Introduction
Chronic cough has been arbitrarily defined as a cough lasting more than 8 weeks [
1]. The distinction between acute and chronic cough is important as the epidemiology and the aetiology of these differ. Chronic cough is a very common symptom with which patients present to both primary and secondary care [
2,
3]. A diagnostic approach involving the systematic use of specialist investigations to elucidate the underlying cause of the cough has been advocated [
4]. Whilst this approach has been shown to result in successful diagnosis and treatment of a majority of chronic cough patients, it does require the routine use of investigations which may be uncomfortable for the patient and which may not be easily accessible to physicians in primary or secondary care [
4,
5].
Many studies from around the world have repeatedly confirmed that the commonest causes of chronic cough are gastro-oesophageal reflux (GOR), asthma and related syndromes and upper airways disease (post nasal drip syndrome, rhinitis and sinusitis) [
6‐
8]. These diagnoses accounted for up to 93% of chronic cough in these studies. It is our experience that these conditions present with differentiated symptoms, which suggests that it should be possible to make a diagnosis from the history and examination.
Based on this premise we hypothesised that in a large proportion of patients with chronic cough the diagnosis can be made simply based on history, examination and simple tests that are commonly available and treatment of the underlying condition can be instituted in a timely fashion without undertaking expensive investigations. In cases where the diagnosis is not clear from initial assessment, sequential trial of therapy for the three commonest conditions should cause resolution of symptoms in a significant proportion of patients.
To test these hypotheses we designed a clinical protocol for the management of cough based on the history, examination, chest radiograph and spirometry tests, which would be available to all clinicians seeing patients with chronic cough. A protocol based on clinical assessment and routinely available simple tests would be particularly useful in primary care.
Discussion
Our results indicate that it is possible to successfully manage a majority of patients with chronic cough based on history, examination and simple tests. Patients in this study had all been seen previously by at least one physician and included tertiary referrals. A combination of treatment of the likely diagnoses and sequential therapeutic trials for the commonest causes of cough led to resolution of symptoms in two-thirds of our patients. This in fact compares favourably with other studies that have used an investigational approach in management of cough patients [
4,
12,
13]. Our results suggest that it would be possible to successfully manage a significant proportion of patients with chronic cough without the need for specialist referral or investigations. This is a management approach which is both effective and cost saving.
A previous study looking into managing patients using an approach of a sequence of trial-and-error treatments based on a presumptive hierarchy of possible diagnoses has shown that this is a feasible option [
14]. In this study chronic cough was defined as that lasting more than 4 weeks and a trial of treatment of a week was much shorter than in our study. The commonest cause for chronic cough, based on response to treatment, was found to be post nasal drip syndrome. Previous epidemiological studies have shown wide variation in specific aetiologies of chronic cough [
7,
8]. In this study a response rate of over 90% was demonstrated although arguably in view of the definition of chronic cough used, in a proportion of patients their cough may have been self limited. The results of both our study and this report support this approach to managing patients with chronic cough.
Since this study was performed in a specialist cough clinic, one might question whether the sample is representative of the general practice population. Although we know that cough is a common problem in primary care, there are few studies in the literature looking at the causes of chronic cough in this setting. Patient selection may be a problem, but in practice, the fundamental difficulty is how to confirm a diagnosis. This study’s principal aim was to test the efficacy of a diagnostic protocol, which could be used in general practice, rather than to describe the frequency of causes of chronic cough. Since approximately 75% of the patients in our sample were referred from primary care, this group should represent patients which General Practitioners find difficult to diagnose and for whom use of a clinical diagnostic protocol would be beneficial.
Confirmation of diagnosis was based on successful therapeutic trial. The most common diagnosis was GOR related chronic cough. It is well known that there is a large placebo response seen in the treatment of chronic cough. It has been shown the response in cough to acid suppression with proton pump inhibitors is no better than placebo [
15,
16]. Presence of dyspeptic symptoms was a marker of positive response seen in this study, which is what we have observed in our randomised trial as well [
15]. Unlike the classical symptoms of GOR disease where the result of gastric acid exposure lead to mucosal damage and symptoms, non-acid or even gaseous reflux is thought to play a major role in reflux associated cough. Acid suppressive therapy leads to a decrease in gastric acid secretion but does not lead to improvement in reflux events themselves as these are mainly related to transient lower oesophageal sphincter relaxation (TLOSR) [
17,
18]. Further escalation for treatment for reflux used in our clinic includes a trial of pro-kinetics (metoclopramide and domperidone) and drugs which reduce TLOSR (baclofen). Baclofen is known to attenuate pharmacologically induced cough as well [
19,
20]. Although we do not have data from well organised clinical trials to support the above, an empirical trial is justified from our understanding of reflux related cough. Although commonly used for other indications, all the above medications do have their associated adverse effects. Metoclopramide, Domperidone and Baclofen are all associated with CNS side effects. The very commonly used proton pump inhibitors also have associated side effects which include an increased risk of developing pneumonia [
15]. The benefits of a trial of therapy have to be balanced against the possible adverse effects. All our patients had chronic debilitating cough and empirical trial of treatment was discussed with them prior to initiation of therapy.
