Eight patients with severe bronchopulmonary dysplasia underwent cardiac catheterisation. Seven had a pulmonary vascular resistance >3 mm Hg.l-' min.m2 (mean 8-9, range 2-2-13-8). All had raised intrapulmonary shunts (mean 25-6%, range 5-4-50%, normal <5%). Two had a high alveolar dead space, and two had unsuspected congenital heart disease. Epoprostenol (prostacyclin), but not 100% oxygen, caused a significant fail in pulmonary vascular resistance. Death was associated with a high pulmonary vascular resistance and a high shunt. Morphometric studies in three cases showed normal numbers of airways, but increased thickness of bronchial muscle. The numbers of alveoli were reduced and the wails thickened. There was increased medial thickness in small pulmonary arteries with distal extension of muscle. In the oldest child some vessels were obliterated by fibrosis. We speculate that measurements of pulmonary vascular resistance and shunt may have prognostic value; that a trial of pulmonary vasodilators other than oxygen might be worthwhile in patients with poor prognosis; and that abnormalities of the pulmonary circulation contribute to the difficulties of managing patients with bronchopulmonary dysplasia.

No abstract

Pulmonary vascular resistance was measured in air, oxygen, and after administration of vasodilators in 14 children with pulmonary hypertension and congenital heart disease. Lung morphology was examined by light microscopy and assessed quantitatively. In this selected group of patients (a) medial muscle thickness of greater than 20% in the intra-acinar arteries and Heath-Edwards changes of I or II were significantly associated with perioperative death from pulmonary complications after cardiac surgery; (b) children with lower percentage medial muscle thickness had a higher baseline resistance (r = -0.84) associated with Heath-Edwards grade III or higher changes (most of these patients were not offered corrective surgery); (c) when the lowest pulmonary vascular resistance was less than 3 units, Heath-Edwards grading was I or II (n = 4). When the pulmonary vascular resistance was greater than 6 units, however, there was no direct correlation with Heath-Edwards grading (n = 9). Four patients with a resistance of greater than 6 units had only grade I or II changes. Three had a medial muscle thickness above 20%, and were among those who died at or soon after operation. It is concluded that (a) patients with a lowest pulmonary vascular resistance of greater than 6 units have a bad prognosis whatever their lung morphology; and (b) some patients with Heath-Edwards grade I or II will have a high resistance (this group has a high medial muscle mass and a poor prognosis and would not be detected by Heath-Edwards grading alone).

We have obtained dose-response curves for the effects of prostacyclin on the pulmonary and systemic circulations in 20 children (median age 3 years) with pulmonary hypertension complicating congenital heart disease. Results were obtained with the children breathing both air and 100% 212 implying that there may be age-and speciesrelated differences in its effects. We have therefore studied the effects of PGI2 in 20 children with congenital heart disease, both preoperatively and in the context of postoperative pulmonary hypertensive crises. MethodsWe studied 20 patients with pulmonary hypertension complicating congenital heart disease who required evaluation for pos-

trophy and the chest radiograph showed mild cardiomegaly and pulmonary congestion. The thymus obscured details of the aortic arch and there was no obvious tracheal deviation.Cross sectional echocardiography showed normal atrioventricular and ventriculoarterial connections. The right heart chambers were enlarged. The left atrium, mitral valve, and left ventricle were hypoplastic. The end diastolic left ventricular dimension was only 10 mm. A perimembranous ventricular septal defect ofmoderate size was seen. The pulmonary artery was much larger than the ascending aorta, which had a diameter of only 6 mm. The aortic valve was small but otherwise normal. The ascending aorta did not arch posteriorly but simply bifurcated. The descending aorta was to the left of the spine and of normal size. It was connected to the pulmonary artery by way of the ductus arteriosus, which was only slightly constricted. The left subclavian artery arose from its postero-superior extremity.Cardiac catheterisation showed a large left to right atrial and ventricular shunt with equal systolic pressures in both ventricles and pulmonary trunk. There was a 20 mm Hg systolic pressure drop across the ductus arteriosus to the descending aorta. The ascending aorta was not entered. Cineangiography confirmed the echocardiographic details of cardiac anatomy. Separate injections of contrast into the left ventricle (fig la and b) and ductus arteriosus (fig lc) confirmed a lack of continuity between the left common carotid and left subclavian arteries. However, the right branch of the bifurcation of the ascending aorta was not an innominate artery-as 477