“Causes of death in asthma, COPD and non-respiratory hospitalized patients: a multicentric study”

After the epidemics of asthma deaths during the 1990s, numerous International [1, 2] and national [3–5] recent studies have confirmed a consistent decrease in mortality. Among the causes of this decline, the correct implementation of clinical practice guidelines has been postulated [6–8], and particularly, the increasingly universal use of inhaled corticosteroids as maintenance asthma treatment [9, 10].

Unlike COPD, for which mortality has risen during the same period [11], the reduced mortality from asthma has perhaps contributed to a reduced knowledge on the clinical characteristics and circumstances of death in asthmatics. Thus, contrary to COPD patients, in whom numerous recent studies have characterized their causes of death, in asthma patients there has been considerably less research. It is well established that in COPD, serious exacerbations, respiratory failure, cardiovascular diseases, and bronchopulmonary malignancies constitute the principal causes of death [12–14]. Variation in the relative importance of these causes according to the severity of COPD has also been described as lung cancer is the principal cause of death in mild COPD, while respiratory failure is in very severe COPD [15].

In asthma, the possible causes of death are still not well understood, except for the recognized reduced contribution of severe asthma exacerbations. The few studies available are limited to those designed with other epidemiological objectives [16]; others have been conducted in specific population groups (women [17], children [18]), or are based on series of deaths too small to draw strong conclusions [19].

Furthermore, the bronchial and systemic inflammation and the clinical impact of the associated comorbidities in asthma [20–23] have provided evidence for suspecting that, as in COPD [12, 24], asthma patients might be at a greater risk of suffering cardiovascular [25, 26] and cerebrovascular [27] diseases and, therefore at risk of death due to vascular events [17, 19]. However, previous studies exploring this association have reported conflicting results, and few of these have managed to determine a causal relationship between asthma and cardiovascular disease. Also, the method used in some of these studies could be questioned. For example, the diagnosis of asthma was assumed from simple self-reporting by patients [19, 25–27]; and, in others, by their responses collected in a telephonic survey addressed to recognize asthma symptoms [17].

Our study was designed with two objectives: to determine the causes of death of patients with asthma compared with those of patients with COPD; and to establish a possible relationship between deaths from asthma and cardiovascular diseases. In Spain about 68% of deaths occur in hospitals with certain regional geographic variation [28]. We believe that the information collected is representative because most of the Spanish population deaths occur in hospitals.

Observational, retrospective, cross-sectional, multi-centre study conducted in a large group of patients who died while admitted in hospital.

Fieldwork included all deaths in admitted patients during the months of January, April, July and October (months taken as representative of each of the four seasons) of 2008 in thirteen Spanish secondary and tertiary hospitals. The clinical records of all deceased patients were reviewed, and the following data were collected: demographic data; existence of risk factors for cardiovascular diseases (smoking, diabetes, arterial hypertension, obesity, hyperlipidaemia); and specific causes of death, distinguishing specifically lung cancer within neoplastic disorders, the pneumonia within infections; and within cardiovascular disease, peripheral arteriopathy, ischemic cardiopathy, cerebrovascular disease or heart failure. In addition, in patients with asthma, severity and control (both by GINA [6] and GEMA [8] criteria); the frequency of scheduled and unscheduled medical visits, spirometry (best FEV1 recorded) and prescribed treatment were also recorded for each death. The study protocol was approved by a central ethics committee (Hospital of Jerez) following the rules of the Helsinki Declaration.

Previous asthma diagnosis was considered of a high degree of certainty, when the clinical record contained diagnostic confirmatory reports (signed by a pulmonologist or allergy specialist) with objective tests like either significant variability of the spirometric flows performed by applying a bronchodilatator or bronchoconstriction challenge test; alternatively, a label of moderate degree of certainty was given in absence of the above mentioned criteria and any of the following: the patient clinical record reported symptoms compatible with asthma, prescription of asthma drugs or an indication that the attending physician during the hospital admission presumed the asthma diagnosis. Indistinguishable cases from COPD, bronchiectasis or residual pulmonary lesions after tuberculosis co-existed, were excluded from the analysis. A diagnosis of COPD was considered when there was a diagnostic medical report (signed by a pulmonologist) supported by poorly reversible airflow limitation with a post-bronchodilator FEV1/FVC ratio < 0.70 ( GOLD criteria) [29].

