Article Text


Chronic cough and gastro-oesophageal reflux
  1. S S Birring1,
  2. I D Pavord1
  1. 1Institute for Lung Health, Department of Respiratory Medicine, Glenfield Hospital, Leicester LE3 3NS, UK;
  1. G A Fontana2,
  2. M Pistolesi2
  1. 2Dipartimento di Area Critica Medico Chirurgica, Sezione di Medicina Respiratoria, Università di Firenze, Viale G B Morgani, 85 50134 Firenze, Italy;

Statistics from

In their in-depth review of chronic cough and gastro-oesophageal reflux (GOR), Fontana and Pistolesi suggest that proton pump inhibitors are the most effective treatment for GOR related cough.1 However, most studies evaluating the efficacy of proton pump inhibitors are not placebo controlled and do not use objective markers of cough severity to assess treatment responses. We are aware of two double blind, randomised, placebo controlled trials of proton pump inhibitors, the results of which are summarised in table 1.2,3 These studies show no clear benefit of proton pump inhibitors in patients with chronic cough and GOR and are consistent with the negative findings of a randomised placebo controlled study of ranitidine in a similar patient population.4 In our experience, only a small proportion of patients with chronic cough have a sustained improvement with proton pump inhibitors. Others have shown that there are no clear clinical indicators or findings on investigation that predict a treatment response.2 Further randomised controlled trials of proton pump inhibitors are required using both objective and subjective markers of cough severity such as quality of life questionnaires,5 cough monitors,6 and cough reflex sensitivity7 before we can accept that GOR is an important cause of chronic cough.

Table 1

 Summary of randomised, double blind, placebo controlled trials of proton pump inhibitors in patients with chronic cough and pathological GOR on 24 hour oesophageal pH studies

In many ways the current data in chronic cough are consistent with findings in asthma, with good evidence of a higher than expected incidence of oesophageal dysfunction and GOR8,9 but little high quality evidence of a causal association between GOR and airway symptoms or dysfunction.10 The interesting question is why oesophageal dysfunction occurs so commonly in patients with a diverse range of airway diseases. One possibility is that this is a manifestation of a global abnormality of upper aerodigestive reflexes,7 perhaps due to inflammation and dysfunction of embryologically related structures. Further studies are required to investigate this possibility.


Authors’ reply

After reading the letter by Drs Birring and Pavord concerning our review of chronic cough and gastro-oesophageal reflux (GOR),1 it appears that there is much upon which we agree. For instance, we agree that (1) the evidence in favour of anti-acid treatment for GOR related cough is not striking; (2) we need more well designed studies to definitely assess the effectiveness of acid suppressant treatment on GOR related chronic cough; and (3) the reasons why oesophageal dysfunction is so frequent in patients with airway diseases deserve to be further investigated, and the complex interactions between the airway and the proximal gastrointestinal tract await clarification.

On the other hand, we disagree with the opinion expressed by Drs Birring and Pavord that the three randomised, placebo controlled studies2–4 on the effectiveness of anti-acid treatment in patients with GOR related cough attained negative results, and we believe that the figures reported in their table do not satisfactorily summarise the outcomes of these studies. Notably, the studies by Kiljander et al2 and Ing et al3 were based on a remarkably similar experimental design. In the study by Kiljander et al2 patients with GOR related cough were randomised into two treatment groups to receive either omeprazole or placebo. After a washout period, patients were then crossed over to the other treatment. The outcome measures also included the cough score. In the patients who received placebo first and omeprazole second, the cough score at the end of the omeprazole period significantly improved compared with that recorded at both baseline and the end of the placebo period; in those who received omeprazole first and placebo second, the improvement in cough score—albeit detectable at the end of the omeprazole period—reached statistical significance only after completion of the placebo period. The lack of a significant improvement in cough scores after the omeprazole period in patients who received omeprazole first was probably due, as concluded by the authors,2 to a worsening of cough during the omeprazole period in two out of nine patients in this group. The lack of symptoms in patients during placebo treatment in those who received omeprazole first probably reflected omeprazole induced acid suppression which extended into the placebo phase resulting in a “carryover” effect.4 The same phenomenon may explain the findings by Ing et al3 who reported a significant fall in the cough score during ranitidine induced acid suppression compared with both baseline conditions and the placebo period when placebo was taken as the first drug, but no significant difference was found between ranitidine and placebo cough scores when ranitidine was taken as the first treatment.

Taken together, the results of the studies by Kiljander et al2 and Ing et al,3 rather than demonstrating the failure of anti-acid therapy, raise the question of whether the randomised, double blind, placebo controlled study design is the optimal method for evaluating the effectiveness of acid suppressant agents in the treatment of GOR related cough. The crossover trial raises the problem of an unbiased estimate of the treatment effect when differences occur because of the different sequences in which treatments are applied. Factors that can invalidate a crossover trial also include non-uniform pharmacological carryover effect, failure to return patients to their baseline conditions before the crossover, and non-uniform conditions of the patients over time.4 When these problems can be anticipated, a parallel study design involving a large number of patients may be preferable.4

In the study by Ours et al,5 six (35%) of the 17 patients with abnormal oesophageal pH monitoring studies responded to acid suppression. Thus, while the study confirms that there are patients with GOR related cough who can be successfully treated with antireflux therapy consisting only of acid suppression, failure of treatment in the remaining 11 patients may be due to a number of reasons such as resistance to omeprazole,6 the presence of another undiagnosed or inadequately treated cause of cough, and the possibility that factors including (but not limited to) acid in the gastric refluxate such as pepsin and bile salts7 may be implicated in the mediation of cough by GOR. Gastro-oesophageal dysmotility may also have a role.8 It is to be hoped that the development of new technologies such as simultaneous monitoring of intraoesophageal impedance and pH will allow us to appreciate fully the diagnostic accuracy and reliability of 24 hour oesophageal monitoring and the role of non-acid components of refluxate in GOR related cough.


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.