Abstract
Rationale: Comorbidities contribute to disease severity and mortality in patients with chronic obstructive pulmonary disease (COPD). Comorbidities have been studied individually and were mostly based on self-reports. The coexistence of objectively identified comorbidities and the role of low-grade systemic inflammation in the pathophysiology of COPD remain to be elucidated.
Objectives: To cluster 13 clinically important objectively identified comorbidities, and to characterize the comorbidity clusters in terms of clinical outcomes and systemic inflammation.
Methods: A total of 213 patients with COPD (FEV1, 51 ± 17% predicted; men, 59%; age, 64 ± 7 yr) were included prospectively. Comorbidities were based on well-known cut-offs identified in the peer-reviewed English literature. Systemic inflammatory biomarkers were determined in all patients. Self-organizing maps were used to generate comorbidity clusters.
Measurements and Main Results: A total of 97.7% of all patients had one or more comorbidities and 53.5% had four or more comorbidities. Five comorbidity clusters were identified: (1) less comorbidity, (2) cardiovascular, (3) cachectic, (4) metabolic, and (5) psychological. Comorbidity clusters differed in health status but were comparable with respect to disease severity. An increased inflammatory state was observed only for tumor necrosis factor (TNF) receptors in the metabolic cluster (geometric mean [lower and upper limit]; TNF-R1, 2,377 [1,850, 3,055] pg/ml, confidence, 98.5%; TNF-R2, 4,080 [3,115, 5,344] pg/ml, confidence, 98.8%) and only for IL-6 in the cardiovascular cluster (IL-6, 3.4 [1.8, 6.6] pg/ml; confidence, 99.8%).
Conclusions: Multimorbidity is common in patients with COPD, and different comorbidity clusters can be identified. Low-grade systemic inflammation is mostly comparable among comorbidity clusters. Increasing knowledge on the interactions between comorbidities increases the understanding of their development and contributes to strategies for prevention or improved treatment.
Comorbidities in patients with chronic obstructive pulmonary disease (COPD) have an important impact on disease severity and survival. Yet, comorbidities have mostly been studied separately, whereas combinations are probably common. Furthermore, most studies used self-reported comorbidity, probably underestimating its true prevalence. Systemic inflammation has been suggested as an etiologic factor in the development of multiple comorbidities, but its role remains to be elucidated.
This study evaluated 13 comorbidities based on validated objective measurements and illustrated the high prevalence of multimorbidity in patients with COPD. Five distinct comorbidity clusters were identified using a nonparametric regression technique. Although disease severity was comparable among comorbidity clusters, it was markedly different in terms of health status. Mostly, systemic inflammatory markers were comparable among the clusters. This study expands the understanding of the cooccurrence of chronic conditions in patients with COPD and sheds new light on the possible cause and pathophysiologic role of systemic inflammation in COPD.
Chronic obstructive pulmonary disease (COPD) is a major global health concern, causing considerable morbidity and mortality around the world. Although defined by the presence of chronic airflow limitation, COPD is nowadays considered a complex, heterogeneous, and multicomponent condition (1). It is increasingly recognized that the presence of other chronic conditions (comorbidities), such as cardiovascular disease, depression, osteoporosis, anemia, and diabetes, substantially contributes to the severity of the disease (2). Comorbidities not only affect symptom burden, functional performance, and health status in patients with COPD (3), but also the risk of hospitalization (4) and mortality (4, 5).
To date, most studies on the impact of comorbidities in COPD used data on self-reported concurrent chronic conditions. This may limit the internal validity of these studies (4–13). Indeed, self-reported data are most probably underestimating the true prevalence of comorbidities in COPD (14). Moreover, most comorbidities have been studied separately, whereas most elderly have two or more chronic morbidities (15). Overlap in self-reported comorbidities has been shown in primary care patients with COPD (11). Although some associations among objectively identified comorbidities, such as vascular stiffness and osteoporosis (16), have been studied, it is currently unknown whether and to what extent comorbidities cluster in COPD. Also, it remains unclear whether or not patients with different comorbidity profiles differ in disease severity, pharmacologic treatment, or other relevant clinical outcomes.
Persistent low-grade systemic inflammation may be the link between COPD and comorbidities (17). Simultaneously elevated inflammatory markers have been associated with an increased risk of comorbidities (18). However, the ECLIPSE study reported that chronic systemic inflammation is a feature in only a small proportion of the patients with COPD (19). Intriguingly, the risk of chronic systemic inflammation was not increased in patients with self-reported cardiovascular comorbidity, whereas older age and obese body mass index (BMI), among other factors, were associated with this feature. These observations shed new light on the role of chronic inflammation in the development of comorbidities in COPD and on the possible cause of systemic inflammation in COPD.
Therefore, the present study investigated the frequency of 13 clinically relevant and objectively identified comorbidities and the clustering of these comorbidities in a well-characterized cohort of patients with COPD. In addition, potential differences in disease severity and clinical characteristics, inflammatory status, and current pharmacologic treatment among comorbidity clusters were explored. Some of the results of this study have been previously reported in the form of an abstract (20).
Please see the online supplement for details and references.
