INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Chronic cough is a common and, at times, difficult management problem. Irwin (1) and others (2) have identified asthma, rhinitis, and gastro-esophageal reflux (GER) as the most common diagnoses associated with chronic cough. Frequently patients have more than one of these conditions. The underlying mechanisms of cough in these conditions are not defined. Cough is not a universal finding with these conditions, and the reason only some patients with these conditions develop chronic cough is not clear. It is possible that there is a single abnormality that predisposes some patients with these conditions to develop chronic cough. Alternatively, cough may be a nonspecific response to diverse pathogenetic abnormalities in these conditions.

Hyperresponsiveness of the upper airway (UAHR) (6, 7) resulting from inflammatory mediator release could be a common mechanism of cough in these patients. The upper airway has abundant cough receptors, and stimulation of these receptors leads to cough, reflex laryngeal closure, and intermittent upper airway obstruction (6). Release of inflammatory mediators (8) could cause hypersensitivity of the upper airway to histamine, which has been described in patients with symptoms resembling asthma (including cough) (6, 7), in patients with ACE inhibitor cough (9), and in rhinitis (10). Inflammation is a feature of the most common conditions which are associated with chronic cough. In asthma and rhinitis, there is mucosal inflammation with eosinophils and mast cells (10). Proximal reflux of acid in GER results in posterior laryngitis (13) and hypersensitivity of the cough reflex (4). In addition, ACE inhibition, which is also associated with chronic cough, and UAHR (9) could enhance the tussive effects of tachykinins, which are normally inactivated by ACE in lung tissue (14). The presence of inflammatory mediators in the upper airway could enhance the sensitivity of afferent nerves. We therefore hypothesized that airway inflammation induced by these diseases would cause hyperresponsiveness of the upper airway to histamine provocation and that this would be a unifying mechanism to explain the occurrence of chronic cough in association with asthma, rhinitis, GER, and ACE inhibitor therapy.

The alternative possibility is that chronic cough is characterized by several distinct and unrelated features. In this regard, the literature attributes chronic cough to a number of diverse causes such as asthma (1), upper airway diseases (5), airway inflammation (12), aspiration (5, 15), and psychogenic causes (16). A number of well-defined features accompany these conditions, namely bronchial hyperresponsiveness, UAHR, eosinophilic bronchitis (12), microaspiration, and excessive psychiatric symptoms. The aims of this study were to examine these mechanisms (airway inflammation with eosinophils, UAHR, BHR, microaspiration, psychiatric morbidity) of chronic cough in an unselected clinic population and to test the hypothesis that hyperresponsiveness of the upper airway to histamine was a unifying mechanism leading to chronic cough.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Thirty consecutive adult patients (20 female) referred to the respiratory outpatient clinics between May and November 1995 with chronic cough as the predominant symptom gave informed consent to be enrolled in the study. All of the patients had a cough for more than 4 wk, a normal chest X-ray prior to inclusion, and all had a forced expiratory volume in 1 s (FEV₁) > 80% of predicted. Twenty comparison subjects (10 male and 10 female) with a similar age distribution but without cough were also studied. Control subjects were recruited from staff members, relatives of non-cough patients in the respiratory clinic, and volunteers responding to an advertisement within the hospital (Table 1).

The subjects attended on two occasions. The first visit was for a clinical assessment and completion of a standardized psychiatric questionnaire, the symptom check list-90 (SCL-90R) (17). The second visit was within 2 wk for laboratory assessment including allergy skin prick testing, histamine provocation challenge with flow volume loop, and induction of sputum with nebulized hypertonic saline for cytological analysis (11). The study was approved by the Hunter Area Research Ethics Committee and each subject gave written informed consent.

Methods

The subjects had a full history taken, and a physical examination was performed. The history was recorded on a standardized questionnaire, and previous medical problems, medications, smoking history, amount of sputum, and hemoptysis were recorded. Subjects rated the frequency, severity, and limitation related to cough using 7-point Likert scales (18), where 7 corresponded to no symptoms, 4 to moderate symptoms, and 1 to maximum symptom intensity. Complications of chronic cough were recorded, including rib fracture, syncope, headache, sleep disturbance, urinary incontinence, and social disruption. Symptoms of intercurrent disease such as post-nasal drip, GER, and asthma were reported on the 7-point scale. A standard self-administered psychiatric symptom questionnaire (SCL-90R) was also administered. Cough was not included as a question on the SCL-90R. Skin prick tests to 20 common allergens, using histamine as a positive control, and saline as a negative control were performed on all patients. A single positive test (wheal > 3 mm) defined the subjects as allergic.

