Risk of cardiovascular events after an exacerbation of chronic obstructive pulmonary disease: results from the EXACOS-CV cohort study using the PHARMO Data Network in the Netherlands

Chronic obstructive pulmonary disease (COPD) is a heterogeneous respiratory condition, characterised by chronic respiratory symptoms due to abnormalities of the airways and/or alveoli that cause persistent, often progressive airflow obstruction [1]. Exacerbations of COPD negatively impact prognosis, often leading to more frequent and severe exacerbations, accelerated loss of lung function, and amplifying risk of mortality [2].

Persons with COPD have a 2 to threefold increased risk of cardiovascular (CV) disease compared to persons without COPD, regardless of age, sex, smoking status or other confounders [3]. In particular, time periods following an exacerbation have been identified as periods wherein patients are at an increased risk of incident severe CV events [4]. Although many studies have examined the cardiovascular risk in COPD patients, to our knowledge, these have focused on prevalent, severe COPD patients, i.e. either from randomized controlled trials or selected because they had a CV event. The association between exacerbations and cardiovascular risk has not been studied in newly diagnosed patients. Importantly, newly diagnosed patients may already be experiencing exacerbations [5] and this may confer an increased risk of cardiovascular disease already early after diagnosis.

The aims of this study were to determine in a population of patients newly diagnosed with COPD, (1) the association between a moderate or severe exacerbation and a composite of non-fatal severe CV events or all-cause death; (2) the association between a moderate or severe exacerbation and the individual types of CV events of interest separately; (3) the association with the composite outcome for moderate and severe exacerbations separately; and (4) the association with the outcome with the exacerbation recurrence, for each of the subsequent first, second, and third exacerbation.

This study is part of the EXAcerbations of COPD and their OutcomS on CardioVascular disease (EXACOS-CV) programme. The overall study design and analytical approaches were detailed previously [6].

The source data contained a population of 1,636,513 individuals, including 8020 (0.5%) newly-diagnosed COPD patients (Additional file 1: Fig. S1). Mean age at COPD diagnosis was 65.4 years (SD, 10.7); 47% were female. Baseline characteristics of the study population are further presented in Table 1.

During a median follow-up duration of 36 months (interquartile range 21–52 months), 2234 (28%) patients experienced at least one exacerbation; 1869 (23%) patients had at least one moderate exacerbation, and 743 (9%) patients had at least one severe exacerbation. Patients who were exposed to at least one exacerbation during study follow-up were more likely to be prescribed with COPD, cardiac and/or metabolic medication in the year preceding the first diagnosis of COPD (Table 1).

5270 (66%) patients were followed until 31 December 2019, whereas others were censored due to the end of data availability (n = 1658, 21%), a first non-fatal severe CV event (n = 631, 8%), and death without a prior severe CV event (n = 461, 6%). Among the 1092 patients who experienced an outcome of interest, the majority (n = 646, 59%) experienced the outcome prior to a first exacerbation or censoring; 33% (n = 360) had the CV event within the 1–365 days following an exacerbation, and 8% (n = 86) > 365 days following an exacerbation.

The frequency and proportion of patients who experienced at least 1 event of each outcome category separately (i.e., ignoring any CV event from other categories) are reported in Additional file 1: Table S2. ACS, HF decompensation, cerebral ischaemia and arrhythmias were each observed in about 2–3% of patients throughout follow-up, with HF decompensation being most common. Acute pulmonary oedema and cardiac arrest were rare (< 0.5%). In total, 567 patients (7%) died during follow-up, with or without a previous CV-event.

Crude incidence rates for the composite outcome are presented in Additional file 1: Table S3. Compared to unexposed periods, exposed periods following a moderate or severe exacerbation were associated with increased risk of the composite outcome. The increased risk decreased gradually over time, but persisted for the entire year [adjusted HR were 15.3 (95% CI 11.8–20.0) in days 1–7; 7.0 (95% CI 4.9–10.0) in days 8–14, 4.8 (95% CI 3.5–6.5) in days 15–30; 1.9 (95% CI 1.5–2.3) in days 31–180, and 1.3 (95% CI 1.0–1.8) in days 81–365] (Table 2).

When examining the association for the time to the first non-fatal severe CV event, exposed periods following a moderate or severe exacerbation event were associated with increased risk of a non-fatal severe CV event, which persisted for the entire year (Table 3).