The current analysis was based on data collected as part of the CIRO CO-morbidity (CIROCO) study, an observational single-center study. All subjects provided written informed consent and the study was approved by the local ethics and review boards (MEC 10-3-067). Patients with moderate to very severe COPD (Global Initiative for Chronic Obstructive Lung Disease [GOLD] grades II–IV [2]), aged 40–80 years, and in a clinically stable state were prospectively recruited between November 2007 and November 2010 during the initial evaluation of a comprehensive pulmonary rehabilitation program at CIRO+ (21). Patients with a history of asthma, α1-antitrypsin deficiency, any previous lung surgery, active inflammatory disease, acute myocardial infarction within the last 6 months, any known bone disease other than osteoporosis, current or recently (i.e., <5 yr before the study) treated malignant disease, or use of high-dose systemic glucocorticosteroids (i.e., >10 mg prednisolone) were excluded from the study.
At study entry, patients’ demographics, smoking status, medications, and long-term oxygen use were documented. In addition, the following assessments were performed: lung function (post-bronchodilator spirometry, static lung volumes, and carbon monoxide transfer factor); body composition (BMI, fat free mass index, and bone mineral density at the hip, lumbar spine, and whole-body using dual-energy absorptiometry scan); vascular status (carotid intima-media thickness [c-IMT], cardiac infarction injury score [CIIS] based on a resting 12-lead ECG, and peripheral blood pressure); functional exercise capacity (6-min-walk test); dyspnea (modified Medical Research Questionnaire); and psychological and health status (Hospital Anxiety and Depression Scale [HADS] and St. George’s Respiratory Questionnaire). For evaluation of the generalizability of the results, the (age-adjusted) Charlson comorbidity index was assessed to compare with existing patient cohorts. Venous blood was obtained in fasting state for assessment of hematology, chemistry, and inflammatory status (leukocytes, C-reactive protein [CRP], tumor necrosis factor-α receptors I and II [TNF-RI, TNF-RII], interleukin [IL]-6, and IL-8). Finally, updated-BODE and Framingham risk scores were calculated as prognostic markers.
Thirteen comorbidities were identified in all patients based on predefined cut-offs as suggested by relevant international societies (see online supplement): chronic kidney disease (estimated glomerular filtration rate < 60 ml/min); anemia (hemoglobin level < 8.1 mmol/L in men and <7.5 mmol/L in women); hypertension (systolic blood pressure >140 mm Hg or diastolic pressure >90 mm Hg); obesity (BMI ≥30 kg/m2); underweight (BMI <21 kg/m2); muscle wasting (fat free mass index <16 kg/m2 for men or <15 kg/m2 for women); hyperglycemia (fasting glucose level >5.6 mmol/L); dyslipidemia (triglyceride level >1.7 mmol/L or high-density lipoprotein cholesterol level <1.03 mmol/L in men or <1.29 mmol/L in women); osteoporosis (t-score less than −2.5); symptoms of anxiety and depression (HADS score ≥10 points); atherosclerosis (c-IMT >0.9 mm); and myocardial infarction (CIIS ≥20).
All statistical analyses were performed using Viscovery Profiler 5.3 by Viscovery Software GmbH (www.viscovery.net; Vienna, Austria). Self-organizing maps (SOMs, also referred to as Kohonen maps) were used to create an ordered representation of the comorbidity data. The SOM method can be viewed as a nonparametric regression technique that converts multidimensional data spaces into lower dimensional abstractions. A SOM generates a nonlinear representation of the data distribution and allows the user to identify homogenous data groups visually.
Patients have been ordered by their overall similarity concerning their present comorbidities and also by the degree of its presence given by parameters from which the comorbidities are calculated. Based on the created SOM model, clusters have been generated using the SOM-Ward Cluster algorithm of Viscovery, a hybrid algorithm that applies the classical hierarchical method of Ward on top of the SOM topology. Summary variables on comorbidities and clinical characteristics for the study sample and for each cluster are presented as mean ± standard deviation for quantitative variables, and percentage for discrete variables. Data of the inflammatory markers were log-transformed to mitigate the effects of extreme values. Here, summary variables are presented as geometric mean with lower and upper limit. Viscovery automatically identified for each cluster the comorbidities, clinical characteristics, and inflammatory markers that differ significantly from the average of the whole study sample of 213 patients using the integrated two-sided t test with a confidence of 95%.
Of 255 patients prospectively recruited from patients admitted to CIRO, 42 were ineligible. The final study population consisted of 213 patients with moderate to severe COPD, substantial smoking history, moderately impaired diffusion capacity, and increased static lung volumes (Table 1). In 30 patients, data on one of the selected comorbidities were missing: 13 patients lacked acceptable quality c-IMT measurement, 6 did not fill out the HADS, and 11 lacked an acceptable ECG, impeding to determine the CIIS. Data on the self-reported (age-adjusted) Charlson comorbidity index are in the online supplement (see E1).
| All Subjects (n = 213) | |
| Age, yrs | 63.6 ± 7.0 |
| Male, % | 59 |
| BMI, kg/m2 | 26.2 ± 5.1 |
| FFMI, kg/m2 | 17.0 ± 2.4 |
| mMRC dyspnea grade | 2.1 ± 1.09 |
| Current smoker, % | 28 |
| Pack-years | 46 ± 26 |
| Long-term oxygen therapy, % | 17 |
| FEV1, L | 1.40 ± 0.54 |
| FEV1, % predicted | 51.2 ± 16.9 |
| FEV1/FVC | 0.40 ± 0.11 |
| ITGV, % predicted | 148 ± 33 |
| TLCO, % predicted | 56 ± 17 |
| 6MWD, m | 470 ± 106 |
| SGRQ, total score | 51.3 ± 17.5 |
| Updated BODE score | 2.9 ± 2.5 |
| Framingham 10-yr risk, % | 9.4 ± 6.7 |
All comorbidities were present in the COPD population, although the frequencies ranged from 5–54% (Figure 1). Hyperglycemia, atherosclerosis, hypertension, dyslipidemia, and osteoporosis were the five most prevalent comorbidities.