The cause of death was determined according to certification that existed in medical history signed by the doctor who treated the patient, no death was excluded. A total of 2,826 deaths were included in the analysis. Of these, 1,583 (56%) were men. The mean ± SD age was 75 ± 14 years (74 ± 14 in men vs 77 ± 15 in women, p = 0.01). The most frequent causes of death observed in the total sample were: 746 (27%) malignancies, of which 173 (23%) were lung cancer; 710 (25.7%) cardiovascular diseases (258 heart failure, 208 cerebrovascular disease, 191 ischemic cardiopathy, and 53 peripheral arterial disease); and 597 infections (21%), of which 332 (54.6%) pneumonia. Farther 251 deaths were observed gastrointestinal origin (9%) and 206 from respiratory failure (72 COPD patients, 5 asthmatics and 131 from other causes and unclear diagnoses).

The main results were not substantially modified when asthmatics were divided according to the degree of certainty of the asthma diagnosis.

The main finding of the present study was that cardiovascular diseases are the most frequent cause of death in patients with asthma. And compared with COPD patients, deaths due to exacerbations and malignancies were less frequent. These findings are particularly relevant given the limited information on this topic.

Our study was designed to include deaths of patients with asthma of any age and both sexes; consequently it probably is more representative to the total population of asthmatics than previous studies carried out in specific sub-groups. For instance, de Marco et al19 evaluated the causes of death of 47 asthmatics aged 20 to 44 years, and found that 34% died due to malignancies and 17% of accidents. Similarly, Camargo et al [17] analyzed deaths of 87 women with asthma (all nurses) aged 30 to 55 years, of whom 31% of deaths were malignancies and 22% cardiovascular diseases.

Another key results from our study comes from the inclusion of deaths of COPD patients. Whereas in asthma cardiovascular diseases constitute the most frequent cause (29.3%), malignancies were the most frequent cause (26.5%) in COPD, likely associated with a higher cumulative tobacco exposure. Similarly, patients with asthma died less frequently than patients with COPD of respiratory failure or during exacerbation of their disease (6.0% vs. 25.5%), due to lung cancer (1.2% against 14.5%) and due to pneumonia (6% against 10.3%). The finding that patients with asthma, in contrast to COPD patients, die relatively infrequently from an exacerbation of their disease, was well-documented in several earlier series. Silverstein et al [16] found in a cohort of 2,499 persons with asthma that only 4% of the 140 deaths recorded during 14 years of follow-up were by an episode of exacerbation. This frequency is very similar (4.6%) out from 87 deaths reported by Camargo et al [17]. Contrary in COPD, serious worsening constitutes one of the most frequent causes of death. In the TORCH study [12] this was determined as the cause of 35% of deaths in 875 deceased COPD patients. Similarly, lung cancer constitutes a frequent cause of death in COPD: 21% in the TORCH study [12], and 38% in the Lung Health Study [13]; as above mentioned, it is not a frequent cause of death in asthma. Finally in terms of differences in the frequency of pneumonia as cause of death in these two diseases. It had been suggested that patients with COPD treated with inhaled corticosteroids have a greater risk of contracting pneumonia, but do not die of it [12, 30], and some authors have suggested a potential beneficial effect, or at least not a deleterious effect, of the outpatient use of inhaled corticosteroid in COPD or asthmatics patients who developed pneumonia [31, 32]. In our series, there is a considerable percentage of patients with COPD who died of pneumonia, but not higher than in the group of patients without previous lung pathology or asthmatics patients.

