Impact of Cough Across Different Chronic Respiratory Diseases

----------

--

Original Research COUGH

u

Impact of Cough Across Different Chronic Respiratory Diseases* Comparison of Two Cough-Specific Health-Related Quality of Life Questionnaires Liam Polley, M B ; Nurman Yaman, MB; Liam Heaney, MD; Chris Cardwell, PhD; Eimear Murtagh, M B ; John Ramsey, MB, MA; Joseph MacMahon, M B ; Richard W. Costello, MD; and Lorcan McGaruey, MD

Background: Cough is a prominent symptom across a range of common chronic respiratory diseases and impacts considerably on patient health status. Methods: We undertook a cross-sectional comparison of scores from two cough-specific healthrelated quality of life (HRQoL) questionnaires, the Leicester Cough Questionnaire (LCQ), and the Cough Quality of Life Questionnaire (CQLQ),together with a generic HRQoL measure, the EuroQol. Questionnaires were administered to and spirometry performed on 147 outpatients with chronic cough (n = 83), COPD (n = 18), asthma (n = ZO), and bronchiectasis (n = 26). Results: There was no significant difference in the LCQ and CQLQ total scores between groups (p = 0.24 and p = 0.26, respectively).Exploratory analyses of questionnaire subdomains revealed differences in psychosocial issues and functional impairment between the four groups (p = 0.01 and p = 0.05, respectively). CQLQ scores indicated that chronic coughers have more psychosocial issues than patients with bronchiectasis (p = 0.03)but less functional impairment than COPD patients (p = 0.04). There was a significant difference in generic health status across the four disease groups (p = 0.04), with poorest health status in COPD patients. A significant inverse correlation was observed between CQLQ and LCQ in each disease group (chronic cough r = - 0.56, p < 0.001; COPD r = - 0.49, p = 0.04; asthma r = - 0.94, p < 0.001; and bronchiectasis r = - 0.88, p < 0.001). There was no correlation between cough questionnaire scores and FEV, in any group, although a significant correlation between EuroQol visual analog scale component and FEV, (r = 0.639, p = 0.004) was observed in COPD patients. Conclusion: Cough adversely affects health status across a range of common respiratory diseases. The LCQ and CQLQ can each provide important additional information concerning the impact of cough. (CHEST 2008; 134:295302) Key words: cough; health status; questionnaire Abbreviations: ANOVA = analysis of variance; CQLQ = Cough Quality of Life Questionnaire; df = degrees of freedom; EQ-5D = EuroQoL five dimensional comfone,nt; EQZAS ,= EuroQol visual analog scale; HRCT = highresolution CT; HRQoL = health-related quality o life, LCQ Leicester Quality of Life Cough Questionnaire; SGRQ = Saint George Respiratory Questionnaire

c

hronic cough is not a trivial symptom and is associated with significant impairment in health status.] Health-related quality of life (HRQoL) questionnaires provide one means of measuring health status and are increasingly used in clinical studies. Two cough-specific quality of life questionnaires have been developed and validated: the Leicester Cough Questionnaire (LCQ),2 and the Cough Qualwww.chestjournal.org

ity of Life Questionnaire (CQLQ).3Both have been evaluated in patients with acute and chronic cough, but there is little information regarding the measureFor editorial comment see page 226

ment of cough-specific health status across a range of common respiratory diseases in which cough is often CHEST I 134 I 2 1 AUGUST, 2008

295

a prominent symptom. It is also unclear whether both questionnaires measure similar aspects of cough-specific quality of life. Therefore we undertook a cross-sectional comparison of scores for the LCQ, CQLQ, and a generic quality of life questionnaire, the EuroQol: obtained in four distinct patient groups: chronic cough, bronchiectasis, asthma, and COPD. In this study, our primary aim was to determine the extent of correlation between the two coughspecific health status questionnaires and a general health-related quality of life measure. As a secondary aim, we sought to assess the impact of cough on health status across common chronic respiratory d’iseases.

