Secundum atrial septal defects (ASD) are common congenital cardiac lesions identified in adulthood. Patients with ASD and left to right shunts are at risk of developing pulmonary arterial hypertension (PAH).1 2 The chronic exposure of the pulmonary vasculature to increased blood flow in patients with ASD may produce histological changes in the intima and media of the pulmonary arteries and arterioles with resultant luminal narrowing and development of pulmonary hypertension.3 However, hyperkinetic pulmonary blood flow in patients with ASD does not uniformly result in PAH and the mechanism that predisposes some patients to develop PAH is unknown. Understanding the factors responsible for the development of PAH in this group of patients is clinically important because its presence adversely influences outcomes and is associated with increased morbidity and mortality.4^–8

Although data suggest that pulmonary arterial pressure can improve after closure of ASD,9 there remain a subset of patients with PAH in whom improvements after ASD closure are not seen.10 The late effects of percutaneous ASD closure in patients with PAH have not been well studied. The aim of this study was to examine the late outcomes in patients with moderate or severe PAH undergoing percutaneous secundum ASD closure and to determine the factors associated with improvements in PAH in late follow-up.

PATIENTS AND METHODS

Four hundred and eighty-four consecutive patients underwent successful transcatheter closure of secundum ASD at the Toronto General Hospital between 1 April 1999 and 31 November 2004. This was a retrospective review of all patients with moderate or severe PAH who returned for early follow-up. Clinical and echocardiographic data from baseline evaluation, early follow-up and late follow-up were obtained by chart review. Early follow-up was obtained at the time of the initial follow-up visit at our centre. These visits are typically scheduled 6 weeks after the procedure. Late follow-up (follow-up at least 6 months after the procedure) was obtained by chart review and by contacting the referring doctors. The study protocol had received approval from the hospital’s research ethics board.

Baseline clinical data included age at the time of device closure, gender and the presence of contributory risk factors for PAH. Information about functional class (according to the New York Heart Association), atrial arrhythmias and the use of cardiac drugs (including drugs used to treat pulmonary hypertension) was collected at baseline and at follow-up. Transthoracic or transoesophageal echocardiography, or both, was performed in all patients for pre-procedure assessment. The maximum diameter of the ASD and associated anomalies were determined by transoesophageal echocardiography. Left and right ventricular systolic function, right ventricular size and the degree of tricuspid regurgitation were assessed by transthoracic echocardiography. The left ventricular ejection fraction was calculated using the Quinones method. Right ventricular systolic dysfunction was assessed visually and graded as normal or hypokinetic. Right ventricle size was assessed visually and/or from the inlet end diastolic diameter from four-chamber view. The right ventricle was considered enlarged when the inlet measurement was >45 mm.11 The degree of tricuspid regurgitation was evaluated by colour Doppler.12 Right ventricular systolic pressure (RVSP) was estimated from tricuspid regurgitation jet velocity, when present.13 Right atrial pressure was estimated according to caval dimensions. After the procedure, residual intra-atrial shunting was examined by colour Doppler. Pulmonary hypertension was graded as mild (40–49 mm Hg), moderate (50–59 mm Hg) or severe (⩾60 mm Hg) according with the RVSP. Normalisation of RVSP was defined as RVSP <40 mm Hg.13

Statistical analysis

Data analysis was performed using the SPSS software (version 11.5; SPSS 2002,Chicago, Illinois, USA). Mean (SD) values were presented and data not normally distributed was presented as a median and range. Comparisons between variables were performed using Student t test, χ² or Fisher exact test as appropriate. Comparisons between serial echocardiographic measurements in late follow-up were performed using non-parametric analysis of variance tests (Friedman tests). A p value <0.05 was considered significant.

