Presence and number of APVCs, demographic data, medical and surgical history were recorded. Clinical, electrocardiographic and invasive hemodynamic data at the time of diagnosis were noted. At latest follow-up, electrocardiographic data and New York Heart Association (NYHA) class was collected. Data were also collected if any of the following occurred: atrial arrhythmia, sick sinus syndrome, pacemaker implantation, heart failure, transplant, death. The presence of arrhythmia was assessed on ECG during an outpatient clinic visit or using Holter assessment when clinically indicated. Heart failure was defined as signs and/or symptoms of heart failure requiring medical therapy [10].

Date and indication for catheterization, right atrial pressure, systolic and end-diastolic right ventricular (RV) pressure, systolic, diastolic and mean pulmonary artery (PA) pressure, pulmonary capillary wedge pressure (PCWP) and systolic and end-diastolic left ventricular (LV) pressure were collected at the time of diagnosis (prior to surgical repair). An estimated oxygen consumption (VO₂) using the LaFarge-Miettinen formula was used for cardiac output (CO) calculations with the Fick principle [11] and indexed for body surface area (BSA). PVR indexed was calculated using the formula PVR indexed = (mean PA pressure – PCWP/cardiac index (CI)), and expressed in WU.m². PAC was calculated as PAC = (stroke volume indexed/PA pulse pressure) and expressed in mL/mmHg.m². The RC time is the product of PVR and PAC (not indexed) and expressed in sec [12]. Shunt ratio was calculated using the Fick principle.

Data were analysed using SPSS® for Mac (version 23, SPSS, Chicago) and tested for normality with the Kolmogorov-Smirnov test. Descriptive data for continuous variables are presented as means ± SD in case of a normal distribution or as medians with interquartile ranges (IQR) in case of a non-normal distribution. Descriptive data for discrete variables are presented as frequencies and percentages. Follow-up duration was calculated from date of diagnosis until date of death or date of censure, which was September 1^st, 2019. Kaplan-Meier survival curves were plotted to visualize outcome starting at birth. For between-group comparisons, the independent Student t-test, Chi-square or Fischer-exact test was used as appropriate. The relationship between PVR indexed and PAC indexed was performed using an inverse model, as described before [8]. All tests were two-sided and P < 0.05 was considered statistically significant.

There were 136 patients with SVSD included in the study between January 1976 and September 2019. Median age at diagnosis was 14 (IQR 5–48) years, 47% of patients were male. One hundred and twenty-eight patients (94%) underwent surgical repair at a median age of 13 (IQR 5–43) years, whereas 8 patients did not undergo repair because of elevated PA pressures (and PVR) precluding safe SVSD closure. One hundred twenty (94%) patients underwent intracardiac baffle repair using a patch to baffle blood from the orifices of the abnormal pulmonary veins to the left atrium. Eight (6%) patients underwent a Warden procedure. Of the 128 patients who underwent surgical repair, thirty-three patients (26%) were <5 years, 87 (68%) were 5–60 years old and 8 (6%) were >60 years (the oldest being 79 years). APVC was present in 91% of patients. Seventy-two (53%) had one APVC, 41 (30%) two, and eight (6%) ≥3 APVC present. Eight (6%) patients required re-intervention (1 due to pericardial effusion, 3 (2%) due to residual leak, 1 (1%) to close a patent ductus arteriosus and 3 unknown). There were 3 (2%) patients with some degree of caval vein obstruction. Patients who did not undergo repair were older (72 (IQR 52–79) vs. 12 (IQR 5–43) years; p < 0.0001), had higher mean PA pressure (40 (IQR 35–62) vs. 21 (IQR 17–25) mmHg; P < 0.0001), higher PVR indexed (15 (IQR 9–28) vs. 3 (IQR 2–6) WU.m²; P < 0.0001) and lower PAC indexed (0.51 (IQR 0.40–0.86) vs. 1.86 (IQR 1.41–2.61) mL/mmHg.m²; P < 0.0001) but similar PCWP (11 (IQR 3–6) vs. 11 (IQR 6–15); P = 0.665).