Almost all subjects (97.7%) had one or more comorbidities and more than half of the patients had at least four comorbidities (Figure 2). The frequencies of different individual comorbidities in patients with each of the 13 specific comorbidities are shown in Figure 3. For example, muscle wasting and osteoporosis were highly prevalent in underweight patients, whereas they were almost absent in obesity.
Figure 3. The frequencies of objectively identified comorbidities in patients with chronic obstructive pulmonary disease (COPD) with each of the 13 selected specific comorbidities. In subjects in which the comorbidity mentioned in the row is present, the prevalence of the other comorbidities mentioned in the column is shown. For interpretation, the table is colored: blue, less than 20% prevalence; green, 20–40% prevalence; yellow, 40–60% prevalence; and red, more than 60% prevalence.
[More] [Minimize]Five clusters with a significantly different comorbidity profile were identified (Figure 4, Table 2). Cluster 1 (less comorbidity cluster) had significantly fewer comorbidities compared with the other clusters. Cluster 2 (cardiovascular cluster) had a significantly higher prevalence of hypertension and atherosclerosis compared with the other clusters. However, dyslipidemia, hyperglycemia, underweight, and muscle wasting were less prevalent compared with the other clusters. Cluster 3 (cachectic cluster) had a higher prevalence of underweight, muscle wasting, osteoporosis, and renal impairment, whereas obesity and atherosclerosis were less often identified. This cluster 3 had a high number of comorbidities. Cluster 4 (metabolic cluster) also had a higher number of comorbidities. This cluster contained more patients with obesity, atherosclerosis, dyslipidemia, hyperglycemia, and hypertension, but fewer patients with anxiety, underweight, muscle wasting, and osteoporosis. Finally, cluster 5 (psychological cluster) had the highest proportion of patients with psychological disorders (anxiety and depression), but also the highest prevalence of myocardial infarction. A detailed description of the five clusters is presented in Table 2.
| Comorbidities | Cluster 1: Less Comorbidity | Cluster 2: Cardiovascular | Cluster 3: Cachectic | Cluster 4: Metabolic | Cluster 5: Psychologic |
| N | 67 | 49 | 44 | 33 | 20 |
| Number of comorbidities | 2.5 ± 1.4* | 3.8 ± 1.7 | 4.2 ± 1.4† | 4.4 ± 1.1† | 4.1 ± 1.8 |
| Renal impairment, % | 16 | 24 | 45† | 9 | 5 |
| Anemia, % | 9 | 4 | 2 | 3 | 5 |
| Hypertension, % | 3* | 98† | 43 | 100† | 5* |
| Obesity, % | 30 | 14 | 0* | 61† | 15 |
| Underweight, % | 0* | 0* | 66† | 3* | 0 |
| Muscle wasting, % | 12* | 10* | 98† | 0* | 20 |
| Hyperglycemia, % | 52 | 41* | 43 | 91† | 60 |
| Dyslipidemia, % | 42 | 16* | 25 | 67† | 40 |
| Osteoporosis, % | 27 | 37 | 52† | 0* | 35 |
| Anxiety, % | 5* | 28 | 26 | 0* | 84† |
| Depression, % | 6* | 23 | 7 | 6 | 68† |
| Atherosclerosis, % | 56 | 67† | 12* | 81† | 53 |
| Myocardial infarction, % | 2* | 11 | 7 | 13 | 32† |
Figure 4. Multimorbidity clusters in chronic obstructive pulmonary disease (COPD). These panels were generated using Viscovery (Vienna, Austria) software. The Viscovery program placed all subjects on a specific position on the map based on their profile of comorbidities. The more subjects resemble in terms of their comorbidity the closer they are on the map. Contrarily, the more they differ the further they are away from each other. When looking at a comorbidity, subjects “raise a red flag” if the comorbidity is present and “a blue flag” when absent. In this way the maps can be interpreted. By drawing lines on the map, the Viscovery program could identify five different clusters of patients with COPD with a significant different profile of comorbidities (95% confidence interval).
[More] [Minimize]FEV1 (% predicted), smoking history, functional exercise capacity, long-term oxygen use, and updated BODE scores were not different among the comorbidity clusters (Table 3).
