Idiopathic pulmonary arterial hypertension (IPAH) is a severe and progressive disease leading to a gradual increase in right ventricular pressure, right heart failure, functional incapacity, and death. The symptoms are non-specific and therefore presentation is late. Children present with shortness of breath, fatigue, failure to thrive, and even syncope. The disease remains incurable but recent advances in medical treatment have improved survival and quality of life in both children and adults.¹ In adult practice atrial septostomy is an additional strategy in the treatment of selected patients in World Health Organization (WHO) functional class III and IV with recurrent syncope or significant right heart failure refractory to medical treatment.^2–5 The procedure permits right to left shunting and so decompresses the right heart and can increase left ventricular preload during an acute rise in pulmonary vascular resistance. Observed beneficial effects include prolonged survival with deferral of transplantation and improved quality of life, symptoms, and haemodynamic function.^6,7,8,9,10 Atrial septostomy has, however, been associated with a high risk of intraprocedural and postprocedural mortality, up to 30% in several series.^7,8,9,10 Experience is far greater with adults than with children and most of the adult patients undergoing the procedure were critically ill.

Many issues still require clarification, particularly in children. These include the optimal timing of the procedure, associated morbidity and mortality, and long term effectiveness. To define better the role of atrial septostomy in the treatment of children with severe IPAH we analysed our experience with 20 children referred to the UK Service for Pulmonary Hypertension, the youngest being 3 months of age.

METHODS AND RESULTS

Patient population

We reviewed, retrospectively, 20 children with severe pulmonary arterial hypertension who underwent an interventional cardiac catheterisation study to create an interatrial communication. Nineteen patients had IPAH, nine boys and 10 girls. One girl had pulmonary hypertension associated with HIV (table 1). The mean age at septostomy was 8.4 years (range 0.25–17 years).

All patients were notably symptomatic, nine in WHO functional class III and 11 in class IV (table 1). Thirteen patients (65%) had recurrent syncope and seven others (35%) were deteriorating in right ventricular failure despite treatment. Three complained of chest pain and palpitations. In all 20 patients the systemic arterial oxygen saturation was at least 90%. Seven older children could perform a six minute walk test. The mean distance walked was 326 m (range 160–432 m). The mean haemoglobin concentration was 125 g/l (range 100–178 g/l). Seventeen children were receiving specific pulmonary hypertensive treatment as a single or dual therapy; four of these were newly referred and being treated only with nifedipine. Three children were untreated before presentation to our service.

The ECG showed signs of right ventricular hypertrophy in all patients, with a strain pattern in 13 (65%). The P wave was ⩾ 0.25 mV in nine patients. Transthoracic echocardiography showed a tricuspid regurgitant jet velocity of at least 4 m/s in 17 children and 3.1–3.4 m/s in the remainder. Right ventricular systolic function was assessed with a semiquantitative grading system where 1 indicates normal and 2 mildly, 3 moderately, and 4 severely impaired.¹¹ Seven patients had moderately or severely impaired right ventricular function (table 2).

Cardiac catheterisation (table 2) was performed under general anaesthesia. The mean pulmonary artery pressure approached or exceeded the mean systemic arterial pressure in 15 patients. The mean right atrial pressure was 9.3 mm Hg (range 3–19 mm Hg). The mean pulmonary vascular resistance was 23 units·m². The pulmonary arterial pressure and pulmonary vascular resistance were fixed in 19 patients and did not fall in response to acute vasodilator testing with a combination of 65% oxygen and nitric oxide. Case 12 had had several cardiac arrests before and on admission to hospital but she had a positive vasodilator response (to prostacyclin), the pulmonary artery pressure falling to one third systemic arterial pressure.

The mean time from diagnosis to septostomy was 2.5 years (range one month to 16 years). Eight of 20 (40%) patients had the septostomy performed within six months of the diagnostic cardiac catheterisation. Three children were so severely symptomatic when they presented that it was thought safer to begin treatment with intravenous epoprostenol for a week or more and then perform an atrial septostomy at the same time as the cardiac catheterisation study, under the same anaesthetic. The clinical condition of three other children admitted for atrial septostomy, particularly right ventricular function on echocardiography, gave cause for concern and therefore epoprostenol was started the day before the procedure and maintained subsequently.

Technique of atrial septostomy

In all children the procedure was performed electively, under general anaesthesia. All the interventional procedures were carried out under fluoroscopic and echocardiographic guidance.

