Intubation is necessary in 7 to 20% of patients with severe acute cardiogenic pulmonary oedema despite optimal treatment. This study evaluated the usefulness of parameters largely available in clinical practice to predict the need for intubation in a population of acute cardiogenic pulmonary oedema patients treated with medical therapy and continuous positive airway pressure. The present retrospective cohort study involved 142 patients with severe acute cardiogenic pulmonary oedema who were admitted to coronary care or the intensive care unit of a university hospital and were treated by an in-hospital protocol. Physiological measurements and blood gas samples were evaluated at 'baseline' (just after admission), 'early' (one to three hours after beginning treatment) and 'late' (eight to 10 hours after beginning treatment). Twenty-two patients (15.5%) required intubation. A systolic blood pressure at admission lower than 140 mmHg was significantly associated with a higher risk for intubation, while hypercapnic patients or those with a reduced left ventricular ejection fraction at admission did not show a worse prognosis. A simple score based on largely available parameters (1 point for each: age >78 years, systolic blood pressure <140 mmHg at admission, arterial blood gas acidosis and heart rate >95 bpm at early time) is proposed. The rate of intubation according to this score ranged from 0% (score of 0) to 90% (score of 3). Our study found that simple parameters available in clinical practice are significantly associated with the need for intubation in acute cardiogenic pulmonary oedema patients treated with continuous positive airway pressure and medical therapy. A simple score to evaluate the need for endotracheal intubation is proposed.