MATERIALSAND METHODS

hensive diagnostic protocol, the details of which have been published elsewhere.5 In brief, after history and physical examination, chest radiography and spirometry were arranged in all patients. Where indicated, 24-h esophageal pH monitoring and/or bronchoprovocation challenge testing were requested. Suspected asthmatic cough or gastroesophageal reflw associated cough were treated according to our established management protocol that was based on existing treatment guidelines.6 Patients with normal spirometry results and no evidence of bronchial hyperreactivity received 2 weeks of oral prednisolone to exclude a steroid-responsive cough. Patients with persisting upper airway symptoms despite intensive nasal therapy underwent formal ear, nose, and throat assessment and/or CT scan of sinuses. Diagnoses were considered on the basis of a consistent history and/or investigation but were only accepted as contributing to cough when the patient reported satisfactory improvement or complete resolution after a period of diagnosis-specific therapy. A satisfactory improvement was recorded when the patient reported that the cough had subsided to the extent that it was no longer troublesome. The diagnostic categories identified through this study are presented in the “Results” section below.

HRQoL Questionnaires Study Population The study was conducted over a 1-year period between August 2003 and August 2004. Consecutive unselected patients with chronic cough referred to a Specialist Cough Clinic at the Belfast City Hospital were recruited together with patients with clinically stable asthma, COPD, or bronchiectasis attending general respiratory outpatient review clinics. The latter three patient groups were recruited consecutively and not because they reported cough as a prominent or troublesome symptom. All cough patients were lifetime nonsmokers, had cough persisting for > 8 weeks, and had no history of a recent upper respiratory tract infection. All study subjects were instructed to complete the three HRQoL questionnaires in the following order: LCQ, CQLQ, and EuroQol. Ethical approval was obtained from the Research Ethics Committee, Queen’s University Belfast, and all patients gave written informed consent.

Procedure After an initial explanation, the questionnaires were completed unaided by the study subjects. At the same visit, FEV, was measured in all subjects using a dry wedge spirometer (Vitalograph; Buckingham, UK). All chronic cough patients participating in the study subsequently underwent evaluation for cough based on our compre*From the Regional Respiratory Centre (Drs. Polley, Yaman, and MacMahon) and Department of Radiology (Drs. Murtagh and Ramsey), Belfast City Hos ital, Respiratory Research Group (Drs. Heaney and McCarveJ, and De artment of Epidemiology and Public Health (Dr. Cardwell), T i e Queen’s University of Belfast, Northern Ireland; and Department of Medicine (Dr. Costello), Royal Colle e of Surgeons in Ireland, Beaumont Hospital, Dublin 9, Iregnd. The authors have no conflicts of interest to disclose. Manuscript received January 22,2007; revision accepted October 25, 2007. Reproduction of this article is rohibited without written permission from the American College otChest Physicians (www.chestjoumal. org/misc/re rintsshtml). CorresponLnce to: Lorcan McGarvey, MD, Department Medicine, The Queen’s University of Beyayt, Northern lrelan e-mail: 1.mcgarve @qub.ac uk DOI: 10.1378/c~est.07-0141 296

The CQLQ: The CQLQ is composed of 28 items presented in a 4-point Likert response format. The lowest overall score is 28 (no adverse effects of cough on quality of life), and the maximum score is 112 (most adverse effects of cough on quality of life). The CQLQ is subdivided into six domains: physical complaints, psychosocial issues, functional abilities, emotional well-being, extreme physical complaints, and personal safety fears. The LCQ: The LCQ consists of 19 items in a 7-point Likert response format allocated to three domains; physical, psychological, and social. A higher score indicates better health status. EuroQol: The EuroQol is a generic HRQoL questionnaire consisting of two parts, the EuroQol five dimension component (EQ-5D) index, which rates mobility, self-care, usual activities, paiddiscomfort, and anxietyldepression; and the EuroQol visual analog scale (EQ-VAS), which contains a vertical rating scale (0 = worst possible health, to 100 = best possible health). The EuroQol has been extensively used in a variety of clinical settings and is widely recognized as a generic measure of HRQoL.4 Radiologic Assessment of Bronchiectasis Seventy We wished to determine the association between the radiologic extent of bronchiectasis and cough-related health status. Highresolution CT (HRCT) scans performed on the study subjects with bronchiectasis within the previous year were assessed by two independent radiologists (E.M. and J.R.) using the scoring system described by Bhalla et al.7 In brief, this system grades the scans on morphologic changes including severity of bronchiectasis, peribronchial thickening, and mucus pluging. HRCT scores are then subtracted from 25 to give an overall score. Thus, lower scores indicate more extensive radiographic change. These scores were correlated with lung function and scores obtained from the HRQoL questionnaires.