Catheterization data was available in 87 (64%) patients and is summarized in Tab. 1. Fifty (57%) patients had mean PA pressure >20 mmHg and 30 (34%) patients had mean PA pressure ≥25 mmHg and would therefore have pulmonary hypertension according to the new and old guidelines, respectively [13,14]. Patient and cardiac catheterization data of patients who did not undergo repair or who had significantly elevated PA pressures prior to repair are summarized in Tab. 2. Of 8 patients who did not undergo repair, 4 (50%) had Eisenmenger syndrome and 4 (50%) had elevated PA pressures in the setting of heart failure. The youngest patient with Eisenmenger syndrome was 36 years, whereas unrepaired patients with heart failure were all >70 years of age. In patients who underwent repair, there was a linear relationship between mean PA pressure (R² = 0.23; P < 0.0001), wedge pressure (R² = 0.24; P < 0.0001), PVR indexed (R² = 0.16; P = 0.0039), PAC indexed (R² = 0.07; P = 0.0226) and age (Fig. 2). There was an inverse hyperbolic relationship between PVR indexed and PAC indexed with an RC time of 0.39 (IQR 0.26–0.53) (Fig. 3) and a weak inverse correlation between RC time and wedge pressure (R²–0.08; P = 0.041). When comparing patients with a catheterization <10 years, 10–50 years and >50 years of age, systolic PA pressure, mean PA pressure, PCWP, PVR indexed, and PAC indexed all changed significantly, whereas transpulmonary gradient and diastolic PA pressure tended to change (Tab. 1).

During a median follow-up time of 31 (IQR 17–55) years, 12 (9%) patients died, 1 (1%) underwent heart-lung transplant, 13 (10%) developed heart failure, 19 (14%) had atrial arrhythmia, 6 (4%) sick sinus syndrome and 7 (5%) required pacemaker implantation. Survival was 100% at 30 years, 94% at 50 years and 78% at 70 years (Fig. 4). When comparing patients with age at diagnosis <10 years, 10–50 years and >50 years of age, the proportion of patients with mean PA pressure ≥25 mmHg, NYHA ≥2, atrial arrhythmia (atrial fibrillation and/or flutter) and heart failure increased with age, whereas the proportion of patients with sick sinus syndrome (SSS) did not (Tab. 3).

With catheterization data available in 87 (64%) patients, PA pressures ≥25 mmHg were present in 34% of patients, and the prevalence of mean PA pressure ≥25 mmHg increased with age. For most cases an increased PA pressure was not a contra-indication for surgical repair of the SVSD since it related to increased pulmonary blood flow due to a left-to-right shunt with PVR still being low. Interestingly, although we could demonstrate that PA pressures increased with age, this rather represented an increase in PCWP and a mild increase in PVR indexed with increasing age. The presence of severe PAH with shunt reversal and arterial desaturation (Eisenmenger syndrome) in 4 (3%) patients occurred rather irrespective of age in a few patients, which would support an underlying susceptibility for developing pulmonary vascular disease in some patients [15]. One patient underwent heart-lung transplant (no pulmonary vasodilators prior to transplant) and 3 patients received pulmonary vasodilator therapy (bosentan monotherapy in 2, bosentan and sildenafil in 1). None of the patients with Eisenmenger syndrome were younger than 30 years of age. Therefore, we assume that a routine cardiac catheterization is not strictly indicated in young patients if anatomy is clear and there are no signs of pulmonary hypertension on non-invasive cardiac imaging. However, in adults, we believe threshold for cardiac catheterization should remain low in order to screen for the presence of pulmonary (arterial) hypertension. In older patients (age >50), cardiac catheterization (along with CT angiography and/or magnetic resonance imaging for anatomical evaluation) provides valuable information on both pulmonary artery pressures, left-sided filling pressures and magnitude of the left-to-right shunt. This is important since these patients often present with complex hemodynamics and a full assessment is needed to see whether surgical repair is feasible. Closure of the defect may not be indicated since this could increase LV filling pressures and worsen systemic heart failure [16,17]. Our data, with increasing pulmonary artery pressure, PVRi and PCWP, indicates the need for a cardiac catheterization is reflected in our data where the proportion of patients who underwent cardiac catheterization increased with age at diagnosis. Similarly, surgical repair was deferred in 4 patients >70 years of age because of elevated PA pressures in the setting of left heart failure.

