| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the North West Lung Centre, South Manchester University Hospitals Trust, Manchester; and Aintree Chest Centre, University Hospital Aintree, Liverpool, UK.
Correspondence to: Jaclyn Smith, MD, PhD, North West Lung Research Centre, Wythenshawe Hospital, Southmoor Rd, Wythenshawe, Manchester M23 9LT, UK; e-mail jacky.smith{at}manchester.ac.uk
Abstract
Background: The relationships between objective cough rates, cough reflex sensitivity, subjective estimates of cough frequency, and cough-related quality of life in patients with COPD are poorly understood.
Subjects: Twenty-six patients with stable COPD who complained of cough (76.9% men; mean ± SD age, 68.7 ± 6.9 years; mean FEV1, 54.2 ± 12.0% of predicted; median smoking history, 42.4 pack-years [range, 13 to 135 pack-years]).
Methods: Subjects performed a cough challenge test, ambulatory daytime and overnight sound recordings, scored the severity of cough (0 to 5 score and visual analog scale) for each recording period, and completed a cough-specific quality of life questionnaire (CQLQ). Coughs were counted manually and quantified in terms of cough seconds per hour (cs/h): the number of seconds within the recording that contain cough.
Results: Overall median time spent coughing was 7.5 cs/h (range, 2.7 to 23.1 cs/h; daytime median, 12.4 cs/h [range, 3.3 to 40.4 cs/h]; overnight, 1.9 cs/h [0.0 to 19.0 cs/h]) [p = <0.01]. Median log concentration of citric acid eliciting five coughs (C5) was – 0.9 mol/L (range, – 1.5 to 0.0 mol/L). Day time but not overnight time spent coughing was significantly correlated with log C5 (log C5 r = – 0.49, p = 0.02, and r = – 0.20, p = 0.37, respectively). Subjective cough scores and visual analog scales were moderately associated with objective time spent coughing: daytime (r = 0.37, p = 0.03, and r = 0.41, p = 0.03) and overnight (r = 0.48, p = <0.01, and r = 0.5, p = 0.01), respectively.
Conclusions: Subjective measures of cough and cough reflex sensitivity are statistically related to time spent coughing in patients with COPD, but with low-to-moderate levels of correlation. These measures have insufficient predictive value to substitute for objective time spent coughing; however, in conjunction with the CQLQ, they may provide a qualitative dimension to the assessment of cough.
Key Words: ambulatory monitoring • chronic bronchitis • citric acid • cough challenge
Cough is a key symptom in COPD, and is one of the most commonly reported symptoms along with breathlessness and sputum production.1 COPD is the term given to a heterogeneous group of patients with chronic airflow limitation. For the COPD subgroup of chronic bronchitis, cough is the defining symptom. In an epidemiologic study2 in the Netherlands, reporting of cough in smokers was a useful tool for case finding for COPD, where it was a better predictor of airflow obstruction than breathlessness or wheeze.
Cough in COPD is probably driven by a combination of mucus hypersecretion and airway inflammation. Cough reflex sensitivity to citric acid and capsaicin in patients with COPD may be normal34 or increased.56 The majority of the available information about cough in COPD is based on subjective reporting, which may be unreliable. For example, a study7 of patients with chronic cough and asthma suggested that subjective reporting correlated with cough counts only in chronic cough patients and then only during the daytime. In children with recurrent cough, a weak but significant correlation was found between 24-h cough counts and cough scores.8 This study8 also found a significant moderate correlation between cough count and cough reflex sensitivity to capsaicin (the concentration causing two coughs but not the concentration causing five coughs), but the correlation was lost on repeat testing 1 week later.
In spite of its importance for patients, cough is rarely used as an outcome measure in the assessment of treatments for COPD, and then it is only measured subjectively using diary card scores or indirectly as a part of quality of life assessments. Two different cough-specific quality of life questionnaires have been developed in subjects presenting with chronic dry cough.910 However, there has been no validation against objective measures of cough, and the relationships between different subjective measures are unknown.