When analysing categories of severe CV events as separately, the risk of ACS, HF decompensation, and arrhythmia increased substantially following moderate or severe exacerbations. This risk decreased over time but remained significantly higher than in the unexposed period over the first 180 days. Adjusted HR were observed from 9.9 (95% CI 5.0–19.4) in days 1–7 to 1.9 (95% CI 1.1–1.3) in days 31–180 for acute coronary syndrome; adjusted HR from 27.4 (95% CI 17.3–43.4) in days 1–7 to 2.3 (95% CI 1.4–3.6) in days 31–180 for HF decompensation; and adjusted HR from 9.4 (95% CI 4.4–20.2) in days 1–7 to 4.0 (95% CI 1.5–10.6) in days 15–30 for arrhythmia. For cerebral ischaemia, the risk following an exacerbation was increased, on average, in the first 30 days following exacerbations, [adjusted HR 3.5 (95% CI 1.2–10.6) in days 1–7 to adjusted HR 2.4 (95% CI 0.8–7.2) in days 15–30], but less pronounced and estimated with uncertainty due to low event counts. The risk of all-cause death remained increased over the first 180 days after an exacerbation (Table 3).

After a moderate exacerbation, the risk of the composite outcome was increased and was stable over the first 30 days [adjusted HR 2.5 (95% CI 1.3–4.8) in days 1–7 to HR 2.6 (95% CI 1.7–4.2) in days 15–30]. After this 30-day period, the adjusted HR gradually decreased to 1.6 (95% CI 1.3–2.1) in the 31–180 days period and 1.3 (95% CI 0.9–1.7) in the 181–365 days period after a moderate exacerbation (Table 2).

After a severe exacerbation, the risk of the composite outcome increased substantially and remained significantly higher than in the overall unexposed periods up to the entire year [adjusted HR were 48.6 (95% CI 36.9–64.0) in days 1–7; 20.0 (95% CI 13.3–30.1) in days 8–14; 11.3 (95% CI 7.7–16.5) in days 15–30; 2.8 (95% CI 2.0–3.8) in days 31–180; and 1.6 (95% CI 1.0–2.6) in days 181–365].

Adjusted HR for day 1–7 period were 15.2 (95% CI 10.3–22.4) following first exacerbation, 22.3 (95% CI 14.3–34.6) following second exacerbation, and 20.2 (95% CI 10.7–38.0) following a third exacerbation (Table 4). Irrespective of whether the exacerbation was a first, second or third, the risk of the outcome decreased following exacerbation but remained significantly higher than the overall unexposed period over at least 180 days.

To our knowledge, this study is the first to report on the association between exacerbations of COPD and the risk of severe CV events or death in patients newly-diagnosed with COPD. We showed a substantially increased risk of a non-fatal severe CV event or death in the months following a moderate or severe exacerbation even in this cohort of patients with a new incident COPD diagnosis. This increase in risk was not constant over time: it was 15 times higher (for the composite outcome of a severe CV event or death) for 1–7 days following the onset of an exacerbation, and then decreased gradually during the subsequent periods of time, although risk remained significantly increased up to one year. The risk for individual types of non-fatal severe CV events was also increased following an exacerbation, including ACS, HF decompensation, cerebral ischaemia, and arrhythmias. The most prominent increase in risk was observed for HF decompensation.

Unique to this study is the fact that we specially explored CV risk after moderate exacerbations. When stratifying by the type of exacerbation, severe exacerbations were associated with a substantially increased risk of any severe CV event or death. The association for moderate exacerbations was less strong, though still significant for the first 180 days. Although the CV risk associated with moderate exacerbation was lower in magnitude compared to a severe exacerbation, moderate exacerbations are much more frequent, highlighting the clinical importance of this finding. Moreover, the elevated risk of a CV event or all-cause death was already present following the first exacerbation following new COPD diagnosis, again illustrating that COPD should be approached with urgency.