The atrial septostomy was carried out by using different techniques (table 3). Graded balloon septostomy was carried out in eight cases, blade septostomy in two cases, and a combination of a blade plus graded balloon septostomy in another three. More recently, a graded balloon approach was followed by the introduction of a custom made fenestrated atrial septal defect (ASD) Amplatzer device, which was implanted in seven patients. Patients who had a fenestrated device implanted and were not already anticoagulated were started on warfarin with a target international normalised ratio between 1.8 and 2.5.

Procedural complications

A communication was successfully created in all but one patient (case 20). The mean size of the communication was 7 mm, as assessed by two dimensional and colour flow echocardiography immediately after the procedure (table 3).

The procedure was uneventful in 16 cases. Major complications occurred in four patients (20%) and all eventually made a good recovery. Case 3 had arrhythmias and a cardiac arrest just before the procedure, was promptly intubated, and was mechanically ventilated until a successful graded balloon septostomy was carried out 13 days later. Case 6 had presented with syncope. Implantation of a 5 mm fenestrated ASD Amplatzer device was complicated by a cardiac arrest, and the patient required mechanical ventilation for 13 days. In case 17 insertion of a fenestrated device was complicated by hypoxaemia and acidosis requiring emergency reintubation and ventilation for five days. Patient 20 was the only patient in whom the procedure failed. The systemic arterial oxygen saturation fell from 98% to 58% during graded balloon septostomy and a 14 mm ASD occluder Amplatzer device was promptly inserted. This patient had a high right atrial pressure of 19 mm Hg. She was ventilated for 24 hours and subsequently discharged on epoprostenol.

Outcome

All the patients were discharged home. Fifteen patients were given continuous epoprostenol infusion. Seven were taking bosentan but in combination with epoprostenol in four (table 1). The patients were reviewed regularly at 3–6 monthly intervals. Syncope was abolished in all except one child (case 1). He had familial IPAH and died two months after the procedure. The seven other patients in right heart failure improved clinically, as judged by a reduction in hepatomegaly and peripheral oedema. The mean systemic arterial oxygen saturation decreased from 97.8% before septostomy to 90% after 3–6 months (p < 0.01). Functional capacity improved as shown by a favourable shift in WHO functional class (p < 0.001) (fig 1). Six children had a shift from WHO class IV to III, and four from WHO class III to II. There was no change in the six minute walk test done by the older children. The only change seen in the ECG was a reduction of the P wave voltage in four patients. Echocardiography showed an improvement in right ventricular function in seven patients and no deterioration in the rest except case 1, who died (p < 0.03). The velocity of the tricuspid valve regurgitant jet was unchanged. At rest the flow across the atrial communication was generally bidirectional.

In case 19, chest pain did not improve after the first procedure and a second, successful procedure was done seven months later, resolving symptoms. In case 17 a second procedure with a fenestrated device insertion was done 16 months after the first improving right ventricular function.

After a mean follow up of 2.1 years (one month to 6.7 years) clinical improvement has been maintained in 18 of the 19 patients with a successful atrial septostomy; the remaining patient is also well and at college. Bosentan and sildenafil were given to two children before bosentan was shown to decrease the plasma concentration of sildenafil.¹² Two children have since had bilateral lung transplantation (cases 5 and 9). In addition to the baby who died at 5 months of age, one patient (case 10) died suddenly five years later after a severe haemoptysis.

DISCUSSION

The present study shows that atrial septostomy is safe and effective in children and young people with IPAH. This is the youngest series of patients yet reported. There were no fatalities and the procedure was successful in 19 of 20 patients, although four patients had a significant morbidity. Syncope was abolished and right heart failure improved. Echocardiographic assessment of right ventricular function improved significantly in seven children and did not deteriorate in the remainder. WHO functional class also improved significantly. Right to left interatrial shunting caused a mean reduction in systemic arterial oxygen saturation of 7.8 percentage points. Closure of one atrial communication led to the introduction of a fenestrated device in subsequent interventions. Seventeen of the 19 children who had an atrial septostomy are alive after a mean follow up of 2.1 years. Two have had a successful bilateral lung transplantation.

Thirteen children in the present series had recurrent syncope, including a child who presented at 2 weeks of age. Children with syncope or right heart failure have a poor prognosis. Syncope can occur at rest, likely caused by a pulmonary hypertensive crisis. It is, however, believed to be more common on exercise when systemic vasodilatation occurs without an appropriate increase in cardiac output. Atrial septostomy is thought to protect against syncope by ensuring transatrial flow of blood to maintain left ventricular output, particularly when the pulmonary arterial pressure rises acutely, and to decompress the right heart and so improve right heart failure. Oxygen transport to the tissues improves due to an increase in cardiac output, despite a modest reduction in systemic arterial oxygen saturation. Ensuring an adequate packed cell volume is obviously important. We did not evaluate the haemodynamic effect of atrial septostomy immediately after the procedure because we were anxious not to prolong use of the anaesthetic in these sick children. But studies on older patients have shown an immediate increase in cardiac index ranging from 15–58%, a fall in systemic arterial oxygen saturation similar to that seen in the present study, with no change in pulmonary arterial pressure.^2,8,10 Fifteen adults were recatheterised between 18 and 27 months after atrial septostomy in three studies.^2,9,10 Cardiac index had increased in one study and right atrial pressure had decreased in two studies. The right atrial pressure is generally higher in adults than the mean of 9.3 mm Hg found in the present paediatric study. The studies on adult patients showed that haemodynamic improvement correlates with clinical improvement and we certainly saw evidence of sustained clinical improvement in the children.^2,9,13

