A group of 60 leading Spanish pulmonologists specializing in COPD were invited by the scientific committee to participate in this study. All are members of the COPD working group of the Spanish Society of Pneumology and Thoracic Surgery (SEPAR), and were initially proposed on the basis of their experience, publications in recent years, participation in research projects on the subject, and geographic diversity.

Experts were defined as professionals seeing more than 100 patients with COPD/year or who had published more than 1 article in the previous year or had made more than 2 communications to congresses on COPD in the previous year. The selection of participants was not random and was conducted mainly in response to criteria of availability and willingness to participate.

The statements included in the Delphi questionnaire were derived from a series of questions that were investigated in a literature review. The results of the review were shared among the experts and debated. The steps followed are set out in detail below, in Fig. 1.

Fig. 1.

Phases in the development of the expert document panel on COPD exacerbators.

(0.22MB).

Median values and the first and third percentiles for each of the questionnaire items were calculated as measures of central tendency and dispersion.

The internal consistency of the questionnaire was determined with the Cronbach's alpha statistic, values above 0.75 being considered acceptable. The correlation coefficient (ri) was calculated as a measure of reliability, and was considered good when values were between 0.4 and 0.75, and excellent above 0.75. Both values were calculated for the total questionnaire and for each of the blocks (Annex 2).

The correlation between rounds was measured using the Spearman correlation coefficient, in blocks and for the overall questionnaire (Annex 3). Similarly, concordance between rounds for each question was analyzed qualitatively using the kappa index, pooling the scores in 3 groups.1–9 Concordance was considered to be moderate when the score was between 0.41 and 0.60, good between 0.61 and 0.80, and very good for values above 0.80. The coefficient of variation (CV) was calculated for each questionnaire and in each round. The criterion for not needing consecutive rounds was that the relative increase from the previous round [(current CVround−previous CVround)/previous CVround] did not exceed 10%.

The level of statistical significance for all the estimators was established at P<0.05. Analyses were performed using SPSS software v24.0 (IBM Corp. Released 2016. IBM SPSS Statistics for Windows, Version 24.0. Armonk, NY: IBM Corp.)

• •

  For a patient to be considered an exacerbator, they must have had at least 2 moderate exacerbations in the previous year, defined as requiring at least outpatient treatment with systemic corticosteroids and/or antibiotics (87.9%, agreement).

• •

  A patient with 2 or more severe exacerbations in the last year is defined as a “severe exacerbator” (91.4%, agreement).

• •

  Patients with COPD-asthma phenotype (asthma-COPD overlap [ACO]) have mostly eosinophilic exacerbations and respond better to steroids than other phenotypes (87.9%, agreement).

• •

  A study of cellularity (eosinophilia) and microorganisms in respiratory samples is needed to characterize some exacerbations (88.3%, agreement).

• •

  Chronic bronchial infection aggravates lung function decline in exacerbator patients (93.1%, agreement).

• •

  Long-acting bronchodilators must not be withdrawn in COPD exacerbators (93.1%, agreement).

• •

  Patients who continue to have exacerbations despite optimal bronchodilator therapy should be evaluated for other treatable traits, in order to offer more personalized management (96.6%, agreement).

• •

  Bronchodilator therapy impacts on exacerbations, acting on lung function by reducing hyperinsufflation (91.4%, agreement).

• •

  The care and management of the exacerbator patient must be coordinated within the framework of an integrated primary care/pulmonology care program (93.1%, agreement).

• •

  Exacerbators should always be monitored by the respiratory medicine department in specific integrated multicomponent programs (94.8%, agreement).

• •

  As a general rule, exacerbators should be followed up in specialized outpatient clinics (91.5%, agreement).

In contrast, the panel of experts failed to reach consensus on the following statements:

• •

  A patient who has at least 3 mild exacerbations per year (change in the intensity of symptoms, change of treatment without the need for systemic corticosteroids and/or antibiotics, without being admitted to hospital or treated in the emergency department) should be considered an exacerbator.

• •

  Peripheral blood eosinophilia>2% is a marker of non-infectious inflammatory exacerbations.

• •

  Oral drugs in COPD exacerbators should only be used after inhaled triple therapy.

• •

  Exacerbation in a severe exacerbator is a criterion for admission, as it is an indirect marker of disease severity.