| Clinical Characteristics | Cluster 1: Less Comorbidity | Cluster 2: Cardiovascular | Cluster 3: Cachectic | Cluster 4: Metabolic | Cluster 5: Psychologic |
| N | 67 | 49 | 44 | 33 | 20 |
| Age, yrs | 62.1 ± 6.8* | 67.2 ± 5.8† | 62.5 ± 7.2 | 63.1 ± 7.3 | 62.8 ± 6.8 |
| Male, % | 60 | 65 | 43* | 79† | 45 |
| mMRC dyspnea grade | 1.99 ± 1.01 | 2.29 ± 1.21 | 1.73 ± 0.9* | 2.12 ± 1.11 | 2.84 ± 1.12† |
| Current smoker, % | 30 | 16* | 45† | 15 | 35 |
| Pack-years | 44 ± 20 | 45 ± 26 | 49 ± 30 | 51 ± 34 | 42 ± 16 |
| LTOT, % | 13 | 18 | 18 | 15 | 25 |
| 6MWD, m | 474 ± 102 | 446 ± 133 | 496 ± 101 | 473 ± 91 | 459 ± 74 |
| FEV1, % predicted | 52.7 ± 17.4 | 50.9 ± 17.7 | 48.3 ± 16.3 | 54.2 ± 16 | 48.3 ± 15.4 |
| ITGV, % predicted | 143 ± 33 | 148 ± 29 | 166 ± 34† | 134 ± 33* | 146 ± 28 |
| TLCO, % predicted | 60 ± 16† | 57 ± 18 | 44 ± 13* | 60 ± 14 | 55 ± 14 |
| SGRQ total score | 47.6 ± 15.3* | 56.5 ± 17.2† | 45.8 ± 19.4* | 49.9 ± 16.1 | 65.9 ± 12.5† |
| SGRQ symptoms, score | 49.1 ± 18.1* | 58.8 ± 20.7 | 55.5 ± 23.1 | 52.8 ± 20.2 | 69.9 ± 14.4† |
| SGRQ activity, score | 68.3 ± 20.2 | 70.2 ± 22.0 | 60.4 ± 24.9* | 66.4 ± 20.5 | 83.5 ± 13.9† |
| SGRQ impact, score | 36.3 ± 17.9 | 43.6 ± 21.2 | 35.1 ± 21.5 | 39.6 ± 18.1 | 52.6 ± 16.6† |
| Updated BODE score | 2.4 ± 2.6 | 3.4 ± 3.3 | 3.0 ± 1.8 | 2.6 ± 2.3 | 3.1 ± 1.9 |
| Framingham 10-yr risk, % | 8.6 ± 6.6 | 11.5 ± 6.6† | 7.6 ± 6* | 11.9 ± 7.3† | 6.6 ± 4.5 |
However, patients in cluster 1 (less comorbidity cluster) had a lower mean age, a higher mean diffusion capacity, and a better mean health-related quality of life. Moreover, a higher proportion of patients used cholesterol-lowering drugs, antiaggregates, or oral antidiabetics (see Table E2). Cluster 2 (cardiovascular cluster) included older patients who were less frequently active smokers but with a worse quality of life and a higher predicted cardiovascular risk according to the Framingham risk score. Cluster 3 (cachectic cluster) had a higher proportion of women and active smokers. In addition, patients in this cluster had on average more static hyperinflation and lower diffusion capacity. Surprisingly, they also had on average a lower dyspnea score, a better health-related quality of life, and a lower cardiovascular risk prediction score. In line with this, a lower proportion of patients in this cluster used cholesterol-lowering drugs or antihypertensive therapy, in particular, angiotensin-converting enzyme inhibitors or angiotensin receptor blockers. Patients in cluster 4 (metabolic cluster) were more frequently male, had less static hyperinflation, and had on average a higher cardiovascular risk prediction score. Also, the proportion of patients using long-acting β-agonists and inhaled corticosteroid inhalers or short-acting rescue medication was lower in this cluster (see Table E3). Patients in cluster 5 (psychological cluster) reported on average more dyspnea and a worse health-related quality of life. Patients in this cluster more frequently used benzodiazepines and nitrates, as also short-acting “rescue” bronchodilators.
Table 4 shows circulating levels of inflammatory markers in the five identified clusters. Figure E1 shows the data point distribution of all inflammatory markers for each cluster. There were no significant differences in CRP, IL-8, or leukocytes among comorbidity clusters. TNF-R1 and TNF-R2 were significantly increased in the metabolic cluster compared with the whole COPD population. TNF-R1 was lower in the cachectic cluster and IL-6 was increased in the cardiovascular cluster.
| Cluster 1: Less Comorbidity | Cluster 2: Cardiovascular | Cluster 3: Cachectic | Cluster 4: Metabolic | Cluster 5: Psychologic | |
| CRP, ng/ml | 2,286 (844, 6,188) | 3,380 (947, 12,062) | 2,005 (677, 5,938) | 3,860 (1,073, 13,886) | 2,519 (767, 8,283) |
| IL-6, pg/ml | 2.4 (1.3, 4.3) | 3.4 (1.8, 6.6)* | 2.2 (1.1, 4.7) | 2.7 (1.6, 4.5) | 2.2 (1.3, 3.6) |
| IL-8, pg/ml | 12.3 (8.2, 18.6) | 12.9 (9.3, 17.9) | 12.1 (7.8, 18.7) | 10.8 (7.6, 15.2) | 11.1 (6.6, 18.7) |
| TNF-R1, pg/ml | 2,013 (1,508, 2,689) | 2,229 (1,513, 3,285) | 1,896 (1,434, 2,505)† | 2,377 (1,850, 3,055)* | 2,133 (1,685, 2,699) |
| TNF-R2, pg/ml | 3,417 (2,454, 4,758) | 3,698 (2,399, 5,701) | 3,302 (2,478, 4,401) | 4,080 (3,115, 5,344)* | 3,419 (2,675, 4,371) |
| Leukocytes, ×109/L | 7.3 (5.6, 9.5) | 7.1 (5.5, 9.4) | 7.0 (5.3, 9.1) | 7.2 (5.9, 8.7) | 7.3 (6.0, 8.9) |
In a cohort of moderate to very severe patients with COPD, five clusters of clinically important and objectively diagnosed comorbidities were identified. Although disease severity was comparable among comorbidity clusters, they were markedly different in terms of health status. Mostly, inflammatory markers were comparable among the clusters, although increased levels of TNF receptors were observed in the metabolic cluster and increased levels of IL-6 in the cardiovascular cluster.
