Low-dose azithromycin prophylaxis in patients with atrial fibrillation and chronic obstructive pulmonary disease

Low-dose azithromycin prophylaxis is associated with improved outcomes in people suffering frequent exacerbations of chronic obstructive pulmonary disease (COPD), but the use of macrolides in patients with cardiovascular disease has been debated. To investigate the risk of adverse events after COPD exacerbations in patients with atrial fibrillation (AF) treated with azithromycin prophylaxis. Retrospective cohort study within the TriNetX Platform, including AF patients with COPD exacerbations. Risks of primary and secondary outcomes were recorded up to 30 days post-COPD exacerbations and compared between azithromycin users and azithromycin non-users. The primary outcomes were the risks for a composite of (1) cardiovascular (all-cause death, heart failure, ventricular arrhythmias, ischemic stroke, myocardial infarction, and cardiac arrest), and (2) hemorrhagic events (intracranial hemorrhage (ICH), and gastro-intestinal bleeding). Cox-regression analyses compared outcomes between groups after propensity score matching (PSM). After PSM, azithromycin users (n = 2434, 71 ± 10 years, 49% females) were associated with a lower 30-day risk of post-exacerbation cardiovascular (HR 0.67, 95% CI 0.61–0.73) and hemorrhagic composite outcome (HR 0.45, 95% CI 0.32–0.64) compared to azithromycin non-users (n = 2434, 72 ± 11 years, 51% females). The beneficial effect was consistent for each secondary outcomes, except ICH. On sensitivity analyses, the reduced risk of adverse events in azithromycin users was irrespective of smoking status, exacerbation severity, and type of oral anticoagulation. Azithromycin prophylaxis is associated with a lower risk of all-cause death, thrombotic and hemorrhagic events in AF patients with COPD. The possible role of azithromycin prophylaxis as part of the integrated care management of AF patients with COPD needs further study. Supplementary Information The online version contains supplementary material available at 10.1007/s11739-024-03653-0.


Introduction
Chronic obstructive pulmonary disease (COPD) in patients with atrial fibrillation (AF) represents one of the most common comorbidities and is responsible for significantly increased risk of cardiovascular events and death [1][2][3][4].Studies that have investigated the risk of cardiovascular events in COPD patients showed that is highest during the first 30 days after acute exacerbations [5,6].Low-dose oral azithromycin treatment is recommended by international guidelines in people with COPD who continue to exacerbate despite optimal inhaled therapy due to their proven anti-inflammatory properties and subsequently reduced exacerbation rates [7,8].Drugs targeting anti-inflammatory pathways have previously been associated with reduced cardiovascular event rates but concerns regarding the cardiac safety of macrolide antibiotics including azithromycin have previously been raised [9][10][11].We, therefore, investigated cardiovascular outcomes of COPD exacerbations in AF patients receiving low-dose azithromycin prophylaxis.

Study design
This was an observational study conducted within TriNetX, a global federated health research network with access to electronic medical records (EMRs) from academic and community hospitals covering approximately 80 million individuals, mainly located in the United States.Within this network, available data include demographics, healthcare utilization data, e.g., emergency department, inpatient and outpatient attendance, diagnoses using International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) codes, laboratory results Logical Observation Identifiers Names and Codes, LOINC, and medications.More information can be found in the supplementary material.

Study population
The searches on the TriNetX online research platform were performed on the 13th of February 2023.Within the Tri-NetX platform, we included adults diagnosed with COPD exacerbation (ICD-10 code J44.1).Azithromycin users were defined as individuals who had been prescribed azithromycin at least three months prior to the index event and were still prescribed it at the index event.Non-users were those with no history of azithromycin use in the 12 months prior to the index event.Both azithromycin users and non-users were concomitantly treated with inhaled corticosteroids (ICS: fluticasone, budesonide, or beclomethasone) in the year before the index event.The baseline index event was defined as the most recent COPD exacerbation reported in the TriNetX platform.
At the time of the search, 63 participating health care organizations had data available for individuals who met the study inclusion criteria.Characteristics registered before the index event were considered the baseline characteristics.The clinical outcomes were identified via ICD-10-CM codes (Supplementary Table 1).All-cause death was recorded using specific variable code within the TriNetX platform.