The concept and definition of what we understand by exacerbator is clearly forms the foundation of subsequent recommendations. Exacerbators or exacerbator phenotypes have previously been defined in the literature,14,17 and have even have been included in CPGs.12,17 However, these definitions are rather controversial with regard to the frequency of events required, the minimum intensity of exacerbations, if a single hospitalization automatically defines a patient with differential features, and the stability of the concept over time. An exacerbator was defined as a patient who had at least 2 moderate (or more severe) exacerbations in the previous year, defined as episodes that require at least ambulatory treatment with systemic corticoids and/or antibiotics. This definition reached a consensus of 87.9% in the first round. Similarly, the experts also consider that a patient who has needed hospital admission in the previous year is an exacerbator (75.9% consensus). This consensus definition is, then, virtually identical to the GesEPOC proposal, and thus reinforces the definition made in those guidelines.17 The proposal refers to the number of exacerbations in the previous year and not to the patient's accumulated history. This factor is important, since it introduces a certain dynamic component. A patient may be an exacerbator at a certain moment, but with time and perhaps with treatment may cease to be defined as such. In the Eclipse study, only 12% of patients classified as exacerbators in the first year of follow-up continued to be exacerbators over the entire 3-year follow-up period.14

One of the most innovative aspects of this document is that it differentiates the “severe exacerbator” as one who has had at least 2 severe exacerbations in the last year. To the best of our knowledge, this definition had not been published previously and could potentially have implications in clinical practice. Repeated severe exacerbations have proven to be an independent adverse prognostic factor,9 but this is not the case with less severe exacerbations. The clinical profile of these cases also differs, and their greater use of resources20,21 means that preventive and therapeutic strategies need to be intensified and personalized (case management) in this subgroup of patients. In general, exacerbations become more frequent and severe as COPD progresses. However, the rate at which they occur seems to reflect an independent susceptibility phenotype14 that needs to be identified.

Another innovative aspect of this consensus document is the proposal to differentiate “worsening of symptoms without exacerbation” from a proper exacerbation. This concept, which achieved 82.8% consensus, was defined as “the absence of exacerbation or signs of decompensation of any of the patient's comorbidities”. Some authors have proposed the term “unstable COPD” to define a worsening of symptoms that is not accompanied by other biological changes, such as tachycardia, tachypnea, desaturation and/or increase in biomarkers.22

The working group reached consensus in identifying exacerbator 4 subtypes (infectious, inflammatory, comorbid, and mechanical). This approach has considerable clinical implications, in that 4 different mechanisms are suggested, with important connotations regarding prevention. Patients with repeated infectious exacerbations usually present neutrophilic inflammation with a significant bacterial burden during stable phases.23,24 In these patients, treatment with inhaled antibiotics can be very useful. There is also evidence that patients with this profile are up to 6 times more likely to have a bacterial-type exacerbation.23 Another proposal is to include patients with predominantly Th2 inflammation in stable phase in the group called “inflammatory exacerbators”. We suggest that these patients, identified in some guidelines as patients with asthma and COPD overlap (ACO)17 or by the presence of increased peripheral eosinophilia (>300eosinophils/mm3),16 may have more eosinophilic exacerbations and a better response to inhaled steroids.23 An important proposal of the consensus document is to define a comorbid exacerbator group and a mechanical exacerbator group (characterized by a worsening of symptoms due to hyperinsufflation with no increase in inflammation). Some studies suggest that the presence of comorbidity, particularly cardiovascular, is associated with an increase in systemic inflammation and risk of exacerbations,25,26 so it follows that the optimization of treatment of these concomitant diseases can be extremely useful. With regard to the mechanical exacerbator, the experts propose the existence of a group of exacerbators characterized by emphysema and signs of air trapping/hyperinsufflation. Functional alterations with a low level of inflammation would be predominant in these patients. Patients with this profile, labelled by some CPGs as exacerbator phenotype with emphysema, have a lighter infective burden and poorer response to inhaled steroids.17

In the opinion of the experts, several considerations must be taken into account to understand these subtypes and predisposing factors. In the first place, eosinophilia must be determined in respiratory samples and microorganisms must be identified. Eosinophilia is easy to analyze in a standard blood test, a procedure that is recommended in all exacerbators, regardless of their severity. Ideally, eosinophilia in sputum should also be analyzed, but as very few centers are equipped for this, this recommendation cannot be generalized. The microbiological study of sputum, however, is not systematically recommended, but it is especially indicated in cases with suspicion of chronic bronchial infection. Similarly, airway and pulmonary parenchymal lesions detected in imaging tests (mainly HRCT) are useful for identifying exacerbation patterns and determining the differential diagnosis of COPD. It is interesting to note that these procedures are considered helpful for recognizing the exacerbation subtype (emphysema, wall thickening, bronchiectasis, cardiovascular, etc.). Moreover, mucus hypersecretion is identified as a factor that favors lung function decline in exacerbators, perhaps due to inflammation.

