Percutaneous dilatational tracheostomy is an increasingly accepted procedure for bed-side tracheostomy. The exact positioning of the endotracheal tube, the localization of the point for puncturing the trachea and damage to the endotracheal tube and the cuff as well as to the bronchoscope due to the puncturing process are technical problems which can endanger the course of the operation. In a prospective randomized study, we examined whether use of the laryngeal mask airway (LMA) is a real alternative to the endotracheal tube during tracheostomy. Of 48 consecutive patients only 43 fulfilled all criteria for this study: PaO2 > 100 mmHg, PaCO2 < 45 mmHg (in patients with head injury < 35 mmHg) under intermittent positive pressure ventilation (IPPV) with a mean ventilation pressure of < 25 mmHg and an FiO2 of 1.0. Patients with intestinal obstruction, hemorrhages of the mouth and nose and unfavourable anatomic conditions were not included in this study. Three more patients had to be excluded from the study because of technical problems. In 21 patients tracheostomy was performed using an endotracheal tube (ET group) and in 19 patients using a LMA (LM group). After positioning of the endotracheal tube or the LMA, tracheostomy was performed in the usual way. Arterial blood gases (PaO2 and PaCO2) were investigated before positioning of the endotracheal tube or the LMA, five minutes after this procedure and five minutes after the end of tracheostomy. Mean arterial pressure (MAP), heart frequency (HF) and peripheral oxygen saturation (SpO2), endexpiratory CO2 partial pressure (PetCO2) and minute ventilation volume (MVV) were registered every 60 seconds. The ET group and LM group did not differ regarding basic diseases, age and severity of illness. Before the beginning of tracheostomy, there were no differences in MAP, HF, SpO2, PetCO2 and PaCO2 between the two groups. Before tracheostomy, only PaO2 was significantly higher in the LM group than in the ET group. Immediately before the insertion of the tracheal cannula and five minutes after the end of tracheostomy, there were no differences in the measured parameters of the two groups. An increase in PetCO2 and a decrease in minute ventilation volume were observed in both groups. Regarding technical complications, the LMA is a safe alternative to the endotracheal tube. The choice of method should depend on the basic disease and the patient's ventilation requirements at the time of tracheostomy, while there is still a call for safe instruments guaranteeing sufficient sealing of the respiratory tract during the dilatational tracheostomy and simultaneous avoidance of technical problems during puncturing of the trachea and widening of the point of puncturing.