The correlation between laryngo-pharyngeal reflux and dysphonia has been evaluated in patients without significant laryngoscopic findings and without vocal misuse. Studies were performed, using a validated questionnaire on typical reflux symptoms as well as instrumental means, e.g. videolaryngoscopy, multi-electrode 24-hr oesophageal pH monitoring, vocal acoustic analysis, gastro-oesophagoscopy, on 62 patients (51 male, 11 female) with dysphonia for > or = 3 months, selected from 350) consecutive patients presenting with voice disorders. Standard criteria were: absence of laryngeal neoformation (benign or malignant) and correct use of voice. Anti-reflux treatment was prescribed in all selected patients. A group of 62 selected patients without laryngo-pharyngeal disease were studied as controls. Mean values of the harmonic to noise ratio and maximum phonation time were pathological in all patients with dysphonia and significantly correlated (p = 0) with the entity of the larynx alteration. The 24-hour pH monitoring revealed gastro-oesophageal reflux in all cases with a clear prevalence of episodes in the upright, compared to supine, position. From a multiple regression analysis of pH-metric values, considered important in predicting maximum phonation time and harmonic to noise ratio alteration. the significant predictors (p < 0.01) were those parameters indicating the existence of a laryngo-pharyngeal reflux disease: in an upright position, the prevalence of the number of refluxes and of time of pH < 4. In conclusion, the association between electro-acoustic reliefs and laryngoscopic data, as well as an alteration in maximum phonation time and harmonic to noise ratio in patients with pH-metric indicative parameters of laryngo-pharyngeal reflux disease led to the hypothesis of a possible correlation between entity and duration of the reflux and dysfunction of the arytenoid muscles, upon which chronic vocal fatigue, with consequent laryngeal compensatory stress, depends.