Gated blood-pool imaging in mechanical left ventricular assistance following myocardial infarction in dogs.

Gated blood-pool scanning has been used to assess the physiological effects of left ventricular assistance following acute myocardial infarction in dogs. Both left atrial-aortic and left ventricular-aortic bypass improve survival up to 6 hr after acute coronary occlusion, compared with control animals. The initial measurement of left ventricular ejection fraction (LVEF) following occlusion appears predictive of the need for ventricular assistance. Assisted animals maintain their level of left ventricular function until 6 hr in contrast with control animals, whose function steadily deteriorates over this period. There was no demonstrable recovery in regional ventricular function in the infarcted territory over this period, and in all assisted animals the right ventricular function deteriorated more rapidly than in controls. These results demonstrate the efficacy of left ventricular assistance in terms of improved survival and preserved left ventricular function. Gated blood-pool ventriculography has also been shown to be a suitable technique for monitoring the physiological changes during left ventricular assistance.

Fourteen years' experience with the Björk-Shiley tilting disc prosthesis.

Between 1970 and 1984, 1,574 Björk-Shiley valve prostheses have been implanted in 1,171 patients in Glasgow. Between 1970 and 1980, 1,023 standard disc prostheses were implanted in 729 patients: 184 had aortic valve replacement, 323 mitral, and 222 multiple valve replacement. Between 1980 and 1984, 551 convexo-concave valves were implanted in 442 patients: 125 had aortic, 228 mitral, and 89 multiple valve replacement. Concomitant coronary artery bypass grafting was performed in 9.2% of the latter group as compared with 2.7% of the earlier group. The overall hospital (30 day) mortality for the group with the standard disc prosthesis was 10.4% and for patients with the convexo-concave prosthesis, 7.0%. Cumulative follow-up extends to 4,125 patient-years--standard disc group 3,378 patients-years, range 4 to 14 years (mean 5.2 years); convexo-concave group 747 patient-years, range 1.5 to 5.3 years (mean 1.9 years). The late mortality is 3.8% per patient-year--standard disc group 2.9% per patient-year and convexo-concave group 4.3% per patient year (no significant difference). Actuarial survival rate at 12 years for the whole group, excluding operative deaths, is 69.4% +/- 6.3%. The freedom from all valve-related complications at 12 years is 65.9% +/- 6.5%. there is no significant difference in the incidence of most major complications between the standard disc and convexo-concave prostheses apart from the occurrence of mitral valve thrombosis (p less than 0.05) in the standard disc group and outlet strut fracture (p less than 0.005) in the convexo-concave group. These results confirm the excellent long-term performance of the Björk-Shiley tilting disc prosthesis.

Björk-Shiley and Carpentier-Edwards valves. A comparative analysis.

Between January 1977 and December 1982, 986 Björk-Shiley and 744 Carpentier-Edwards valves were implanted in 774 and 620 patients, respectively, at the same institution. All Björk-Shiley patients and 57% of patients with a Carpentier-Edwards valve in the mitral position received long-term anticoagulation. Mean follow-up was 3.2 years (range 0 to 8.8) in the Björk-Shiley patients and 3.5 years (range 0 to 8.2) in the Carpentier-Edwards group. There was no significant difference between the two groups in hospital mortality (Björk-Shiley 7.6%; Carpentier-Edwards 6.0%), overall incidence of embolism (Björk-Shiley 1.4 per 100 patient-years; Carpentier-Edwards 1.6% py), endocarditis (Björk-Shiley 0.6% py; Carpentier-Edwards 0.8% py), periporsthetic leak (Björk-Shiley 1.6% py; Carpentier-Edwards 1.4% py), anticoagulant-related complications (Björk-Shiley 0.3% py; Carpentier-Edwards 0.1% py), valve failure (Björk-Shiley 0.78% py; Carpentier-Edwards 0.68% py), reoperation for complication (Björk-Shiley 1.68% py; Carpentier-Edwards 1.22% py), and late mortality (Björk-Shiley 3.1% py; Carpentier-Edwards 3.0% py). Actuarial freedom from valve-related events was similar in the two groups. In the aortic position, freedom from embolism was significantly better in the Björk-Shiley group than the Carpentier-Edwards group (Björk-Shiley 99% at 3 and 5 years; Carpentier-Edwards 96% and 92% at 3 and 5 years; p = 0.023). In the mitral position, the overall incidence of reoperation was higher in the Björk-Shiley group (1.78% py) than in the Carpentier-Edwards group (0.48% py) (p = 0.004). Actuarial analysis shows this difference to be confined to the first 6 years of follow-up. The commonest indication for reoperation was valve failure in both groups. However, when analysis is confined to this indication, the difference between the reoperation incidence in the mitral position becomes insignificant (Björk-Shiley 0.85% py; Carpentier-Edwards 0.29% py; p = 0.085). This study confirms the satisfactory performance of both the Carpentier-Edwards and Björk-Shiley valves in the short and middle term and indicates no clear-cut advantage for either prosthesis.

