OBJECTIVE: To simultaneously record tracheal and oesophageal pH in mechanically ventilated children to determine: (1) the feasibility and safety of the method; (2) the incidence of gastro-oesophageal reflux (GOR) and pulmonary contamination; and (3) their associated risk factors. DESIGN: Prospective study. SETTING: Paediatric intensive care unit in a university hospital. PATIENTS: Twenty mechanically ventilated children (mean age 6.7 years) who met the following inclusion criteria: endotracheal tube with an internal diameter of 4 mm or more (cuffed or uncuffed), mechanical ventilation for an acute disease, no treatment with antiacids, prokinetics, or H2-receptor blockers, and no nasogastric or orogastric tube. METHODS: The tracheal antimony pH probe was positioned 1 cm below the distal end of the endotracheal tube. The oesophageal antimony pH probe was positioned at the lower third of the oesophagus. pH was recorded on a double channel recorder and analysed with EsopHogram 5.01 software and by examination of the trace. The following definitions were used: GOR index, percentage of time pH < 4; pathological GOR, GOR index > 4.8%; tracheal reflux, fall in tracheal pH < 4, 4.5, or 5, or a decrease of one unit from baseline, in both cases preceded by an episode of GOR. The results were analysed statistically by Fisher's exact and the Kruskal-Wallis test. RESULTS: The procedure was well tolerated and the median duration of analysable recording was 6 hours (range 5-22.6). Pathological GOR was observed in eight (40%) children. GOR was more frequent with an uncuffed endotracheal tube than with a cuffed one (p = 0.01). Tracheal reflux (pH < 4) was observed in four children (20%) without clinical evidence of pulmonary aspiration. Episodes of tracheal reflux were associated with a GOR index > 10% (p < 0.01) and were more frequent with a maximal inspiratory pressure of < 25 cm H2O (p = 0.03), but were not related to the indication for mechanical ventilation, whether the endotracheal tube was cuffed or not, age, or drug treatment. CONCLUSIONS: Simultaneous tracheal and oesophageal pH monitoring was feasible in the setting of this study. Tracheal reflux can occur without pathological GOR, and GOR may occur without tracheal reflux. Further prospective studies in larger groups of patients are now justified.