We have developed a system for objectively monitoring cough using a digital sound-recording device, microphone, and manually counting coughs. This technique has excellent agreement when validated against cough counting from video recordings,11 and been demonstrated to be responsive to change in a study12 of patients with cystic fibrosis. We quantify cough in terms of time spent coughing: the number of seconds per hour that contain at least one explosive cough sound.13
The aim of this study was to determine the relationships between established measures of cough in patients with COPD (ie, cough reflex sensitivity, subjective diary scores, cough-specific quality of life questionnaire [CQLQ], and objective cough monitoring). We hypothesized that current measures of cough are poor surrogates for objective measurement of time spent coughing. Hence we predicted that cough reflex sensitivity, subjective cough scores, and cough-related quality of life would be weakly associated with objective time spent coughing.
Materials and Methods
Patients with physician-diagnosed stable COPD (FEV1 > 30% and < 75% of predicted) and who complained of chronic cough (daytime cough score 
2; Table 1
)9 were recruited. Current smokers, patients with exacerbations within the last month, and patients receiving angiotensin-converting enzyme inhibitors were excluded. Of 48 patients contacted, 26 met the inclusion criteria and completed the study. The study was approved by the Local Research Ethics Committee, and all subjects gave written, informed consent.
|
Cough Reflex Sensitivity
A citric acid cough challenge test was performed to assess the sensitivity of the cough reflex.6 Briefly, six ascending concentrations from 0.03 to 1 mol/L were delivered as 12-µL, single-breath inhalations (Mefar dosimeter; Mefar; Brescia, Italy) with three inhalations of normal saline solution randomly interspersed. The operator and patient were blinded to the position of the placebo inhalations in the challenge test sequence. Following each inhalation, the number of coughs in the subsequent minute was counted by two experienced observers. The concentration of citric acid causing two coughs (C2) and the concentration of citric acid causing five coughs (C5) were noted.
Objective Time Spent Coughing
Daytime ambulatory cough recording was commenced after the cough challenge was completed. A digital recording device (Nomad Jukebox 3; Creative Technology; Singapore) capable of making a 10-h continuous recording and worn in a pouch around the waist was used. A lapel microphone (ECM-1025; AOI Electronics; Tokyo, Japan) was attached to the clothing approximately 30 cm from the mouth. The patients were then allowed to go home wearing the cough-recording equipment. They were visited in the evening at home to set up the overnight cough-recording device at the bedside. They were asked to commence this recording at their usual bedtime and stop the recording on rising the next morning. Cough recordings were transferred from the digital recorders to a personal computer and archived on compact disk. Recordings were compressed with custom-written software to remove silences and low-amplitude sounds. For each recording period, coughs were counted manually in "cough seconds": the number of seconds per hour in which at least one explosive cough sound was present.13 If several cough sounds occurred successively, after one inspiration (peel of coughing), the number of seconds in which these sounds fell were counted. This measure therefore encompasses an estimate of the length of peels of coughs. All objective cough data are presented as time spent coughing per hour. For this measure, we have a very high level of agreement between observers (intraclass correlation coefficient, 0.999; p = <0.001).14
Subjective Measures of Cough
Patients were asked to subjectively assess the frequency of their cough for the two recording periods using a scoring system (Table 1)7 and a 100-mm VAS: 0 mm = no cough, and 100 mm = worst cough. Additionally, cough-related quality of life was measured using a self-completed questionnaire (CQLQ),9 which consisted of 28 questions in a 4-point Likert scale answer format. The minimum overall score is 28 (good quality of life), and maximum score is 112 (poor quality of life). The CQLQ is subdivided into six domains representing psychosocial issues, physical complaints, functional abilities, emotional well-being, extreme physical complaints, and personal safety fears.
Statistical Analysis
Both day and night cough rates were positively skewed but became normally distributed with logarithmic transformation. For clarity, the median and interquartile ranges are reported for day and night coughs. However, for the correlations between times spent coughing and other measures of cough, the log10 transform was used.
Subjective measures of cough were also skewed but could not be improved by transformation; therefore, Spearman correlations were used to examine the relationships between subjective and objective measures of cough. The primary outcomes of the study were the correlations between existing measures of cough and the objective cough rate. Comparisons of day and night cough rates and effects of medication were exploratory analyses and hence should be interpreted with caution. Statistics were performed using statistical software (SPSS version 11.0; SPSS; Chicago, IL; Prism 4.0; Graphpad Software; San Diego, CA).
