Effectiveness of Ivermectin/Doxycycline combination in COVID-19: a systematic review and meta-analysis

This systematic review and meta-analysis aimed to assess the ecacy of the Ivermectin/Doxycycline combination for the treatment of coronavirus disease 2019 (COVID-19). We to August 26, 2021 for relevant studies. We included studies reporting at least one of the outcomes of interest: all-cause mortality; time to clinical recovery; hospital stay and viral clearance. The logarithm of risk ratios or mean differences and their corresponding standard errors for each outcome were pooled using a random-effects model. The risk of bias was assessed using the Cochrane Collaboration's tool for randomized clinical trials and the Newcastle-Ottawa Scale for cohort studies. I 2 = 91.48%). Based on low-quality evidence, this meta-analysis showed that Ivermectin/Doxycycline combination is accompanied with shorter time of clinical recovery in COVID-19 patients. However, it did not reduce all-cause mortality, viral clearance, and hospital stay signicantly. Not only the number of the studies are limited but also they ranked methodologically medium to low with limited participants. To assess the exact effective dose and ecacy of this combination therapy, high-quality and large-scale randomized clinical trials are needed.


Introduction
Since December 2019, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) causing Coronavirus disease 2019 (COVID- 19) has become a global health issue (1). About 19% of involved patients undergo progressive worsening lead to pneumonia and its complications; however, most of the involved cases recover after a period of illness (2,3).
Ivermectin is an antiparasitic drug that has been proposed as a COVID-19 therapeutic agent due to its antiviral effects (20,21). In-vitro studies have demonstrated that Ivermectin can dramatically decrease viral replication (9,22). However, clinical trials studying the clinical outcome of Ivermectin showed controversial ndings (23)(24)(25)(26).Recent efforts have revealed that a combination of Ivermectin, with broad-spectrum antiviral effects, and Doxycycline, a Tetracycline antibiotic with anti-in ammatory properties, may be bene cial and promising. Some clinical trial studies assessed the e cacy of the combination of Ivermectin/Doxycycline (IVE/DOXY) in COVID-19 treatment (27)(28)(29). Because of the scarcity of randomized controlled trials (RCTs) and inconclusive observational studies, reliable data to further shed light on the bene ts and harms is needed. To the best of our knowledge, no meta-analysis has been conducted to summarize the ndings of these studies and prepare a better insight. Therefore, we aim to review systematically the previous studies and perform a meta-analysis to measure the effectiveness of the combination of IVE/DOXY in COVID-19 treatment.

Search strategy
A systematic search of the published or unpublished studies earlier than August 26, 2021 was conducted using keywords ("COVID-19" OR "SARS-CoV-2" OR "Coronavirus" "2019 nCoV" OR "SARS CoV 2 " OR "Severe Acute Respiratory Syndrome Coronavirus 2" )AND ("Ivermectin" OR "Doxycycline" OR "Stromectol" OR "Mectizan" OR "Eqvalan "OR "Ivomec" OR "Vibramycin" OR "Atridox" OR "Doryx" OR "Hydramycin" OR "Oracea" OR "Periostat" OR "Vibravenos") in PubMed, Web of Science, Scopus, ClinicalTrials.gov, and Google Scholar databases by 2 independent investigators (M.O. and A.A.) to identify related articles. All keywords were selected from the Medical Subject Headings (MeSH) database. No language and time restrictions were imposed. After removing duplicates, a manual search was also performed to identify probable lost articles in electronic searches.
This study was registered on www.crd.york.ac.uk/Prospero (registration number: CRD42021272400) (30) and was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) (31).

Inclusion criteria:
Studies comparing the clinical e cacy of the IVE/DOXY combination and its comparators for the treatment of COVID-19 and explicitly reporting at least one of the outcomes of interest, namely all-cause mortality, time to clinical recovery, hospital stay, and viral clearance (time to negative PCR), were included in our study. Considering a speci c control arm by the articles was not a necessity to include.

Exclusion criteria
The exclusion criteria were as follows: I) case reports; II) single-arm studies; III) studies that did not report outcomes for IVE/DOXY in COVID-19; IV) in-vitro studies; V) review articles.

