Abstract
Background Clarithromycin clinical efficacy has not been described in COVID-19.
Research question Is oral clarithromycin beneficial for treating patients diagnosed with COVID-19?
Study and methods An open-label non-randomized trial in 90 patients with COVID-19 of moderate severity was conducted at four study sites in Greece between May and October 2020. Ninety participants with respiratory tract infections received clarithromycin 500 mg every 12 hours for 7 days; another 90 standard-of-care (SOC) propensity score-matched concurrent controls received azithromycin plus hydroxychloroquine. The composite primary endpoint was defined for patients with upper respiratory tract infection as: (a) no need for hospital re-admission or (b) lack of progression into lower respiratory tract infection and, for patients with lower respiratory tract infection, as at least 50% decrease of the score of respiratory symptoms at the end-of-treatment (EOT) without progression into severe respiratory failure (SRF). The incidence SRF at the test-of-cure (TOC) on day 14 was a secondary endpoint. For clarithromycin-treated patients, viral load of SARS-CoV-2, biomarkers, the function of mononuclear cells, and safety were assessed; biomarkers were also measured in SOC comparators.
Results The primary endpoint was attained in 86.7% of patients treated with clarithromycin (95% CIs 78.1-92.2%) and 73.3% of concurrent SOC comparators (95%CIs; 63.4-81.4%). The odds ratio for the primary endpoint with clarithromycin treatment in univariate analysis was 2.36 (95%CIs 1.09-5.08; P: 0.039). Results were confirmed after multivariate stepwise logistic regression analysis (odds ratio 3.30; 95% CI 1.10-9.87; P: 0.033). At the TOC visit, the incidence of SRF was 12.2% (n = 11 ; 95%CIs 6.9-20.6%) among patients treated with clarithromycin (odds ratio for SRF 0.38; 95%CIs 0.17-0.84) versus 26.7% (n= 24; 95%CIs 18.6-36.6%) among concurrent SOC comparators (P: 0.023). Clarithromycin use was associated with decreases in circulating levels of C-reactive protein, of tumour necrosis factor-alpha and of interleukin (IL)-6; by an increase of the ratio of Th1 to Th2 mononuclear responses; and by suppression of SARS-CoV-2 relative viral load. No safety concerns were reported. Patients starting clarithromycin with the first five days from symptoms onset achieved better responses.
Interpretation Clarithromycin treatment is associated with early clinical improvement in patients with moderate COVID-19. Modulation of the Th1/Th2 responses is proposed as the mechanism of action.
Trial Registration ClinicalTrials.gov, NCT04398004
INTRODUCTION
The progression of pneumonia by the novel coronavirus SARS-CoV-2 (also known as COVID-19) is unpredictable. Some patients deteriorate into severe respiratory failure (SRF) through a process of complex immune dysregulation which is often accompanied by unfavourable outcome1. From the very beginning of the COVID-19 pandemic azithromycin, a macrolide drug was introduced in the treatment algorithm due to its anti-inflammatory properties in the attempt to attenuate the hyper-inflammatory reaction of the host and prevent the development of SRF2. Azithromycin was given in combination with hydroxychloroquine based on the results of an open-label trial showing better viral containment when both drugs were used in combination.
The anti-inflammatory properties of the macrolide class of antibiotics are reflected in the guidelines by the American Thoracic Society in 2019 clearly arguing in favour of the combination of β-lactam antibiotics with macrolides for the empirical management of community-acquired pneumonia (CAP)3. These guidelines are influenced by observational studies showing that the addition of a macrolide decreases the risk of death from severe CAP4-8.
Two double-blind, randomized, placebo-controlled clinical trials of our group have reported that intravenous administration of clarithromycin is associated with substantial reduction of the risk of death among critically ill patients with Gram-negative sepsis9, 10. The first trial was conducted among patients with sepsis developing after ventilator-associated pneumonia; survival by day 90 was 40% in the placebo group and 57% in the clarithromycin treatment group11. We have also recently shown that 28-day mortality of severe CAP was 20.8% among patients treated with a combination of β-lactam and clarithromycin and 33.8% when treated with a combination of β-lactam and azithromycin12.