Statistical Analysis Data analysis was performed using statistical software (SPSS version 14; SPSS; Chicago, IL). Analysis of variance (ANOVA)was used to compare the questionnaire scores between the four groups (asthma, COPD, bronchiectasis, and cough). When ANOVA indicated there was a significant difference between the four groups, the Duncan test was used to identifjr groups that differed significantly. Original Research

Multiple linear regression was used to test the group term after adjustment for sex and age (as continuous) using the extra sums of squares principle.",g Specifically, for each outcome variable of interest, the extra sums of squares associated with the group variable after adjustment for age and sex was calculated by subtracting the sums of squares due to the regression of age and sex from the sums of squares due to the regression of age, sex, and group (comprising three indicator variables, one for three of the four group) and dividing by the difference in degrees of freedom between these two models (three). This quantity was then divided by the residual mean square from the larger model and compared to the F distribution with the appropriate degrees of freedom (three, number of observations minus three). A Wald test based on the regression coefficient from multiple linear regression was used to test the sex term after adjustment for group and age (fitted as a continuous variable in the regression model) for total LCQ and CQLQ. Pearson correlation coefficient was used to measure the association between CQLQ and LCQ scores and compared between the four groups using the method described by Snedecor and Cohran."'

RESULTS Table 1contains baseline characteristics of the 147 patients recruited to the four groups as well as test statistics for comparisons of these characteristics between the groups. Diagnostic categories in the chronic cough group were as follows: gastroesophageal reflux, 35%; asthma, 21%; idiopathic cough, 12%; postnasal drip, 7%; postviral, 6%; eosinophilic bronchitis, 6%; and other, 13% (eg, angiotensinconverting enzyme inhibitor-related cough, pulmonary fibrosis). There was a significant difference in age (ANOVA p = 0.001) between the four groups. COPD patients (mean age, 64.4 years) were the oldest and were significantly older (Duncan p < 0.05) than patients in the asthma and chronic cough group (mean age, 51.6 years and 53.9 years, respectively). There was a difference in the proportion of female patients in the four groups (p = 0.001). The chronic cough group (67.5% female) contained a significantly (Bonferroni p < 0.05) greater proportion of female patients than

Table 1-Characteristics

the COPD and asthma groups (16.7% female and 35% female patients, respectively). There was a significant difference (ANOVA p < 0.001) in mean percentage of predicted FEV, between the four groups. In particular, the COPD (mean, 42.3) patients had significantly lower lung function than all other groups (Duncan p < 0.05 for all comparisons) and both patients with asthma (mean, 75.2) and patients with bronchiectasis (mean, 73.2) had significantly lower lung hnction than those with chronic cough (mean, 102.4).

HRQoL Questionnaires (Between-Disease Group Comparison) The comparisons of the questionnaire scores between groups of patients, and calculated test statistics. are shown in Table 2.

Cough-Specqic Questionnaires: Although total scores on the LCQ were lower (suggesting poorer health status) in chronic cough patients compared to other respiratory diseases, there was no significant difference in mean score between the four groups (ANOVA p = 0.24). Total CQLQ scores were highest (suggesting worst health status) in COPD patients, but again there were no significant difference in the mean scores between the four groups (ANOVA p = 0.26). Unadjusted, exploratory analysis of the CQLQ subdomains revealed a significant difference in the psychosocial domain between the four groups (ANOVA unadjusted p = 0.01). Duncan tests revealed that the chronic cough group had a significantly (Duncan unadjusted p < 0.05) higher psychosocial score (mean, 12.6) compared with both the asthma and bronchiectasis group (mean, 10.2 and 10.0, respectively). There was evidence of a borderline significant difference in the CQLQ functional domain between the four groups (unadjusted p = 0.05). Significant differences (Duncan unad-

of Study Participants by Patient Group*

Characteristics

Asthma

Bronchiectasis

COPD

Patients, No. Female gender, No. (%) Age, Yr Cough duration, mo Bronchiectasis score FEV, % of predicted

