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.