Based on the above evidence, it seems likely that treatment of COVID-19 with oral clarithromycin may attenuate the hyper-inflammatory responses of the host and decrease the risk of progression into SRF. We conducted the ACHIEVE trial (Anti-inflammatory Clarithromycin to Improve SARS-CoV 2 Infection Early) to investigate whether early administration of oral clarithromycin in patients with upper or lower respiratory tract infection (RTI) due to SARS-CoV-2, may attenuate the inflammatory burden, modulate the immune response of the host and result in early clinical improvement. We compared the efficacy of clarithromycin with a group of propensity-score matched concurrent comparators receiving standard-of-care treatment (SOC) with azithromycin and hydroxychloroquine. The efficacy of clarithromycin was also studied in relation to the delay on onset since start of symptoms and to the lineage of the infecting strain of SARS-CoV-2.
METHODS
Trial oversight
ACHIEVE is an open-label non-randomized trial conducted in four study sites in Greece (EudraCT number 2020-001882-36; National Ethics Committee approval 45/20; National Organization for Medicines approval ISO 36/20; ClinicalTrials.gov registration NCT04398004). The study sites were medical departments in tertiary hospitals outside the territory of Athens. Concurrent comparators receiving azithromycin plus hydroxychloroquine/chloroquine were hospitalized at the same time period in five other medical departments in tertiary hospitals of Athens. Those five departments participate in the registry of the Hellenic Sepsis Study Group (https://hssg.sepsis.gr) but they did not participate in the ACHIEVE trial. The trial was sponsored by the Hellenic Institute for the Study of Sepsis (HISS) under the financial support of Abbott Product Operations AG. Funding bodies had no role in the design, conduct, analysis and interpretation of data, and decision to publish. The laboratory of Immunology of Infectious Diseases of the 4th Department of Internal Medicine at ATTIKON University General Hospital served as a central laboratory for the study. The initial draft of the manuscript was written by the first and the last author. All authors vouch for the adherence of the trial to the protocol and first and last author vouch for the accuracy and completeness of the data and analysis.
Patients
Enrolled patients were adults (age ≥18 years) with confirmed infection by SARS-CoV-2 virus by real-time PCR reaction of nasopharyngeal secretions; and infection of the upper respiratory tract (URTI) or of the lower respiratory tract (LRTI). The URTI was defined as the acute presentation of at least two of the following signs: a) core temperature ≥ 37.5°C; b) new onset of cough; c) chills or rigor; and d) total absolute lymphocyte count less than 1,500/mm3. The LRTI was defined as the presence of infiltrates compatible with lower respiratory tract infections in chest X-ray or in chest computed tomography accompanied by at least one of the following: a) new onset of cough or worsening cough; b) dyspnea; c) respiratory rales compatible with lung infection; and d) total absolute lymphocyte count less than 1,500/mm3. Exclusion criteria were: age below 18 years; intake of any other macrolide and of hydroxychloroquine for the current episode of infection under study; ratio of partial oxygen pressure to the fraction of inspired oxygen (pO2/FiO2) less than 150; need for mechanical ventilation (MV) or non-invasive ventilation under positive pressure (NIV); neutropenia (<1,000/mm3); any intake of corticosteroids at a daily dose ≥ 0.4mg/kg prednisone or equivalent the last 15 days; presence of any contraindication for one macrolide; QTc interval at rest electrocardiogram ≥500 msec; history of known congenital long QTc syndrome; and pregnancy or lactation.
In order to derive the most appropriate ones among concurrent comparators treated at the same time period with azithromycin and hydroxychloroquine, the inclusion and exclusion criteria of the ACHIEVE trial were applied to all available concurrent comparators. Patients meeting the inclusion criteria and not meeting the exclusion criteria were eligible for inclusion in the comparator SOC group. Patients who were administered biological disease modifying agents (bDMARDs) because of severe COVID-19 were excluded from participation in the SOC group. Written informed consent was provided by the patient or legal representative before screening.