Of interest, we report two cases where the SVSD was repaired despite the presence of a significantly elevated PA pressure and PVR. One patient underwent surgical repair with fenestration of the interatrial septum (in order to decrease shunt ratio) because of refractory cyanosis. The patient is doing reasonably well with persistent cyanosis (84% on room air), single pulmonary vasodilator and diuretic therapy. The second patient had elevated PA pressures in the presence of a large left-to-right shunt (confirmed on cardiac magnetic resonance imaging). After surgical repair the patient developed combined pre- and postcapillary pulmonary hypertension in the setting of heart failure and died a few years later. These patients reflect the limited prognosis observed in patients with persistent pulmonary hypertension after shunt closure [18].

Our data appears to be in line with data in patients with an atrial septal defect (ASD) type secundum who have a progressive increase in PA pressures with increasing age [19,20]. This study provides additional information by showing that there is also an age-related increase in PCWP. Although not unexpected, it underscores the importance of a full cardiac catheterization since a substantial portion of older SVSD patients will have post-capillary or combined pre- and post-capillary pulmonary hypertension in the setting of heart failure. In contrast, the occurrence of Eisenmenger in our dataset was rare, occurring in 3% of cases and absent <30 years of age, which appears in line with earlier reports [21].

Our data also supports an inverse hyperbolic relationship between PVR and PAC in patients with SVSD, indicating that substantial declines in PAC occur before increases in PVR are observed [8,22]. RC time does not change to a great extent with age, which is in line with previous reports [8]. Tedford et al. showed that the relationship between PAC and PVR is sensitive to changes in PCWP, as an elevation in PCWP lowers PAC for any given PVR (Fig. 1). We also observed a weak correlation between PCWP and RC time. This could be important in older patients where the presence of an atrial shunt will initially serve as a pop-off, leading to an increased left-to-right shunt and the accompanying increase in wedge pressure could cause increased pulsatile RV load further contributing to RV failure [8]. It is known that the RV is able to sustain volume load quite well, but poorly tolerates increases in pressure afterload. We have previously shown that patients with an unrepaired secundum type ASD have a steeper increase in PA pressure during exercise [3], which was related to a lower RV contractile reserve during exercise [23]. This has also been observed in PAH [24], chronic obstructive lung disease [25] and heart failure [26]. Muneuchi et al. [27] albeit in a different cohort of young infants (age 2–4 months) indicated a decrease in RC time following VSD repair (and therefore decrease in shunt ratio). It must be highlighted, however, that some authors question the relative impact of the changes observed in RC time [28].

Although we, in accordance to others [1], report good results after surgical repair of SVSDs, overall clinical status (NYHA) deteriorates and co-morbidities increase with advancing age. This finding and the inability to repair some older patients with heart failure underscore the importance of early diagnosis and intervention.

Sinus node dysfunction, although more common than after ASD repair [29] was rather infrequent occurring in 5% of patients and independent of age of diagnosis/repair. Notably, but not unexpected, was the occurrence of atrial fibrillation and flutter in 15% of patients, with a clear relationship with older age at diagnosis (42% >50 years of age) [19,30]. Surgical scar, atrial dilatation, and PA pressures probably all contribute to the occurrence of atrial arrhythmias [20,30]. Heart failure is common, occurring in 10% of patients, with a prevalence of 30% in those >50 years of age at diagnosis. Although it is reassuring that heart failure is rare at a younger age, it complicates management later in life and underscores the need for early diagnosis and treatment [31]. Nevertheless, it is worth noting that overall outcome is reasonable with 100% survival at 30 years and 78% survival at 70 years of age.

The current study is limited by its retrospective, single center design. Some information, such as ECG, echocardiography and CMR was not available in all patients. There was no standard ambulatory monitoring for atrial arrhythmias, which could therefore be underreported. PAC was measured using the ratio of stroke volume over pulse pressure, which correlates well with other methods to measure PAC²². Pulmonary hemodynamics were not available in all patients. For the purpose of this study, we evaluated PAC using the pulse pressure method since parameters were available at the time of right heart catheterization. Patients with isolated abnormal pulmonary venous connection (i.e., in the absence of a sinus venosus defect) were not included in the current analysis.