SCL-90R. The Symptom Checklist, 90-item revised (SCL-90R), a self-rating scale providing the quality and intensity of psychological distress, was used. Each of the 90 items were rated along a 0-4 point scale indicating the degree to which the subject had been distressed by the symptoms within the past week. The items are clustered into nine clinical axes of psychopathology, namely: (1) somatization; (2) obsessive-compulsive; (3) interpersonal sensitivity; (4) depression; (5) anxiety; (6) hostility; (7) phobic anxiety; (8) paranoid ideation; (9) psychoticism. For each subject, the raw scores were referred to the appropriate normal values for conversion to standard T scores. Results were reported in each of the 9 dimensions as the number of subjects whose scores exceeded the expected normal range of values.

Bronchial provocation testing. A histamine bronchial provocation challenge was performed (19) with recordings of the maximal inspiratory and expiratory flow volume loops. The tests were performed using a Medical Graphics 1070 spirometer (St. Paul, MN), connected to a microprocessor for flow volume loop analysis. The flow volume loop was recorded after inhaling normal saline diluent as a control and repeated until a reproducible loop was obtained. Doubling doses of histamine (up to 8 µmol) were then administered by a hand held DeVilbiss No. 40 nebulizer (Somerset, PA), and the flow volume loop recorded 1 min after inhalation. The best inspiratory and expiratory flows from at least three reproducible loops were used in the analysis. The test was terminated at a cumulative dose of 3.91 µmol of histamine or if a fall in forced expiratory volume in one second of greater than 20% was recorded prior to the maximum dose. A dose of 400 µg of salbutamol or 500 µg of terbutaline was administered by pressurized metered dose inhaler with spacer attached and the flow volume loop repeated after 10 min. Abnormal BHR was recorded if a 20% fall in FEV₁ was noted at or below 3.91 µmoles of histamine. Hyperresponsiveness of the upper airway was noted if a reproducible diminution of the maximum mid-inspiratory flow rate (FIF₅₀) of 20% or greater was recorded at any stage during the test. If the subject was not able to produce sputum during these tests, hypertonic sodium chloride (4.5%) was given by ultrasonic nebulizer until sputum could be expectorated. In every case, sputum for cytological analysis was obtained by one of the methods outlined.

Sputum analysis. Sputum was analyzed and performed as described elsewhere (11). Sputum volume and the size and number of plugs were recorded. Cellular portions of sputum were selected at inverted microscopy and the cells were dispersed by mixing with 1:10 dithiothreitol (Sputolysin; Calbiochem, La Jolla, CA). The cell suspension was filtered through 50 µm nylon gauze and a total cell-count was performed in a Neubauer hemocytometer, and cytocentrifuge preparations (Shandon Cytospin II, Sewickey, PA) were prepared from 100 µl of the cell suspension, which was adjusted to a concentration of 1 × 10  ⁶ cells /ml. Two slides were fixed with methanol and stained with Chromotrope 2R (BDH Chemicals, Poole, UK) for estimation of eosinophils, which were expressed as the percentage of 400 nucleated cells. Two slides were fixed with Carnoy's solution and stained with 0.5% toluidine blue in 0.7 M hydrochloric acid at pH 0.1, and 1,500 nucleated cells were counted from each slide to obtain a differential count of metachromatic cells. Two slides were stained with oil-red-O for assessment of lipid-laden macrophages as described (20, 21). The amount of intracellular lipid was measured as follows: 100 consecutive alveolar macrophages were counted and the amount of lipid per cell graded from 0 (no lipid droplets) to 4 (presence of many confluent intracellular oil-red-O obscuring the nucleus); the grades for each cell were summed to give a lipid index ranging from 0 to a maximum 400.