Trial interventions
Enrolled patients received one tablet of 500 mg of clarithromycin every 12 hours for 7 days. All other drugs except macrolides and hydroxychloroquine/ chloroquine phosphate were allowed. Study visits were done daily starting from day 1 before start of treatment until day 8 after end of treatment (end-of treatment visit, EOT). The last test-of-cure (TOC) visit was done on day 14. For patients discharged earlier from hospital, EOT and TOC visits were done by phone calls. Recorded baseline information was severity assessed by APACHE II score, SOFA score and pneumonia severity index (PSI), comorbidities and laboratory information. On each visit follow-up, the state of the patient was assessed and the respiratory symptoms score (RSS) was calculated (see Supplementary methods). Fifteen ml of whole blood was collected on day 1 before start of treatment and on EOT into EDTA-coated tubes and sterile and pyrogen-free tubes for the isolation of peripheral blood mononuclear cells (PBMCs), serum and plasma. PBMCs were stimulated for cytokine production. Biomarkers and cytokines were measured in plasma, serum and supernatants of PBMC cultures (see Supplementary methods).
Nasopharyngeal swabs were collected from SARS-CoV-2 positive patients using the 350C UTM© mini 1ml collection and transport system (COPAN, Italia, S.p.A.) on day 1 before start of treatment, on day 4 and on EOT. The collected samples were kept at −80°C until further processing. The viral load was expressed as the Ct of the reaction positivity and as the relative log2 copies of the E gene and of the RdRp gene. Next generation sequencing was done for viral isolates (see Supplementary Methods). The genome sequences of SARS-CoV-2 were assigned to viral lineages according to the dynamic nomenclature system proposed by Rambaut et al13. To this effect, the FASTA files generated by the SOPHiA DDM tool were uploaded to the pangolin 2.0 web application (https://pangolin.cog-uk.io/).
Blood was also sampled from concurrent SOC comparators on day 1 before start of treatment and on day 8 for the measurement of biomarkers.
Outcomes
The primary study endpoint was described on EOT day 8. This was a composite endpoint defined differently for patients with URTI and for patients with LRTI. For patients with URTI the primary endpoint was defined as either no need for hospital re-admission in case of earlier discharge or as lack of progression into lower RTI. For patients with LRTI the primary endpoint was defined as any least 50% decrease of the RSS from the baseline score of day 1 provided that the patient has not progressed into SRF or had died by EOT. SRF was defined as any decrease of the pO2/FiO2 below 150 mmHg necessitating MV or NIV. The same endpoints were captured for the SOC concurrent comparators. Secondary outcomes were the development of SRF by TOC visit; hospital readmission until TOC; the change of viral load from baseline in respiratory secretions on day 4 and EOT; the change of circulating biomarkers; and the changes of cytokine responses from monocytes and from Th1 and Th2 lymphocytes.
Study exploratory endpoints were the impact of time delay from the onset of COVID-19 symptoms on the achievement of the primary endpoint; and the association with the lineage of SARS-CoV-2.
The study endpoints were defined for concurrent SOC comparators using the same definitions as for patients treated with oral clarithromycin in the ACHIEVE study.
Adverse Events (AE) (Common Terminology Criteria for Adverse Events, version 4.03) and Severe Adverse Events (SAE) (see Supplementary Appendix) were captured.
Statistical analysis
Qualitative data were presented as percentages with confidence intervals (CI) and quantitative data as means and SD; serum biomarkers and cytokines were expressed as means and SD. After applying the inclusion and exclusion criteria of the ACHIEVE study, the available concurrent SOC comparators treated with azithromycin and hydroxychloroquine were selected. Propensity score 1:1 matching was applied to finally select the best matched comparators to the 90 patients treated with clarithromycin. Matching variables were: age, Charlson’s comorbidity index (CCI), admission PSI; the frequency of URTI and LRTI; and admission C-reactive protein. These matching variables were selected because they were all covariates measured before treatment and they are related to the treatment and outcome.
Comparisons between groups for the primary endpoint were performed by the Fisher’s exact test using confirmatory forward stepwise logistic regression analysis (IBM SPSS Statistics v. 25.0). Paired comparisons were done by the Wilcoxon’s signed rank test. Using the median of the time interval between the start of COVID-19 symptoms and start of clarithromycin, patients were divided into those with an early start and into those with a late start of clarithromycin. The achievement of the primary endpoint was compared between patients with early and late start by the binomial test. Any two-sided P value <0.05 was statistically significant.