RESULTS

Table 1 shows the baseline clinical data. In total, 11% (54/484) of patients referred for percutaneous ASD closure had moderate or severe PAH. The mean (SD) age was 59 (15) years (range 20–79 years). Fifty-nine per cent (32/54) of the patients had limitation of functional capacity (New York Heart Association class ⩾II) and 22% (12/54) of patients had a history of atrial arrhythmias at the time of closure (persistent atrial fibrillation n = 6, paroxysmal atrial tachyarrhythmia n = 6). Ninety-six per cent (52/54) of patients had evidence of right ventricle dilatation and 41% (22/54) had more than mild right ventricle dysfunction. Two patients (4%) with associated coronary artery disease had left ventricle dysfunction (left ventricle ejection fractions of 40% and 48%, respectively). Fourteen patients (26%) had systemic hypertension and preserved left ventricular systolic function. Of these patients, six patients had left ventricular hypertrophy documented on echocardiography. Measurements of left ventricular wall thickness were not available in two patients. Because this was a retrospective review, assessment of diastolic function was incomplete in many patients. Cardiovascular drugs alone or in combination were taken by 63% (34/54) patients. Fourteen patients (26%) were taking oral anticoagulants. One patient with shunt-related plexiform pulmonary arteriopathy had an ASD closure after 1 year of continuous intravenous prostacyclin treatment; this case has been reported previously.14

The mean (SD) baseline RVSP was 57 (11) mm Hg (range 50–95). Sixty-three per cent (34/54) of patients had moderate PAH and 37% (20/54) had severe PAH before the device closure. There was no relationship between mean RVSP and gender (female 58 (9) mm Hg and male 58 (11) mm Hg, p = 0.91). The RVSP correlated with age at time of intervention (r = 0.73, p = 0.05), but not with the defect size (r = 0.66, p = 0.06) or stretch diameter (r = 0.07, p = 0.6). There was no difference in the mean RVSP in patients with or without systemic hypertension (53 (21) vs 51 (16), p = 0.79). Nineteen per cent (10/54) of patients had additional potential contributory causes of PAH including respiratory disease (chronic obstructive lung disease n = 4; sarcoidosis n = 1, and tuberculosis with pulmonary fibrosis n = 1), thromboembolic disease (n = 2) and left ventricular dysfunction (n = 2). Women were no more likely than men to have secondary causes of PAH (25% vs 40%, p = 0.29). The RVSP was not significantly different between patients with and without secondary causes of PAH (60 (13) mm Hg vs 57 (9) mm Hg, p = 0.27).

Early follow-up

Early follow-up (mean (SD) 2.3 (1.2 months, median 2 months, range 1–7) was available in all patients. There were no early deaths. Twenty per cent (11/54) of patients remained in New York Heart Association class II. Twenty per cent (11/54) of patients had atrial arrhythmias in early follow-up (persistent atrial fibrillation n = 6; paroxysmal atrial tachyarrhythmia in patients with a history of tachyarrhythmia n = 3 and new onset paroxysmal atrial tachyarrhythmia n = 2). Overall, the RVSP decreased from 57 (11) mm Hg to 51 (17) mm Hg (p = 0.003) (table 2, fig 1). Five patients had increases in their RVSP >10 mm Hg (compared with baseline) at early follow-up. Twenty-six per cent (14/54) of patients had normalisation of RVSP: 13 patients with moderate PAH before closure and one patient with severe PAH before closure.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Changes in right ventricular systolic pressure. Baseline, early and late follow-up.

Late clinical follow-up

Late clinical follow-up data (mean 31 (15) months, median 32 months, range 8–63) was available in 45 (83%) patients. Seventy-six (34/45) per cent of the patients were female and the mean age was 59 (15) years. The mean (SD) baseline RVSP was 58 (10) mm Hg. Baseline characteristics of these nine patients without late follow-up were similar to those with late follow-up data available; 75% were female (p = 0.60), the mean (SD) age was 60 (16) years (p = 0.78) and the mean (SD) RVSP was 59 (6) mm Hg (p = 0.52).

Two deaths occurred during late follow-up: one patient had recurrent pulmonary thromboembolism and the other patient died of a non-cardiac cause (bowel obstruction). Two months after ASD closure, a 35-year-old man with severe PAH at baseline, developed a pulmonary embolus and received thrombolysis treatment. After an initial period of stabilisation he had a recurrent fatal pulmonary embolism. Twenty per cent (9/45) of patients had documented atrial arrhythmia (persistent atrial fibrillation n = 4; paroxysmal atrial tachyarrhythmia n = 5). No new onset atrial tachyarrhythmia was documented at late follow-up; however, two patients with persistent atrial fibrillation were lost from follow-up.

