If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
a Postal Address: STMU Shifa College of Medicine, Sector H-8/4, Islamabad, Pakistan. Tel. : +92 333 6741 967
Ali Haider Bangash
Correspondence
Corresponding author: Dr Ali Haider Bangash, MBBS, Shifa International Hospital, Islamabad, Pakistan. Medical student member of the American College of Physicians (ACP), American Society of Clinical Oncology (ASCO), European Academy of Neurology (EAN), Congress of Neurological Surgeons (CNS), North American Spine Society (NASS), Society for Neuro-Oncology (SNO), Alzheimer's Association International Society to Advance Alzheimer's Research and Treatment (ISTAART), International Parkinson and Movement Disorder Society & Spine Intervention Society.
Azithromycin has been adopted as a component of the COVID-19 management protocol throughout the global healthcare settings but with a questionable if not downright unsubstantiated evidence base.
Objectives
In order to amalgamate and critically appraise the conflicting evidence around the clinical efficacy of Azithromycin (AZO) vis a vis COVID-19 management outcomes, a meta-analysis of meta-analyses was carried out to establish an evidence-based holistic status of AZO vis a vis its efficacy as a component-in-use of the COVID-19 management protocol.
Methods
A comprehensive systematic search was carried out through PubMed/Medline, Cochrane and Epistemonikos with a subsequent appraisal of abstracts and full-texts, as required. The Quality of Reporting of Meta-analyses (QUOROM) checklist and the Assessment of Multiple Systematic Reviews (AMSTAR) methodology were adopted to assess the methodological quality of the included meta-analyses. Random-effects models were developed to calculate summarized pool Odds Ratios (with 95% confidence interval) for the afore determined primary and secondary outcomes.
Results
AZO, when compared with best available therapy (BAT) including or excluding Hydroxychloroquine, exhibited statistically insignificant reduction in mortality [(n= 27,204 patients) OR= 0.77 (95% CI: 0.51-1.16) (I2= 97%)], requirement of mechanical ventilation [(n= 14,908 patients) OR= 1.4 (95% CI: 0.58-3.35) (I2= 98%)], induction of arrhythmia [(n= 9,723 patients) OR= 1.21 (95% CI: 0.63-2.32) (I2= 92%)] and QTc prolongation (a surrogate for torsadogenic effect) [(n= 6,534 patients) OR= 0.62 (95% CI: 0.23-1.73) (I2= 96%)].
Conclusion
The meta-analysis of meta-analyses portrays AZO as a pharmacological agent that does not appear to have a comparatively superior clinical efficacy than BAT when it comes to COVID-19 management. Secondary to a very real threat of anti-bacterial resistance, it is suggested that AZO be discontinued and removed from COVID-19 management protocols.
The emergence of Coronavirus Disease 2019 (COVID-19) pandemic has led to several strategies adopted by the scientific community to contain the devastating effect of this novel virus, one of them being ‘drug repurposing’ that essentially is exploring such medications that stood the test of time for non-COVID ailments to be incorporated into the COVID-19 management protocols secondary to the anti-viral, immunomodulatory and/or anti-inflammatory properties of such pharmacological agents.
One of them happens to be Azithromycin (AZO), a macrolide that disrupts the aminoacyl-tRNA transition primarily in gram-negative bacteria and is explored for its potential effect on containing suppurate secondary bacterial pneumonia in COVID-19 patients.
Several studies attempted to establish its effect on COVID-associated mortality as well as secondary in-patient outcomes including but not limited to arrhythmic induction and QTc prolongation providing a mixed cumulative inference with some pointing in favor whereas others pointing against its efficacy and adoption.
Azithromycin for the prevention of rehospitalisation and death among Kenyan children being discharged from hospital: a double-blind, placebo-controlled, randomised controlled trial.
RECOVERY Collaborative Group Azithromycin in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial.
In order to counter and amalgamate the conflicting evidence, a meta-analysis of meta-analyses was conducted to explore the clinical efficacy of AZO for COVID-19.
Methods
Search strategy
The systematic review was carried out in accordance with the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) guidelines
with a prospective study protocol (PSP) guiding the objectives, search strategy and planned analyses developed and subsequently adopted rigorously. Being a systematic review of meta-analyses, ethics approval was exempted from being sought.
Three investigators (ADKY, AHB & SYA), independently, carried out the literature search through Cochrane Database of Systematic Reviews, PubMed/Medline and Epistemonikos in accordance with the said PSP specifically looking for such meta-analyses where researchers have tried to establish the efficacy of AZO vis a vis COVID-19 management — by critically appraising and adding together the evidence from primary research studies including randomized controlled trials (RCTs) and prospective as well as retrospective cohort studies. The reference lists of the selected, eligible meta-analyses were searched for relevant publications followed by manual search of those publications over Epistemonikos to identify if their evidence has been incorporated in such a meta-analysis that could not be identified during the primary search.