20 7 (35)" 51.6 -t 17.5"

26 13 (50)",h 57.5 2 11.8".b

18 3 (16.7)" 64.4 2 9.7"

75.2 (22.7)"

10.4 2 4.15 73.2 (24.5)"

42.3 ( 16.9)h

Chronic Cough 83 56 (67.5)b 53.9 2 13.3" 36 (4-240) 102.4 (17.7)'

p Value

< 0.001t O.OOl$

< 0.001p

*Data are presented as mean 2 SD or mean (range) unless otherwise indicated. t X 2 = 19.21,d f = 3, comparing the proportion of female patients in the four groups. $ANOVA F test = 3.83, df = 3,139, comparing the mean age between the four groups. PANOVA F test = 59.38, df = 3,140, comparing the mean percentage of predicted FEV, between the four groups. ".h,'Denote groups that are significantly different using the Duncan test to control for multiple comparisons for the variables age and % predicted FEV, and Bonferroni test to control for multiple comparisons for gender. Groups with the same superscript are not significantly different (p > 0.05) from each other, while groups with different superscripts are significantly different (p < 0.05) from each other. www.chestjournal.org

CHEST I 134 I 2 I AUGUST, 2008

297

Table 2-Questionnaire

Asthma

Bronchiectasis

Scores by Group

COPD

Unadjusted Comparison of Means*

Chronic Cough

Adjusted Comparison of Meanst I

I I

Measures

No.

Mean(SD)

No.

Mean(SD) No.

LCQ Physical Psycological Social CQLQ Physical Psychosocial Functional Emotional Extreme physical Personal safety EQ-5D EQ-VAS

20 20 20 20 18 18 18 18 18 18

13.3(5.0) 4.6(1.8) 4.6(1.8) 59.8 (21.8) 20.4 (8.0) 10.2 (4.1)” 10.4 (4.5)“ 5.6(2.0) 7.1 (2.5);’.”

26 26 26 26 25 25 25 25 25 25

14.1 (4.8) 4.5(1.5) 4.9 (1.8) 4.8(1.7) 58.7(19.5) 19.0(7.3) 10.0 (4.0)” 10.7 (4.2)”

18

6.1 (2.7)

25

20 20

4.1(1.6)

0.63 (0.38)“ 26 60.2 (18.2)“.” 26

Mean (SD) No.

6.2 (2.2)“

18 13.2(4.4) 18 4.0(1.4) 18 4.6(1.7) 18 4.6(1.6) 18 68.6 (12.7) 18 23.1 (5.4) 18 11.0 (3.3)“.” 18 13.2 (3.2)” 18 6.6(2.2) 18 8.3 (2.8)”

75 76 79 79 68 74 75 73 72 73

7.0(2.7)

18

72

5.8(1.8)

0.70 (0.32)“ 18 67.4 (19.0)h 18

6.3(2.3)

0.45 (0.31)’’ 73 53.8 (17.4)“ 73

F

P

F P Value Value

Mean (SD)

cy

12.2 (4.2) 4.3(1.2) 4.1 (1.7) 3.9(1.7) 62.4 (15.1) 19.8(7.6) 12.6 (4.0)” 10.2 (4.1)“ 6.0(2.2) 7.6 (2.8)a,”

3,135 3,136 3,139 3,139 3.125 3,131 3,132 3,130 3,129 3,130

1.43 0.55 1.78 2.22 1.35 1.28 3.86 2.62 .83 2.69

0.24 0.65 0.16 0.09 0.26 0.29 0.01 0.05 0.48 0.05

3,129 3,130 3,133 3,133 3,120 3,125 3,126 3,124 3,124 3,124

0.83 1.74 1.06 1.25 2.51 1.94 2.88 3.56 1.07 3.64

3,129

1.70

0.17

3,123

1.38 0.25

0.76 (0.26)” 3,133 68.2 (22.2)”.” 3,133

5.48 2.85

0.001 3,127 0.04 3,128

7.3(2.3)

