Discussion

To our knowledge, the FRANCISCO registry is the first long-term prospective multi-centre source of data on children who have isolated pmVSD with LV volume overload but no PAH. While the treatment of muscular, outlet, and inlet VSDs is well standardised,^{Reference Hoffman and Rudolph1,Reference Penny and Vick13} the management of pmVSD remains controversial. No clear guidelines exist about the indications for closure in patients who have asymptomatic isolated pmVSD with LV volume overload but no PAH. It has been suggested that closure of small pmVSDs may be warranted if Qp/Qs exceeds 1.5 or if echocardiography shows left atrial or ventricular dilation.^{Reference Eroglu, Atik, Sengenc, Cig, Saltik and Oztunc9,Reference Kidd, Driscoll and Gersony32,Reference VanAuker, del Nido, Tacy, Sigfusson and Cape33} Paediatric cardiology textbooks still advise VSD closure when Qp/Qs exceeds 2, but the indications have changed for adults. Many studies have reported delayed complications in patients who have asymptomatic unrepaired VSD with LV volume overload but no PAH.^{Reference Gabriels, De Backer and Pasquet16,Reference Baumgartner, Bonhoeffer and De Groot28,Reference Soufflet, Van de Bruaene and Troost29} We estimated that the cumulative incidence of events will rise 10% at 10 years, but there is a lack of data to support this assertion on this specific sub-type of VSD. These studies are biased by heterogeneity regarding the anatomical and physiological types of VSDs in the included patients. Thus, the incidence of complications in the specific population of patients who have pmVSD with LV volume overload but no PAH is still unknown. We, therefore, expect that our prospective study focusing on this specific population will provide new insights into the occurrence of long-term cardiovascular complications.

Furthermore, the exact definition of cardiac volume overload determined using non-invasive means in this population remains unclear in most studies. Recently, non-invasive echocardiographic measurements of LVEDD z-score values have been identified as reliable for diagnosing and monitoring LV volume overload. Thus, the FRANCISCO study will bring more accurate information on the incidence and type of complications in patients with pmVSD who meet the LVEDD z-score criterion for LV volume overload.^{Reference Kampmann, Wiethoff and Wenzel20–Reference Harmon, Sisco, Dutro and Cua22}

Currently, pmVSD with LV volume overload but no PAH is managed either by watchful waiting or by closure, depending on the clinical symptoms and practice patterns at each centre. The closure is often advocated to avoid late LV dysfunction related to cardiac volume overload. However, this complication has not been well described. Other complications of long-term VSD shunting are infective endocarditis, valvular or sub-valvular complications, heart failure, and arrhythmias. The incidence of infective endocarditis was about 1.8% (n = 4) in a cohort of 222 patients with unrepaired VSD aged 30 ± 10 years at last follow-up^{Reference Gabriel, Heger and Innerhofer5}; 2 of the 4 patients required aortic valve replacement. Aortic regurgitation by aortic cusp prolapse due to the Venturi effect (Laubry-Pezzi syndrome),^{Reference Piazza, Santoro and Russo30} pulmonary or aortic stenosis, and tricuspid regurgitation have been reported in 21, 6, and 3.7%, respectively, of patients with unrepaired pmVSD.^{Reference Erdem, Ozbarlas, Küçükosmanoğlu, Poyrazoğlu and Salih11,Reference Penny and Vick13,Reference Gabriels, De Backer and Pasquet16,Reference Ellis, Moodie, Sterba and Gill31,Reference Kidd, Driscoll and Gersony32} In the retrospective Belgian pmVSD registry of adults, aortic regurgitation affected 21% of patients with unrepaired pmVSD.^{Reference Gabriels, De Backer and Pasquet16} However, in these studies, the relationship of complications with the shunt volume was not assessed. Severe shunting and anatomical characteristics such as the aorto-septal angulation may be linked to the occurrence of sub-valvular complications.^{Reference VanAuker, del Nido, Tacy, Sigfusson and Cape33} Moreover, the anatomical features of the sub-aortic rim seem to have a major influence on the feasibility of percutaneous closure. The FRANCISCO study will obtain details on this point, as the echocardiographic core lab will specifically analyse the anatomical features of pmVSD and their associations with outcomes.

Dilation of the left atrium and LV is associated with the occurrence of heart failure and arrhythmias. Paroxysmal atrial arrhythmia has been reported in 1–3% of patients,^{Reference Gabriels, De Backer and Pasquet16} and heart failure in less than 5%. Some reports suggest that a long-standing VSD may lead to disturbed diastolic function and decreased compliance of both ventricles.^{Reference Neumayer, Stone and Somerville34} Eisenmenger syndrome has been reported in 15 (6%) of 266 patients with pmVSD,^{Reference Gabriels, De Backer and Pasquet16} with risk factors being Down syndrome and large VSD size. However, the occurrence of PAH during the follow-up of patients who have pmVSD with LV volume overload but no PAH is unusual, although the exact incidence is unknown. Our study may also offer new information on this point. Overall, the incidence of complications in patients with unrepaired pmVSD followed up for more than 1 year is considered to be low, and prophylactic pmVSD closure remains controversial in France for pmVSD with LV enlargement but no PAH, as the benefits may not outweigh the risks of surgical or percutaneous closure.^{Reference Padiyath, Makil and Braley35}

