This trial prospectively compares two methods of percutaneous tracheostomy, both routinely used in ICU: the Ciaglia progressive dilational tracheostomy and the Griggs forceps dilational tracheostomy. One hundred patients were randomized using a single-blinded envelope method to receive progressive or forceps percutaneous tracheostomy performed at the bedside. Operative time, the occurrence of hypoxaemia or hypercapnia and complications were recorded. The progressive technique took longer than the forceps technique (median 7 (range 2-26) vs. 4 (1-16) minutes, P = 0.0005). Hypercapnia occurred in both groups but was more marked with the progressive technique (56 (16) vs. 49 (13) mmHg, P = 0.0082). Minor complications (minor bleeding, transient hypoxaemia, damage to posterior tracheal wall without emphysema) were also more frequent with the progressive technique (31 vs. 9 complications, P < 0.0001). Six major complications occurred with the progressive technique, none with the forceps technique (P = 0.0085): tension pneumothorax, posterior tracheal wall injury with subcutaneous emphysema, loss of airway with hypoxaemia, loss of stoma with impossible re-catheterization, and two conversions to another technique. In conclusion, progressive dilational tracheostomy took longer, caused more hypercapnia and more minor and major difficulties than forceps dilational tracheostomy.