Discussion
Azithromycin presents numerous characteristics that could confer a clinical benefit in the treatment of COVID-19. Its potential antiviral, immunomodulatory and antibacterial properties could be of interest in all the three proposed stages of COVID-19. However, despite all these promising benefits, the use of azithromycin in the treatment of SARS-CoV-2 pneumonia remains unclear.
Unfortunately, studies carried out on the potential value of azithromycin have been mostly accompanied by its prescription in association with hydroxychloroquine/chloroquine. This constitutes a great limitation and hampers the assessment of the potential benefits of the macrolide, especially given the recent negative benefit-risk balance of hydroxychloroquine and chloroquine.
In other viral infections as influenza, although the available evidence is of low quality, azithromycin has proven to be useful when given early in the disease. In the early stage of COVID-19, where the use of antivirals might be indicated, azithromycin could reduce the number of complications, including the need for hospitalization. However, most studies have focused on advanced forms of the disease, so outpatients have been misrepresented.
Gautret et al . assessed the early treatment with hydroxychloroquine and azithromycin in patients with mild disease (mostly asymptomatic or with upper respiratory tract infections)[8]. This was the first study assessing the efficacy of this combination. Although a higher virological clearance was observed, this assumption should be taken with caution given the limitations. This was a non-randomized open-label clinical trial that only included 36 patients. Only 6 patients were treated with combination therapy, without adequate controls. From a total of 26 patients treated with hydroxychloroquine, 6 were lost in follow-up: 3 because were transferred to ICU, 1 died, 1 decided to leave and 1 stopped the treatment due to nausea. Patients in the treatment group had higher viral loads, so a likely benefit was easier to demonstrate[5]. Finally, baseline clinical data were lacking, and no clinical outcomes or safety data were reported. Thereafter, Gautret et al . and Million et al . showed positive results without significant safety issues when given this combination in a higher number of mild patients at day 5-6 of the symptom onset[72,73]. Again, the lack of control group, however, prevents the attribution of any benefits to this therapy.
A recent review concluded that this combination should be used in symptomatic high-risk outpatients, mainly based on the study performed by Barbosa et al, where although treated patients were sicker presented a reduction in the need of hospitalization[87]. However, this is a pre-print study open-label study and was performed by a telemedicine healthcare team, so it may not be applicable to other settings. The study of Guerin et al . seems to confirm the potential benefit of this strategy[79]. Interestingly, this study showed that azithromycin alone presented similar outcomes when compared to combination therapy[79]. Again, it was a pre-print study with a small sample size and the outcome was a subjective measure. Other limitations include that the time of treatment initiation from symptom onset was day 1 in 41 % of patients, while the rest initiated within 15 days and 1 in the azithromycin alone group in day 40. The lack of data concerning the viral load is another limitation that prevents the evaluation of the potential antiviral activity of azithromycin. Azithromycin was well tolerated in these studies with no associated cardiovascular events, suggesting that these toxicities may be more evident in sicker patients[72,73,79,80,87].
The use of azithromycin in the first stage of COVID-19 has been therefore poorly studied, with a low number of patients included and studies with many flaws. However, the available evidence suggests a potential benefit of its use alone or in combination that requires further study.
Based on the first study of Gautret et al ., combination therapy began to be used in the second stage of COVID-19[8]. After adjusting for confounding factors, this combination was associated with an increased risk of mortality and adverse events[78]. Other studies also reported a lack of benefit of the treatment with hydroxychloroquine[75,77]. Again, these findings must be interpreted with caution given the many limitations of the included studies. Some of them presented low sample sizes so were underpowered. None of the studies was a placebo controlled randomized clinical trial (all were observational studies), so they were not designed to assess the efficacy of these regimens. Despite the efforts to control for confounding factors, in observational studies even the best adjustment methods can miss major systematic biases[89]. Among confounding factors, the use of other therapies such as antivirals, immunomodulators (specially glucocorticoids) and anticoagulation therapy were not either described or adequately controlled. This is of upmost importance given recent evidence showing clinical benefits with the use of remdesivir, corticoids or anticoagulation therapy[4,90,91]. Another important issue is that azithromycin was given alone, when reported, in 29-37 % of patients in the control groups. Given the potential benefits associated with this macrolide, this may have also been a potential confounding factor.
All these studies have evaluated robust and objective clinical outcomes as in-hospital mortality or need for intubation. However, other outcomes such as time to clinical recovery, time to symptom resolution or length of stay were not analyzed and could offer another vision of the treatment.
The time from the onset of symptoms until the initiation of treatment is another important issue. Only one study reported these data and treatments were initiated late (8 days)[75]. This could have underestimated combination treatment efficacy as was not initiated when it should be more active. One might think that by this time azithromycin should show clinical benefits since there might already be some hyperinflammation. However, patients included in these studies presented mild disease with a low incidence of comorbidities, which prevents demonstrating the potential benefit of azithromycin in this setting.