Cardiovascular drugs were used by 44% (20/45) of patients. One patient, described above, was receiving intravenous prostacyclin therapy before closure and bosentan treatment after closure.15 In two additional patients pulmonary vasodilator therapy was started after closure. One patient, a 45-year-old woman without any secondary causes of pulmonary disease, developed severe PAH 1 year after ASD closure. Her RVSP increased from 51 to 105 mm Hg and treatment was subsequently started with bosentan. After 7 months of treatment, her RVSP had decreased to 88 mm Hg. The second patient was a 39-year-old woman with chronic obstructive lung disease. Two months after closure her RVSP had increased from 50 mm Hg to 85 mm Hg. After 26 months of bosentan treatment her RVSP had decreased to 53 mm Hg.

Late echocardiographic follow-up

In a subset of patients late follow-up echocardiograms were available (n = 39). Three patients with clinical follow-up did not have a recent echocardiogram and three additional patients had no tricuspid regurgitation at the time of the echocardiogram and therefore the RVSP could not be calculated. Overall, the RVSP decreased from 58 (10) mm Hg at baseline to 44 (16) mm Hg in late follow-up (p = 0.004) (table 2, fig 1). Three patients had increases in their RVSP >10 mm Hg (compared with baseline) in late follow-up including one patient who subsequently died. Improvements in the RVSP were not related to the duration of follow-up (p = 0.3). Two patients with associated cardiac lesions (one restrictive ventricular septal defect, one severe pulmonary regurgitation) had regression in their RVSP from 53 mm Hg and 75 mm Hg to 40 mm Hg and 38 mm Hg, respectively. In patients with moderate PAH at baseline the mean RVSP decreased from 51 (2) mm Hg to 38 (17) mm Hg. Similarly, in patients with severe PAH at baseline the mean RVSP decreased from 70 (11) mm Hg to 48 (22) mm Hg. Although the overall mean RVSP decreased at late follow-up, only 44% (17/39) of patients had normalisation (<40 mm Hg) of the RVSP: 15 patients with moderate PAH before closure and two patients with severe PAH before closure. There was no difference in the mean RVSP at late follow-up in patients with or without systemic hypertension (43 (18) mm Hg vs 41 (20) mm Hg, p = 0.70).

The characteristics of patients with residual RVSP ⩾60 mm Hg (6/39) were compared with patients with RVSP <60 mm Hg (33/39) and the results are shown in table 3. In comparison with patients with RVSP <60 mm Hg, patients with residual PAH were more frequently symptomatic after device closure (15% vs 67%, p = 0.02) (table 3).

DISCUSSION

Transcatheter closure in patients with secundum ASD and pulmonary hypertension can be successfully performed in selected subjects and is associated with good outcomes. Early improvements in pulmonary artery systolic pressures are seen in patients with moderate or severe PAH undergoing transcatheter ASD closure. Continued improvement in pressures occurs in late follow-up. Despite an overall improvement in the mean RVSP, many patients do not have complete normalisation of pressures.

Consistent with previous studies which reported prevalence rates of PAH of 6–17% in patients with ASD,4 7 15 only a small portion of patients (11%) referred to our centre for ASD closure had PAH (7% moderate PAH, 4% severe PAH). A subset of these patients had additional contributing causes of PAH. The preponderance of female patients (3.2:1) in our study is higher than the normally reported female:male ratio of 2:1 typically reported for patients with ASD. In a study by Sachweh et al, 75 patients with shunt-related PAH underwent perioperative pulmonary biopsy. In that study, women were more likely to have more severe histological PAH-related changes in the lungs.3 This observation may, at least in part, explain the increased female to male ratio observed in our study.

