Discussion

There are rare descriptions of isolated atrial septal defects with cyanosis in children, but in the literature the majority of these were classified as sinus venosus defects with straddling of the superior caval vein and subsequent streaming towards the left atrium. ^{Reference Abadeer, Luceri and Abraham5} The anatomical basis in these cases is the position of the atrial septal defect, which is basically the connection of the left atrium and the pulmonary veins, in which the superior caval vein also drains. ^{Reference Van Praagh, Geva and Lock6} Rarely, this can be detected as early as in fetal life, ^{Reference Vassallo, Pascotto and Pisacane7} but is sometimes diagnosed as late as in adulthood. ^{Reference Shapiro, Al-Sadir and Campbell8}

If the interatrial communication is very large, cyanosis can occur due to mixing on atrial level. This has been described, i.e., in a child with 18q deletion chromosomal anomaly. ^{Reference Kim, Park, Han, Yoon and Choi9}

True anomalous drainage of the inferior caval vein to the left atrium has rarely been described. ^{Reference Aliter, El-Haddad, Gallo and Al-Halees4} In these patients, the mechanism of cyanosis is obvious. This entity can be detected antenatally. ^{Reference Vassallo, Pascotto and Pisacane7}

The position and size of the eustachian valve can be a determinant for direction of shunt. There are reports of a floppy eustachian valve directing the flow from the inferior caval vein to the left atrium. ^{Reference Manuel, Ghosh and Alex1–Reference Raffa, Al-Ibrahim, Kayali, Sorefan and Rustom3} This has been described as early as in 1976. ^{Reference Maillis, Cheng, Meyer, Crawley and Lindsay10} The very first description of atrial septal defects with cyanosis dates from 1949, where Selzer and Lewis postulate functional drainage of systemic venous blood into the left atrium. ^{Reference Selzer and Lewis11} These assumptions were based on postmortem examinations.

Cyanosis can also occur due to additional defects, which may not be easily detected or correctly interpreted, like tricuspid regurgitation, ^{Reference Magoon, Choudhury, Karanjkar and Singh12} in which streaming through the interatrial communication leads to clinical cyanosis. An atrial septal defect with bilateral caval veins of which the left one drained through a coronary sinus leading to intermitted cyanosis has been described. ^{Reference Atik, Tanamati and Barbero-Marcial13} This entity could also have been missed easily. Of course in these cases the atrial septal defect is not isolated.

Right-to-left shunt over an isolated atrial septal defect has been described in adults with platypnea–orthodeoxia syndrome, but the shunt direction in these patients can be dependent on body position. ^{Reference Kenny, Murphy, Clough and Ashley14}

Shunt reversion in Eisenmenger’s syndrome can lead to right-to-left shunt, too, but here the underlying are changes in pulmonary vascular resistance with impaired right ventricular filling and subsequent high right atrial pressures. ^{Reference Schwerzmann and Pfammatter15} Persistent hypertension of the newborn can be a potentially fatal course of cyanosis with an isolated atrial septal defect. ^{Reference Chong, So, Fok and Gerlis16} The mechanism in both these diseases lies in high pulmonary vascular resistance, even if the cause is different.

To our recent knowledge, there is only a single case in the literature describing an anomaly of the diaphragm as the reason for right-to-left shunt through an atrial septal defect. ^{Reference Shely, Loitz, Fox and Wells17} In this case, a congenital diaphragmatic herniation into the pericardial cavity lead to altered streaming of the inferior caval vein return through the defect towards the left atrium. We assume a comparable mechanism in our case: due to the eventration of the diaphragm, the cause of the inferior caval vein was altered resulting in a different, more oblique angle of the inferior caval vein towards the fossa ovalis. Streaming was therefore not directed towards the body of the right atrium, but directly through the atrial septal defect. As contrast/dye was injected into the upper half of the body during contrast echocardiography and computed tomography angiography, shunting could not be determined during these investigations. It may be worth performing contrast echocardiography both from the upper and lower extremity in cases with otherwise unexplained mechanism of cyanosis.

Deterioration during cardiac catheterisation probably occurred as soon as the tricuspid valve was passed. Herewith, there was increased tricuspid regurgitation and hence less forward flow through the right ventricle and the pulmonary arteries, which received superior caval vein blood only. As soon as this flow was further reduced due to the catheter splintering the tricuspid valve, cardiac output was not sufficient any more. It remains unclear why this could not be re-established after the catheter was removed. Extracorporeal membrane oxygenation stabilised the patient until operation.

As soon as the atrial septal defect was closed surgically, cyanosis was solved and did not recur.