Multiple studies have shown that comorbid chronic conditions occur more frequently in patients with COPD than in subjects without COPD and contribute to adverse clinical outcomes in these patients (1, 4–13). Moreover, the frequency of comorbidities was shown to be independent of the degree of airflow limitation (1, 8). In the present study, comorbidity clusters also had similar degree of airflow limitation.
Previous studies used self-reported comorbidities (1, 4–13) or healthcare databases that were not specifically designed to evaluate comorbidities (4, 9, 10, 13). These methods most probably underestimate the true prevalence of comorbidities in COPD. Here, 13 clinically relevant and well-known comorbidities were diagnosed using predefined internationally accepted cut-offs in a moderately sized, well-characterized sample of patients with COPD. This study confirms the high frequency of individual comorbidities reported previously.
In the general population, there is a growing public attention to the concept of multimorbidity, because the presence of multiple chronic morbidities is associated with poor outcome and increased health care use (22). Only a few studies have investigated the coexistence of two or more comorbidities in individuals with COPD. For example, Siebeling and colleagues (11) reported an overlap in (self-reported) diabetes, musculoskeletal conditions, and cardiovascular disease in primary care patients with COPD. The present study is the first to identify clusters of objectively identified comorbidities in patients with COPD. The unbiased approach to identify these comorbidity clusters is a major strength of the current analyses. Indeed, the use of self-organizing maps may be the most compact way to represent a data distribution, because data dependency can be understood easily if one is familiar with the map visualization.
Cluster analyses have been used before in patients with COPD (23–26). Only two studies (24, 25) used comorbidities in their cluster analyses, which also included all types of other clinical characteristics. Both studies identified a comorbidity-predominant subtype of patients with milder respiratory status but a higher prevalence of obesity, cardiovascular disease, and diabetes. In the PAC-COPD cohort, also higher levels of systemic inflammatory markers were identified in that cluster (24). These clusters seem to resemble the metabolic cluster in the present study.
In total, five distinct clusters were recognized based on the comorbidity profiles. Patients in the cachectic cluster had a lower diffusion capacity and more static hyperinflation compared with the other clusters, which probably is consistent with a higher amount of emphysema. Previously, osteoporosis and renal impairment (27, 28) were associated with emphysema, and osteoporosis was linked to muscle wasting (29). This suggests that there may be a common pathophysiologic pathway for the cooccurrence of these chronic conditions.
Contrary to the cachectic cluster, a metabolic cluster was identified. These patients had less severe pulmonary impairment compared with the other clusters, but more pronounced metabolic disorders and a low-grade systemic inflammation. The identification of these two distinct clusters is reminiscent of the historical description of the pink puffer (emphysematous type with a cachectic impression) and the blue bloater (chronic bronchitis type with a metabolic impression) (30).
Despite comparable increased cardiovascular risk according to the Framingham risk score, the cardiovascular cluster, characterized by increased blood pressure and “subclinical” atherosclerosis, was clearly different from the metabolic cluster. The former cluster had significantly less obesity, dyslipidemia, or hyperglycemia, and a different inflammatory marker is increased (IL-6 in the cardiovascular cluster, compared with TNF receptors in the metabolic cluster). This suggests that different pathways may be involved in the development of cardiovascular comorbidity in COPD.
A large subgroup of patients with significantly less comorbidity was clustered. These patients were younger and a higher proportion was treated with oral antidiabetics and cholesterol-lowering drugs. This may partially explain why these patients had less comorbidity.
The psychologic cluster, including a high proportion of patients with increased symptoms of anxiety and depression, and ischemic heart injury, is considered an important finding. Indeed, anxiety was reported to be the strongest independent predictor of mortality in patients with COPD (5). The (unbiased) recognition of a psychopathologic cluster is consistent with growing literature as mental conditions are recognized as a relevant factor of multimorbidity in the general population (31). The high prevalence of anxiety and depression after myocardial infarction may explain the high proportion of patients with ECG abnormalities in this cluster (32). Whether and to what extent the higher prevalence of ECG abnormalities in anxious patients with COPD may explain the worse prognosis in these patients remains unknown.
A direct link between (chronic low-grade) systemic inflammation and comorbidities in patients with COPD remains equivocal. For example, bivariate associations among various systemic inflammatory biomarkers and pulmonary cachexia (33), subclinical atherosclerosis (34), metabolic syndrome (35), and ischemic heart disease have been reported in patients with COPD (36). Moreover, CRP, fibrinogen, and leukocyte count were associated with increased risk of comorbidities in patients with COPD (18). Then again, many studies showed comparable circulating levels of inflammatory biomarkers in patients with COPD with and without comorbidities, such as cachexia (37), osteoporosis (38), cardiovascular disease (39), and depression (40). So, a causal relationship between low-grade systemic inflammation and comorbidities in COPD has not yet been proved. The current findings also show that the possible interactions between biomarkers of systemic inflammation and comorbidities in patients with COPD are very complex, if present at all.