Clinical diagnosis. Clinical diagnoses were made according to the following criteria: rhinitis was considered to be present when there was a history of current (past week) symptoms of post-nasal drip, nasal stuffiness, or sneezing rated  4/7 on the 7 point scale; asthma was diagnosed when there was a history of cough and current variable airflow obstruction, evidenced by a PD₂₀FEV₁ to histamine of < 4 µmol, or a > 20% improvement in FEV₁ to bronchodilator; GE reflux was diagnosed when symptoms of indigestion, waterbrash, or bolus sensation were present at a severity of  4 on a scale of 1 to 7; sinusitis was diagnosed by a history of facial pain and a response to antibiotics. Subjects using ACE inhibitors, either recurrently or previously at the time of coughing were considered to have ACE inhibitor cough.

Analysis. The provocation dose of histamine causing a 20% fall in FEV₁ (PD₂₀FEV₁) or forced inspiratory flow (PD₂₀FIF₅₀) was estimated by linear interpolation of the log dose-response curve. The proportion of subjects reaching a PD₂₀ in each group is reported with 95% confidence intervals (CI) and was compared using Fisher's exact test. The maximum percentage fall from baseline in FEV₁ and FIF₅₀ and the FIF₅₀/FEF₅₀ ratio was compared using Student's t test. Sputum cell counts (eosinophils, mast cells, and oil-red-O-positive cells) were compared using a t test. Sputum eosinophil counts greater than 3.2% which were outside the 95% confidence limits of published normals, were considered elevated (11). The SCL-90R questionnaire was scored by computer program on nine clinical axes (somatization, obsessive-compulsive, interpersonal, depression, anxiety, hostility, phobic anxiety, paranoia, psychotic) as well as a general severity index, a positive symptom index, and a positive symptom total score. The proportion of subjects scoring in the clinically significant range on each of the axes was compared between cough patients and controls using chi-squared analysis. Significance was accepted at the p < 0.05 level.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Clinical Characteristics

Most patients with chronic cough had symptoms consistent with rhinitis or gastroesophageal reflux or both (Table 2). Of the 30 patients, 28 had significant symptoms, with a score   4/7 of at least one of post-nasal drip, nasal discharge, nasal stuffiness or sneezing in the week preceding the interview. Twenty-one patients reported significant indigestion, waterbrash, or bolus sensation. Five of the patients had been previously on an ACE inhibitor when the cough developed, and two patients were still on treatment up to and including the study. One patient also had symptoms and signs suggestive of left ventricular failure and mixed mitral valve disease. None of the patients satisfied the clinical criteria for sinusitis.

The cough had been present for an average of 228 wk (range 5-2,500 wk). The majority of the patients had been on unsuccessful trials of medications, with 25 having received inhaled corticosteroids, 21 antibiotics, 13 inhaled ₂ agonists, 10 histamine-type 2 receptor antagonists, 8 antacids, 6 oral corticosteroids, 4 inhaled cromoglycate, 2 nasal decongestants, 2 nasal corticosteroids, and 1 omeprazole.

Most (28/30) patients had experienced several complications from coughing. These included sleep disturbance (17/ 30), social disruption (16/30), urinary incontinence (15/30), presyscope (14/30), headache (13/30), syncope (2/30), rib fracture (2/30), and back pain (2/30).

Bronchial Provocation Testing

Bronchial hyperresponsiveness was present in seven of the patients with chronic cough (Tables 3, 4). None of the control patients had BHR (p < 0.05). Hyperresponsiveness of the upper airway was present in 12 of the 30 (40%, 95% CI 22%- 57%) cough patients (Table 4). Some subjects developed marked, but brief, upper airway obstruction in response to inhaled histamine with a maximum fall in FIF₅₀ of 90% from baseline. The prevalence of UAHR was less in the comparison group (4/20, 20%, 95% CI 2%-37%), but this failed to reach statistical significance (Table 4; p > 0.05). The maximum percentage fall in FIF₅₀ during the histamine challenge was also similar in those with cough (20.9%, SE 3.7) compared with controls (13.5%, SE 3.1; p > 0.05). There was no significant difference in the maximum FEF₅₀/FIF₅₀ ratio during the challenge between the two groups: cough 1.15 (SE 0.32) and controls 0.86 (SE 0.07). These findings refuted the hypothesis that UAHR was a common unifying mechanism in chronic cough.