RESULTS
Trial conduct
Patients participating in the ACHIEVE open-label trial were enrolled between May 17 2020 and October 8 2020. Patients participating in the SOC comparator arm were enrolled within the same time frame. After propensity matching, 90 concurrent SOC comparators were selected. The propensity score of patients enrolled in the ACHIEVE trial was 0.357 ± 0.094; the respective score of the concurrent SOC comparators was 0.360 ± 0.099 (P: 0.829). The study flow chart is shown in eFigure 1. None of the 180 participants received bDMARDs. The administered dose of azithromycin was 500mg daily for seven days and of hydroxychloquine 500 mg every 12 hours for seven days. Baseline demographics of patients receiving clarithromycin and of concurrent comparators are shown in eTable 1. No difference was found between the two groups, although there was a trend for patients enrolled in the ACHIEVE trial to present with a higher CCI.
Study endpoints
The primary endpoint was met in 78 patients treated with clarithromycin (86.7% 95% CIs 78.1-92.2%). Sixty-six patients among the concurrent SOC comparators met this primary endpoint (73.3%; 95%CIs; 63.4-81.4%). In univariate analysis the OR for positive response with clarithromycin treatment was 2.36 (95%CIs 1.09-5.08; P: 0.039) versus concurrent SOC comparators. Nine other variables showed an association with the endpoint in univariate logistic regression. However, in a multivariate step-wise logistic regression analysis, only three factors retained a robust independent association with the endpoint (eTable 2), including assignment to clarithromycin treatment (odds ratio 3.30; 95% CIs 1.10-9.87; P: 0.033). For the two other factors - SOFA score and PSI score - the analysis showed that higher values were associated with lower number of positive responders (eTable 2).
The changes of the RSS score over time among patients treated with clarithromycin are shown in supplementary Figure 1. Comparisons with the baseline values before start of treatment demonstrate early positive response with clarithromycin treatment, as this is reflected in the significant decrease of RSS within the first 24 hours.
At the TOC visit, the incidence of SRF was 26.7% (n= 24; 95%CIs 18.6-36.6%) among SOC concurrent comparators and 12.2% (n= 11; 95%CIs 6.9-20.6%) among patients treated with clarithromycin (odds ratio for SRF 0.38; 95%CIs 0.18-0.84; P: 0.023). Multivariate logistic regression analysis revealed that the only independent variable associated with protection from SRF at TOC was the intake of clarithromycin (odds ratio 0.22; 95% CIs 0.06-0.79; P: 0.022; supplementary Table 1). The same analysis indicated that baseline SOFA score, PSI and CRP were independent variables associated with progression into SRF at TOC visit.
The change of baseline (before start of the study drug) viral load on day 4 of treatment with clarithromycin and at the EOT visit with clarithromycin are shown in eFigure 2. The viral load was expressed as both the Ct of RT-PCR positivity and as relative log2 copies to the reference gene. The viral load was expressed for both the constitutive E gene and the specific RdRp gene. The time of Ct to positivity was increased over study visits so as to be considered as indirect index of the decrease of the copies of E gene and RdRp gene. The relative log2 copies of both genes were decreased over-time. NGS analysis of viral isolates did not show any single nucleotide variant (SNV) in either the E gene or in the RdRp gene.
Clarithromycin treatment was associated with attenuation of pro-inflammatory responses. On day 8 the serum concentrations of C-reactive protein (CRP), tumour necrosis factor-alpha (TNFα) and interleukin (IL)-6 were lower among patients treated with clarithromycin than among concurrent SOC comparators. No differences were found in IL-10 (eFigure 3).