The modified Norwood procedure for hypoplastic left heart syndrome: early to intermediate results of 120 patients with particular reference to aortic arch repair.

BACKGROUND: Classic first-stage Norwood repair of hypoplastic left heart syndrome uses a homograft patch enlargement to obtain an unobstructed aorta and coronary arteries. Because of possible disadvantages of the homograft, such as lack of growth, degeneration and calcification, and availability, we have tried to repair the aorta without patch supplementation. METHODS: Between February 1993 and September 1997, 120 patients, aged birth to 47 days (median 4 days) and weighing 1.7 to 4.4 kg (median 3.1 kg), underwent first-stage palliation for hypoplastic left heart syndrome. The diameter of the ascending aorta ranged from 1.5 to 8.0 mm (median 3.0 mm). Eight patients had an aberrant right subclavian artery arising from the descending thoracic aorta. In 95 patients (group I), all duct tissue was excised and the descending aorta was anastomosed to the aortic arch, which had been opened back into the ascending aorta. Then to this confluence was anastomosed the proximal main pulmonary artery. In the remaining 25 patients (group II), continuity of the aortic arch was maintained and the repair was performed with a Damus-Kaye-Stansel anastomosis. The size of the systemic-to-pulmonary shunt was 3 mm in 48 patients, 3.5 mm in 70, and 4.0 mm in 2. RESULTS: Circulatory arrest time ranged from 19 to 105 minutes (median 54 minutes). A homograft patch was necessary for the arch reconstruction in 18 patients (15%); 9 group I patients (10%) and 9 group II (36%) (P =.001). There were 82 hospital survivors (68%); 69 group I patients (73%) and 13 group II (52%) (P =.04), 71 patients without a patch (70%) and 11 with a patch (61%) (P >.2). By multiple logistic regression, the aberrant right subclavian artery was a significant risk factor for hospital death (P =.008). There were 6 late deaths. Sixteen of 71 patients (23%) who underwent second-stage palliation had a neoaortic arch obstruction develop, with a peak gradient greater than 10 mm Hg; 14 group I patients (23%) and 2 group II (22%) ( P >.2), 15 without a patch (23%) and 1 with a patch (17%) (P >.2). Overall survivals were 57% at 1 year and 55% at 2 years. CONCLUSION: The modified Norwood procedure for first-stage palliation of hypoplastic left heart syndrome is possible in the majority of patients without the use of exogenous materials and does not result in an increased incidence of neoaortic arch obstruction. Repair of the aorta without patch supplementation may improve the potential for long-term growth of the new aorta.

False aneurysm formation: a complication following the modified Blalock-Taussig shunt.

False aneurysm formation is a rare complication of the modified Blalock-Taussig shunt. A patient is described in whom this complication arose 11 months after operation. Death resulted from rupture of the aneurysm into the right lung with associated massive hemoptysis. The onset of hemoptysis in patients with a functioning modified Blalock-Taussig shunt may be the first evidence of a developing false aneurysm.

Intraoperative ultrasonographic troubleshooting after the arterial switch operation.

BACKGROUND: Less than perfect coronary artery translocation accounts for the majority of perioperative deaths after the arterial switch procedure for transposition of the great arteries. Some types of coronary arterial anatomy are associated with a higher risk of death. METHODS: Prospective epicardial ultrasound examination of all neonates with failing left ventricle or difficulty in weaning off cardiopulmonary bypass was performed after completion of the arterial switch operation during a 2-year period from March 1994 to February 1996. The aim was to identify any mechanical, and potentially remediable, factors accounting for ventricular failure. RESULTS: Four neonates fulfilling the above criteria were identified during a 2-year period when epicardial echocardiography was routinely applied. In 2 patients coronary arterial problems in the form of kinking of the proximal left coronary artery (1 patient) and extrinsic compression of the artery by the neo-pulmonary trunk (1 patient) were identified and successfully corrected. In 2 other patients, supravalvar aortic stenosis was recognized, leading to prompt revision. CONCLUSIONS: Epicardial echocardiography has an important "troubleshooting" role in the subgroup of patients with a suboptimal hemodynamic result after the arterial switch operation. Patients with unusual coronary anatomy should routinely be candidates for such studies.