Our results confirm previous studies showing the risk of a severe CV event is increased in the days and months following an exacerbation [4, 11‐13]. Previous studies were conducted in selected patients included in clinical trials, or used a self-controlled design selecting patients on the occurrence of a CV outcome [9, 14]. In this study with a narrow definition of exacerbation, over a quarter of COPD patients experienced an exacerbation within 3 years after diagnosis, exposing them to increased risk of CV events early in their disease trajectory. Our results thus highlight the burden of exacerbations at a very early stage of the disease. In addition, supporting evidence suggested that CV risk was higher following severe exacerbations than following moderate exacerbations [10, 12, 14, 15]. Importantly, we report here that the risk of severe CV events or all-cause death is immediately increased following a moderate exacerbation. The clinical relevance of experiencing significant risk of CV events following a first, moderate exacerbation is of paramount urgency, especially given that the majority of exacerbations in newly-diagnosed people living with COPD are of moderate severity. The importance of a single exacerbation for future cardiopulmonary risk has also recently been highlighted in the study by VanFleteren et al. showing that one exacerbation increased future risk for exacerbations and hospitalisation [16]. The fact that we showed that this already occurs early in the COPD disease trajectories leads to the hypothesis that early preventive treatment may be beneficial and this would be interesting for future research.

Information on specific cardiovascular events has been scarce up till now. Pooled results from a recent meta-analysis showed that the risk of stroke was 1.7 times higher and the risk of acute myocardial infarction was 2.4 times higher in the 1–3 months following an exacerbation [4]. Information on other CV events, such as HF decompensation and new-onset AF following exacerbation of COPD, has been limited. One study, based on a self-controlled case series reported an increased 30-days and 1-year incidences of AF and HF after 30 days and after 1 year following an exacerbation [11]. The current study design strengthens these findings in respect to granularity of the differential effects of exacerbations on the various CV events. We showed that people with COPD are most vulnerable in the first 30 days after an exacerbation throughout the whole spectrum of cardiovascular diseases. Risks for HF and ACS were highest suggesting that the acute effects of an exacerbation on the CV system could be drivers of such CV events. Our analyses adjusted for key observed confounders of the association between post-exacerbation time periods and the occurrence of a CV event, including comorbidities, among other factors. Nonetheless, unmeasured confounding may play a role and so our results should be interpreted as associations rather than causal effects.

By using real world GP data, our results increase the external validity of the findings of CV risk after exacerbations found in the randomized controlled trials. All citizens in the Netherlands are registered at a GP practice, except those living in a nursing home or hospice. The primary care information of patients included in the GP data can be considered representative for the general population [17], although linkage to hospital and pharmacy information could only be made for a subpopulation. Some limitations should be mentioned as well. First, some CV events could have been misclassified as exacerbation of COPD. For instance, a patient can be admitted for a severe exacerbation but actually have an HF decompensation, since symptoms of HF decompensation can mimic those of exacerbation of COPD. Such a misclassification, should it occurred, would have mainly impacted the day 1–7 period. Second, moderate exacerbations were strictly defined using the combination of a GP visit for COPD and a dispensation of oral corticosteroid. This strict definition tends to underrepresent moderate exacerbations: exacerbations managed with antibiotics only or modification of inhaled medications were missed by definition. Third, smoking history was missing for approximately 50% of the study population, and could therefore not be included as a potential confounder. Fourth, as cause of death is unknown in the PHARMO Data Network, it was not possible to include CV death in the composite outcome of severe CV events and all-cause death was included instead. Fifth, some less severe COPD patients might have been excluded from our study population because the diagnosis of COPD could not be confirmed.

This study has several clinical implications. The findings highlight the impact of COPD beyond respiratory function and symptoms. Patients with COPD are at high risk of cardiopulmonary events such as exacerbations and CV events. It emphasizes the importance of a treatment paradigm that includes prevention, early monitoring, and treatment of this cardiopulmonary risk from the start of COPD diagnosis. In clinical practice, severe CV events are often under-diagnosed in patients experiencing exacerbations of COPD [18‐20]. There is a need to optimize COPD management across specialties focusing on reducing risk of exacerbations. Future research should further clarify whether the number of exacerbations affect the magnitude of the CV risk, and should further examine the risk of cerebral ischaemia following exacerbations of COPD.

In this population of newly diagnosed COPD patients, we observed a substantial increase in the risk of severe CV events or all-cause death after a moderate or severe exacerbation. The increase in risk was strongest in the month after the onset of an exacerbation and concerned all categories of severe CV events, i.e., ACS, HF, arrhythmias, or cerebral ischemia. These findings highlight the high cardiovascular risk in people living with COPD, even among those that are early in their disease trajectory.