The proposed mechanism of action of clarithromycin is associated with the increase of the Th1 response from circulating PBMCs compared to an attenuation of the Th2 response. More precisely, we were able to show using stimulation with heat-killed C.albicans for 5 days that the targeted production of IFNγ, which reflects the Th1 response, was increased (eFigure 4A). Under the same conditions, the targeted production of IL-6 that reflects the Th2 response was decreased (eFigure 4B). The ratio of Th1/Th2 response from PBMCs at the EOT was greater among patients with lower viral load (eFigures 4 C and 4D). On the contrary, no change of the monocyte function was found (supplementary Figure 2)
Exploratory study endpoints
Using the median of distribution, patients treated with clarithromycin were divided into those who started treatment within the first 5 days from start of symptoms (n= 48, range 1-5 day) and into those who started treatment 6 days or later from start of symptoms (n= 42, range 6-12 days). The primary endpoint was achieved in 91.7% (44 out of 48 patients; 95% CIs 80.4-96.7%) and in 81.4% (34 out of 42 patients; 95%CIs 66.7-90.0%) respectively (P: 0.021). The respective percentages for concurrent SOC comparators were 72.9% (35 out of 48 patients; 95% CIs 59.0-83.4%) and 73.8% (31 out of 42 patients; 95%CIs 58.9-84.7%) respectively (P: 1.00).
The lineage of SARS-CoV-2 viral isolate could be determined following quality control amputation in isolates coming from 49 patients. They were matched as B1.1 (n= 30, 61.2%); B1.1. with 75% matching to the parent B1.1 (n= 2, 4.1%); B1.1. with 50% matching to the parent B1.1 (n= 8, 16.3%); B1.1.2 (n= 3; 6.1%); B1 (n=2; 4.1%), B1.1.25 (n= 2; 4.1%), B1.1.10 (n=1, 2.1%); and B1.1.70 (n=1, 2.1%). Viral isolates B1.1 and B1.1 with 75% matching to the parent B1.1 were analysed together as viral isolates of the B1.1 lineage. As such, analysis included 32 viral isolates of the B1.1 lineage and 17 isolates of the non-B.1.1 lineage. Patients starting early clarithromycin had an achievement of the positive primary endpoint of treatment response similar for both types of viral lineage. However, among patients starting late clarithromycin the achievement of the positive treatment response was significantly greater among those infected by the non-B1.1 SARS-CoV-2 viral isolates (supplementary Table 2).
Safety
The adverse events (AEs) and serious adverse events (SAEs) that were captured during the study period of 14 days are listed in eTable 3. Reported events probably or possibly-related to the study drug with incidence greater than 1% were: increase of aminotransferases (Grade I 18.8%; Grade II 1.1%); diarrhea (mild 12.1%; moderate 1.1%); mild vomiting in 1.1%, Grade I increase of bilirubin in 2.22%; and mild allergic reaction in 1.1%) (eTable 3). No drug-related SAEs were reported.
DISCUSSION
Based on previous experience with clarithromycin9-12, we sought to investigate its potential in reducing the hyper-inflammatory response in moderate COVID-19 and contribute to better clinical outcomes than currently available treatment approaches. The ACHIEVE open-label trial enrolled patients with moderate severity of COVID-19, compared with propensity-matched concurrent SOC comparators hospitalized during the same time-period. Analysis revealed that clarithromycin treatment was associated with a significantly greater clinical benefit at the EOT visit of 86.7%, as compared with 73.3% achieved by the concurrent SOC comparators. This early benefit shown at the end of the 7-day course with clarithromycin was further reflected in the incidence of SRF at the TOC visit of day 14, when treatment with clarithromycin provided 78% relative decrease of the risk for SRF compared to the SOC regimen. The benefit was pronounced for patients starting clarithromycin within the first 5 days from the onset of Covid-19-related symptoms. The benefit of clarithromycin was also linked with decrease of CRP and IL-6 at the EOT.
The current study provides insight into the mechanism of action of clarithromycin. Emerging data suggest that active virus replication and persistence in the upper respiratory tissue are associated with the severity of the illness14, 15. Patients treated with clarithromycin showed remarkable reduction of the viral load in the nasopharynx which was associated with clinical improvement. At the end of treatment with clarithromycin, the Th1 cell responses were increased whereas the Th2 cell responses were attenuated. The Th1 cells are responsible for the containment of the viral replication in viral infections. Specifically, in COVID-19 the inhibition of the Th1-mediated immune responses has a critical impact on the viral replication and the severity of the illness16. Patients with lower viral load at the EOT visit had greater Th1/Th2 ratio as this was expressed by modulation of the production of IFNγ and of IL-6 by circulating lymphocytes. IFN-γ has a clear role in host defence against viruses. By producing large amounts of IFNγ, Th1 cells induce activation and M1 polarization of macrophages. IFNγ-stimulated macrophages produce IL-12, a critical cytokine that induces proliferation of Th1 cells, and support continuous production of IFNγ. Furthermore, IFNγ inhibits the proliferation of Th2 lymphocytes and of Th2-derived IL-4, IL-6 and IL-1016.
