COPD Assessment Test as a Screening Tool for Anxiety and Depression in Stable COPD Patients: A Feasibility Study

Abstract Anxiety and depression are common comorbidities in chronic obstructive pulmonary disease (COPD) patients but are often under-diagnosed. We aimed to assess the suitability of the COPD Assessment Test (CAT) in screening anxiety and depression in patients with COPD. Stable COPD patients from a cross-sectional observational study were assessed by CAT. Anxiety and depression were identified using the Generalized Anxiety Disorder questionnaire (GAD-7) and Patient Health Questionnaire (PHQ-9), respectively. Logistic regression analysis and receiver operating characteristic curve analysis were used to identify factors associated with anxiety or depression and to calculate the predictive values. A total of 530 stable COPD patients were enrolled and of those, the proportions of anxiety and depression were 17.0% and 21.5%, respectively. The adjusted odds ratios of the CAT for the presence of anxiety and depression were 1.094 (95%CI: 1.057–1.131) and 1.143 (95%CI: 1.104–1.183), respectively. The CAT score had a significant predictive value for the presence of anxiety (AUC = 0.709) and depression (AUC = 0.791) with an optimum cutoff score of 15. However, the psychometric properties of CAT were undesirable, presenting high negative predictive value (NPV) but low positive predictive value (PPV). Among CAT items, analysis further showed that non-respiratory CAT components were superior to respiratory components in identifying both anxiety and depression. Our results indicated that CAT is more useful to exclude anxiety and depression rather than detect them.


Introduction
Chronic obstructive pulmonary disease (COPD) is a complex respiratory disease with high prevalence [1], and it often coexists with several comorbidities that may have a significant impact on morbidity and mortality [2].Anxiety and depression are common comorbidities of COPD.The prevalence of anxiety ranged from 7% to 50%, and that of depression ranged from 10% to 57% among stable COPD patients in primary care settings or respiratory clinics [3].Patients with anxiety and depression are prone to having a higher risk of exacerbations, hospitalization, and mortality [4][5][6][7][8].Therefore, early identification of anxiety and depression is of great importance.
There are several valid and specific questionnaires to screen for anxiety and depression, such as the Hospital Anxiety and Depression Scale (HADS) [9], the Hamilton Anxiety/ Depression Rating Scale [10,11], and the Beck Anxiety/ Depression Inventory [12,13], but none of them are routinely used in clinical practice of COPD patients due to inadequate psychiatric knowledge and time constraints of pulmonologists.
The COPD assessment test (CAT) is a brief and widely used questionnaire to assess COPD health status [14].In the development phase of CAT, items that patients used to describe their COPD symptoms were gathered, including respiratory symptoms, systemic symptoms, limitations in daily activities, social life, and emotional health [15].Although the item of anxiety was removed from the finalized version, the CAT might still be able to reflect patients' emotional status and was more time-saving than psychiatric rating scales.
CAT contains eight items including respiratory components (cough, phlegm, chest tightness, and breathlessness going up a hill/stairs) and non-respiratory components (activity limitation at home, confidence leaving home, sleep, and energy) [14,16].Previous studies indicated that CAT score was associated with anxiety and depression in COPD [17][18][19][20], implying that the CAT score may have some value in screening anxiety and depression in COPD patients.Thus, we hypothesized that CAT may be a convenient screening tool to detect or exclude anxiety and depression in COPD patients.To test this hypothesis, we used the Generalized Anxiety Disorder questionnaire (GAD-7) and Patient Health Questionnaire (PHQ-9) to evaluate the feasibility of CAT in screening anxiety and depression in patients with stable COPD.We also investigated which components of CAT (respiratory components or non-respiratory components) had better abilities to identify anxiety and depression.