Antihistamines may have activity on a number of tussive mechanisms [
21]. Central reduction of cough threshold may be associated with sedative activity, whereas blockade of mast cell derived histamine may be an effective treatment of rhinitis or asthma [
22,
23]. Antihistamines may also work on attenuating the cough reflex itself by an effect on the transient receptor potential ion channels [
24]. Because of this lack of specificity a precise diagnosis cannot necessarily be relied on in this study. We suggest that the response to therapy experienced by patients was unlikely to be simply a placebo effect in view of the failure of previous treatment and the long duration of cough.
We assessed response to therapy based on subjective report of the patient and a graduated numeric response scale. Cough-specific quality-of-life scales have been validated in the assessment of chronic cough. Other descriptive scores or visual analogue scales have not been validated. We have shown that various forms of subjective assessment of cough are reproducible at 8 weeks and correlate well with both objective cough counting as well as cough-specific quality-of-life questionnaire scores [
25]. Hence although the scale used by us in this study cannot reflect scores on other subjective means of assessment, such as the visual analogue scale, it would be appropriate to subjectively quantify the severity of cough.
As in other studies two thirds of our patients were female. GOR was more common amongst our patients while fewer had rhinitis as compared to previous studies [
6,
7]. All patients had previously been tried on some form of therapy and it was only when these did not succeed that they were referred to the cough clinic. This may well have altered the mix of patients we saw compared with the general population. Studies have shown that chronic cough is slow to resolve on therapy and this may explain why therapeutic trials of less than eight weeks have limited success [
4,
26]. In a significant proportion of patients no underlying diagnosis was established despite a sequential trials of treatment and extensive investigations. This is the experience of other centres as well. Recently the “cough hypersensitivity syndrome” has been proposed as a unifying diagnosis for patients with chronic cough which may explain the pathogenesis of cough in this group of patients [
27]. Our group of patients did not have another clinic review once they had reported improvement in their cough. Hence long term follow up data is not available and is a limitation of this study.
It is of interest to note that although 31% of patients gave a history of an upper respiratory tract infection preceding the onset of the cough only 4% had a diagnosis of post-infective cough at discharge. In general, respiratory tract infections cause acute cough more commonly than chronic cough, but prolonged post viral cough does occur. This often resolves spontaneously and therefore may have been selected out of our study population in the time between referral and attendance at clinic. However, since viruses seem to cause cough by lowering the threshold of the cough reflex, [
28,
29] it is not uncommon in our experience to see patients in whom an upper respiratory tract infection seems to unmask a cough which is due to a different cause but was previously sub-clinical.
In contrast to previous studies which did not find information on the character, timing and complications of cough useful in identifying the underlying cause, [
30] we have shown that important diagnostic information is contained in the presenting complex of symptoms which are usually associated with the common causes of cough. For instance dyspeptic symptoms, if present, would suggest a diagnosis of reflux related cough. However quite often dyspeptic symptoms are absent in patients with reflux related cough. Other symptoms which would point to this diagnosis include: cough on phonation, on rising from bed, associated with certain foods or with eating in general [
31]. There is considerable overlap of associated symptoms which patients with chronic cough have, however based on the presentation a correct diagnosis can be established in a large proportion. We think that airway reflux leading to an enhanced cough reflex is the most common cause of chronic cough and we have termed this the cough hypersensitivity syndrome. The validated Hull Airway Reflux Questionnaire is a useful tool to diagnose this [
27].
Competing interests
None of the authors have any competing interests relevant to this study.
Authors’ contributions
All the authors were involved in the management of patients seen in our cough clinic. CFE drafted the initial manuscript. All authors have read and approved the final manuscript.