The findings of the ACHIEVE study should be interpreted in the light of certain limitations. The study was designed in mid-March 2020 at the beginning of the pandemic in Greece. It was decided to endorse an open-label and single-arm design in an attempt to include (and benefit) as many patients as possible since no SOC was framed at that time period. The azithromycin plus hydroxychloroquine concurrent comparators were optimally matched without differences in baseline severity and co-administered treatment. When the study was started, most of the patients with COVID-19 hospitalized in the Greek Hospitals were co-administered azithromycin and hydroxychloroquine as off-label treatment option. In the following months, emerging data showed that hydroxychloroquine does not affect the natural course of COVID-1917, 18. As such, the comparison between the clarithromycin group and the azithromycin plus hydroxychloroquine group should be conceived as comparison between clarithromycin and azithromycin.
In the randomized COALITION trial from Brazil, patients with moderate to severe COVID-19 were randomized to treatment with placebo (n=227), hydroxychloroquine (n=221) and a combination of hydroxychloroquine and azithromycin (n=217). The primary study endpoint was the distribution of the 7-point ordinal scale after 15 days. The achievement of the primary endpoint was similar between the three groups of treatment19. However, the design of the COALITION trial was different than the ACHIEVE trial where participants had less severe disease. Furthermore, the median time from symptoms onset to start of treatment was seven days in the COALITION trial contrary to five days in the ACHIEVE trial. The same group of investigators performed the COALITION II trial were patients were allocated to treatment with standard-of-care with (n=214) or without azithromycin (n=183). Although the two groups did not differ in the achievement of the primary endpoint after 15 days, patients allocated to azithromycin treatment had a greater chance of being at a greater point of the scale after 7 days of treatment20. Both COALITION trials evidenced the safety of azithromycin.
In conclusion, the results of the ACHIEVE trial clearly indicate that clarithromycin treatment in patients with moderate COVID-19 is associated with early clinical improvement and containment of viral load. This is associated with an increase of the ratio of Th1/Th2 response. To the best of our knowledge this is the first study to show the anti-inflammatory impact of clarithromycin on COVID-19. Further studies are needed to better define its future role in mild and moderate COVID-19.
Data Availability
Date are available from the corresponding author on request.
Funding and disclosure
The study is supported by Abbott Operation Products AG.
P. Panagopoulos has received honoraria from GILEAD Sciences, Janssen, and MSD.
G. Poulakou has received independent educational grants from Pfizer, MSD, Angelini, and Biorad.
H. Milionis reports receiving honoraria, consulting fees and non-financial support from healthcare companies, including Amgen, Angelini, Bayer, Mylan, MSD, Pfizer, and Servier.
E.J. Giamarellos-Bourboulis has received honoraria from Abbott CH, Angelini Italy, InflaRx GmbH, MSD Greece, XBiotech Inc., and B·R·A·H·M·S GmbH (Thermo Fisher Scientific); independent educational grants from AbbVie Inc, Abbott CH, Astellas Pharma Europe, AxisShield, bioMérieux Inc, Novartis, InflaRx GmbH, and XBiotech Inc; and funding from the FrameWork 7 program HemoSpec (granted to the National and Kapodistrian University of Athens), the Horizon2020 Marie-Curie Project European Sepsis Academy (granted to the National and Kapodistrian University of Athens), and the Horizon 2020 European Grant ImmunoSep (granted to the Hellenic Institute for the Study of Sepsis).
The other authors do not report any conflict of interest.
The authors would like to thank the patients, families, clinical, laboratory and research staff who contributed to the trial.