The primary outcome, as determined prior to implementing the search strategy, was reduction in mortality. Secondary outcomes included requirement of mechanical ventilation and torsadogenic effect.
The search strategy included MeSH terms "COVID-19", "Azithromycin", "Standard of Care" and "Respiration, Artificial" and is concisely summarized in the ‘Supplementary Material’.
Study selection
Three investigators (ADKY, SMMA & HFK) independently, reviewed the cohort of eligible articles with independent consideration of the abstracts and full-texts, as required, followed by undertaking discussions to reach a consensus in those cases where disagreements arose. The updated PRISMA flow diagram (Figure 1) represents the study selection process transparently.
Figure 1PRISMA flow-chart illustrating search, selection and inclusion of meta-analyses.
Manual extraction of the required data, in accordance with the pre-determined “Characteristics of studies” table, was carried out independently by three investigators (AHB, SZ & AUK). The extracted variables included publication year with the range of years of included studies, number of primary studies with the number of Included RCTs and respective meta-analysis numerical results. Three investigators (SYA, MK & NA), independently, authenticated the developed “Characteristics of studies” table.
Statistical analysis
Three investigators (ADKY, AHB & SYA), independently, adopted the Cochrane Review Manager (v. 5.4) to develop Mantel-Haenszel Random-effects models in order to calculate summarized pool Odds ratio (with 95% confidence intervals) for the primary (reduction in mortality) and secondary outcomes (requirement of mechanical ventilation, induction of arrhythmia & QTc prolongation).
I2 statistic was interpreted to comment on the heterogeneity with the conduction of sensitivity analysis explored to inspect heterogeneity.
Quality assessment and risk of bias
Three investigators (ADKY, AKK & RM), independently, adopted the Quality of Reporting of Meta-analyses (QUOROM) checklist to establish the degree of thorough reporting.
Moreover, three investigators (AHB, OA & AF), independently, adopted the Assessment of Multiple Systematic Reviews (AMSTAR) methodology to assess the methodological quality of the included meta-analyses.
A comprehensive search carried out through PubMed/Medline, Cochrane Database of Systematic Reviews and Epistemonikos yielded 80 publications with no publications identified via manual search. After removing duplicates, the abstracts and, where required, full-texts of 71 publications were considered in accordance with the aforementioned inclusion eligibility criteria. Only 2 meta-analyses satisfied the inclusion eligibility criteria and were, therefore, included in the review. The underlying causes of exclusion of publications are concisely indicated in the relevant component of the PRISMA flow diagram (Figure 1).
The characteristics of the included meta-analyses are demarcated accordingly. (Table 1 & Table S1)
AZO is termed “well-tolerated” and “safe to use” but is reported not to have a better clinical outcome for COVID-19 patients when compared with the best available therapy.
Kamel AM et al. discouraged incorporation of AZO as a component of COVID-19 management protocol and recommended its discontinuation after undergoing a thorough meta-analysis by considering only RCTs to assess the efficacy of AZO against the best available therapy (BAT) vis a vis factors including but not limited to requirement for invasive mechanical ventilation and length of hospital stay.
Mangkuliguna G et al. termed AZO “well-tolerated” and “safe to use” but reported them not having a better clinical outcome for COVID-19 patients when compared with BAT. Their conclusions were driven by the findings of an in-depth meta-analysis carried out by critically appreciating the findings from a group of studies including RCTs that compared AZO with BAT vis a vis ICU transfers and ECG changes.
The QUOROM analysis, carried out to assess the degree of comprehensiveness of reporting in the included meta-analyses, and AMSTAR analysis, carried out to explore internal validity and methodological quality (Table S2), indicated Kamel AM et al. falling short of exploring the publication bias using standard analysi(e)s. Mangkuliguna G et al. was found to have neither commented on the handling of missing data nor highlighted the data required to calculate effect sizes and confidence intervals in intention-to-treat analyses. Both of the included meta-analyses did not provide the list of excluded studies. That being indicated, both meta-analyses were mutually agreed to be of sufficiently high quality vis a vis the degree of comprehensiveness of reporting as well as internal validity and methodological quality. A substantial overlap at 20% between the primary studies considered in the included meta-analyses was recognized on CCA analysis. (Table S3) Since only 2 meta-analyses met the inclusion criteria and were subsequently incorporated in the meta-analysis of meta-analyses, sensitivity analysis was not carried out as the threshold for studies to carry out such an exploratory analysis was not met.