Results
Complete data were collected on 26 patients (20 men; mean ± SD age, 68.7 ± 6.9 years; mean FEV1, 54.2 ± 12.0% of predicted; median smoking history, 42.4 pack-years [range, 13 to 135 pack-years]). The duration of coughing was a median of 9 years (range, 2 to 72 years), and 69% described sputum production. All patients were receiving inhaled medication: short-acting B2-agonists (n = 26), inhaled corticosteroids (n = 20; median equivalent beclomethasone dose of 800 µg/d [range, 200 to 4,000 µg/d]), long-acting ß-agonists (n = 19), and anticholinergics (n = 16).
Objective Cough Monitoring
The median total time spent coughing was 7.51 cough seconds per hour (cs/h), with a wide range between patients (2.67 to 23.11 cs/h). The time spent coughing was much higher during the day than at night (Fig 1
; Table 2
): median time during the day spent coughing, 12.4 cs/h [range, 3.3 to 40.4 cs/h]; median night at night spent coughing, 1.9 cs/h (range, 0.0 to 19.0 cs/h) [p = <0.001, Wilcoxon rank test]. There was a significant moderate correlation between time spent coughing during the day and at night (r = 0.40, p = 0.04): patients coughing more during the day tended to cough more at night. There were no correlations between time spent coughing and lung function (FEV1, r = – 0.14; p = 0.95; Fig 2
) or smoking history (r = 0.29, p = 0.16).
|
|
|
There was a significant inverse correlation between log total time spent coughing and log C5 (Fig 2) but not log C2 (r = – 0.46, p = 0.02, and r = – 0.28, p = 0.16, respectively). Cough reflex sensitivity correlated significantly with time during the day spent coughing (log C5, r = – 0.49, p = 0.02) but not with time at night spent coughing (r = – 0.20, p = 0.37).
Subjective Cough Scores:
Most patients had moderate cough scores (Table 2). Subjective measures were substantially higher by day than night for both cough scores and VAS. Both cough scores and VAS correlated moderately with objective time spent coughing: cough scores (day r = 0.37, p = 0.03; and night r = 0.48, p = <0.01) and VAS (day r = 0.41, p = 0.03; and night r = 0.50, p = 0.01). Correlations tended to be stronger for overnight than for daytime scores (Fig 3
) and for VAS than for cough scores.
|
|
Discussion
Cough is a frequent symptom for many patients with COPD, and this is the first study to demonstrate the relationships between objective cough measurements and subjective measures of cough severity and cough-related quality of life. This study shows that it is possible to objectively measure spontaneous cough by day and by night. Currently, this is a very laborious and time-consuming task. A number of groups151617181920 have been attempting fully automated cough counting systems with varying success; at this stage, none have been fully validated or are patient/investigator friendly. The assessment of cough in the quiet of the night is less difficult than by day, when extraneous noise and especially speech provide many false-positive results and still provide a major technical challenge. We have used customized interactive computer software; however, in this study all coughs were manually validated.
Quantifying the cough sound signal is also difficult; currently there is no standard method for quantifying cough. Trying to count the explosive cough sounds in a long peel of coughs can be difficult, both manually and automatically. If only the number of peels of coughs is counted (rather than the explosive sounds), then the length of the peel is disregarded when this may be of importance. We have addressed these difficulties by introducing "cough seconds" as a method for quantifying cough; this provides an estimate of the time spent coughing and takes into account the length of peels of coughing. We have found previously21 that agreement between manual cough counters is slightly better for cough seconds than explosive cough sounds, and there is a strong linear relationship between cough seconds and explosive cough sounds. This study was not designed to compare cough seconds with other methods for quantifying cough. Although cough seconds look promising as a method for quantifying cough, it is possible that other methods may correlate better with subjective measures.
In this study, we investigated patients with stable COPD selected for the symptom of cough. The patients were typical of COPD patients seen in respiratory outpatient clinics. The overnight time spent coughing was similar to those we have reported earlier in a group of unselected COPD patients.22 Furthermore, the cough reflex sensitivity was similar to that found by Wong and Morice6 in a group of COPD patients also unselected for cough.