The surgical management of extravalvular aortic root infection.

Fifteen patients with extravalvular aortic root infection and associated infective endocarditis underwent urgent operation for this condition at Glasgow Royal Infirmary between 1977 and 1983. Four patients (26.7%) subsequently died between 1 and 68 months after operation. All patients underwent aortic valve replacement, with debridement or suture closure of abscess cavities as indicated. Three patients also required insertion of permanent pacing systems for complete heart block. Three other patients required further surgical intervention. Aortic valve rereplacement was done on two occasions in 1 patient and repair of a periprosthetic leak in the second patient; the third required one additional procedure to close an aorto-right atrial fistula postoperatively. Ten out of 11 survivors are in New York Heart Association Class I, and the remaining patient is in Class II. Aggressive surgical therapy without the need for complicated reconstructive procedures of the aortic root is effective in the management of extravalvular aortic root infections.

Intraoperative epicardial echocardiography in assessing pulmonary artery banding procedures.

BACKGROUND: Accurate perioperative assessment of the adequacy of pulmonary artery banding has hitherto relied on oximetry and direct pressure measurements. We report the use of epicardial Doppler echocardiography for assessing banding procedures. METHODS: Six consecutive infants (3 male, 3 female; median age, 17 days) who underwent pulmonary artery banding for complex defects were studied by intraoperative epicardial Doppler echocardiography. RESULTS: Excellent visualization of the relation of the band to the pulmonary valve and branch arteries was obtained, enabling or confirming optimal positioning. Doppler echocardiographic estimation of the pressure drop across the band was possible in each patient and resulted in further tightening of the band in 2 patients (33%) with satisfactory increases in transband pressure gradients. Intraoperative echocardiographic estimates of the transband gradients (median, 57 mm Hg; range, 52 to 71 mm Hg; mean +/- standard deviation, 59 +/- 7 mm Hg) correlated well with those obtained at subsequent transthoracic echocardiography (median gradient, 66 mm Hg; range, 52 to 67 mm Hg; mean +/- standard deviation, 63 +/- 6 mm Hg). CONCLUSIONS: Epicardial echocardiography is a valuable adjunct to direct pressure and saturation measurements for adjustment of final band size and aids in confirming accurate band placement.

Endocarditis with multiple intracardiac shunts: identification and repair.

An 8-year-old boy who suffered from Hirschsprung's disease had development of tricuspid valve endocarditis that progressed to aortic root abscess formation, development of a fistulous communication between aorta and right atrium, atrial and ventricular septal defects, and a left ventricle to right atrium defect. Several surgical procedures were required. Operation consisted initially of closure of the septal defects and aortic valve repair. This was followed by homograft replacement of the aortic valve for persistent infection, and further closure of a left ventricle to right atrium fistula.

Univentricular heart with systemic outflow obstruction: palliation by primary Damus procedure.

In 24 consecutive infants (19 male and 5 female) with complex forms of single-ventricle physiology and systemic outflow obstruction, a modified Damus operation without the use of exogenous material was undertaken in conjunction with creation of an aortopulmonary shunt 3.5 mm in diameter. The median age at operation was 6 days (range, 1 to 170 days) and the median weight, 3.4 kg (range, 2.6 to 4.6 kg). There were nine early deaths. All 15 survivors (median follow-up, 6.5 months) were clinically well without major systemic ventricular dysfunction or atrioventricular or arterial valve regurgitation. Ten of them have undergone a superior vena cava-pulmonary shunt (one death), and 1 has required patch angioplasty of the aortic arch and innominate artery with revision of the aortopulmonary shunt. The 4 other survivors are awaiting a cavopulmonary shunt. Univariate analysis yielded the chronologic rank for an individual procedure (higher risk of death early in the series), presence of aortic arch atresia, and presence or absence of transposition of the great arteries as predictors of death. This aggressive surgical approach provides excellent early palliation, and because the operation prevents abnormal ventricular hypertrophy from pressure or volume overload, systemic ventricular function is optimally conserved for a future Fontan-type procedure.