Exacerbators, especially severe exacerbators, have a clearly increased risk of death.9 There is no clear evidence to indicate which mechanisms induce this increase in mortality. In the opinion of the experts, exacerbators are at increased risk of death due to a persistent systemic inflammatory response that causes an increase in cardiovascular events during and after the exacerbation (with an agreement of 84.5%). This statement may clearly have important therapeutic connotations, and is formulated on the basis of indirect data. Persistent systemic inflammation in COPD has been linked to increases in all-cause death and more frequent exacerbations.27 However, it is not clear whether the phenomenon is a cause or a consequence of repeated exacerbations. An increase in systemic inflammation and oxidative stress is known to occur during exacerbations, and these mechanisms could influence any of the extrapulmonary manifestations of COPD (nutritional, metabolic or cardiovascular changes).28 The persistence of residual systemic inflammation has also been seen to increase with time, especially in the case of repeated exacerbations.29,30 In this setting, reducing systemic inflammation should be beneficial; however, this has not been fully demonstrated. Recent data from the IMPACT STUDY31 suggest a reduction in mortality in the group of severe patients with frequent exacerbations treated with triple therapy. However, these data should be interpreted with caution, as they were generated in a secondary analysis. In patients with a cardiovascular profile, the use of combined LABA/ICS therapy showed no benefits in terms of survival.32

The working group identifies long-acting bronchodilators as the main preventive therapy for exacerbations in patients with no criteria for ACO. It is clear that the effect of this treatment on exacerbations affects lung function, mucus production, and mucociliary clearance. In the opinion of the experts, this treatment is sufficient in low-risk symptomatic patients and in well-controlled exacerbators. The document proposes the use of single-agent bronchodilators in COPD patients with a maximum of 2 outpatient exacerbations (non-exacerbators) and mild or moderate obstruction. Dual bronchodilation is the treatment of choice in exacerbators, with the exception of patients with ACO. If exacerbations persist despite optimal bronchodilator treatment, other treatable traits must be evaluated, and patient management must be personalized (inhaled corticosteroids [ICS] in cases of peripheral blood eosinophilia, nebulized antibiotics in cases of persistent bronchial infection, low-dose macrolides in patients with bronchiectasis, etc.).

Almost 90% of participants agreed that the use of ICS should be recommended to prevent exacerbations that persist despite dual bronchodilation. The strongest evidence to date comes from the TRIBUTE33 and IMPACT31 studies, which demonstrated the superiority of triple therapy compared to dual bronchodilation in the prevention of exacerbations. This benefit is more pronounced in the presence of significant peripheral eosinophilia and less marked in cases with eosinopenia.31,33 The same biomarker has also been proposed to guide ICS withdrawal. Experts recommend evaluating the discontinuation of ICS in patients who are stable for at least a year, with no severe obstruction, and eosinophilia<300c/mm3, provided dual bronchodilation is maintained. Finally, dose tapering is seen as a way to optimize treatment with ICSs.

The working group reached consensus regarding the usefulness of integrated care programs. In general, these services improve quality of life and exercise capacity and reduce hospital admissions and length of stay. These benefits were described in a Cochrane review published in 201334 (26 trials with 2997 patients) and in various publications.35,36 However, a Dutch study conducted in primary care with a 24-month follow-up found no benefit over standard care, with the exception of a greater level of integration and an improvement in activities of daily living.37 Despite this controversy, the experts support the use of specific integrated multicomponent programs, in which pharmacological treatment, respiratory rehabilitation, health education, and case management are optimized, and the social aspects of the disease are addressed. This approach must be coordinated between primary care, which is much closer to the family environment of the patient, and respiratory care. However, given the complexity, heterogeneity, and high risk of these patients, we believe that exacerbators, especially severe cases, must undergo protocolized follow-up led by the pulmonology department. In less severe cases, monitoring should be primarily performed at the primary care level. However, these patients are still considered to be at risk, and therefore must receive optimized multidimensional assessment, treatment and follow-up.

It was expected that some of the statements would not reach consensus. COPD is a dynamic field, and our knowledge of this pathology is continuously being updated. There is no doubt that new research will help to clarify the controversial issues that did not achieve consensus. Furthermore, the consensus that were reached must always be treated with caution until they are endorsed by scientific evidence.

Another possible limitation of this document is that, due to the characteristics inherent to the method, the results can only be taken as expert opinions. The participants were exclusively specialists in pulmonology. This fact may also have influenced some of the statements, and other specialties, such as primary care, or the viewpoint of the patient, should ideally have been included. We believe, however, that the number of respondents and their distribution across Spain, together with the high degree of participation achieved, add great value to this document.