Outcomes
The primary outcome was the 30-day risk of i) a composite outcome of all-cause death, acute heart failure, severe ventricular arrhythmia, ischemic stroke, and myocardial infarction, and ii) a composite of hemorrhagic events of intracranial hemorrhage (ICH) and gastro-intestinal (GI) bleeding.The secondary outcomes of interest were the 30-day risk of each component of the primary outcomes.The 30-day risk of primary and secondary outcomes was further assessed in different time frames: 1-7 days, 8-14 days, and 15-30 days after the acute COPD exacerbation [5,6,12].The 30-day interval was selected to assess the impact of azithromycin prophylaxis on reducing the risk of cardiovascular events associated with COPD exacerbation.Moreover, we performed a number of some sensitivity analyses to assess the robustness of our primary findings.
First, to further investigate the anti-inflammatory efficacy of macrolides we used another "active comparator" by restricting the azithromycin non-users to those receiving inhaled corticosteroids (i.e., akin to GOLD E group [7]), and roflumilast, a selective phosphodiesterase-4 inhibitor.
Second.given a previous study reported macrolides may be less effective in reducing the risk of COPD exacerbation in smokers, we performed a further analysis restricted only in patients with labeled as current smokers [13].
Third, we investigated the risk of primary outcome considering only those patients with severe COPD exacerbations requiring emergency department access.
Fourth, we analyzed the risk of adverse events subdividing patients based on the type of oral anticoagulant in warfarin users and non-vitamin k antagonist anticoagulant (NOAC) users.

Statistical analyses
Baseline characteristics were compared using chi-squared tests for categorical variables and independent-sample t-tests for continuous variables.Absolute standardized mean differences (ASD) were used to show the distribution of demographic and clinical data among the groups and calculated as the difference in the means or proportions of a particular variable divided by the pooled estimate of standardized differences for that variable.Propensity score matching (PSM) 1:1 was used to control the differences in the comparison cohorts.Cohort matching was performed for age at index event, sex, ethnicity, arterial hypertension, diabetes, ischemic heart disease, heart failure, cerebrovascular disease, and cardiovascular medications (β-blockers, antiarrhythmics, diuretics, lipid lowering drugs, antianginals, calcium channel blockers, angiotensin-converting enzyme inhibitors, anticoagulants, and platelet aggregation inhibitors).These variables were chosen because they may influence clinical outcomes.Any baseline characteristic with a Std diff.< 0.100 was considered well matched.After PSM, Cox-regression proportional hazard models were used to calculate hazard ratios (HRs) with 95% confidence intervals (CIs) for the risk of primary and secondary outcomes in azithromycin users compared to azithromycin non-users.
Sensitivity analyses were performed as described above.All tests were two tailed and p-values of ≤ 0.05 were taken to indicate statistical significance.All analyses were performed in the TriNetX platform which incorporates R (v4.3.1,R Foundation for Statistical Foundation, Vienna, Austria).

Results
The initial cohort of AF patients with COPD was composed by 2442 azithromycin users (71.4 ± 10.3 years, 49.1% females) and 58,817 non-azithromycin users (73.0 ± 10.3 years, 47.5% females) (Table 1).Before PSM, azithromycin users were younger, with a higher prevalence of dyslipidemia and hypertension compared to azithromycin non-users (Table 1).After PSM, 2434 patients were successfully matched for each group and no significant differences were found for all the demographic and clinical variables considered (Table 1).