Cough was much more frequent by day and virtually stopped at night. Time spent coughing was independent of lung function but was related to cough reflex sensitivity. However, the correlation was at best moderate and was only predictive for 24% of the variation in time spent coughing and then only during the day. Coughing is thought to be largely suppressed overnight by sleep,23 and cough reflex sensitivity is reduced.24 It is possible that spontaneous coughing during the day is in part related to exposure to airborne irritants, explaining the association with cough reflex sensitivity. Had cough reflex sensitivity been measured overnight, this may have correlated with objective cough rate.
When objective coughing was compared to cough reflex sensitivity in children with recurrent cough,8 the relationships were inconsistent. In a very small group of adults with chronic cough, there was a significant, strong correlation between cough reflex sensitivity and daytime cough frequency,25 but it remains to be seen whether this can be replicated in larger numbers of patients. As a consequence of the limited sample size in this study, there is insufficient power to detect weak correlations between different measures of cough. The sample size is currently restricted by the extremely time-consuming nature of the task of manual cough counting.
Subjective estimates of cough frequency (cough scores and VAS) correlated moderately with time spent coughing, more strongly for overnight than during the day. It is likely that subjective cough scores are affected by confounding factors not measured in this study, such as the intensity of coughing, how much attention the subjects pay to their coughing, or even their mood at the time. Cough-related quality of life was also more strongly related to objective cough monitoring overnight than during the day. Although time spent coughing was reduced overnight, we speculate that any nighttime cough may be associated with poorer sleep quality, worse mood, and poorer quality of life scores. Patients may therefore pay more attention to coughing overnight, leading to more accurate subjective scoring.
Cough quality of life measurements provide an added dimension to objective cough counts. The CQLQ has only previously been studied in chronic cough patients of mixed etiology and has not been compared to objective cough counts. The relationships between time spent coughing and quality of life are complex, and are likely to vary from individual to individual. For example, a subject with low time spent coughing but urinary incontinence associated with each episode of coughing is likely to have a markedly different quality of life from a high-frequency cougher who is continent. Subjects with chronic cough seem to have much higher cough frequencies than COPD patients14 but have similar CQLQ scores.26 Hence, cough-related quality of life measurement is likely to be complimentary to objective measures of cough.
The differences in time spent coughing in patients receiving different bronchodilators should be interpreted with caution as a secondary outcome measure. Subjective cough frequency and severity have been found to improve with ipratropium bromide.27 Patient selection bias could explain our findings (patients with persisting cough accumulate more medications), but adverse effects of ipratropium on cough reflex sensitivity5 and cough clearance28 have been reported. The evaluation of the efficacy of currently available COPD medications for cough needs addressing. We have recently demonstrated in a double-blind randomized cross-over trial that codeine was no more effective than placebo for cough in COPD.29 This underscores the need for novel and effective antitussive drugs.
In summary, monitoring of cough in stable COPD is laborious but feasible and can be used to define the relationships between objective coughing, cough reflex sensitivity, subjective scores, and cough-related quality of life. Due to their low predictive value, these measures cannot be used as surrogates for objective time spent coughing, but are of use in understanding patient perceptions of cough, potential mechanisms causing coughing, and the effects of coughing on quality of life. These relationships are not likely to be consistent between different disease states or different patient populations, and further work is required. Understanding the differences between subjective cough reporting and objective time spent coughing will provide the opportunity to determine the effectiveness of current treatments and potential new antitussives on the symptom of cough in a variety of diseases.
Acknowledgements
We thank the patients who took part in the study, and we are very grateful to Dr. Richard Irwin and Mrs. Cynthia French for their permission to use the CQLQ.
Footnotes
Abbreviations: cs/h = cough seconds per hour; CQLQ = cough-specific quality of life questionnaire; C2 = concentration of citric acid causing two coughs; C5 = concentration of citric acid causing five coughs; VAS = visual analog scale
No financial or other potential conflicts of interest exist for any of the authors.
Financial support was provided by the North West Lung Centre Research Fund.
Received for publication January 16, 2006. Accepted for publication March 27, 2006.
References
This article has been cited by other articles:
|
|
 |
|
  S. Decalmer, A. A. Woodcock, J. A. Smith, R. Tutuian, and D. O. Castell Patient Misreporting May Lead to Underestimation of Cough Events Chest, July 1, 2007; 132(1): 358 - 359. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) and dashed line indicate night.