Value Value

df

0.48 0.16 0.37 0.29 0.06 0.13 0.04 0.02 0.37 0.02

5.63 0.001 2.94 0.04

*F test from comparison of means using ANOVA. fAdjusted for age (continuous) and sex by comparing a regression model containing age, sex, and group to a regression model containing age and sex using the extra sums of squares principle as described in “Statistical Analysis” section. “,”.‘Denote groups that are significantly different using Duncan test to control for multiple comparisons. Groups with the same superscript are not significantly different (p > 0.05) from each other, while groups with different subscripts are significantly different (p < 0.05) from each other.

justed p < 0.05) were observed between the COPD group (mean, 13.2) and the other groups (mean of asthma, bronchiectasis, and chronic cough groups, 10.4, 10.7, and 10.2 respectively). Significant differences (Duncan unadjusted p < 0.05) were also observed in the extreme physical domain between bronchiectasis (mean, 6.2) and those with COPD (mean, 8.3) or chronic cough (mean, 7.6).

Generic EuroQol: A significant difference in the EQ-5D questionnaire scores between the four groups was noted (ANOVA unadjusted p = 0.001). The COPD group (mean, 0.45) had a significantly lower score (unadjusted p < 0.05) [poorer health status] than the other three groups (mean of asthma, bronchiectasis, and chronic cough group, 0.63, 0.70, and 0.76, respectively). There was also a significant difference in mean EQ-VAS scores between the four groups (ANOVA unadjusted p = 0.04). Specifically, there was a significant difference (Duncan unadjusted p < 0.05) between the COPD (mean, 53.8) and bronchiectasis groups (mean, 67.4). Adjustments for Age and Sex Table 2 also shows tests (and test statistics) for the comparison of means between the four groups after adjustment for age and sex using linear regression. The significant differences in mean observed in the 298

psychosocial, functional, and extreme physical domain of the CQLQ questionnaire scores between groups remained significant after adjustment for age and sex (adjusted p = 0.04, p = 0.02, and p = 0.02, respectively). Also, the significant differences in the mean of EQ-5D and EQ-VAS between the four groups remained significant after adjustment for age and (adjusted p = 0.001 and p = 0.04). The results of comparisons between specific groups after adjustment for age and sex were similar to the results of the tests for multiple comparisons prior to adjustments and therefore have not been presented.

Concurrent Validity There was a significant difference in the correlation coefficients for the association between CQLQ and LCQ in the four disease groups (x2= 21.7, degrees of freedom [dfl = 3, p =
Asthma

COPD

(r = -0.94. P = <0.001)

(r = -0.49, P = 0.04)

201

. . . '

15

.*

*

l01

0.

(r = -0.88, P = <0.001)

f

z *

(r = -0.56,P = <0.001)

25 -

c

0

20 -

0

.

*.

* * -1 .****.**** * . .

15-

.

0

100

.

lo: 5

0

0.0.

5-

20

40

60

80

100

0

20

40

60

80

100

Total CQLQ score FIGURE1. Correlation between total CQLQ and total LCQ scores by group comparison.

and EQ ( r = 0.48, p < 0.001). The correlation between CQLQ total scores and EQ scores in all patients was r = - 0.53 (p < 0.001).

there was a significant correlation between EQ-VAS and FEV, ( r = 0.65, p = 0.004).

Gender Diferences in HRQoL Questionnaire Scores

HRQoL Questionnaires and Lung Function

Female patients had significantly (independentsample t test p = 0.04) higher (worse health status) total CQLQ scores (mean, 65.1; SD, 16.2)than male patients (mean, 59.0; SD, 17.2).After adjustment for

There was no significant correlation between total scores on either the LCQ or CQLQ and FEV, in any subject group (Table 3). In patients with COPD,

Table 3-Table of Correlation CoeJJicientsDisplaying Concurrent Validity of Questionnaires and the Association Between HRQoL and Lung Function Asthma Variables Concurrent validity of' questionnaires CQLQ total and EQ-5D CQLQ total and EQ-VAS LCQ total and EQ-5D LCQ total and EQ-VAS HRQoL and lung function CQLQ total and FEV,, % predicted LCQ total and FEV,, % predicted E Q J D and FEV, % predicted EQ-VAS and FEV, % predicted

www.chestjournal.org

I

r (p Valiie)

NO.