Atrial septostomy was a relatively safe procedure in the present paediatric study but has not always been so in adult practice. A summary report of published studies concluded that 13% died during or immediately after the procedure, and the 30 day survival was 82%.¹³ The current WHO exclusion criteria for atrial septostomy are a right atrial pressure > 20 mm Hg, systemic arterial oxygen saturation of < 90%, predicted one year survival of < 40%, and pulmonary vascular resistance > 55 units·m². The only instance in which we almost contravened these recommendations in the present study was in attempting an atrial septostomy in a 16 year old with a right atrial pressure of 19 mm Hg. Septal puncture led to massive right to left interatrial shunting and severe desaturation, necessitating prompt insertion of an Amplatzer occluder device. Although we adhered to the recommendations, three other children either became very hypoxaemic or developed a tachyarrhythmia, as others have reported.¹⁴ These patients are fragile. They must be kept well hydrated to maintain an adequate right heart filling pressure and, if they have right heart failure, diuresis must be gradual. Those who were notably symptomatic with poor right ventricular function on echocardiography started epoprostenol treatment before catheterisation. General anaesthesia was used in all patients and required a high level of expertise to ensure good analgesia, cardiovascular stability throughout the study, and safe extubation.

The atrial septostomy technique was modified with experience during the course of the study. Blade septostomy has a low closure rate of 3% in adults but there is a risk of septal laceration and fatal hypoxaemia.^14,15 The only patient in the present series in whom the communication closed had had a blade plus graded balloon septostomy. Graded balloon septostomy is preferable and was successful in our hands although there is a 15–17% risk of later closure in adult patients.^{6,7,8,9,10,14,16} A fenestrated device is likely to ensure patency of an interatrial communication for a longer period and it offers the possibility of redilatation of the fenestration. Two models are available, one for use in small children (> 10 kg) with a single 5 mm fenestration and a larger device with two 5 mm fenestrations. Device insertion requires anticoagulation to keep the international normalised ratio between 1.8 and 2.5.

The optimal timing of atrial septostomy is much discussed. In general we carried out the procedure earlier in the clinical course of the disease than is usual in adult practice, although the indications for atrial septostomy were the same as in adults: syncope, refractory right heart failure, or both. Syncope was either the presenting feature or occurred relatively soon after presentation in the children, whereas it is an uncommon mode of presentation in adults and tends to occur later in the course of the disease in 40% of adults.¹⁷ The children and young people in right heart failure were deteriorating despite treatment; atrial septostomy offered a logical additional treatment option, and it seemed there was no advantage to be gained by waiting once the patient had been stabilised. Indeed the risk increases with time and the mortality risk is high in patients with end stage severe pulmonary arterial hypertension.^3,7 What is less clear is the role and therefore the timing of atrial septostomy in patients with severe pulmonary hypertension and a fixed pulmonary vascular resistance who have no history of syncope or evidence of right ventricular failure and whose main symptom is exercise limitation. We took a pragmatic approach in four children with this clinical picture. The risk of anaesthesia is increased in children with severe pulmonary hypertension, and when both atrial septostomy and haemodynamic studies were carried out under general anaesthesia it seemed advisable to carry out both procedures under the same anaesthetic when the child first presented. There were no procedural complications and these patients have since done well. Since the presence of a naturally occurring atrial communication is known to confer a survival advantage in patients with IPAH it would in fact be rational to consider making an atrial communication relatively early in the clinical course when the patient is stable, and certainly not to wait until deterioration is advanced when the risk of mortality is known to be significant.^2,7

In conclusion, atrial septostomy improved symptoms and quality of life in a group of severely symptomatic children with severe IPAH. The procedure was safe and effective. Complications can, however, arise in these fragile patients, and the WHO recommendation that atrial septostomy be performed only in experienced centres is re-enforced. The atrial communication can close spontaneously, careful follow up is essential, and fenestrated devices may help ensure longer patency of the atrial communication.