However, clinical studies were reporting contradictory results on the efficacy and safety of HCQ when used for treating COVID-19 patients. us, systematic review and meta-analysis of the existing studies was performed to explore the efficacy, safety, and tolerability of HCQ with or without AZ in COVID-19 patients. e finding of this meta-analysis suggested the use of HCQ did not result in a rapid viral clearance. It also failed to improve survival and rate of disease progression. e pooled prevalence of mortality was higher in patients exposed to HCQ with or without AZ. Moreover, HCQ exposure carried a significant risk of adverse effects and a sizable proportion of patients ended up with drug discontinuation.
On the other hand, a combination of HCQ and AZ did not result in increased risk of QTc prolongation, improved survival, or preventing admission to ICU. However, the pooled proportion of observational studies indicated an alarming rate of QTc prolongation among patients receiving HCQ with or without AZ. e finding of this study showed an absence of rapid viral clearance (OR � 0.78; 95% CI: 0.39-1.56) in patients treated with HCQ compared to the standard care. However, the review by Yang et al. [35] found better virologic efficacy with statistically insignificant results. e finding of this review was based on the data from a single study [8], which had a high risk of bias. More importantly, the former review compared HCQ + AZ with the standard care (control group). is arm of the study had a virologic cure rate of 100% reported from 6 patients. Nevertheless, the current review considered the arm which compared HCQ alone versus the control group, which reported a virologic cure rate of 57% in the HCQ arm versus 12.5% in the control arm. e study by Shamshirian et al. [37] had also questioned the virologic potency of HCQ + AZ (RR � 2.15, 95% CI: 0.31-14.77). Moreover, the optimal duration for virologic clearance is not well known. Gautret et al. [8], Jun et al. [21], and Tang et al. [20] reported a virologic cure after 6, 7, and 28 days of treatment, respectively. erefore, drawing a conclusion from such findings irrespective of the incurred heterogeneity may be erroneous. Of note, all the included studies were open-labeled, non-randomized, and the authors did not describe the type of treatment given for the control group. Similarly, HCQ could not demonstrate improved survival (OR � 1.26; 95% CI: 0.66-2.39) among COVID-19 patients. Gautret et al. [8] also reported an episode of death in the HCQ group, but not in the control group. Previous reviews [37][38][39] reported either increased mortality among HCQ groups or no statistically significant difference. As compared to our review, earlier reviews included a limited number of studies, while Sarma et al. [15]    combined result of mortality and clinical worsening. Our finding was not different from former studies [15,39]. is review also found a 5.8% pooled prevalence of mortality from five observational studies [11,12,16,27,29] among COVID-19 patients treated with HCQ with or without AZ. In three [11,16,27] of these studies, respiratory and multiorgan failure was the cause of death. ere was no reported death due to adverse effects. In all of the included studies, patients had active comorbidity. ere was significant heterogeneity among the studies (I 2 � 92.28%, p ≤ 0.001), which could be attributed to the differences in disease severity and age of the participants. Yet, meta-regression analysis did not produce any evidence.
Furthermore, HCQ therapy was not significantly associated with slowing the composite end point of disease progression (OR � 0.9; 95% CI: 0.36-2.29), unlike the finding by Chen et al. [22] (OR � 0.17; 95% CI: 0.04-0.86). e finding by Chen et al. [22] may not be dependable as it included limited number of patients (31 in each arm). e subgroup analysis result of observational studies also indicated HCQ therapy was not significantly affecting disease progression (OR � 1.15; 95% CI: 0.45-2.95). Analogous results were reported by the preceding studies [15,38,40]. However, one review indicated that HCQ may prevent progression to severe disease among COVID-19 patients [35]. Since the previous review included a limited number of studies, its finding must be interpreted cautiously. Our finding was in agreement with reviews by Sarma et al. [15] and Shamshirian et al. [37]. Nonetheless, these studies reported better outcomes of radiological progression in the HCQ arm, a finding generated from two poorly designed controlled trials [21,22]. e safety of HCQ with or without AZ has been questioned by several studies [37,41]. Diarrhoea, vomiting, blurred vision, rash, and headache were commonly reported adverse effects. In this review, the risk of adverse effects among COVID-19 patients treated with HCQ was 2.35 (OR � 2.35; 95% CI: 1.15-4.8). Shamshirian et al. [37] also found an increase in the odds of the adverse effect among patients exposed to HCQ (OR � 3.55, 95% CI: 1.61-7.82). Similar findings were also reported by other studies [40][41][42].