18 18 20 20

-

18 20 20 20

-

COPD

Bronchiectasis I 1

I

No.

r ( p Value)

0.68 (0.002) 0.74 (< 0.001) 0.66 (0.002) 0.64 (0.002)

25 25 26 26

- 0.69 (< 0.001) - 0.52 (0.01)

0.12 (0.62)

25

0.13 (0.60) 0.06 (0.81) 0.16 (0.51)

26 26 26

I

No.

-

- 0.16 (0.46)

18

-

0.25 (0.23) 0.40 (0.04) 0.28 (0.17)

18 18 18

0.67 (<0.001) 0.52 (0.01)

'

'No.

0.30 (0.23) 0.24 (0.33) 0.60 (0.008) 0.43 (0.07)

63 63 68 67

0.00 (0.99)

66

- 0.17 (0.17)

0.46 (0.06) 0.20 (0.44) 0.65 (0.004)

73 71 70

-

r (p Value)

18 18 18 18

-

Chronic Congh r (p Value) -

0.43 (< 0.001) 0.47 (< 0.001) 0.40 (0.001) i 0.44 (0.001)

0.06 (0.63) 0.32 (0.12) 0.24 (0.04)

CHEST / 134 / 2 /AUGUST, 2008

299

age and diagnostic group (as female patients were younger than male patients, and a preponderance of female patients were in the chronic cough group), the mean difference in CQLQ between male and female patients remained significant (from regression model containing group, age, and sex, the regression coefficient for female patients was 8.I (SE, 3.19; t = 2.55; p = 0.012). This regression coefficient corresponded to an average increase in female patients of 8.1 (95% CI, 1.8 to 14.4) after adjustment for age and sex. Female patients had significantly (independent-sample t test p = 0.03) lower (worse health status) total LCQ scores (mean, 12.1; SD, 4.2) than male patients (mean, 13.7; SD, 4.6). This difference in mean was little altered after adjustment for age and diagnostic group (adjusted difference in mean, - 1.7; 95% CI, - 3.3 to - 0.02) and remained significant (regression coefficient for female patients, - 1.7; SE, 0.83; t = - 2.0; p = 0.05).

Bhalla Score in Bronchiectasis Group HRCT scans were available for scoring in 17 of the bronchiectasis patients. Bhalla scores correlated with FEV, (r = 0.55, p = 0.02); and although there was a relationship trend between Bhalla scores and EQ-VAS, this failed to reach significance (r = 0.47, p = 0.059). There was no relationship between Bhalla scores and cough-specific health status on either LCQ or CQLQ scores.

DISCUSSION This is the first study to compare the impact of cough on HRQoL across a range of common chronic respiratory diseases using validated cough-specific quality of life questionnaires. We have demonstrated that the magnitude of cough-specific HRQoL impairment is similar among all respiratory disease groups studied and is greatest among female patients. Generic measures of HRQoL in COPD patients were significantly lower than chronic cough, asthma, and bronchiectasis patients, and this appeared to be driven more by impaired lung function than cough. We demonstrated sigdicant cross-validation between coughspecific questionnaires across each of the respiratory diseases studied. Cough is a common feature in asthma and COPD, and a cardinal symptom in bronchiectasis, but there is little information regarding its impact on health status, and any adverse effects may be overlooked in these conditions. A number of groups have reported cough-specific HRQoL scores in bronchiectasis]1 and COPD patients.I2 This is the first study to 300