Study design and population
The Cohort Study for COPD in China (COMFORT study) is an ongoing multicenter prospective observational study to investigate the clinical characteristics of COPD subjects in China.Details of the study design can be found at http:// www.chinacopd.com/#/hot(ClinicalTrials.govID: NCT03044847).We used baseline data between December 2016 and January 2021 from a single center (Beijing Chao-Yang Hospital) within the COMFORT study.Patients who went to the outpatient clinic of Beijing Chao-Yang Hospital were included in this analysis.The inclusion criteria were as follows: (1) 40-75 years old; (2) diagnosed with COPD based on persistent respiratory symptoms, risk factors (e.g.cigarette smoking, biomass, or occupational exposure), and post-bronchodilator forced expiratory volume in 1s/ forced vital capacity (FEV 1 /FVC) <0.7 according to the GOLD criteria [21].We also used lower limit of normal (LLN) of Chinese reference values to define COPD in a sensitivity analysis [22].The exclusion criteria were as follows: (1) patients with acute exacerbation of COPD in the previous 3 months; (2) patients with other lung diseases such as uncontrolled tuberculosis or bronchiectasis; and (3) patients with contraindication for spirometry owing to uncontrolled high blood pressure, severe angina, myocardial infarction or stroke in the last 3 months, aortic aneurysm, or retinal detachment.The study protocol was approved by the Ethics Committee of Beijing Chao-Yang Hospital (No. 2016-KE-183) in accordance with the tenets of the Declaration of Helsinki.Written informed consent was obtained from all patients.

Data collection
Age, sex, smoking history, education level, annual household income, comorbidities, and the number of exacerbations in the past 12 months were collected based on patient reports.Ever-smokers were defined as current or former smokers with at least 10 pack-years of tobacco exposure.Exacerbations were defined as worsening of respiratory symptoms that result in a change of at least one of these medications, namely antibiotics, corticosteroids, and/or bronchodilators, or necessitate a visit to the emergency room or hospitalization [21].Spirometry was performed using a MasterScreen spirometer (VIASYS Healthcare GmbN, Germany) and quality-control checks for the measurement results were based on the ATS guideline [23].COPD severity was categorized according to the GOLD criteria (GOLD 1: FEV 1 ≥80% predicted; GOLD 2: 50%≤FEV 1 <80% predicted; GOLD 3: 30%≤FEV 1 <50% predicted; and GOLD 4: FEV 1 <30% predicted).Dyspnea was assessed by the modified Medical Research Council (mMRC) dyspnea scale [24].

Measurement of anxiety and depression
Anxiety was measured using the GAD-7, which is a self-reported questionnaire that proved valid in both general populations and primary care centers [25,26].GAD-7 contains seven items about generalized anxiety symptoms in the previous 2 weeks.GAD-7 scores range from 0 to 21, with scores of 0-4, 5-9, 10-14, and 15-21 defined as no, mild, moderate, and severe anxiety, respectively.The Chinese version of GAD-7 has been validated in general hospital outpatients [27].In the present study, we defined a GAD-7 score of ≥5 as clinical anxiety, based on previous reports [28,29].
PHQ-9 was used to assess the presence of depression.It comprises nine items based on the diagnostic criteria of the Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV) for major depressive disorder in the previous 2 weeks.Each item was rated on a 4-point Likert scale from 0 (not at all) to 3 (nearly every day), with a maximum total score of 27.Depression symptoms were categorized into no (0-4), mild (5-9), moderate (10)(11)(12)(13)(14), and severe (15-27) [30].The Chinese version of the PHQ-9 demonstrated good sensitivity and specificity in screening depression [31].A PHQ-9 score of ≥5 was defined as depression in this study [30,32].

COPD assessment test
The CAT is a valid disease-specific questionnaire to evaluate the health status among COPD individuals and is recommended by the Chinese national guidelines on COPD [14,33].It consists of eight items: cough, phlegm, chest tightness, breathlessness going up a hill/stairs, activity limitation at home, confidence leaving home, sleep, and energy.Each item ranges from 0 to 5, presenting a total score of 0 to 40.Higher CAT scores indicate more COPD symptoms and poorer control of the disease [34].The CAT score was classified into 0-10, 11-20, 21-30, and 31-40 for different symptom levels in this study.

