The usefulness of doppler-echocardiography for the assessment of pulmonary arterial hypertension in patients with chronic respiratory failure was evaluated in 24 consecutive patients with chronic obstructive lung disease. Seventeen of these 24 patients (71 p. 100) who had tricuspid valve regurgitation analysable by the continuous wave doppler technique were selected as study group; they included 15 men and 2 women aged from 33 to 78 years (mean 63 years). The highest maximum velocity value (method A) or the maximum velocity value averaged on several cycles (method B) of the tricuspid regurgitation jet was used to calculate the right ventriculo-atrial pressure gradient, using Bernouilli's equation. Right atrial pressure was determined by three methods: haemodynamic measurement, clinical evaluation or attribution of an arbitrary 10 mmHg value. The pulsed doppler study of the pulmonary ejection flow included measurement of the acceleration time and calculation of the acceleration time/ejection time ratio. The usual echocardiographic parameters were measured. Catheterization was performed 2.5 days on average after the doppler study. Correlations between doppler examination and catheterization to evaluate the right ventricular systolic pressure were significant (p less than 0.001) and better with method B than with method A. Depending on the method employed to evaluate the right atrial pressure, the correlation coefficients obtained with method B were: 0.93 (haemodynamic measurement), 0.91 (clinical evaluation) and 0.88 (arbitrary value of 10 mmHg). The right ventricular systolic pressure evaluated by doppler ultrasound using method B and by clinical evaluation of the right atrial pressure was 47 +/- 12 mmHg (22 to 70 mmHg), as against 51 +/- 13 mmHg (28 to 74 mmHg) measured by catheterization.(ABSTRACT TRUNCATED AT 250 WORDS)