Chronic low-grade systemic inflammation is a hallmark of obesity in the general population (41), suggesting a systemic origin of inflammation. Increased levels of systemic inflammation have also been reported in patients with COPD, in particular in patients who are obese (19, 42–44). This may partially explain the low-grade systemic inflammation in the metabolic cluster, and its lesser presence in the cachectic cluster. Similarly, the increased IL-6 in the cardiovascular cluster might be caused by the significant older age in that cluster. Indeed, a wealth of data indicates that normal aging is associated with low-grade systemic inflammation, including IL-6 (45).
The present study included GOLD grade II–IV patients. Therefore, the current finding should not be extrapolated to GOLD grade I patient subsets. Moreover, patients were recruited in a tertiary care pulmonary rehabilitation setting, which may also limit the external validity of the current findings. However, the following arguments do support the generalizability of our findings to a broader population. The prevalence of objectively identified renal impairment in our study was similar to results from patients with COPD attending the pulmonary medicine outpatient facilities of 15 centers located throughout Italy (46). Furthermore, the Charlson comorbidity index score in the present study (1.6 ± 0.9 points; age-adjusted, 3.5 ± 1.3 points) was comparable with the PAC-COPD cohort (2 ± 1.3 points) (24), and the ESMI cohort (3.1 ± 2 points), respectively (47). Furthermore, the prevalence of self-reported cardiovascular comorbidities (as recorded in the Charlson comorbidity index) in the present study was comparable with previous studies (1, 5, 19). Thus, the prevalence of comorbidities in the present study seems comparable with data obtained in the secondary outpatient care setting.
Another limitation considering comorbidities might be that patients with active malignant diseases, active inflammatory diseases, or recent myocardial infarction were excluded from participation in the study because these patients are not suitable for pulmonary rehabilitation.
Although the elected biomarkers in the present study are well studied in COPD, it should be mentioned that the concept of systemic inflammation is more complex than that which can be captured in a short series of biomarkers. The strength of measuring comorbidities in an objective manner brings with it the limitation of having evaluated only a limited number of comorbidities. Because of logistical constraints, chronic heart failure, obstructive sleep apnea, and degenerative joint disease were not assessed, although it is known that these conditions are common in COPD (5). Also, it is desirable to validate the current findings in an external replication cohort. In addition, it would be interesting to evaluate clusters of comorbidities in subjects with another index disease (e.g., diabetes mellitus) and in the general population. Furthermore, the longitudinal stability of comorbidity clusters and the prognostic outcome of the different clusters are interesting future research questions.
This is the first study that objectively identified multiple comorbidities in a well-characterized cohort of patients with COPD. The self-organizing maps are new in the field of COPD research and provide easy-to-use graphs of the five independent comorbidity clusters that were identified. The current findings will increase the awareness among physicians that multiple comorbidities can be present in patients with COPD. This will also enhance treatment strategies and facilitate the combination of future treatment guidelines for different comorbidities. Interestingly, the degree of airflow limitation, exercise capacity, and score on the updated BODE index were similar among the comorbidity clusters. This emphasizes the fact that comorbidities are an additional clinical attribute in patients with COPD, which cannot be predicted by the aforementioned clinical outcomes. In addition, low-grade systemic inflammation was mostly comparable among comorbidity clusters. This shows that the presumed association between systemic inflammation and comorbidities in patients with COPD is more complex than assumed at present.
| 1. | Agusti A, Calverley PM, Celli B, Coxson HO, Edwards LD, Lomas DA, MacNee W, Miller BE, Rennard S, Silverman EK, et al.. Characterisation of COPD heterogeneity in the ECLIPSE cohort. Respir Res 2010;11:122. |
| 2. | Vestbo J, Hurd SS, Agusti AG, Jones PW, Vogelmeier C, Anzueto A, Barnes PJ, Fabbri LM, Martinez FJ, Nishimura M, et al.. Global strategy for the diagnosis, management and prevention of chronic obstructive pulmonary disease, GOLD executive summary. Am J Respir Crit Care Med 2013;187:347–365. |
| 3. | von Leupoldt A, Taube K, Lehmann K, Fritzsche A, Magnussen H. The impact of anxiety and depression on outcomes of pulmonary rehabilitation in patients with COPD. Chest 2011;140:730–736. |
| 4. | Mannino DM, Thorn D, Swensen A, Holguin F. Prevalence and outcomes of diabetes, hypertension and cardiovascular disease in COPD. Eur Respir J 2008;32:962–969. |