Statistical analysis
Continuous variables were described as means and standard deviations (SDs) for normally distributed data and as medians and interquartile ranges (IQRs) for non-normally distributed data.Categorical variables were expressed as number and percentage (%).The differences between two groups (anxiety vs. non-anxiety, depression vs. non-depression) were assessed by Student's t-test or Mann-Whitney U test for normally or non-normally distributed data, and by chi-square test or Fisher's exact test for categorical variables.Group comparisons between CAT categories or GOLD stages were performed by Pearson's chi-square test and linear-by-linear association.Univariate and multivariable logistic regression analyses were performed to identify the factors associated with anxiety or depression.Relevant factors included in the multivariable analyses were age, sex, BMI, smoking status, education level, annual household income, FEV 1 %predicted, and exacerbation history.Odds ratios (ORs) were calculated with 95% confidence intervals (CIs).The receiver operating characteristic (ROC) curve was drawn to analyze the predictive value of the total CAT score and CAT component scores for the presence of anxiety or depression.Operating characteristics including sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), accuracy, and Youden index were shown for identifying the optimal cutoff value.The comparisons between the AUCs were performed using Delong test [35].SPSS version 23.0 (IBM, Armonk, NY, USA) and MedCalc version 20.0 (MedCalc Software, Ostend, Belgium) were used for statistical analyses.p < 0.05 was considered to indicate statistically significant differences.

Patient characteristics
A total of 530 patients were enrolled in this study with a mean age of 64.9 ± 7.8 years.Demographic and clinical characteristics of the patients are presented in Table 1.According to the GAD-7 score and PHQ-9 score, 90 (17.0%) patients had anxiety symptoms, while 114 (21.5%) patients had depression symptoms.Characteristics of patients classified by anxiety status or depression status are shown in Table 2. Patients with depression symptoms had a higher proportion of women, lower FEV 1 , higher CAT score, higher mMRC dyspnea score, and reported more exacerbations in the preceding year than those without depression symptoms.Similar results were observed in patients with anxiety symptoms compared to those without anxiety in terms of CAT, mMRC, and exacerbation history.The FEV 1 %predicted was lower in the depression group than the non-depression group (53.2 ± 23.5% vs. 60.7 ± 20.1%, p = 0.002), but no significant differences were found between the anxiety group and non-anxiety group with respect to FEV 1 %predicted (58.2 ± 22.6% vs. 59.3 ± 20.8%, p = 0.658).In addition, 503 patients met the COPD diagnostic criteria by FEV 1 /FVC < LLN underwent a sensitivity analysis; the characteristics of these patients are presented in Table E1, and the comparisons of the characteristics after stratifying by anxiety or depression status are shown in Table E2 (Supplementary material).

Correlations of the CAT and GOLD stage with anxiety and depression
The proportions of anxiety and depression symptoms categorized by CAT groups and GOLD stages are shown in Figure 1.The proportion of mild, moderate, and severe anxiety according to the GAD-7 score did not differ with different GOLD stages (x 2 =11.3, p = 0.254 by chi-square test).However, among patients stratified by CAT score, the proportion of mild, moderate, and severe anxiety significantly increased with increasing CAT scores (x 2 =46.2, p < 0.001 by chi-square test and p < 0.001 by linear-by-linear association).As for depression, the proportion of mild, moderate, and severe depression increased with increasing CAT scores  (x 2 =133.0,p < 0.001 by chi-square test and p < 0.001 by linear-by-linear association), and also increased with increasing GOLD stages (x 2 =34.0, p < 0.001 by chi-square test and p < 0.001 by linear-by-linear association).