Three studies analyzed the effect of this macrolide alone. Geleris etal . did not find any clinical benefit with the use of this macrolide[77]. However, they did not demonstrate any clinical benefit either with remdesivir, which has recently shown significant clinical benefits in a randomized controlled trial[4]. This fact raises concerns about the conclusions of this study, given that both azithromycin and remdesivir were assessed as potential covariates without showing specific data of patients that received them[77]. In the study of Rosenberg et al . reporting data on the sickest patients to date, this macrolide was associated with a trend towards reduction in mortality[78]. Another pre-print study showing potential benefits with azithromycin alone was withdrawn. As commented when evaluating the outcomes with combination therapy, caution is advised given the multiple limitations. As observational studies, other unmeasured confounding factors may have been present[78].
In the second phase of COVID-19, the combination therapy, after adjusting for confounding factors, has been associated with an unacceptable risk of cardiovascular toxicity and arrhythmias. However, patients in the treatment groups were sicker, which may have affected safety outcomes despite adjusting for confounding variables. The rate of treatment discontinuation was not systematically reported, and treatment regimens were different across the studies.
Azithromycin, without concomitant hydroxychloroquine/chloroquine treatment, does not seem to confer the same risk of adverse events[27]. This may suggest that the main driver of toxicity in this setting is the use of other drugs and not azithromycin by itself. If this macrolide is considered, when possible oral route should be preferred due to lower peak levels that have been associated to a lower risk of cardiac toxicity[92].
Despite all the limitations, the treatment with hydroxychloroquine and azithromycin does not seem to offer any benefit in the second stage of COVID-19 and, on the contrary, has been associated with an increased risk of adverse effects and mortality. Therefore, until more data are available, its use should not be recommended outside from clinical trials. Azithromycin alone in addition to standard of care may provide additional benefits without safety concerns that need to be validated in clinical trials.
The third and more severe stage of COVID-19 is characterized by the development of hyperinflammation and cytokine storm. In this setting, other immunomodulatory therapies as corticosteroids or anti-IL6 have been proposed[18,90]. Azithromycin’s immunomodulatory effects may therefore play a role, given its ability to reduce cytokine expression among other properties. In other diseases as CAP, the immunomodulatory activity of azithromycin observed in vitro and in animal models has been demonstrated in high-quality clinical studies without safety issues. Furthermore, the potential benefits of azithromycin in severe lung injury and ARDS when initiated early in the disease have been demonstrated[19,63]. These benefits may be translated into patients with COVID-19, as a recent pre-print study showed that the cytokine profile in plasma (IL-1β, IL-1RA, IL-6, IL-8, IL-18, and TNFα) of severe COVID-19 patients did not differ from that found in other ARDS and sepsis of other causes[93]. In addition, its potential antifibrotic activity may be useful in ARDS or in patients who develop lung fibrosis. Recent evidence has demonstrated that COVID-19 can cause microvascular damage with endotheliitis, suggesting that therapies that stabilize the endothelial cells may be of interest[20]. Azithromycin may be useful since it has shown to stabilize and maintain the epithelial cells integrity[7].
In spite of all these potential benefits in critically ill patients, these patients have also been misrepresented. In all but one of the previous studies, patients admitted to the ICU at the time of treatment initiation were excluded. This is important since, at least in CAP, the beneficial immunomodulatory protective effect seems to be more evident in the most severe patients[53]. Unfortunately, the its potential usefulness in COVID-19 induced lung injury, ARDS or fibrosis remains unknown.
Concerning its bacterial activity, a recent meta-analysis showed that 7 % (14 % if admitted to ICU) presented bacterial co-infections, which was lower than with other viruses like influenza[94].Mycoplasma pneumoniae was found in 42 % of confirmed co-infections, although they were diagnosed serologically through the detection of IgM, which may have overestimated the rate of infections[94]. Unlike in influenza, where this macrolide reduced the rate of bacterial superinfections, the potential antibacterial benefit of azithromycin in the setting of COVID-19 has not been studied[55].
Azithromycin has demonstrated clinical benefits in other settings due to its antiviral and immunomodulatory action. However, in the treatment of COVID-19 it has been poorly studied, mainly in combination with hydroxychloroquine. Moreover, it has been studied in a very specific subgroup of patients, with other subgroups where it may offer the greatest clinical benefits being misrepresented. Although the paucity of data and associated limitations, azithromycin has shown promising results that deserve further study and may play a role in the treatment of COVID-19. The upcoming clinical trials will show whether this macrolide, alone or in combination, may be useful and which patients benefit most from it in the treatment of COVID-19.
Acknowledgements : We would like to thank Inés de Mena Urrutia for the thoughtful revision of the manuscript.
Conflict of interest: None to declare.
Funding: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.