Management and outcome of infants and children with right atrial isomerism.

OBJECTIVES: To assess the current results and outcome of surgery in infants and children with right atrial isomerism and complex congenital heart disease. SETTING: Tertiary referral centre. METHODS: 20 consecutive children with right atrial isomerism and complex congenital heart disease underwent surgery over a 6 year period between August 1987 and July 1993. The results and outcome were analysed according to age, presentation, and surgical procedures. RESULTS: Patients were divided into two groups depending on age at presentation and initial surgery: group A comprised 11 patients who required surgical intervention in the first month of life (mean age 5 days); and group B comprised nine patients who required initial surgical intervention after the first month of life (mean age 6.8 months). Seven (64%) of the 11 patients in group A had obstructed pulmonary venous drainage and ten (91%) had pulmonary atresia. There were seven early deaths (64%), including the five patients who required systemic to pulmonary artery shunt and simultaneous repair of obstructed pulmonary veins. The long-term survival rate in this group was 18% (two of 11). Pulmonary venous obstruction was present in two (22%) of the nine patients in group B and four (44%) had pulmonary atresia. There were no early deaths. One patient died after a second palliative procedure. There was one late sudden death. Four patients had a Fontan operation with no deaths. Two of the remaining three patients meet the Fontan criteria. The long-term survival rate in this group was 78% (seven of nine). CONCLUSIONS: Surgical management of patients with right atrial isomerism who have complex congenital heart disease carries a high mortality and remains palliative. The overall survival rate was 45% (nine of 20); 18% in patients requiring surgery in the first month of life (group A) and 78% in patients requiring surgery after the first month of life (group B) (P < 0.001). Of the total of 20 patients, nine were potential candidates for a Fontan operation. Seven of these have undergone a Fontan procedure with five survivors.

Severe pulmonary stenosis and interruption of the aortic arch.

A case of severe pulmonary stenosis associated with a subpulmonary ventricular septal defect and interruption of the aortic arch is described in a 5-month-old girl. This combination defies explanation by current theories of the development of obstructive abnormalities of the aortic arch secondary to reduced aortic flow during fetal life, and, to the best of our knowledge, is the first reported case with this particular association.

Post-operative blood loss following cardio-pulmonary bypass in children.

Significant post-operative bleeding may occur following cardio-pulmonary bypass, sometimes necessitating early re-exploration. Two questions were asked by this study. Firstly, in children undergoing cardio-pulmonary bypass, can pre-operative, operative and immediately post-operative variables be used to predict post-operative blood loss? Secondly, in the same patient group, can the same variables be used to predict the surgical intervention of re-exploration for excessive bleeding within 12 h of surgery? In a prospective study of patients operated on in a 9-month period, bleeding fell rapidly in the first 3 h after surgery and early re-exploration for bleeding was necessary in 8 of the 206 patients. The only significant predictor of post-operative blood loss was minimum oesophageal temperature, low values being associated with greater blood loss (correlation r = -0.19 (P < 0.05)). With regard to re-exploration for bleeding, a long bypass time (P < 0.05) and abnormal pre-operative coagulation indices (P < 0.01) were significant and independent predictors. The data were also used to create a practical nomogram relating re-exploration for bleeding to hourly post-operative blood loss.

Congenital left ventricular inflow obstruction: is the outcome related to the site of the obstruction?

Between 1978 and 1987, 39 patients aged 1 day to 15 years underwent surgery for symptomatic left ventricular inflow obstruction. Four diagnostic groups were identified: cor triatriatum (6 patients), supravalvar mitral membrane (SVMM) with a normal mitral valve (7 patients), SVMM with an abnormal mitral valve (9 patients) and mitral stenosis (17 patients). Associated cardiac anomalies occurred in 26 patients (67%). There were 8 deaths (21%), 3 in patients with SVMM and an abnormal mitral valve and 5 in patients with mitral stenosis. Survival for patients with normal mitral valves was significantly better than that for patients with abnormal mitral valves (13/13 vs 18/26, P less than 0.05). There was also high morbidity in patients requiring prosthetic mitral valve replacement. These data suggest that the outcome of surgical treatment for left ventricular inflow obstruction may be predicted according to the site of the obstruction. This is best determined preoperatively by cross-sectional echocardiography which allows optimal planning of surgical strategy.

Total anomalous pulmonary venous connection: outcome of surgical correction and management of recurrent venous obstruction.