Watchful waiting is supported by the good tolerance of the shunt after 1 year of age. Moreover, there is evidence that shunt severity and LV dilation tend to decrease after 1 year.^{Reference Kleinman, Tabibian, Starc, Hsu and Gersony12,Reference Viswanathan and Kumar36} LVEDD z-score values decreased in 29 (88%) of 33 patients with LV volume overload, falling below 2 in 26 (79%) patients.^{Reference Kleinman, Tabibian, Starc, Hsu and Gersony12} The FRANCISCO study will provide additional data in a larger population by assessing the functional status and LVEDD z-score changes during follow-up.

Another argument in favour of watchful waiting is that surgical or percutaneous closure carries risks. Despite recent advances in paediatric cardiac catheterisation and surgery, pmVSD closure remains a complex invasive procedure given the close relationship of the defect with the aortic and tricuspid valves and conduction pathway. However, currently, surgical closure of isolated VSD is effective, rarely results in complications, and has a mortality rate of less than 1%.^{Reference Scully, Morales, Zafar, McKenzie, Fraser and Heinle37,Reference Mavroudis, Gevitz, Ring, McIntosh and Schwartz38} One of the complications of greatest concern is cAVB, which has been reported in less than 1% of children^{Reference Mavroudis, Gevitz, Ring, McIntosh and Schwartz38} Long-term morbidity after surgical VSD closure has also been assessed. In 174 patients with 40 years’ follow-up after VSD closure, mortality was not significantly different from that in the general population.^{Reference Menting, Cuypers and Opić39} However, the event-free survival rate was significantly lower than in the general population.^{Reference Goldberg40} Symptomatic arrhythmia, heart failure, and infective endocarditis occurred in 11, 4, and 3% of patients.^{Reference Menting, Cuypers and Opić39} The most frequent complication was mild-to-moderate aortic regurgitation, which developed in 21% of patients. In the Belgian registry, pmVSD closure was followed by atrial arrhythmias in 4% and by cAVB requiring pacemaker implantation in 4% of patients.^{Reference Gabriels, De Backer and Pasquet16} The need to perform a right atriotomy contributes to the occurrence of re-entrant intra-atrial arrhythmias. In patients who had had surgical closure in early childhood and were evaluated in young adulthood, peak oxygen uptake, ventilatory anaerobic threshold, maximal workload, and peak heart rate were lower compared to those in controls.^{Reference Heiberg, Petersen, Laustsen and Hjortdal41} However, myocardial protection and post-operative care have improved substantially over time, and VSD closure is now rarely performed at a very early age. The FRANCISCO study will supply prospective follow-up data on the incidence of complications and changes in functional status in patients with pmVSD and LV volume overload managed by surgical closure versus watchful waiting.

Systematic reviews and meta-analyses have established that transcatheter VSD closure is feasible using a coil or occluder. A pooled estimate of successful device implantation of about 96.6% has been reported.^{Reference Yang, Tai, Khin and Quek42,Reference Santhanam, Yang, Chen, Tai, Rajgor and Quek43} However, residual shunting, valvular defects, arrhythmias, and early or delayed cAVB were observed in some cases. Reports of cAVB both during the procedure and during long-term follow-up with the first Amplatzer device put a halt to its widespread use. A substantial rate of severe haemolysis related to residual shunting has limited the use of coils.^{Reference Houeijeh, Godart and Jalal44} Softer occluders are being developed and may be of interest, although strong clinical evidence will need to be obtained before considering their widespread use.^{Reference Pillai, Rangasamy and Balasubramonian45–Reference Haas, Kock and Bertram47} The newest devices seem to carry a lower risk of cAVB, but follow-ups remain short and cAVB may be delayed for several years.^{Reference Shah, Saraiya, Nikam and Jha18,Reference Ghosh, Sridhar and Sivaprakasam48,Reference Fraisse and Bautista-Rodriguez49} In most centres in France, a surgical closure is the preferred method when the closure is indicated. The FRANCISCO study will also determine the potential influence of anatomical pmVSD features on the safety of percutaneous closure and may help to design new and safer devices.

Although some meta-analyses and reviews have assessed the risk/benefit ratio of percutaneous versus surgical pmVSD closure, most studies were retrospective and did not specifically include patients older than 1 year with LV volume overload but no PAH. The FRANCISCO study will provide new answers about the management of this specific population. We also expect to identify anatomical features that predict complications. We acknowledge that a randomised controlled trial would provide a higher level of evidence about the respective merits of current treatment options. Nevertheless, given the small number and heterogeneity of current guidelines, our prospective cohort reflecting current practices in an industrialised country will offer useful data to guide therapeutic decisions.