compare these diseases with the condition for which the tools were designed to evaluate, namely chronic cough. The CQLQ and the LCQ have been designed to identify areas of life that are disturbed by cough. This is important as it gives prominence to the patient’s view of those factors influencing quality of life. Although in this study we found that the impact of cough was of similar magnitude across all of these conditions, more exploratory analyses revealed differences in certain questionnaire subdomains, in particular on the CQLQ. Chronic cough patients had significantly more psychosocial issues than patients with bronchiectasis; this domain includes questions relating to self-consciousness and social embarrassment. COPD patients had significantly more functional complaints than chronic cough patients. That cough patients had significantly more psychosocial issues than patients with bronchiectasis was unexpected given that patients with bronchiectasis often report impairment in domains that measure how their disease impacts on social activities, employment, and causes anxiety, fatigue, and embarrassment.13 In our study, there was no association between cough-specific HRQoL and either the radiologic severity of bronchiectasis or lung function impairment, although the small numbers in this disease group do limit the interpretation of this result. In bronchiectasis, extent of radiologic disease appears to correlate well with degree of lung function impairment,l4 and measures of health status relate best with extent of systemic inflammation and frequency of exacerbations.15 We suggest that although chronic cough is a common symptom for patients with bronchiectasis, it is not the major factor influencing overall health status. The St. George Respiratory Questionnaire (SGRQ) is the best-validated and most widely used measure of health status in COPD. Patients who report cough have higher SGRQ scores (worse health status) than those without cough. The frequency of cough seems to account for a significant proportion of the symptom subdomain of the SGRQ; however, it is less important than wheeze, anxiety, and breathlessness in determining the total SGRQ score.lGIn our study, COPD patients had similar cough-specific health status but significantly poorer generic health status than that measured in bronchiectasis and chronic cough. Although clinically stable, these COPD patients had moderate-to-severely impaired lung function (mean FEV,, 42.2 2 17% of predicted). In contrast to generic measures of HRQoL, coughspecific quality of life scores did not correlate with lung function. In our study, we suggest that the poor health status reported by COPD patients is influenced more by lung function impairment than cough Original Research

severity. However, we acknowledge that this suggestion may not be the case for COPD patients with milder disease. Cough is identified by asthmatics as a symptom that significantly interferes with activities of daily life.17 Cough is also recognized by physicians as a very important symptom in determining asthma control, although less so than shortness of breath and wheeze.I8 In the asthmatic group in our study, we observed a highly significant correlation between both cough-specific questionnaires and generic HRQoL measures, suggesting that cough may be a particularly important factor in overall quality of life. We observed strong correlation between the two cough-specific questionnaires in the asthmatic and bronchiectasis groups, and although significant the correlation in the cough patient group was weaker. This may be a result of the difference in how the questionnaires were developed. The CQLQ was designed in North America and the LCQ in the United Kingdom, and there may be differences in how cough is perceived in these two populations. For example, European subjects may be more reluctant to answer questions about how cough impacts on their personal hygiene. In this study, we have found that urinary and fecal incontinence was reported in 25% and 3%, respectively. This finding contrasts that from a surveylg of UK residents responding to a cough questionnaire in which incontinence was reported in 55% of women. The CQLQ, which addresses the important issues of urinary and fecal incontinence, may be a better tool to assess the psychosocial impact of coughing. The weaker correlation, particularly in chronic cough and COPD patients, may suggest that both questionnaires provide important additional information concerning the impact of cough. Female patients are recognized to have poorer cough-specific HRQoL than male patients, in particular having more psychosocial problems.20-22 In this study, there were significantly fewer female patients in the COPD and asthma groups compared to the chronic cough group. Although the gender imbalance may be a limitation of the study, in particular because female patients are more likely to have urinary incontinence, we suggest that the impact of cough may be even greater in female patients with asthma or COPD and requires further study. There are a number of potential limitations to our current study; the relatively small subject numbers in the asthma and COPD groups do limit the interpretation of differences in questionnaire scores between disease categories. Another possible limitation of the study was that chronic cough patients were new referrals, whereas those in the other disease groups were clinically stable and recruited from follow-up www.chestjournal.org

clinics. However, all the cough patients had been attending either primary or secondary care clinics prior to referral, and the median duration of cough was 3 years. Therefore, we do not believe this invalidates our study conclusions. The inclusion of subjective measures of cough severity such as a visual analog scale or the determination of individual cough thresholds using either capsaicin or citric acid may have enhanced our study, although the relationship of these with cough-specific quality of life has been previously reported.2 In conclusion, the LCQ and CQLQ offer a means to assess the adverse effects of cough on HRQoL in common respiratory conditions. Subjective measures such as these are likely to best reflect cough severity from a patient’s perspective. Further work is required to determine the performance and responsiveness of these tools to therapeutic intervention in these respiratory diseases.

