Bronchocutaneous fistula (BCF) can originate at different levels, from the major airways to the peripheral lung. Little is published on the influence of the fistula origin or the ventilatory effect of the airleak. This study evaluates relative CO2elimination via fistulas of various size and how different ventilatory modes influence both the quantity and quality, i.e., oxygen and CO2content, of the airleak. We created BCF with five polyethylene tubings of different diameters (tube 1, 3.0 mm; tube 2, 4.0 mm; tube 3, 5.1 mm; tube 4, 6.4 mm; tube 5, 9.8 mm) in nine dogs. Six modes of ventilation were used with each tubing: spontaneous breathing (SB), pressure support (PS), high frequency (HF), assisted controlled with inspiration set at 20% (AC20) and at 67% (AC67) of the respiratory cycle, and AC20 with an end-expiratory pressure of +10 cm H2O (PEEP). For each ventilatory mode, the fistula air flow (Vf), CO2, and oxygen partial pressure of fistula air (Pfco2and Pfo2) and arterial blood were measured. Vf was measured for all tubes, while gas analysis was done for tubes 1, 3, and 5 only. As expected, VF increased with tubing size. Vf was higher with AC67 and PEEP than with the other ventilatory modes. Pfco2was not significantly influenced by the tube size and Vf. Fistula air alveolization was increased only with HF ventilation. Air leaked via the fistula contributed significantly to gas exchange: even when expiration was totally via the fistula, the arterial gases remained unchanged. We conclude that: a) ventilatory modes which increase airway pressures, such as PEEP and AC67, increase the fistula air flow; b) the airway level of a fistula does not influence its major contribution to gas exchange; and c) HF ventilation may increase the alveolization of air leaked through a fistula.