OBJECTIVE Many heat and moisture exchangers with filter (HMEF) have been developed. In-house data from companies provide some information about their performances; unfortunately, to our knowledge, no comparative evaluation in clinical conditions has been undertaken of these newer products. The aim of this study was to compare the efficiency of two HMEFs, one hydrophobic and one hygroscopic, on humidifying capacity and the rate of bronchial colonization and ventilator-associated pneumonia in ICU patients. METHODS Prospective, randomized study. METHODS ICU of a university hospital. METHODS All patients who required mechanical ventilation for > or = 24 h during the study period. METHODS On admission to the ICU, patients were randomly assigned to one of two groups. In one group, the patients were ventilated with a hygroscopic device (Humid-Vent Filter Light HMEF; Gibeck; Upplands Vaesby, Sweden). The condensation surface was made of paper (Microwell) impregnated with CaCl2. The filter membrane was made of polypropylene. In the other group, the patients were ventilated with a hydrophobic device (Pall BB100 HMEF). The condensation surface was made of a hydrophobic resin with a hydrophylic layer. The filter membrane was made of ceramic fibers. In both groups, HMEFs were changed daily. RESULTS Both groups of patients were similar for the tested characteristics, including parameters of mechanical ventilation. Sixty-six patients were ventilated for 11.7+/-11 days with the Humid-Vent Filter Light HMEF and 70 patients for 12.2+/-12 days with the Pall BB 100. Patients ventilated with the Humid-Vent Filter Light underwent 6.0+/-3.0 tracheal aspirations and 1.7+/-2.0 instillations per day, and those with the Pall BB 100, 6.0+/-3.0 and 1.6+/-2.0 per day, respectively (not significant [NS]). Abundance of tracheal secretions, presence of blood, and viscosity, evaluated by semiquantitative scales, were similar in both groups. No difference in the rate of atelectasis was observed between the two groups (7.5% and 7.1%, NS). One episode of tracheal tube occlusion was observed with the Humid-Vent Filter Light HMEF, and one with the other HMEF (NS). One patient in each group (NS) was switched to an active heated humidifier because of very tenacious bronchial secretions despite repeated instillations. Tracheal colonization was observed at a rate of 67% with the Humid-Vent Filter Light and 58% with the Pall BB 100 (NS). A small, but NS difference was observed in the rate of ventilator-associated pneumonia: Humid-Vent Filter Light, 32% (27.1 per 1000 ventilator days); and Pall BB 100, 37% (30.4 per 1000 ventilator days). Bacteria responsible for tracheal colonization and pneumonia were similar in both groups. Three patients in each group died from their nosocomial pneumonia. CONCLUSIONS Despite differences in their components, the two HMEFs tested achieved similar performances in terms of humidification and heating of inspired gases. Only one episode of endotracheal tube occlusion was detected and very few patients (one in each group) had to be switched to an active heated humidifier. No difference was observed either in the rate of tracheal colonization or of ventilator-associated pneumonia. These data show that the hygroscopic HME (Humid-Vent Filter Light) and the hydrophobic HME (Pall BB 100) are suited for use in ICU patients.