REFERENCES 1 French CL, Irwin RS, Curley FJ, et al. Impact of chronic mugh on quality of life. Arch Intern Med 1998; 158:1657-1661 2 Birring SS, Prudon B, Carr AJ, et d. Development of a symptom specific health status measure for patients with chronic cough: Leicester Cough Questionnaire (LCQ). Thorax 2003; 58:339-343 3 French CT, Irwin RS, Fletcher KE, et al. Evaluation of a cough-specific quality-of-life questionnaire. Chest 2002; 121: 1123-1131 4 EuroQol: a new facility for the measurement of health-related quality of life; the EuroQol Group. Health Policy 1990; 16:199-208 5 McGarvey LP, Heaney LG, Lawson JT, et al. Evaluation and outcome of patients with chronic non-productive cough using a comprehensive diagnostic protocol. Thorax 1998; 53:738-743 6 Morice AH, Fontana GA, Sovijarvi AR, et al. The diagnosis and management of chronic cough. Eur Respir J 2004; 24:481-492 7 Bhdla M, Turcios N, Aponte V, et al. Cystic fibrosis: scoring system with thin-section CT. Radiology 1991; 179:783-788 8 Draper NR, Smith H. Applied regression analysis. London, UK: John Wiley and Sons, 1966 9 Kirkwood BR, Sterne JAC. Essential medical statistics, second edition. Oxford, UK: Blackwell Science Ltd, 2003 10 Snedecor GW, Cohran WG. Statistical methods. 6th ed. Ames, IA: Iowa State University Press, 1978 11 Torrego A, Haque RA, Nguyen LT, et d.Capsaicin cough sensitivity in bronchiectasis. Thorax 2006; 61:706-709 12 Smith J, Owen E, Earis J, et al. Cough in COPD: correlation of objective monitoring with cough challenge and subjective assessments. Chest 2006; 130:379385 13 Wilson CB, Jones PW, O’Leary CJ, et al. Validation of the St. George’s respiratory questionnaire in bronchiectasis. Am J Respir Crit Care Med 1997; 156:536541 14 Demirkazik FB, Ariyurek OM, Ozcelik U, et al. High resolution CT in children with cystic fibrosis: correlation with pulmonary functions and radiographic scores. Eur J Radio1 2001; 37:5459 15 Wilson CB, Jones PW, O’Leary CJ, et al. Systemic markers of inflammation in stable bronchiectasis. Eur Respir J 1998; 12:820- 824 CHEST I 134 1 2 I AUGUST, 2008

301

16 Jones PW, Quirk FH, Baveystock CM, et al. A self-complete measure of health status for chronic airflow limitation: the St. George’s Respiratory Questionnaire. Am Rev Respir Dis 1992; 145:1321-1327 17 Juniper EF, Guyatt GH, Epstein RS, et al. Evaluation of impairment of health related quality of life in asthma: development of a questionnaire for use in clinical trials. Thorax 1992; 47:76-83 18 Juniper EF, Guyatt GH, Cox FM, et al. Development and validation of the Mini Asthma Quality of Life Questionnaire. Eur Respir J 1999; 14:32-38

302

19 Everett CF, Kastelik JA, Thompson RH, et al. Chronic persistent cough in the connnunity: a questionnaire survey. Cough 2007; 3:5 20 Bining SS, Pate1 RB, Prudon B, et al. Quality of life in chronic cnugh. [abstract]. Am J Respir Crit Care Med 2003; 167:A135 21 French CT, Fletcher KE, Irwin RS. Gender differences in health-related quality of life in patients complaining of chronic cough. Chest 2004; 125:482-488 22 French CT, Fletcher ICE, Irwin RS. A comparison of gender differences in health-related quality of life in acute and chronic coughers. Chest 2005; 127:1991-1998

Original Research