Review: insufficient evidence exists on the effectiveness of proton pump inhibitors for adults with prolonged non-specific cough associated with gastro-oesophageal reflux
Q What is the efficacy of various treatments for gastro-oesophageal reflux for prolonged non-specific cough in adults and children?
Medline (1966–2004), OldMedline (1951–65), EMBASE/Excerpta Medica (1997–2004), Cochrane Controlled Trials Register (searches current as of April 2005); reference lists of relevant studies; and communication with authors.
Study selection and assessment:
randomised controlled trials (RCTs) of adults and children receiving any gastro-oesophageal reflux disease (GORD) treatment (ie, anti-reflux conservative measures, H2 receptor antagonists, proton pump inhibitors [PPIs], or surgical therapy) for chronic non-specific cough lasting ⩾3 weeks (dry, non-productive cough without any other respiratory symptoms, signs, or systemic illness). 12 RCTs met the selection criteria; 3 trials involved infants and children (n = 198), and 9 trials involved adults (n = 196, mean age 48–58 y). 6 RCTs compared PPIs and placebo; 1 compared PPIs and ranitadine; 1 compared cisapride and a diet intervention; and 1 compared H2 antagonists and placebo. In a quality assessment based on 4 categories (allocation concealment, blinding, reporting of patients by allocated group, and follow up), 3 RCTs were rated as high quality in 3 categories, and 4 RCTs did not have high quality scores in any category. 6 RCTs had Jadad scores >3 out of 5.
failure to cure (proportion of patients not cured). Secondary outcomes included subjective cough indices and adverse effects.
Adults. Meta-analysis of 3 RCTs (n = 49) using a random effects model showed that PPIs (pantoprazole or omeprazole) did not differ from placebo for the proportion of patients not cured (table). Meta-analysis using a fixed effects model also showed that PPIs and placebo did not differ for mean cough scores (3 RCTs, n = 62, standardised mean difference [SMD] −0.51, 95% CI −1.02 to 0.01) or change in cough scores (4 RCTs, n = 77, SMD −0.40, CI −0.86 to 0.06). Subgroup analysis of 2 crossover trials (1 for pantoprazole and 1 for omeprazole) showed that PPIs improved mean change in symptoms more than placebo (n = 70, standard deviation units −0.41, CI −0.75 to −0.07). No study reported any significant adverse effects.
Infants and children. 1 RCT (n = 25) compared thickened with unthickened feeds in infants and found an increased frequency of cough with thickened feeds (p = 0.006). 1 RCT (n = 110) found that a pre-thickened formula reduced percentage of feeds associated with cough/gag/choke more than placebo. 1 controlled trial of 65 children (unclear if randomised) found no difference in cough between cisapride, domperidone, and no treatment.
Insufficient evidence exists on the effectiveness of proton pump inhibitors in adults with prolonged non-specific cough associated with gastro-oesophageal reflux. Evidence on the effectiveness of milk modification in infants is contradictory.
- Ruth Martin Misener, RN, NP, PhD
Chronic cough is one of several extra-oesophageal manifestations associated with GORD and often occurs without the heartburn that characterises the condition.1 Cough is a frequent presenting symptom and has many causes. Assessment and diagnosis can be challenging and, as Chang et al have noted, treatment effectiveness is difficult to evaluate because of the placebo effect and time limited nature of cough.
Chang et al clearly show that there is no consistent evidence that treatment of adults with either PPIs or H2 antagonists resolves GORD related cough. However, subgroup analysis showed an improvement in cough scores with PPI administration. This small effect suggests that PPIs may be of benefit to some patients, and those with longstanding cough symptoms may welcome the potential for symptom reduction. Before beginning on a trial treatment, patients should be informed about the relatively scant evidence of treatment effectiveness. Few adverse effects are associated with PPI use, but the cost of these drugs may be a factor in patient decision making.2 Although there is no supportive or refuting evidence for conservative treatment approaches such as diet, weight loss, and positioning, patients may nevertheless choose these treatments.
Follow up of patients receiving PPIs is important. Non-response to treatment requires re-examination of the accuracy of the diagnosis of GORD and possibly referral for further investigation.1 Finally, although this commentary has focused on the implications of the review for adults, it is also important to note the lack of evidence to support PPI use or milk modification approaches for children with chronic cough.
For correspondence: Dr A B Chang, Royal Children’s Hospital, Brisbane, Queensland, Australia.
Sources of funding: National Health and Medical Research Council Australia and Cochrane Child Health Field Bursary Scheme Canada.