Risk of cardiovascular events during the different time frames.
In time-stratified analyses, the beneficial effect of azithromycin in reducing the risk of cardiovascular outcome was highest during the 1st week (HR 0.57, 95% CI 0.50-0.65),and progressively decreases during the 2nd week (HR 0.69, 95% CI 0.59-0.80)and the 3rd-4th weeks (HR 0.76, 95% CI 0.68-0.88)(Table 3).The same trend could be observed for the risks of all the other secondary outcomes, where greatest effect was observed during the first 7 days after exacerbation (Table 3).
In the analyses performed considering separately warfarin and NOAC users we found that the 30-day risk of composite outcome in azithromycin users was reduced, and no increased risk of hemorrhagic events was found, for both the treatments (Fig. 2).In azithromycin users treated with warfarin, the HR of composite cardiovascular outcome was 0.72 (95% CI 0.53-0.98)when compared to azithromycin non-users, whereas no difference in the number of hemorrhagic events (10 vs 10) was recorded (Fig. 2).Compared to azithromycin non-users, azithromycin users treated with NOAC the HR of composite cardiovascular outcome was 0.69 (95% CI 0.56-0.85),whereas that for hemorrhagic events was 0.30 (95% CI 0.10-0.93)(Fig. 2).

Discussion
In this study, the principal findings are as follows: (1) After COPD exacerbations, AF patients treated with ICS and lowdose azithromycin prophylaxis have a reduced 30-day risk of all-cause death, cardiovascular events and bleeding compared to those treated only with ICS; (2) The highest risk reduction for both the composite cardiovascular and hemorrhagic events was observed during the first week after the COPD exacerbation; and (3)The beneficial effect of lowdose azithromycin prophylaxis was robust across a number of sensitivity analyses.
Long-term macrolide therapy is recommended by multiple guidelines for COPD patients suffering frequent exacerbation with high quality evidence from multiple randomized clinical trials and meta-analyses demonstrating reduced exacerbation rates associated with treatment [14][15][16].However, there is conflicting evidence as to the cardiovascular safety of macrolide therapy.For example, the Federal Drug Administration placed a risk warning for ventricular arrythmias based on an observational study which suggested increased risk of cardiovascular death, but not all-cause death, 10 days after initiation of a course of azithromycin for acute indications when compared to amoxicillin [9].A population based retrospective cohort study of people with non-tuberculous mycobacteria pulmonary disease in South Korea observed increased cardiovascular disease incidence in macrolide-treated patients [17].However, no increased risk was observed in a large Danish national cohort study on COPD patients [18] and no adverse cardiovascular disease signal was reported in the RCT of macrolides in COPD or bronchiectasis [15,19].
COPD represents one of the pivotal factors in determining the risk of death and cardiovascular events in AF patients already treated with OAC [1].The relationship between AF and COPD is bidirectional, the AF onset in COPD patients is associated with a worse quality of life and an increased risk of respiratory failure [20], whereas the COPD exacerbations in AF patients dramatically increase the risk of cardiovascular events [21].This increase is the results of different mechanisms, (i) hypoxia and hypercapnia induce vasoconstriction in the pulmonary circulation with pulmonary hypertension, (ii) the inflammatory response can precipitate thrombotic events, and (iii) the hyper sympathetic stimulation necessary to relax the airways' smooth muscle cells triggers life-threatening arrhythmias and induce ischemic events [22].
In this context, the anti-inflammatory mechanisms of action by which macrolides are thought to reduce exacerbation risk that include improved efferocytosis and suppression of inflammatory cytokines such as IL-6, IL-1 and reduced TNF-α, may provide a suggestive explanation of the protective effect we found in azithromycin users.Indeed, the reduced inflammatory state associated with low-dose azithromycin prophylaxis in AF patients with COPD, could lead to less severe exacerbations and thus to a reduced risk of cardiovascular events and death.Hence, this suggests that azithromycin's anti-inflammatory effect may be additive to that of ICS.Moreover, the reduced risk of adverse events was consistent also in the sensitivity analyses: (i) using as comparator COPD patients treated with ICS and roflumilast, another anti-inflammatory drug that exerts its effect through a different mechanism than azithromycin (i.e., inhibiting phosphodiesterase), (ii) in patients with an additional source of oxidative stress, inflammation, and sympathetic agonism such as the smokers, in which some guidelines do not recommend use of macrolides [7]; and (iii) only in those with severe COPD exacerbations, that were at higher risk per se.These subgroup analyses demonstrate the robustness of our findings.
Beyond the cardiovascular safety profile, another source of possible concerns for macrolides use in AF patients with COPD is the possible presence of drug interactions.Indeed, macrolides are a well-known class of CYP3A4 and p-glycoprotein inhibitors and may perturbate the plasma concentration of both vitamin-K antagonists and NOACs [23].In our study, we showed that not only the risk of cardiovascular events, but also the hemorrhagic risk was reduced in azithromycin users, and this was irrespective of the type of oral anticoagulant (warfarin or NOACs).A partial explanation behind the reduced risk of bleeding in COPD patients with AF azithromycin users could lie in the reduced incidence of collateral complications triggered by COPD exacerbation that could increase the hemorrhagic risk.COPD exacerbations are often associated with hypertensive crisis that could facilitate vascular rupture and hemorrhagic events, the respiratory distress facilitates the onset of dehydration that in turn can facilitate the onset of acute renal failure and reduced OAC clearance, moreover, both the hypoxemia and the hypercapnia can cause delirium in the elderly with subsequent risk of falls and traumatisms.
This body of evidence underlines the complexity of AF patients with COPD and the importance of reducing the background inflammatory state associated with COPD to counteract the risk adverse events.This could be reached through the optimal management not only of COPD but also of all the comorbidities that often coexist in AF patients.
The latest management guidelines for AF management [24,25] advocate this type of holistic approach and proposing the atrial fibrillation better care (ABC) pathway as a possible tool drives the clinical decisions [26].The ABC pathway is based on the 3 main pillars: "A" avoid stroke with oral anticoagulation, "B" better management of the symptoms with a patient-centered symptom-directed decisions on rate or rhythm control; and "C" Cardiovascular risk factor optimization and lifestyle changes [25].Adherence to the ABC pathway was associated with a reduced risk of all-cause death, cardiovascular events, and AFrelated hospitalization in different populations [27,28], even in complex phenotypes [29].One of the main concepts sustained by this integrated approach is that the risk of adverse events in already anticoagulated AF patients is due to all the cardiovascular comorbidities that often coexist in AF patients [26].Although a schematic approach has been proposed to encode the optimal management of common cardiovascular comorbidities [28], no specific treatment associated with a reduced risk of cardiovascular events was found for COPD.In this context, the possible use of low-dose azithromycin prophylaxis in AF patients with COPD could represent the possible missing piece of this puzzle.Indeed, reducing the rate and the severity of COPD exacerbation seems to be associated with an overall risk reduction for all the related adverse events.