The value of CAT for detecting anxiety and depression
Logistic regression analyses were performed to identify factors associated with anxiety and depression (Table 3).In univariate analysis, a higher CAT score was associated with a greater risk of both anxiety and depression (OR = 1.089, 95%CI: 1.059-1.121,p < 0.001; OR = 1.148, 95%CI: 1.113-1.183,p < 0.001; respectively).These associations were still significant in multivariable analysis after adjusting for potential confounders (OR = 1.094, 95%CI: 1.057-1.131,p < 0.001; OR = 1.143, 95%CI: 1.104-1.183,p < 0.001; respectively).Additionally, women, younger age, and smoking history were significantly associated with anxiety; while women, younger age, and COPD exacerbation history were significantly associated with depression.Lung function impairment defined by FEV 1 %predicted showed no relationship with either anxiety or depression in the multivariable analysis.
The operating characteristics of the CAT for identifying anxiety and depression are presented in Table 4.The highest Youden index was achieved at a cutoff score of CAT ≥ 15 for both anxiety and depression.The sensitivity, specificity, PPV, NPV, and accuracy for anxiety were 70.0%, 63.4%, 28.1%, 91.2%, and 64.5%, respectively.The same parameters for depression were 78.1%, 67.6%, 39.7%, 91.8%, and 69.8%, respectively.Similarly, the operating characteristics of CAT did not change much in the sensitivity analysis (Table E3, Supplementary material).

Psychometric properties of CAT stratified by GOLD stage
The screening performance of CAT ≥ 15 for anxiety and depression among patients from different GOLD stages is presented in Table 5.CAT had good specificity (72.4%-79.2%)but poor sensitivity (58.3%-61.4%)for identifying anxiety in GOLD stages I and II, but good sensitivity (83.3%-90.0%)and poor specificity (27.6%-45.0%) in GOLD stages III and IV.Similar results were also observed in the analyses of depression.

Comparison of different CAT components to identify anxiety and depression
ROC curve analysis was performed on each CAT item to determine which component of the CAT could better identify anxiety or depression (Figure 4).Among all items, sleep had the best predictive value for the presence of anxiety (AUC = 0.708, 95%CI: 0.667-0.746),whereas energy had the best predictive value for the presence of depression (AUC = 0.765, 95%CI: 0.727-0.801).Generally, non-respiratory CAT items (activity limitation at home, confidence leaving home, sleep, and energy) showed higher predictive value for the presence of anxiety and depression than respiratory-related items (cough, phlegm, chest tightness, and breathlessness going up a hill/stairs).Statistical comparisons of the predictive value of the total CAT score and the two components within CAT also showed that non-respiratory CAT components had a significantly higher predictive value than that of the respiratory components for both anxiety (AUC = 0.721 vs. 0.651, p = 0.011 by Delong test) and depression (AUC = 0.805 vs. 0.720, p < 0.001 by Delong test).Similar results were also observed in the sensitivity analysis using LLN-defined COPD (Figure E3, Supplementary material).