Data extraction
Studies were screened by 2 independent reviewers (M.O. and A.A.), examining titles and abstracts against prede ned eligibility criteria. After selecting the group of potentially inclusive studies for analysis, their full texts were examined to apply the eligibility criteria. Any discrepancies were resolved by discussion between the two reviewers, or with a third person (A.S.-A.) if consensus could not be reached. Studies that have eligibility criteria were taken for data extraction.
The following information was extracted by 2 independent reviewers (M.O. and A.A.): author, year of publication, study design, country, patients' characteristics, the regimen of IVE/DOXY and comparative agent, the primary outcome (allcause mortality), secondary outcome (time to clinical recovery, hospital stay, and viral clearance or time to negative PCR). If studies did not report the standard deviation (SD) changes from baseline, we calculated standard error (SE) and then converted them into SDs according to the formula provided in the Cochrane Handbook of Systematic Reviews (32).

The assessment of the risk of bias in individual studies
The risk of bias of all included studies was assessed by 2 reviewers (M.O. and A.A.) using the Cochrane Collaboration's tool for randomized clinical trials and the Newcastle-Ottawa Scale (NOS) for cohort studies (32)(33)(34). The Cochrane Collaboration's tool was assessed based on the following sources of bias: 1) random sequence generation, 2) allocation concealment, 3) blinding of participants or personnel, 4) blinding of outcome assessment, 5) incomplete outcome data, 6) selective outcome reporting, and 6) other biases. Studies were categorized as low, high, or unclear risk for each source of biases. Studies that had a low risk of bias for all domains were regarded as good quality; studies had fair quality if one criterion was high risk or 2 criteria were unclear, and if 2 or more of the items were high risk or unclear risk of bias, the studies were listed as poor quality.

Statistical analysis
Comprehensive Meta-Analysis (CMA) software version 2 was used for the meta-analysis. The random-effects model, which considers the between-study variations, was used to derive the summary estimates. For continuous outcomes (time to clinical recovery, hospital stay, and time to negative PCR), we calculated mean differences and their corresponding SEs to be used as effect size for meta-analysis. For dichotomous outcomes (all-cause mortality), we calculated the logarithm of risk ratios and their corresponding SEs. Statistical heterogeneity among included studies was assessed using the I-squared (I 2 ) statistic. Heterogeneity was considered signi cant if the I 2 statistic was more than 25%. Sensitivity analysis was done to explore the extent to which inferences might depend on a particular study. P values of less than 0.05 were considered statistically signi cant.

Search
The primary search was conducted up to August 26, 2021. Figure 1 shows the ow diagram of the included studies in detail. The search strategy initially yielded 1614 studies; after excluding 619 duplicate articles, 995 articles were screened. After examining titles and abstracts, 26 articles were identi ed for full-text review for eligibility. Twenty-one articles were excluded after full-text review according to the exclusion criteria. One study was excluded because of the outcome, not of interest; it reported the number of patients who got negative PCR on days 5, 6 in the IVE/DOXY group and on days 11, 12 in the control group (35). Finally, 4 randomized clinical trials (27,29,36,37) and 1 prospective study (38) ful lled our criteria for inclusion in the systematic review and meta-analysis. The studies included in this metaanalysis are summarized in Table 1. This meta-analysis involved a total of 789 patients, including 399 in the IVE/DOXY group and 390 in the control group. Three studies were conducted in Bangladesh (27,29,36). Two studies were performed in Iraq (37) and India (38).

Risk of bias
The assessed risk of bias using NOS for one prospective study was 4, indicated a high risk of bias (38). According to the Cochrane Collaboration's tool for randomized clinical trials, 2 trials were classi ed as fair (27, 36) and 2 had poor quality (29, 37) (Fig. 2).
Common biases were related to the randomization process and allocation concealment; in fact, two trials randomized patients based on odd/even date of enrollment (37) or registration number (29). The details of random sequence generation and allocation concealment were not reported in one study (27). In addition, two studies did not report any information on the process of blinding (29, 37). Attrition bias was also present in one article (36) because about 10% of randomized patients were not included in the analysis. All the studies reported outcomes and none of them had selective reporting bias. Other biases were not observed in any of the included studies.

All-cause mortality
Among all studies, mortality was observed in only two trials (36, 37). In the rest of the studies, all-cause mortality was reported to be zero in both the IVE/DOXY and control groups. A total of 503 participants were included. The all-cause mortality rate of patients with COVID-19 in the IVE/DOXY group was 0.79% (2/253), which was lower than in the control group (3.6%; 9/250). However, the difference was not statistically signi cant (Log risk ratio=-1.288; 95% CI:-2.671, 0.096; P = 0.068; I 2 = 0%) (Fig. 3).