Limitations
This study was observational and retrospective, and despite successful matching, we cannot rule out residual bias or indication bias related to unmeasured confounders.Outcomes in this study were based on ICD-10 codes and as such are unadjudicated and may be prone to misclassification or under-reporting.Similarly, azithromycin prescriptions were identified using RxNorm codes which may be prone to the same limitations.We did not adjust the analyses for procedures or treatments administered after the index event to avoid the possibility of introducing immortal time bias.Overall, results here are hypothesis generating and robust prospective studies are needed to validate our findings.

Conclusion
Our data suggest that long-term azithromycin prophylaxis is associated with reduced 30-day risk of cardiovascular events associated with exacerbations in AF patients with COPD.Further studies are needed to prospectively validate these findings, understand mechanisms, and delineate those most likely to benefit.

Table 1
Baseline characteristics of azithromycin users and non-users before and after propensity score matching

Table 2
Fig. 1 Risk of composite outcome and hemorrhagic events in azithromycin users compared to azithromycin non-users

Table 3
Risk of early cardiovascular complications after COPD exacerbation in azithromycin users and non-users during the different time frames HR Hazard Ratio, CI Confidence Interval, ICH intracranial hemorrhage, GI Gastrointestinal * No HRs were calculated for ICH in the second and third period because only 1 event occurred