OBJECTIVE: Total anomalous pulmonary venous connection (TAPVC) can be corrected with low mortality and good outcome. If complicated by pulmonary vein stenosis (PVS), either at presentation or secondary to the repair, the long-term outcome is compromised. We have evaluated an institutional experience with TAPVC, with particular regard to the evolving management of PVS. METHODS: Retrospective analysis of 85 consecutive patients with non-isomeric TAPVC undergoing surgical correction over a 10-year period (1988-1997). In addition, three patients were referred to us with secondary PVS, having had their primary procedure elsewhere. Attention was focused on incidence of PVS, and strategies for management. RESULTS: Median age at first operation was 33 days (range 1-533). Site of drainage was supracardiac (43/88), infracardiac (20/88), cardiac (17/88), and mixed (8/88). On presentation, 35% of patients were ventilated. Early mortality was 7% (6/85), with one late non-cardiac death. 82% of the original patients (70/85) are currently well at a median follow-up of 64 months (range 6-119). The incidence of PVS requiring intervention was 11% (9/85). Median time to PVS was 41 days. In these patients, 18 balloon angioplasties, four endovascular stent placements (in two patients), and a further 23 surgical procedures were performed. Of the nine patients undergoing re-intervention after initial surgery at our institution, five (56%) survived. Two of these have no residual obstruction and right ventricular pressure (RVP) < 50% systemic, two have unilateral obstruction and RVP < 50% systemic, and one has bilateral obstruction and RVP 80% systemic. Of the three patients referred to us with secondary PVS, two are alive and well, and one died early after the first re-operation. CONCLUSIONS: Intrinsic obstruction (endocardial sclerosis or thickening) is associated with worse prognosis and earlier re-intervention than extrinsic (anatomical) obstruction. We advocate an early, aggressive approach to the management of patients with TAPVC, especially in the presence of PVS. This complication is most appropriately managed by a combination of re-operation and repeated balloon dilation.

Early results of right ventricular-pulmonary artery conduits in patients under 1 year of age.

OBJECTIVES: Management strategies for the repair of many complex heart defects require the implantation of a valved conduit between the right ventricle (RV) and the pulmonary artery (PA), often using aortic or pulmonary homograft valves. Their limited availability, however, has led to the development and use of new conduits. We retrospectively compared our experience with small homografts in patients of less than 1 year of age with the TissueMed bioprosthetic valved conduit. METHODS: From March 1994 to November 1997 29 patients in their first year of life underwent conduit implantation for complex heart defects. These were retrospectively reviewed in order to determine the incidence of death or conduit stenosis. Seventeen patients received homografts and 12 TissueMed conduits. RESULTS: Diagnoses and operative details including conduit size were similar in the two groups and in all cases complete repair of the underlying defect was carried out. Early post-operative mortality was 4/17 (23.5%) in the homograft group and 3/12 (25%) in the TissueMed group. Echo Doppler evaluation within 1 month of operation showed no right ventricular outflow tract (RVOT) obstruction in any of the survivors. In the TissueMed group 8/9 (77%) survivors have gone on to develop significant RVOT obstruction within 12 months of operation. There have been three late deaths in this group all related to severe RVOT obstruction. Two patients died during an attempt at balloon dilatation and one patient died of progressive right heart failure. Five patients had successful replacement of the TissueMed conduit. One child remains well with no evidence of RVOT obstruction. At operation to replace conduit, or at autopsy, the stenoses were related to the deposition of fibrous tissue at the anastomotic suture lines. In the homograft group none of the survivors developed RVOT obstruction during the first 12 months post-operatively. There was one late death (non-cardiac in origin) and one child is awaiting conduit replacement 40 months after initial implantation for obstruction. CONCLUSIONS: The homograft is a satisfactory conduit for re-establishment of RV-PA continuity in infancy. Further work needs to be undertaken in order to elucidate the mechanisms of early graft failure in bioprosthetic conduits if these are to be a suitable alternative for RV outflow reconstruction in infants.

Complete atrioventricular septal defects: the influence of associated cardiac anomalies on surgical management and outcome.

