Clinical guidelines for chronic obstructive pulmonary disease (COPD) point to depression and anxiety as key comorbidities.1,2 These conditions, associated with higher mortality and more frequent and longer hospital stays for exacerbations,3 are a main determinant of the quality of life of patients with COPD.4

Although current guidelines recommended that patients with COPD undergo specific psychosocial assessment,1,2,5 anxiety and depression frequently go undiagnosed and untreated.3,6 A review article published as recently as 20143 discussed the methodological limitations affecting scientific studies of this association, pointing, among other factors, to the failure to obtain an accurate diagnosis of anxiety or depression and the absence of anxiety or depression as primary endpoints. Furthermore, both prevalence of COPD-related anxiety and depression and patient characteristics differ according to geographical region.7,8 Better designed studies are clearly needed to pinpoint the relationship between COPD and psychiatric comorbidities.9

According to the PubMed database, 6 articles have been published in the last 10years referring to anxiety and depression in patients with COPD in Spain: 4 confirmed the existence of psychiatric comorbidities on the basis of different symptom-related questionnaires10–13; and the other 2 were population studies where COPD, anxiety and depression diagnoses were established based on medical records and patient surveys.14,15 Just 4 of these studies referred to stable patients and just 1 was specifically designed to explore depression in patients with COPD.13

Our aim was to investigate the presence of anxiety or depression in a sample of Spanish patients with stable COPD in order to developing a predictive model of the association between COPD and these comorbidities.

A total of 204 patients were included, 18 (9%) of whom had a previous psychiatric diagnosis of anxiety or depression. Patient characteristics are summarized in Table 1.

Mean HADS scores were 10.98±7.19 overall, 6.26±4.08 for the anxiety subscale and 4.77±3.59 for the depression subscale. The HADS pointed to significant symptoms of anxiety or depression in 79 patients (39%). Of these 79 patients, 15 had been diagnosed prior to inclusion in the study and so were considered confirmed cases; the remaining 64 patients underwent psychiatric evaluation and a diagnosis of depression or anxiety was confirmed in 56 cases. In total, therefore, 74 study patients (36%) had anxiety or depression confirmed by psychiatric diagnosis (Fig. 1) and a fifth of these had a mixed anxiety-depression disorder. Of the total sample of 204 patients, 19% had pure anxiety and 9.8% had pure depression (Fig. 2).

Fig. 1.

Patients with chronic obstructive pulmonary disease and a firm diagnosis of anxiety or depression. Considered as confirmed cases were 18 patients diagnosed prior to inclusion plus 56 patients diagnosed by psychiatric assessment after inclusion.

(0.07MB).

Fig. 2.

Psychiatric diagnoses in patients with stable chronic obstructive pulmonary disease.

(0.06MB).

According to the bivariate analysis (Table 2), patients with anxiety and depression had a higher educational level, lacked home support, were younger and occupationally active and were single, separated or divorced. They had also experienced more exacerbations in the previous year, had higher baseline dyspnoea, covered less distance in the T6MM and scored higher on the BODE index. To a statistically significant degree, anxiety and depression were more frequent in patients treated with 7 or more drugs and in patients with more limited autonomy. The presence of more than 4 comorbidities associated with COPD was not statistically significant (P=.08), nor were statistically significant differences found regarding sex, smoking, alcoholism, presence of cardiovascular comorbidities, FEV1, FVC, FEV1/FVC, hospital admissions for exacerbations or use of home oxygen therapy. A higher proportion of patients with stage I and stage IV airflow obstruction according to the GOLD classification2 presented anxiety and depression (P<.009). Anxiety and depression were also more prevalent in the more symptomatic patients versus the less symptomatic patients (groups D and B versus groups A and C, respectively; P=.065) (Table 3).

For the multivariate analysis, which included all significant variables from the bivariate analysis, an equation each was calculated for the overall BODE index score and for that of individual BODE index parameters (Table 4). Of the 2 resulting models, the overall BODE index offered a better ROC curve value, for an AUC of 0.765 (P<.001), contrasting with an AUC of 0.753 (P<.001) for the individual BODE index parameters. Of those parameters, only a higher level of dyspnea was a risk factor. The logistic regression model for the BODE index had specificity and sensitivity values of 90% and 54%, respectively.

Age in the two models was highly significant (P<.001), for an estimated odds ratio of 0.93. Also significant in both models were frequent exacerbations, a higher education level and lack of home support.

We investigated a sample of 204 Spanish patients with COPD for anxiety or depression. Our sample reflected the profile of patients typically seen in a Spanish pulmonology clinic and, except for the high smoking rate, also reflected the characteristics of the INSEPOC Spanish multicentre population study.14,25 We estimated prevalence of anxiety and depression in patients with COPD, confirmed by psychiatric diagnosis, to be between 30% and 43% (95% CI). Depression was detected in 17% (95% CI: 12%–22%) and anxiety in 26.5% (95% CI: 21%–33%) of the patients in our study. Of the patients with psychiatric comorbidities, 20% (15/74 patients) had a mixed anxiety-depression disorder.

A recent literature review of anxiety in COPD identified studies designed on the basis of structured psychiatric interviews and established diagnostic criteria (ICD-10; Diagnostic and Statistical Manual of Mental Disorders (DSM III and DSM IV)).26 Only 3 of those studies17,18,27—with confirmed COPD diagnoses—explored both anxiety and depression in sufficiently large samples (137, 116 and 143 subjects).