Time to clinical recovery
The meta-analysis of four studies (29, 36-38) which reported the mean number of days to clinical recovery (n = 741 participants) showed a signi cant reduction in time to clinical recovery after IVE/DOXY (Difference in means =-2.427 days; 95% CI:-4.033, -0.820; P = 0.003, I 2 = 91.475%) (Fig. 4). Sensitivity analysis after deleting every individual study successively revealed the same ndings. The heterogeneity was tangibly reduced after removing a study done by

Viral clearance
Two studies assessed the mean number of days to negative PCR or viral clearance outcome (164 participants) (27,38).
The IVE/DOXY group had a shorter period of time to negative PCR than the control group. The difference, however, was not statistically signi cant (Difference in means =-0.768 days; 95% CI:-1.550, 0.013; P = 0.054, I 2 = 91.48%) (Fig. 4) Doxycycline is an antibiotic which can be used to treat atypical bacterial pneumonia and community-acquired pneumonia (39). In mammalian cells, doxycycline has an anti-in ammatory action that is mediated by chelating zinc compounds on matrix metalloproteinases (MMPs) (40). Murine coronaviruses rely on MMPs for cell fusion and viral multiplication, according to prior in vitro research (41). COVID-19's pathologic characteristics are similar to those of previous SARS-CoV infections, in which MMPs play a key role in disease development (42). Doxycycline's pharmacokinetics indicate that it is dispersed in pulmonary tissue following oral administration, with the drug's concentration in the lungs being 18-23% of the serum concentration in humans (43). As a result, doxycycline might be useful in the treatment of COVID-19 infection.
Ivermectin is a well-known antiparasitic medication that is used to treat a variety of parasites all over the world. A metaanalysis demonstrated that ivermectin can not only reduce the number of COVID-19 deaths, but also lead to signi cant reduction in the number of progressive disease cases if started in early clinical course (8). In preclinical models of numerous additional diseases, ivermectin has exhibited immunomodulatory and anti-in ammatory functions. Ivermectin has been shown to inhibit the synthesis of in ammatory mediators such as nitric oxide and prostaglandin E2 in vitro studies (44). In mice given a fatal dosage of lipopolysaccharide, Ivermectin decreased TNF-α, IL-1, and IL-6, and increased survival (45). Subcutaneous ivermectin reduced the IL-6/IL-10 ratio in the lung tissues of Syrian golden hamsters infected with SARS-CoV-2 and prevented clinical deterioration (46). Ivermectin's antiviral activity against SARS-CoV-2 in Vero/hSLAM cells has been proven. However, after oral administration of the medication to patients, the concentrations necessary to suppress viral multiplication in vitro (EC 50 = 2.8µM; EC 90 = 4.4µM) are not attained systemically (22,47). Although the standard dose of a single 200µg/kg oral of ivermectin for treating strongyloidiasis is thought to accumulate in lung tissues (2.67 times that in plasma), there is no evidence that this dosage regimen would result in ivermectin reaching an antiviral concentration in the lungs (48, 49). As a result, alternative administration methods for ivermectin in COVID-19 should be considered, such as utilizing in combination with medicines that increase ivermectin activity through mechanisms such as increasing ivermectin pulmonary penetration (50). Therefore, we aim to perform a systematic review and meta-analysis to determine the e cacy of the IVE/DOXY combination in COVID-19 therapy.
Based on a meta-analysis of ve related studies, the combination of IVE/DOXY signi cantly reduced the time to clinical recovery. Although the mortality rate was lower in the IVE/DOXY group than in the control group, there was no signi cant difference. The IVE/DOXY group had a shorter period of time to a negative PCR than the control group. The difference, however, was not statistically signi cant. Co-administration of ivermectin and doxycycline did not differ in the length of hospital stay.
These ndings, however, should be taken with caution. It must be considered that the low mortality rate which was reported in just two studies is due to that majority of enrolled patients were inclined to mild to moderate. On the other hand, in the Hashim et al. (37)  Although the methodological quality of all studies was medium to low, study's limitations must be taken into account.
There were few trials and limited patients, and adverse events were not reported in the majority of the studies. Thus, Page Table   Table 1 is available in the Supplementary Files section. Figure 1 The owchart of search strategy and study selection.

Figure 2
Study quality and risk of bias assessment using Cochrane collaboration tool.

Figure 3
Page 13/14 Forest plot of studies reported all-cause mortality.

Figure 4
Forest plot of studies that assessed the mean number of days to clinical recovery (time to clinical recovery).

Figure 5
Forest plot of studies that assessed the mean number of days to negative PCR (viral clearance). Figure 6