Paradoxically, since patients with asthma are nowadays more likely to survive with close to normal longevity, it resulted in an increase in prevalence of certain comorbidities associated with the disease having a significant impact on their quality of life and probably also on their specific causes of death [33]. Among others, the cardiovascular diseases are significant, and in particular, cardiac disease and cerebrovascular accidents, especially in the elderly, where their prevalence tends to be similar to that of COPD [8, 20]. Independently of the alleged role by inhaled β2-adrenergic agonist in enhancing cardiac comorbidities and death, both in COPD and asthma [34], an endothelial pathogeny linked to bronchial inflammation has been invoked as possibly responsible for the cardiovascular diseases associated with or coincident with asthma. Some studies have correlated bronchial inflammation with the production of C reactive protein (CRP) in the vascular endothelium of the bronchial tubes, through the stimulation of IL-6. CRP, in turn, has been implicated directly in atherosclerosis and its complications [35, 36]. This would result in a lesion of the vascular endothelium and, in consequence, the development of cardiovascular disease. A similar mechanism has also been suggested in the pathogeny of endothelial lesions of COPD [37]. On the other hand, the possible role of the cysteinyl leukotrienes has been proposed; in addition to constituting one of the principal inflammatory mediators of asthma, these molecules may cause endothelial lesion or dysfunction [38]. Particularly in women, in whom they are promoted by circulating estrogens, cysteinyl leukotrienes could contribute to the greater incidence of cardiovascular diseases observed in these patients. Hence those patients with active asthmatic inflammatory disease, who therefore produce more leukotrienes, would have a greater risk of suffering cardiovascular diseases [39, 40]. The hypothesis is coherent with previous observations particularly in asthmatic women, in whom the association with suffering or dying from cardiovascular, and especially cerebrovascular diseases, appears more consistent [17, 26, 27]. Our results, 78% of female asthma deaths who presented an exceptionally aggressive asthmatic disease support that hypothesis. Of these, 49% were suffering a severe or moderate asthma, 85.5% were not receiving sufficient monitoring of the disease and the mean of previous hospital admissions for asthma was two. All of them were largely modifiable factors. In addition, they had a greater proportion of obese patients (32.8%) and overall a significantly greater BMI (mean of 27.6 kg/cm²), in comparison with the two reference groups as suggested elsewhere [41], it could contribute to a greater risk of accompanying vascular disease. However, the multivariate study carried out did not confirm the association between asthma and death from a cardiovascular disease (OR 0.95; p = 0.9809), nor with any of its variants considered in the study. Perhaps a greater sample size could have identified a statistical association, for example, with death due to heart failure which showed results close to significance, in agreement with Proser et al [42], reporting a strong association of both (OR 4.62).

Although the study is subject to the customary methodological limitations due to the retrospective collection of information from patients clinical records, we consider that, with our choice of study design with two reference populations this potential limitation was lessened. Data quality on causes of death is greater when obtained from hospital charts in comparison when obtained via death certificates of deaths occurring outside the hospital context, in which the nature of the cause of death can often be questionable. Perhaps one strength here is the rigor employed in the assumption of the diagnosis of asthma in all deaths analyzed. Anyway, our conclusions must be tempered by the large number of statistical tests carried out.

To conclude, the most common causes of death of asthmatics in our study were those related to cardiovascular diseases and, unlike COPD patients, there were significantly fewer deaths due to exacerbation and lung cancer. These findings are relevant, as mortality in asthma has been analyzed only partially and mostly within subgroups. We did not observed an association between suffering asthma and death from a cardiovascular disease, but similar studies need to be conducted with larger samples of asthmatics, in which causes of death should be analyzed rigorously. Until such information be available, the present study could provide arguments for adopting a more proactive attitude in routine clinical practice, towards the prevention and treatment of cardiovascular diseases in patients with asthma, particularly in those of advanced age.