Effect of pantoprazole on the course of reflux-associated laryngitis: a placebo-controlled double-blind crossover study.

BACKGROUND

The optimal management of patients with reflux-associated laryngitis is unclear. We performed a placebo-controlled crossover trial in patients with proven reflux disease and associated laryngitis to determine the effect of pantoprazole and to gain information on the natural course of the disease.

METHODS

Sixty-two consecutive non-smoking patients with hoarseness and proven laryngitis were examined. Scores with respect to the larynx and for subjective complaints were determined and 24-h pH-metry to assess acid reflux in the lower oesophagus and pharynx was performed. Patients with pathologic reflux were given the chance to enter a double-blinded randomized crossover trial with pantoprazole 40 mg b.i.d. and placebo for a duration of 3 months each, separated by a 2-week washout period.

RESULTS

Twenty-four of 62 patients showed pathological reflux; 21 patients were included in the study and 14 concluded all parts of the study. Both pantoprazole and placebo resulted in a marked improvement in laryngitis scores (decrease of 8.0 +/- 1.4 versus 5.6 +/- 2.6; no significant difference between the 2 treatments) and symptoms after the first 3 months (decrease of oesophageal symptom score of 2.2 +/- 1.4 versus 5.4 +/- 2.8; decrease of laryngeal scores of 8.3 +/- 3.6 versus 10.3 +/- 3.9; also no significant difference between the 2 treatments). A second pH-metry 2 weeks thereafter proved the persistence of reflux in most of these patients. Switching to pantoprazole led to a further improvement of scores. In the group switched to placebo there was recurrence only in a minority of patients.

CONCLUSIONS

The self-limited nature of reflux-associated laryngitis in non-smokers is largely underestimated. Laryngitis improves despite the persistence of reflux. Pantoprazole may be helpful especially in relieving acute symptoms, but the advantage of long-term treatment over placebo has been greatly overestimated.