In contrast with our finding of a 36% prevalence rate, Kühl et al.,17 reported an overall prevalence of 25%; this lower percentage could be due, as the authors themselves concede, to the fact that the number of patients with generalized anxiety disorder was probably underreported, given that anxiety secondary to the disease itself was not taken into account. Yohannes et al.27 and Laurin et al.18 reported prevalence rates of 47% and 49%, respectively. The fact that these 2 studies included approximately 50% of women might explain their higher prevalence rates; indeed, when stratified by sex, Laurin et al.18 reported a 38% rate of anxiety or depression in men—a rate comparable to that reported in our study.

We found anxiety to be more frequent than depression in our patients, corroborating the findings of Laurin et al.18 Yohannes et al.,27 in contrast, found depression to be more frequent, although this could be attributable to the significant degree of disability in their patients. In fact, disability was the factor most strongly associated with depression according to their own conclusions. In Kühl et al.,17 depression and anxiety were found in roughly equal proportions, although as already mentioned, anxiety was probably underdiagnosed. Overall, the findings of the above studies tend to reinforce the cogency of our own results.

As for the Spanish population, a recent article by Miravitlles et al.13 reported that 74.6% of patients with stable COPD treated at outpatient clinics in Spain had depression as assessed using the Beck Depression Inventory. We would suggest that this questionnaire with its conventional cutoff points is possibly overestimating the presence of depression in patients with stable COPD.

Our risk model points to relative youth, higher education level, higher BODE index score, greater dyspnea, an exacerbator phenotype and lack of home support as predictors of anxiety and depression in patients with stable COPD.

In previous studies based on certain diagnoses, some of which did not specifically analyze age, no association was found with age. In our study, the association may be due to the fact that our sample had more patients aged under 65years. Yohannes et al.,27 in contrast, specifically included subjects aged over 60years (mean age, 73years).

While a higher level of education was a risk factor for developing anxiety or depression in our sample, the few other studies that considered this variable have reported contradictory findings. Chazelle et al.28 found that level of education was associated with income level, which, in turn, was associated with access to, or lack of, private medical insurance. This suggests that access to healthcare may modify the link between educational level and anxiety or depression and so should be taken into account in comparing populations located in different countries. Evidence that a higher educational level is a risk factor for developing anxiety and depression in both the general population29 and in patients with COPD,30 as reported in recent studies in Ontario, echo our own findings within the context of a similar healthcare system.

Our study did not find occupational status to be a predictive factor, although our sample admittedly had only 4 unemployed patients. The higher frequency of anxiety and depression found in occupationally active patients could be due to the younger age of these patients.

Regarding COPD clinical variables, other studies have championed the BODE index score as a predictor of risk.31,32 Our analysis shows that the overall BODE score was more relevant than the scores for individual parameters, for an odds ratio of 3.98 (P<.001) for patients with BODE score ≥4. Of the BODE parameters, only dyspnea was significant in the multivariate analysis. As for FEV1, although we found a significant association when patients were stratified according to the GOLD 2007 classification, this significance was not maintained in the multivariate analysis. A higher proportion of stage I and stage IV patients presented anxiety and depression. The lower age of stage I patients and the higher levels of dyspnoea in stage IV patients may have acted as confounders. Our study confirms the absence of any link between FEV1 and psychiatric comorbidities31,32; it should be noted, however, that the COPD Strategy of the Spanish National Healthcare Plan only provides for psychiatric assessment in patients with poor lung function.33

Our sample included few women (20/204 patients), in line with other similar studies in Spain.14,25 This would explain why no significant difference was found in our study between men and women, in contrast with findings of higher female prevalence reported for other studies.8,34,35 Few of the patients in our study were admitted for hospitalization during the previous year (30/204), and that could influence that hospital admission was not associated with anxiety or depression. We found that not having home support was more important than loss of autonomy, although this conclusion cannot be taken for granted for patients with greater physical limitations. For our population of patients with stable COPD monitored in a pulmonology clinic, neither the number of drugs prescribed nor the number of comorbidities were predictors, although these 2 variables could be relevant in populations with greater comorbidity rates.

Finally, for logistical reasons, the fact that patients who scored below the HADS cutoff point were not assessed by a psychiatrist may have underestimated the prevalence of anxiety and depression in our sample. To overcome this limitation, we sought greater sensitivity by considering total HADS scores as well as HADS subscale scores, given that some authors have reported total scores to be more useful for screening purposes.36,37

Our study included 204 patients with stable COPD and established a prospective psychiatric diagnosis of anxiety or depression using ICD-10 criteria. We found 36% of our study subjects to have anxiety and/or depression. Since only 9% were diagnosed before inclusion in the study, underdiagnosis of these comorbidities in our setting would appear to be very high. Anxiety was more common than depression, and the two conditions coincided in a significant proportion of our patients. Our results indicate that younger patients with higher education levels have an increased risk of developing anxiety and depression. Other risk factors were higher BODE index scores, more exacerbations and lack of home support. Of the individual BODE components, only dyspnoea was an independent predictor. Future studies should take into account access to healthcare resources in comparing samples from different countries.

Research grant from the Neumosur Foundation.

The authors declare that they have no conflicts of interest.