Sputum Analysis

The mean eosinophil count in the cough patients was 3.0% and ranged from 0% to 34% (Table 5). Eosinophils were present in the sputum of 50% of the cough patients (13/26), which was significantly more than the control group (3/16, 19%; p < 0.04). Sputum eosinophilia of greater than 3.2% was present in 6 (20%) of the chronic cough patients, and in two controls who had coexisting rhinitis. The majority of the cough patients with sputum eosinophils had normal bronchial responsiveness. Mast cell counts were not different between the two groups. Oil-red-O scores were similar in the two groups and no subject had a lipid score of greater than 100. 

Psychiatric Symptoms

Forty-one percent of the subjects with chronic cough reported clinically significant symptoms in the somatization domain and 26% in the hostility domain of the symptom check list (Table 6; p < 0.05, versus control). There was no association between cough complications and psychiatric symptoms.

Diagnostic Abnormalities

The majority of cough patients had one or more abnormalities present on testing (21/30, 73%) (Figure 1), whereas abnormalities were detected less frequently in the control subjects (6/20, 30%, p = 0.002). The most frequent abnormalities were UAHR (12/30, 37%) and somatization (10/30, 33%). More than one abnormality was present in nine (30%) patients (Figure 2).

View larger version (19K): [in this window] [in a new window]   Figure 1.   Diagnostic abnormalities in patients with chronic cough (open bars) and controls (hatched bars). Patients with cough had more abnormalities than controls overall (p < 0.05), and in particular had more somatization (SCL-90) and bronchial hyperresponsiveness (BHR) (p < 0.05). UAHR = hyperresponsiveness of the upper airway. Eos = sputum eosinophils > 3.2%.

View larger version (15K): [in this window] [in a new window]   Figure 2.   Venn-diagram demonstrating the overlap between diagnostic abnormalities in patients with chronic cough. See Figure 1 for abbreviations.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study has examined mechanisms that may be involved in chronic cough. We have confirmed the high frequency of nasal disease and symptoms of gastroesophageal reflux in patients with chronic cough. Sputum eosinophilia, bronchial hyperresponsiveness, and clinically significant psychological symptoms were each identified in the cough patients to a greater degree than in controls. Eosinophilic inflammation was present in about half of the cough patients and BHR in 23%. These findings confirm that airway inflammation can occur in the absence of BHR (12), and raise the possibility that it may be relevant to the pathogenesis of cough in some patients.

The results of this study suggest that chronic cough has a diverse pathogenesis, with several independent mechanisms being responsible. There were significant differences in psychological symptoms between the patients with cough and control subjects. Patients with cough reported more psychological symptoms which raises the question of whether psychiatric disease may predispose patients to develop cough, or whether chronic cough may cause psychological symptoms. Most causes of chronic persistent cough occur in association with asthma, rhinitis, and/or gastroesophageal reflux. The mechanisms of cough in these apparently unrelated conditions have not been defined. We postulated that hyperresponsiveness of the upper airway may be an important and common mechanism leading to chronic cough. Although some subjects clearly demonstrated UAHR, this also occurred in control subjects without cough.

Several prior studies have examined upper airway function in chronic cough. Irwin and coworkers (22) selected a group of patients with cough and post-nasal drip and found evidence of upper airway obstruction from inspiratory: expiratory mid-flow ratios without bronchoprovocation. This abnormality improved with treatment. Bucca and colleagues have also studied selected subjects with sinusitis and found an improvement in UAHR with treatment (10). There are several reasons we failed to define such an abnormality in our study. First, we studied an unselected group of subjects. Although there was certainly a tendency for increased upper airway obstruction (Table 4) and UAHR in the cough subjects, this failed to reach significance. These data are consistent with UAHR being relevant in a subgroup with chronic cough but not generally. Second, our study was cross-sectional and did not examine treatment effects. A study of treatment on selected subjects with upper airway obstruction would be a more sensitive way to detect a difference in this variable (22). Third, the control subjects were chosen based upon the absence of cough only, and this meant that the control group included some subjects with rhinitis. This would tend to obscure a relationship between UAHR and cough in a sub-group with rhinitis, and the frequency of UAHR in control subjects is consistent with this. Overall, our data indicate that UAHR is not a single unifying mechanism of cough in unselected subjects. The data support multiple mechanisms operating in these subjects. The results are consistent with the literature in suggesting that UAHR may be a relevant mechanism in a sub-group of subjects, particularly those with chronic cough and rhinitis (7, 10, 22).