| 5. | Divo M, Cote C, de Torres JP, Casanova C, Marin JM, Pinto-Plata V, Zulueta J, Cabrera C, Zagaceta J, Hunninghake G, et al.. Comorbidities and risk of mortality in patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2012;186:155–161. |
| 6. | McGarvey LP, John M, Anderson JA, Zvarich M, Wise RA. Ascertainment of cause-specific mortality in COPD: operations of the TORCH clinical endpoint committee. Thorax 2007;62:411–415. |
| 7. | Antonelli Incalzi R, Fuso L, De Rosa M, Forastiere F, Rapiti E, Nardecchia B, Pistelli R. Co-morbidity contributes to predict mortality of patients with chronic obstructive pulmonary disease. Eur Respir J 1997;10:2794–2800. |
| 8. | Crisafulli E, Costi S, Luppi F, Cirelli G, Cilione C, Coletti O, Fabbri LM, Clini EM. Role of comorbidities in a cohort of patients with COPD undergoing pulmonary rehabilitation. Thorax 2008;63:487–492. |
| 9. | Feary JR, Rodrigues LC, Smith CJ, Hubbard RB, Gibson JE. Prevalence of major comorbidities in subjects with COPD and incidence of myocardial infarction and stroke: a comprehensive analysis using data from primary care. Thorax 2010;65:956–962. |
| 10. | Schnell KM, Weiss CO, Lee T, Krishnan JA, Leff B, Wolff JL, Boyd C. The prevalence of clinically-relevant comorbid conditions in patients with COPD: a cross-sectional study using data from NHANES 1999–2008. BMC Pulm Med 2012;12:26. |
| 11. | Siebeling L, Puhan MA, Muggensturm P, Zoller M, Ter Riet G. Characteristics of Dutch and Swiss primary care COPD patients: baseline data of the ice cold ERIC study. Clin Epidemiol 2011;3:273–283. |
| 12. | van Manen JG, Bindels PJ, Ijzermans CJ, van der Zee JS, Bottema BJ, Schade E. Prevalence of comorbidity in patients with a chronic airway obstruction and controls over the age of 40. J Clin Epidemiol 2001;54:287–293. |
| 13. | Soriano JB, Visick GT, Muellerova H, Payvandi N, Hansell AL. Patterns of comorbidities in newly diagnosed COPD and asthma in primary care. Chest 2005;128:2099–2107. |
| 14. | Vanfleteren LE, Franssen FM, Uszko-Lencer NH, Spruit MA, Celis M, Gorgels AP, Wouters EF. Frequency and relevance of ischemic electrocardiographic findings in patients with chronic obstructive pulmonary disease. Am J Cardiol 2011;108:1669–1674. |
| 15. | Barnett K, Mercer SW, Norbury M, Watt G, Wyke S, Guthrie B. Epidemiology of multimorbidity and implications for health care, research, and medical education: a cross-sectional study. Lancet 2012;380:37–43. |
| 16. | Sabit R, Bolton CE, Edwards PH, Pettit RJ, Evans WD, McEniery CM, Wilkinson IB, Cockcroft JR, Shale DJ. Arterial stiffness and osteoporosis in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2007;175:1259–1265. |
| 17. | Barnes PJ, Celli BR. Systemic manifestations and comorbidities of COPD. Eur Respir J 2009;33:1165–1185. |
| 18. | Thomsen M, Dahl M, Lange P, Vestbo J, Nordestgaard BG. Inflammatory biomarkers and comorbidities in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2012;186:982–988. |
| 19. | Agusti A, Edwards LD, Rennard SI, MacNee W, Tal-Singer R, Miller BE, Vestbo J, Lomas DA, Calverley PM, Wouters E, et al.. Persistent systemic inflammation is associated with poor clinical outcomes in COPD: a novel phenotype. PLoS ONE 2012;7:e37483. |
| 20. | Vanfleteren LE, Spruit MA, Franssen FM, Groenen M, Wouters EFM. Objectified co-morbidities in patients with COPD: the CIROCO study. Am J Respir Crit Care Med 2012;185:A3674. |
| 21. | Spruit MA, Vanderhoven-Augustin I, Janssen PP, Wouters EF. Integration of pulmonary rehabilitation in COPD. Lancet 2008;371:12–13. |
| 22. | Tooth L, Hockey R, Byles J, Dobson A. Weighted multimorbidity indexes predicted mortality, health service use, and health-related quality of life in older women. J Clin Epidemiol 2008;61:151–159. |
| 23. | Burgel PR, Paillasseur JL, Caillaud D, Tillie-Leblond I, Chanez P, Escamilla R, Court-Fortune I, Perez T, Carre P, Roche N. Clinical COPD phenotypes: a novel approach using principal component and cluster analyses. Eur Respir J 2010;36:531–539. |
| 24. | Garcia-Aymerich J, Gomez FP, Benet M, Farrero E, Basagana X, Gayete A, Pare C, Freixa X, Ferrer J, Ferrer A, et al.. Identification and prospective validation of clinically relevant chronic obstructive pulmonary disease (COPD) subtypes. Thorax 2011;66:430–437. |
| 25. | Burgel PR, Paillasseur JL, Peene B, Dusser D, Roche N, Coolen J, Troosters T, Decramer M, Janssens W. Two distinct chronic obstructive pulmonary disease (COPD) phenotypes are associated with high risk of mortality. PLoS ONE 2012;7:e51048. |
| 26. | Burgel PR, Roche N, Paillasseur JL, Tillie-Leblond I, Chanez P, Escamilla R, Court-Fortune I, Perez T, Carre P, Caillaud D. Clinical COPD phenotypes identified by cluster analysis: validation with mortality. Eur Respir J 2012;40:495–496. |
| 27. | Chandra D, Stamm JA, Palevsky PM, Leader JK, Fuhrman CR, Zhang Y, Bon J, Duncan SR, Branch RA, Weissfeld J, et al.. The relationship between pulmonary emphysema and kidney function in smokers. Chest 2012;142:655–662. |