OBJECTIVE: Major associated cardiac anomalies are known to increase the risk of repair of complete atrioventricular septal defects (CAVSDs). The purpose of this study was to examine the effects of such anomalies on the current surgical management of CAVSDs and their influence on outcome following repair. METHODS: We performed a retrospective review of a 100 consecutive non-isomeric patients undergoing repair of CAVSD at our institution, between January 1989 and December 1994; patients with partial or intermediate defects were excluded. Complete atrioventricular septal defect patients with other major cardiac abnormalities (complex) were then compared to those with isolated CAVSDs. RESULTS: There were 15 patients (15%) with associated anomalies; 3 had tetralogy of Fallot, 1 patient had pulmonary atresia, 6 had hypoplastic left or right ventricle, 1 had tetralogy of Fallot and hyperplastic right ventricle, 2 patients had double outlet right ventricle, 1 had hypoplastic aortic arch and 1 patient had aortic coarctation. The median age at operation was similar for both groups (4.2 months), while the median weight was not significantly different for isolated CAVSDs compared to complex (4.2 months vs 3.4 months, P = 0.89), but there was a higher incidence of trisomy 21 (70/85, 82% vs 8/15, 53.3%, P = 0.01). Two of the 85 isolated CAVSD patients (2.3%) had undergone palliative pulmonary artery banding, while 5 of the 15 complex patients (33.3%) had either banding or Blalock-Taussig shunts performed. The technique of CAVSD repair was identical in each group. All complex patients had standard repair of their associated anomalies. Hospital mortality was higher in the complex group (3/15, 20% vs 2/85, 2.3%, P = 0.004); all early deaths in the complex group occurred in patients with a hypoplastic ventricle. Reoperation for left atrioventricular valve regurgitation was required in six isolated CAVSD patients (7.1%) and in one complex (6.6%). CONCLUSIONS: In the absence of significant ventricular hypoplasia, the early results of surgical repair in patients with CAVSDs and associated cardiovascular anomalies are similar to those achieved in patients with isolated CAVSD.

Cavopulmonary anastomosis as long-term palliation for patients with tricuspid atresia.

Prior to 1970, most patients with tricuspid atresia (TA) and reduced pulmonary blood flow were palliated by means of a systemic to pulmonary artery shunt (SPAS). In Birmingham between 1967 and 1988, 19 patients with TA (aged 16 days to 23 years, mean 3.8 years) have undergone side-to-side cavopulmonary anastomosis (CPA) with preservation of pulmonary artery continuity. One had undergone a previous right Blalock-Taussig shunt and in 1 patient with left and right SVCs, bilateral CPAs were performed. Nine patients were less than 1 year old at the time of operation. There were no operative deaths. One late death occurred due to pneumonia. Immediate clinical improvement occurred in all cases. Six symptomatic patients subsequently required further surgery (mean interval between operations 8.9 years, range 4.6-14.2 years). The remaining 12 patients have recently been reviewed. Mean follow-up was 9.5 years (range 5 weeks to 21 years). Exercise tolerance remains good (9 in NY-HA Class 1 or 2) and resting transcutaneous oxygen saturations are higher (mean 85%, range 70%-89%) compared with preoperative values (P less than 0.01). In small children, CPA avoids the complications of SPAS and, in some patients, may itself achieve adequate long-term definitive palliation.

Delay in diagnosis of right atrial myxoma.

Clinical, echocardiographic, and nuclear angiographic findings in a 51-year-old woman who presented with a history of dyspnea are discussed. Initial echocardiography revealed no abnormality. However, a subsequent radionuclide angiogram revealed a filling defect on the right side of the heart. This represented a right atrial myxoma. Radionuclide angiography can provide a useful noninvasive tool in the diagnosis of intracardiac tumors when echocardiography has not been helpful.

Surgical closure of persistent ductus arteriosus (PDA) in infants before 30 weeks gestation.

Between October 1981 and December 1983 21 premature infants of mean gestational age 27.5 weeks (range 26-29 weeks) underwent surgical closure of persistent ductus arteriosus. Mean birth weight was 1080 g. There was no operative mortality. One death in an infant with pseudomonas septicaemia occurred two days after surgery. Twenty infants had features of idiopathic respiratory distress syndrome (IRDS) and required assisted ventilation prior to operation. Six infants had associated bronchopulmonary dysplasia (BPD) and 11 had signs of congestive cardiac failure. All infants presented with clinical features suggesting the diagnosis of PDA and in 18 the left atrial/aortic ratio was increased (mean 1.9:1). In 18 infants a trial of Indomethacin therapy had failed. This experience supports the view that surgical closure of PDA in infants born before 30 weeks gestation can be accomplished safely. We believe that surgical treatment of PDA represents the optimal therapy in this high risk group of infants.