Discussion
In this cross-sectional study, we reported that the proportion of clinical anxiety and depression were 17.0% and 21.5%, respectively, in patients with stable COPD.The CAT score was independently associated with anxiety and depression and had a significant predictive value for these symptoms.However, the psychometric properties of CAT were undesirable, presenting high NPV but low PPV.Furthermore, analysis of components within CAT demonstrated that non-respiratory components can better identify anxiety and depression than respiratory-related components.
COPD is currently considered a complex disease with not only pulmonary symptoms but also systemic comorbidities.Mental health disorders, including anxiety and depression, are known to play an important role in the progression and prognosis of COPD [5].In this study, we observed 17.0% anxiety and 21.5% depression among COPD outpatients using valid screening tools such as GAD-7 and PHQ-9.A high prevalence of comorbid anxiety and depression in COPD was reported in various world regions.The prevalence of clinical anxiety in COPD ranged from 13% to 56.4% in outpatients and 10% to 55% among inpatients [36,37].The prevalence of clinical depression ranged from 10% to 42% in patients with stable COPD and from 10% to 86% in patients with acute COPD exacerbation [38].In addition, Lou et al. [39] reported a prevalence of 18.3% for anxiety symptoms and 35.7% for depression symptoms in stable COPD patients in rural China based on the HADS.Another report from a Chinese multicenter study conducted in tertiary hospitals showed that the prevalence of anxiety and depression was 8.8% and 15.6%, respectively [40].However, a direct comparison between our findings and those studies was difficult because of the differences in populations and the use of different assessment tools for anxiety and depression.
Our present study showed that women are more likely to be depressed.The global prevalence of anxiety and depression is higher in women than men in the general population as well as COPD patients [41,42].A possible explanation is that women patients are less confident in their ability to control respiratory symptoms, which may induce feelings of helplessness and hopelessness that result in the development of anxiety and depression [42].Other risk factors for anxiety and depression in COPD patients including younger age, smoking, and COPD exacerbation history were also involved in this study, showing consistent results with previous studies [39,[43][44][45].These findings suggest more efforts are needed to identify anxiety and depression in COPD patients, particularly in high-risk subjects.bmi: body mass index; fev 1 %predicted: forced expiratory volume in one second in percent of the predicted value; Cat: COPD assessment test; Or: odds ratio; Ci: confidence interval.In the present study, no significant relationship was observed between FEV 1 %predicted and anxiety, which is in concordance with the report from Gudmundsson et al. [46].They found that the prevalence of anxiety status was not different among COPD patients with different GOLD stages.Another study of patients from the general public and pulmonary rehabilitation centers in the Netherlands showed that there was no significant difference in psychological distress between patients with severe or very severe COPD and patients with mild or moderate COPD [47].An increased prevalence of depressive symptoms was observed with increasing GOLD stages in this study.Some previous studies identified impaired lung function (FEV 1 or GOLD severity) as a risk factor for depression [20,48,49].However, other studies showed that FEV 1 %predicted was not independently associated with depression after adjusting for factors including respiratory symptoms and patient-reported outcomes [39,[50][51][52].Our results showed that the relationship between lung function and clinical depression is weak, and FEV 1 %predicted was not associated with anxiety or depression after adjusting for CAT score and other confounders.The results suggest that the relationship between COPD severity and anxiety/depression should be explained better by a clinical measure of disease symptoms rather than the physiologic assessment of lung function.
We found that the CAT score was significantly associated with anxiety and depression independent of age, sex, BMI, and other potential confounders.Previous studies showed that St George's Respiratory Questionnaire (SGRQ) was strongly associated with anxiety and depression in COPD [19,39,45,53].However, the SGRQ questionnaire is too complex and time-consuming, which limits its utility in clinical practice.The CAT was developed as a brief questionnaire that includes key dimensions of COPD disease impact.Hilmarsen et al. [18] reported from COPD patients referred for pulmonary rehabilitation that patients with anxiety and/or depression symptoms had a significantly higher CAT score than those without these symptoms.The present study further suggested that CAT may be useful in screening both anxiety and depression among stable COPD outpatients.
Our study showed that the optimal cutoff score to identify both anxiety and depression was CAT ≥ 15, lower than that reported in a previous study by Harryanto et al. in Australia [17], which suggested a CAT score of ≥20 as an optimal cutoff to identify potential anxiety.It was also lower than a previous study in Korea [54] that recommended using a CAT score of ≥21 for screening for depression in COPD patients.These discrepancies are probably partly due to different study populations and different questionnaire scales and thresholds used for identifying clinical anxiety and depression in different studies.
Our results showed that the PPV of CAT is only 28.1% for anxiety and 39.7% for depression.The suboptimal PPV in this study was probably because of the low prevalence rate of anxiety and depression in COPD patients.In addition, CAT is a disease-specific questionnaire for symptoms of COPD, which limits its specificity for anxiety and depression.However, the NPV of CAT is excellent in this study (91.2% for anxiety and 91.8% for depression), which means in COPD patients with a low CAT score, the possibility of concomitant anxiety/depression is relatively low.These results indicated that CAT is more useful to exclude anxiety and depression rather than detect them.A low CAT score suggested a good health status and good control of symptoms in stable COPD patients [14]; hence, it is less likely for less symptomatic patients to suffer anxiety and depression.
We further analyzed the psychometric properties of CAT stratified by GOLD stages.The results showed that the CAT had good specificity but poor sensitivity in GOLD stages I and II, but good sensitivity and poor specificity in GOLD stages III and IV.These findings indicated that the psychometric properties of CAT were easily affected by the severity of airflow limitation, so it may not be ideal for anxiety/ depression screening in all patients.
The association between anxiety/depression and different CAT components was investigated.We found that all CAT items were associated with anxiety and depression, while non-respiratory components had a higher predictive value than that of the respiratory components.Our findings are consistent with a previous study suggesting that CAT energy score was the most predictive item of clinical depression [54].In clinical practice, physicians typically emphasize respiratory symptoms and overlook systemic symptoms when evaluating the health status of patients with COPD.Our results show that non-respiratory CAT components were more closely associated with anxiety and depression than respiratory-related components.Thus, we suggest that more attention should be focused on these systemic symptoms and an assessment of anxiety and depression should be performed in patients with prominent non-respiratory symptoms of CAT.
Our study has some limitations.First, this is a single-center study which may cause limitation in the generalizability of the results.Second, due to the cross-sectional design, anxiety and depression trajectories and follow-up were not investigated.Third, we assessed anxiety and depression using self-reported questionnaires rather than the gold standard (Diagnostic and Statistical Manual of Mental Disorders-V).However, previous studies have shown the GAD-7 and PHQ-9 were valid in detecting psychiatric morbidities in primary care and among COPD patients [28,55].Moreover, the aim of this study was to screen anxiety and depression symptoms rather than to diagnose psychiatric disorders.