However, one study reported a conflicting finding where HCQ may be safe and effective, though the authors hinted more data are required for a definitive conclusion [15]. A study conducted in China reported 37.8% of ADRs in a cluster of 217 COVID-19 patients. e predominant adverse effects were drug-induced gastrointestinal disorders and liver system disorders [43]. Cardiac side effects including conduction disturbances (bundle-branch block, incomplete or complete atrioventricular block, QT prolongation, and subsequent torsade de pointes) and cardiomyopathy (hypertrophy and congestive heart failure) were also described [44,45].
In this review, although the risk of QTc prolongation above 500 ms among those exposed to HCQ (OR � 1.11; 95% CI: 0.54-2.28) may not appear concerning, HCQ/AZ combination could be more worrisome. A finding generated from five single-arm studies [16,[26][27][28]36], which comprised the data of 607 COVID-19 patients treated with HCQ ± AZ, showed 11.2% (95% CI: 7.0%-15.5%) in the pooled prevalence of QTc prolongation above 500 ms. In addition, a finding generated from other observational studies [11,12,16,[26][27][28] indicated 13.0% (95% CI: 3.8%-22%) of patients had an increase in QTc by more than 60 ms from the baseline. is finding was not similar to the data synthesized from double-arm studies. is is because of the fact that the double-arm section had compared the data of HCQ alone with HCQ + AZ. As adverse cardiovascular sequels, such as myocarditis, acute myocardial infarction, and heart failure, have been reported in COVID-19 patients and these off-label therapies are not familiar to cardiovascular clinicians managing these patients [46], emergency care physicians should outweigh the risk and the benefits. In a nutshell, numerous studies [14,16,17,26,27,46] have raised concerns over the cardiac safety of this combination.
Drug discontinuation due to adverse effects was also a common finding. e current study showed 4.1% of patients discontinued treatment due to an increase in QTc prolongation. Another study also indicated 4.5% of COVID-19 patients treated with HCQ with or without AZ have discontinued treatment due to adverse effects. Similarly, one patient out of 117 discontinued HCQ in patients receiving HCQ with or without AZ after three days due to QTc prolongation [47].
ough the combination of HCQ + AZ was effective in some studies [15,35], the other study [37] found it carries more hazard of death (RR � 3.65; 95% CI: 1.10-12.10) as compared to the control group. Our findings showed a lack of evidence for curbing mortality and intensive care unit admission of this combination (OR � 0.88; 95% CI: 0.55-1.45).
It is irrefutable that our study has several limitations. e inclusion of studies with a high risk of bias and methodological flaws, combining findings from controlled and uncontrolled studies, may limit its generalizability. We also reported more than one outcome, for instance, ICU admission and need of hospitalization under a single heading, namely, disease progression. is may not show the true picture of the real-world. e existence of heterogeneity, uniform treatment of all cases (mild to severe), and inclusion of limited studies with a small sample size for outcomes such  Selection: (1) representativeness of the exposed cohort: A, consecutive eligible participants were selected, participants were randomly selected, or all participants were invited to participate from the source population; B, not satisfying requirements in part (a) or not stated. (2) Selection of the nonexposed cohort: A, selected from the same source population; * B, selected from a different source population; C, no description. (3) Ascertainment of exposure: A, structured injury data (e.g., record completed by medical staff); * B, structured interview; * C, written self-report; D, no description. (4) For a demonstration that the outcome of interest was not present at the start of the study: * A, yes; B, no or not explicitly stated. b Comparibility: for comparability of cohorts based on the design or analysis: * A, study controls for previous injury; * B, study controls for age. c Outcome: (1) assessment of outcome: A, independent or blind assessment stated or confirmation of the outcome by reference to secure records (e.g., imaging and structured injury data); * B, record linkage (e.g., identified through ICD codes on database records); * C, self-report with no reference to original structured injury data or imaging; D, no description. (2) Was follow-up long enough for outcomes to occur? * A, yes (≥3 months); B, no (<3 months). (3) Adequacy of follow-up of cohorts: A * , complete follow-up-all participants accounted for; * B, subjects lost to follow-up unlikely to introduce bias (<15% lost to follow-up, or description provided of those lost * ); C, follow-up rate <85% and no description of those lost provided; D, no statement. d Total is out of 9 stars. Note: ≥7, high-quality study; 5-7, moderate quality study; <5, lowquality study.
as virologic efficacy may result in biased findings. We could not retrieve some of the important findings such as disease severity scale for each study and the treatments used in the standard care setting.

Conclusion
is systematic review, which included a limited number of poorly designed controlled clinical trials and several real-world studies of patients with COVID-19 requiring hospitalization, found that the use of a regimen containing HCQ with or without AZ did not offer clinical benefit. HCQ with or without AZ did not improve the rate of virologic cure, disease progression, and mortality. ese regimes were associated with more adverse effects. erefore, these drug regimens should only be used in a clinical trial setting, and a large pool of data from randomized clinical trials is warranted to have concrete evidence for safety, efficacy, and tolerability [48].