The patients in this study represented consecutive attenders to a tertiary referral clinic. The subjects were selected by the referral process, together with the requirement for a normal chest radiograph and FEV₁ to exclude obvious chronic lung disease. This differs from previous studies of mechanisms in chronic cough (12, 23) where subjects were highly selected and where the importance of the mechanism under study was likely to be overemphasized. To avoid this problem, we studied consecutive patients and evaluated several different mechanisms concurrently. The referral process would tend to exclude patients with milder forms of chronic cough, and those with obvious asthma or GER, and, thus, the results are not likely to be generalizable to primary care settings.

The occurrence of chronic cough due to gastroesophageal reflux is now well established, and mechanisms have been described (1, 13, 23). In this study, we were interested in defining the role of less well-established mechanisms such as aspiration, eosinophilic bronchitis, UAHR, and psychological morbidity. For these reasons, and because of problems with pH monitoring in a hospital setting (24), we relied on the presence of symptoms to detect GER and did not proceed to pH monitoring. We found that symptoms of GER and nasal disease were very common in chronic cough. However, we also found that the cough persisted despite the fact that many of the patients had received specific therapy for GER. In future studies, it will be instructive to examine changes in some of the objective markers of airway inflammation, BHR and pH monitoring in relation to therapy.

Although aspiration of gastric contents is an important cause of respiratory disease in some settings, we found no evidence to support its role in chronic cough. The presence of lipid-laden macrophages was used as a marker of aspiration (20, 21). A lipid score of 100 has good sensitivity (100%, 20) but is relatively nonspecific for aspiration. In this study we found no subjects with a lipid score over 100 and can rule out aspiration as a cause of chronic cough in these subjects. The method of sputum analysis we used ensured that pulmonary macrophages were analyzed (11) and gives results comparable with bronchial washings (25). It is possible that insisting on a normal chest radiograph as an inclusion criterion would have excluded subjects with aspiration since the chest X-ray is frequently abnormal in this setting (20, 21). In any case, the high prevalence of cough and GER occurring without aspiration indicates that there are other mechanisms causing cough in these subjects (23).

The association between psychological symptoms and cough is interesting. The validity of the relationship is supported by an epidemiological association between chronic cough and anxiety (16), and multiple case reports of psychogenic cough in the pediatric and adult literature (26, 27). It is unclear, however, whether psychological distress is a cause or an effect of chronic cough. Persistent cough could contribute to excessive psychological distress because of its ability to disturb sleep and normal work and social activities. Further studies are needed to clarify these issues.

The data in this study suggest that chronic cough results from several independent mechanisms. By systematically evaluating several mechanisms in each patient, we could identify abnormalities in 70% (21/30) of patients who presented with chronic cough. The most frequent abnormalities were UAHR and somatization, whereas BHR and sputum eosinophilia were less common. A third of patients had more than one abnormality indicating that more than one mechanism may be operating to cause cough in these patients. The importance of these mechanisms in causing chronic cough could also be evaluated by treatment trials. Specific treatment of BHR leads to a reduction in BHR (1, 4), an improvement in cough and reduced sensitivity of the cough reflex (4), confirming that BHR is a relevant mechanism in these patients. Treatment of eosinophilic bronchitis with corticosteroids leads to a reduction in sputum eosinophils (28) and an improvement in cough (12, 28), which confirms a role for eosinophilic inflammation. Therapeutic intervention for UAHR in chronic cough has not been systematically studied. The effectiveness of sodium cromoglycate in reducing ACE inhibitor cough (29), which is associated with UAHR (9), suggests that this therapy requires evaluation.

In conclusion, this study has systematically evaluated pathogenetic mechanisms responsible for chronic cough. Hyperresponsiveness of the extrathoracic airway could not be implicated as a single unifying mechanism. The results suggested that several unrelated mechanisms were operating to cause chronic cough, with hyperresponsiveness of the extrathoracic airway and somatization occurring commonly and bronchial hyperresponsiveness and eosinophilic bronchitis being less frequent. These observations confirm the need to systematically evaluate patients with chronic cough for several differing diagnoses, and to base treatment upon this evaluation.