Conclusion
Our results showed that CAT has limited value in screening concomitant anxiety or depression in stable COPD patients; however, a lower CAT score is a good indicator for excluding anxiety and depression.In addition, non-respiratory components among CAT items can better identify anxiety and depression than respiratory-related components.We believe these findings will help to more effectively identify the COPD patients with concomitant anxiety or depression.

Figure 4 .
Figure 4. Comparison of area under the curve (auC) values of individual Cat items, respiratory components, non-respiratory components, and total Cat score for identifying anxiety (a) or depression (b).respiratory components: cough + phlegm + chest tightness + breathlessness going up a hill/stairs.non-respiratory components: activity limitation at home + confidence leaving home + sleep + energy.

Table 1 .
Demographic and clinical characteristics of the enrolled patients.
Data are presented as mean ± sD, median (iQr) or n (%).* a minority of the subjects (5%) refused to report their income.bD: bronchodilator; bmi: body mass index; Cat: COPD assessment test; fev 1 : forced expiratory volume in 1 s; fvC: forced vital capacity; fev 1 %predicted: forced expiratory volume in 1 s in percent of the predicted value; GOlD: Global initiative for Chronic Obstructive lung Disease; mmrC: modified medical research Council dyspnea scale.

Table 2 .
Demographic and clinical characteristics of patients categorized by anxiety or depression status.
Data are presented as mean ± sD, median (iQr) or n (%).significant values were presented in bold.bD: bronchodilator; bmi: body mass index; Cat: COPD assessment test; mmrC: modified medical research Council dyspnea scale; fev 1 : forced expiratory volume in 1 s; fev 1 %predicted: forced expiratory volume in 1 s in percent of the predicted value; GOlD: Global initiative for Chronic Obstructive lung Disease; na: not applicable.

Table 3 .
multivariable logistic regression analysis of the factors associated with anxiety and depression in patients with COPD.entered as a continuous variable.significant associations were presented in bold. *

Table 4 .
Operating characteristics of the Cat for identifying anxiety and depression.

Table 5 .
Operating characteristics of Cat ≥15 for identifying anxiety and depression among patients from different GOlD stages.