Improvement in functional capacity after ASD closure has been previously reported.9 Our study confirms this finding even in this group of patients with pulmonary hypertension. This improvement may in part be due to decreases in the volume load to the right ventricle and subsequent improvement in left ventricular diastolic, and perhaps systolic function, a finding that has been documented after ASD closure.16 17 In addition, patients with volume-overloaded right ventricles may have less functional reserve.18

In this series atrial arrhythmias were relatively common. Earlier observations have suggested that chronic volume overload with varying degree of pulmonary hypertension and right ventricular dysfunction may be factors associated with arrhythmogenesis.19 20 Two previous studies have reported higher systolic pulmonary artery pressures in patients with atrial arrhythmia undergoing ASD surgery.3 21 de Lezo et al demonstrated that late after transcatheter ASD closure there was reduction in the prevalence of atrial fibrillation (41% to 24%) and this change was related to the decrease in RVSP.9 Discrepant findings between their study and ours may be explained by the high number of patients with persistent RVSP >40 mm Hg in our study and the subsequent effect on the right atrium and pulmonary veins. However, the numbers of patients with atrial arrhythmias in our series were small and therefore meaningful conclusions are difficult.

Hyperkinetic pulmonary blood flow in patients with ASD does not uniformly result in PAH and the mechanism which predisposes some patients to develop PAH is unknown. Progression from hyperkinetic PAH (high pulmonary flow) to obstructive pulmonary vascular disease (high pulmonary vascular resistance) in adult patients is probably multifactorial and not solely dependent on the shunt flow. Most likely there is an interaction between genetic predisposition and other contributing factors.22 Similar to previous reports, we did not demonstrate a correlation between the degree of PAH and defect size,1 3 suggesting that additional factors may be contributing to the development of PAH. In our cohort, among patients with moderate or severe PAH, additional potential risk factors for PAH were present in 10/54 (19%). Contributory risk factors including pulmonary vasoconstriction and endothelial dysfunction related to hypoxic states such as chronic lung disease, left ventricular dysfunction and thrombogenic factors have been implicated in, and may contribute to, aggravation of longstanding shunt-relate PAH.23^–25 Furthermore, the correction of intra-atrial shunting alone did not result in normalisation of pressures in many of the patients. Whether normalisation of pressures is necessary or whether reduction in RVSP is adequate is not known.

Previously published data, suggested that pulmonary vasodilator therapy may offer patients with “irreversible” anatomical changes of the pulmonary vascular bed a chance for further improvement of pulmonary pressures.12 26 27 This treatment perhaps should be considered in selected patients before ASD closure, in order to improve outcomes.

LIMITATIONS

Our centre is a tertiary referral centre with patients referred from various regions of Canada. Therefore not all patients are followed up in our clinic. As a consequence late follow-up data were available in 45/54 (83%) patients. However, baseline characteristic of patients lost to follow-up were similar to those of the group of patients with late follow-up data. Assessment of diastolic function was incomplete in many patients and therefore we were unable to examine accurately the potential impact of this contributory cause of PAH. This variable requires further study. Echocardiography was performed on clinical grounds and therefore there was no standardisation of either the machine used or the sonographers performing the study. Unlike cardiac catheterisation, echocardiography cannot distinguish between pre- and postcapillary pulmonary hypertension. Unfortunately, baseline cardiac catheterisation with haemodynamic data was not available in many patients nor is it routine practice at our centre to follow up patients with serial cardiac catheterisation after ASD closure. Therefore, cardiac catheterisation data were limited and not included in this study. Because echocardiography and the cardiac catheterisations were not performed at the same time, a meaningful comparison between the RVSP as measured by the two different modalities is not possible. Finally, this was a retrospective review and pulmonary function tests and computed tomography scans were not available in all patients.

CONCLUSIONS

Transcatheter closure in patients with secundum ASD and pulmonary hypertension can be successfully performed in selected subjects and is associated with good outcomes. Early improvements in pulmonary artery systolic pressures are seen in patients with moderate or severe PAH undergoing transcatheter ASD closure. Continued improvement in pulmonary pressures occurs in late follow-up. Despite decreases in the mean RVSP in late follow-up, many patients do not have complete normalisation of pressures. Larger studies in this field are needed to determine how best to treat these patients.