| 28. | Bon J, Fuhrman CR, Weissfeld JL, Duncan SR, Branch RA, Chang CC, Zhang Y, Leader JK, Gur D, Greenspan SL, et al.. Radiographic emphysema predicts low bone mineral density in a tobacco-exposed cohort. Am J Respir Crit Care Med 2011;183:885–890. |
| 29. | Bolton CE, Ionescu AA, Shiels KM, Pettit RJ, Edwards PH, Stone MD, Nixon LS, Evans WD, Griffiths TL, Shale DJ. Associated loss of fat-free mass and bone mineral density in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2004;170:1286–1293. |
| 30. | Filley GF, Beckwitt HJ, Reeves JT, Mitchell RS. Chronic obstructive bronchopulmonary disease: II. Oxygen transport in two clinical types. Am J Med 1968;44:26–38. |
| 31. | Kirchberger I, Meisinger C, Heier M, Zimmermann AK, Thorand B, Autenrieth CS, Peters A, Ladwig KH, Doring A. Patterns of multimorbidity in the aged population: results from the KORA-age study. PLoS ONE 2012;7:e30556. |
| 32. | Hanssen TA, Nordrehaug JE, Eide GE, Bjelland I, Rokne B. Anxiety and depression after acute myocardial infarction: an 18-month follow-up study with repeated measures and comparison with a reference population. Eur J Cardiovasc Prev Rehabil 2009;16:651–659. |
| 33. | Di Francia M, Barbier D, Mege JL, Orehek J. Tumor necrosis factor-alpha levels and weight loss in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1994;150:1453–1455. |
| 34. | Iwamoto H, Yokoyama A, Kitahara Y, Ishikawa N, Haruta Y, Yamane K, Hattori N, Hara H, Kohno N. Airflow limitation in smokers is associated with subclinical atherosclerosis. Am J Respir Crit Care Med 2009;179:35–40. |
| 35. | Watz H, Waschki B, Kirsten A, Muller KC, Kretschmar G, Meyer T, Holz O, Magnussen H. The metabolic syndrome in patients with chronic bronchitis and COPD: frequency and associated consequences for systemic inflammation and physical inactivity. Chest 2009;136:1039–1046. |
| 36. | Sin DD, Man SF. Why are patients with chronic obstructive pulmonary disease at increased risk of cardiovascular diseases? The potential role of systemic inflammation in chronic obstructive pulmonary disease. Circulation 2003;107:1514–1519. |
| 37. | Broekhuizen R, Grimble RF, Howell WM, Shale DJ, Creutzberg EC, Wouters EF, Schols AM. Pulmonary cachexia, systemic inflammatory profile, and the interleukin 1beta -511 single nucleotide polymorphism. Am J Clin Nutr 2005;82:1059–1064. |
| 38. | Graat-Verboom L, Spruit MA, van den Borne BE, Smeenk FW, Martens EJ, Lunde R, Wouters EF. Correlates of osteoporosis in chronic obstructive pulmonary disease: an underestimated systemic component. Respir Med 2009;103:1143–1151. |
| 39. | Mills NL, Miller JJ, Anand A, Robinson SD, Frazer GA, Anderson D, Breen L, Wilkinson IB, McEniery CM, Donaldson K, et al.. Increased arterial stiffness in patients with chronic obstructive pulmonary disease: a mechanism for increased cardiovascular risk. Thorax 2008;63:306–311. |
| 40. | Quint JK, Baghai-Ravary R, Donaldson GC, Wedzicha JA. Relationship between depression and exacerbations in COPD. Eur Respir J 2008;32:53–60. |
| 41. | Goossens GH. The role of adipose tissue dysfunction in the pathogenesis of obesity-related insulin resistance. Physiol Behav 2008;94:206–218. |
| 42. | Poulain M, Doucet M, Drapeau V, Fournier G, Tremblay A, Poirier P, Maltais F. Metabolic and inflammatory profile in obese patients with chronic obstructive pulmonary disease. Chron Respir Dis 2008;5:35–41. |
| 43. | Breyer MK, Spruit MA, Celis AP, Rutten EP, Janssen PP, Wouters EF. Highly elevated c-reactive protein levels in obese patients with COPD: a fat chance? Clin Nutr 2009;28:642–647. |
| 44. | Rutten EP, Breyer MK, Spruit MA, Hofstra T, van Melick PP, Schols AM, Wouters EF. Abdominal fat mass contributes to the systemic inflammation in chronic obstructive pulmonary disease. Clin Nutr 2010;29:756–760. |
| 45. | Ershler WB, Keller ET. Age-associated increased interleukin-6 gene expression, late-life diseases, and frailty. Annu Rev Med 2000;51:245–270. |
| 46. | Incalzi RA, Corsonello A, Pedone C, Battaglia S, Paglino G, Bellia V. Chronic renal failure: a neglected comorbidity of COPD. Chest 2010;137:831–837. |
| 47. | Almagro P, Cabrera FJ, Diez J, Boixeda R, Ortiz B, Murio C, Soriano JB. Comorbidities and short-term prognosis in patients hospitalized for acute exacerbation of COPD: the Epoc en Servicios de Medicina Interna (ESMI) study. Chest 2012;142:1126–1133. |
Supported by a research grant from Astra-Zeneca.
Author Contributions: Conception and design, L.E.G.W.V., M.A.S., P.L.B.B., E.F.M.W., F.M.E.F. Drafting of the manuscript, L.E.G.W.V., M.A.S., S.G., F.M.E.F. Acquisition and analysis of data, L.E.G.W.V., M.A.S., M.G., S.G., V.P.M.v.E., E.P.A.R., J.O.tR., F.M.E.F. Analysis and interpretation of data, L.E.G.W.V., M.A.S., S.G., F.M.E.F. Drafting the manuscript for important intellectual content, all authors. All coauthors critically revised the article and gave final approval of this version to be published.
This article has an online supplement, which is accessible from this issue’s table of contents at www.atsjournals.org
Originally Published in Press as DOI: 10.1164/rccm.201209-1665OC on February 7, 2013
Author disclosures