Our meta-analysis of meta-analyses reported statistically insignificant reduction in mortality [(n= 27,204 patients) OR= 0.77 (95% CI: 0.51-1.16) (I2= 97%)], requirement of mechanical ventilation [(n= 14,908 patients) OR= 1.4 (95% CI: 0.58-3.35) (I2= 98%)], induction of arrhythmia [(n= 9,723 patients) OR= 1.21 (95% CI: 0.63-2.32) (I2= 92%)] and QTc prolongation (a surrogate for torsadogenic effect) [(n= 6,534 patients) OR= 0.62 (95% CI: 0.23-1.73) (I2= 96%)] upon comparing AZO against BAT. (Figure 2)
Figure 2Pooled Odds ratio (with 95% Confidence Interval) for (a) mortality (b) requirement of mechanical ventilation (c) arrhythmic induction and (d) torsadogenic effect comparing Azithromycin to the Standard of Care.
Our research endeavor happens to be the very first attempt to undergo a comprehensive systematic review of meta-analyses that explores the clinical efficacy of AZO for COVID-19 considering the pertinent primary evidence in its totality, except for a recently published RCT. Oldenburg CE et al. concluded that AZO does not introduce a greater likelihood in an outpatient to become relieved of symptoms of COVID-19 at day 14 when compared with placebo.
It was not considered because of the fact that we included only meta-analyses and not individual RCTs.
A thorough meta-analysis of meta-analyses was carried out that assessed and amalgamated the relevant literature in its entirety. It would be correct to ascertain that AZO, when compared against BAT that may or may not contain Hydroxychloroquine, decreases neither the requirement of mechanical ventilation nor mortality among COVID-19 patients. That being affirmed, AZO was also found not to induce arrhythmia or torsadogenicity to a statistically significant extent. Obliquely, similar findings pointing towards invalid adoption of such pharmacological agents that do not exhibit statistically significant betterment of patient outcomes when observed through the lens of robust RCTs further cement our data-driven deductions.
Concordance between the results of randomized and non-randomized interventional clinical trials assessing the efficacy of drugs for COVID-19: a cross-sectional study.
No meta-analysis of meta-analyses has been undertaken before to explore the under-consideration research question which furthers the novelty of this systematic review as well as the universal implications, both explicitly and implicitly indicated, heralded by the findings.
That being indicated, one of the limitations of our study lies in a considerably high heterogeneity. After recognizing that for all of the analyzed outcomes, the majority of heterogeneity — the observed variance in the effect size — arises from true variance in the effect size secondary to inter-population variabilities that can be attributed to several underlying factors and is conceded as a limitation, those currently unknown factors are suggested as a direction of future research. Such suggested endeavors shall allow us to identify those specific predictors which portend a patient to relatively negative outcomes of COVID-19 — a step towards optimized, individualized medicine.
Moreover, our systematic review could not consider literature disseminated in languages other than English as well as that dispersed via those scientific journals that are not indexed in the searched databases upon which we implemented the search strategy. However, we are confident that such “grey literature” shall be constituting no more than a minuscule proportion of the text that considers the under-consideration research question. In addition to that, as recognized in the supplementary material, there happens to be a significant overlap of primary studies considered by the included meta-analyses that might have introduced a confounding effect. We recognize that as an inherent limitation to the study design adopted and suggest that exploration be undertaken to develop such methodological adjuncts that would minimize the ramifications of this limitation.
The meta-analysis of meta-analyses portrays AZO as such a pharmacological agent that does not appear to have a comparatively superior clinical efficacy when it comes to COVID-19 management upon it being compared with BAT. On the flip side, it was not found to be causing arrhythmic induction or instituting a torsadogenic effect by prolonging QTc, either. However, secondary to a very real threat of anti-bacterial resistance
, it is suggested that AZO be discontinued and removed from COVID-19 management protocols.
Supplementary data
The study protocol along with tables S1 to S3 constitute the ‘Supplementary Material’.
Declaration of Competing Interest
The authors have no conflict of interest.
Quick Look
Current Knowledge: Azithromycin (AZO) is adopted as a component of COVID-19 medical management protocol but its clinical efficacy is equivocal for COVID-19 management with several reported adverse events including QTc prolongation.
What This Paper Contributes To Our Knowledge
This study is a comprehensive systematic review of meta-analyses that assess the current literature around the clinical efficacy of AZO for COVID-19 management exploring its effectiveness on containing COVID-19 associated mortality and requirement of mechanical ventilation. AZO-associated adverse events such as QTc prolongation have also been assessed to critically appraise its rationale for being a component of COVID-19 management protocols and conclusively suggests dropping AZO for COVID-19 management protocols on the grounds of no added benefit when compared against the best available therapy and incurred undeniable threat of antibiotic resistance.
Azithromycin for the prevention of rehospitalisation and death among Kenyan children being discharged from hospital: a double-blind, placebo-controlled, randomised controlled trial.
Concordance between the results of randomized and non-randomized interventional clinical trials assessing the efficacy of drugs for COVID-19: a cross-sectional study.
00069-9&id=gr1.jpg){kind=link}
00069-9&id=gr2.jpg){kind=link}