Superior vena cava flow and management of neonates with vein of Galen malformation
- 1Department of Neonatal Medicine, The Royal Hospital for Sick Children, Glasgow, UK
- 2Department of Neuroradiology, Southern General Hospital, Glasgow, UK
- Correspondence to Dr Anne Marie Heuchan, Department of Neonatal Medicine, The Royal Hospital for Sick Children, Yorkhill, Glasgow G3 8SJ, UK;
- Received 1 August 2011
- Accepted 20 December 2011
Objective Vein of Galen malformation (VGAM) in neonates presents a complex management challenge. Measurement of superior vena cava (SVC) blood flow may provide insights into the haemodynamics of VGAM and the effects of therapeutic intervention.
Methods SVC flow was assessed in 15 neonates with VGAM. SVC flow results, Bicêtre scores (clinical assessment), echocardiographic assessment and clinical outcomes are presented.
Results SVC flows (166–581 ml/kg/min) were significantly elevated at presentation (p<0.001; normal range 55–111 ml/kg/min). Endovascular intervention was undertaken in 12 cases, with nine survivors. SVC flows decreased sequentially with each embolisation, with a median SVC flow at discharge of 124 ml/kg/min (IQR 79–155 ml/kg/min). All cases with SVC flow >400 ml/kg/min (n=5) had an adverse outcome (death or profound neurological damage). Cases with SVC flow <400 ml/kg (n=10) required embolisation before discharge at a median age of 6 days. There were no survivors with Bicêtre scores <8 (n=2) but the predictive value of early Bicêtre score was poor.
Conclusions SVC flow measurements provide insight into the haemodynamic challenges of VGAM and provide additional useful prognostic information.
Vein of Galen aneurysmal malformation (VGAM) is a rare intracranial vascular malformation1 most commonly presenting as cardiac failure in the neonatal period. Endovascular embolisation has significantly improved outcome, but VGAM is still associated with high rates of mortality and morbidity.2 ,3 The challenges in clinical management of VGAM in the neonate are to identify patients with a poor prognosis for whom treatment is inappropriate, to identify cases requiring urgent intervention and to identify cases which may be deferred for later treatment, which is associated with fewer complications.1 This process is classically guided by the Bicêtre score (table 1), although accurate scoring may be difficult, particularly in the first few days of life.1 Early cranial MRI may be used to identify patients with severe underlying ischaemia, but cerebral angiographic patterns do not have good prognostic significance.2 ,3 Echocardiographic studies in patients with VGAM have demonstrated increased left and right ventricular output, but no correlation between cardiac output at presentation and survival.4 Suprasystemic arterial pressures at presentation, particularly if persistent following intervention, are associated with a poor prognosis.4 Reverse flow in the descending aorta in diastole has also been associated with a poor outcome4 suggesting that mortality is due to redistribution of blood from the systemic to the cerebral circulation.
What is already known on this topic
Bicêtre score <8 predicts poor outcome but early Bicêtre scores are unreliable.
Ventricular outputs at presentation are not predictive of poor outcome.
Retrograde flow in diastole in the aortic arch can be a poor prognostic feature.
What this study adds
Superior vena cava flow provides useful prognostic information in vein of Galen malformation presenting in the neonatal period.
Retrograde blood flow in the aortic arch extending into systole is a more specific adverse prognostic indicator than retrograde flow in diastole.
Strategies to address acute changes in postembolisation haemodynamics need to be developed.
Superior vena cava (SVC) flow measurements have been validated in term and preterm infants.5 This measurement reflects the venous return from, or blood flow to, the upper body, about 80% of which is from the brain. Unlike ventricular output, SVC flows are not affected by shunts at ductal or atrial level, both of which are common in VGAM. Our aim was to compare measurements of SVC flow at referral and after endovascular intervention with clinical outcomes in neonates with VGAM.
All neonates referred to our institution for assessment for interventional embolisation of VGAM from 2005 to 2010 were included in this study. Diagnosis was confirmed by MRI before intervention. Decisions to treat were based primarily on Bicêtre score and cranial imaging. The treatment strategy was to reduce the shunt to a level where cardiac failure could be controlled, allowing more definitive treatment at around 6 months of age. Endovascular embolisations were performed by the same interventional neuroradiologist (JB) using 100% N-butyl cyanoacrylate opacified SVC flow, superior vena cava flow (ml/kg/min); VTI, volume time integral (m); mean SVC diameter, average of maximum and minimum SVC diameters over 10 cardiac cycles. with titanium powder. Following embolisation, patients were sedated for a minimal period of 24 h.
SVC flow results, pulmonary arterial pressure and aortic flow patterns were recorded at presentation and, where clinically possible, following embolisation procedures. Measurements were taken with ultrasound scanners with cardiac capability using high-frequency (8–12 MHz) transducers incorporating colour-flow and pulsed-wave Doppler ultrasound. Congenital cardiac malformation was ruled out in all cases. SVC flow was calculated as described by Kluckow and Evans5 (figure 1). All measurements were performed by the same operator (AMH). Pulmonary arterial pressures were estimated from the maximal velocity of the regurgitant pulmonary jet using the modified Bernouilli equation. Ductal flow patterns were assessed using colour and pulse wave Doppler ultrasound. Proximal descending aortic flow was interrogated by Doppler ultrasound immediately proximal to the point of insertion of the ductus arteriosus. Retrograde flow was defined as absent, early diastolic or continuous if extending into systole. The Bicêtre score at presentation and the level of intensive care support and postintervention complications were reviewed from the case notes. Data were analysed using PASW Statistics 18 using appropriate tests of significance for parametric and non-parametric data.
Echocardiographic and clinical data are presented in table 2. The median values and IQR for gestational age and weight were 40 (36–40) weeks and 3.5 (2.7–4.0) kg, respectively, with a male preponderance (80%). Care was reorientated without embolisation in three cases. The remaining 12 had interventional embolisation, with 75% (n=9) survival. The median age at first interventional embolisation was 4.5 (IQR 3–17.2) days, with one to three interventions required in those treated, before discharge.
SVC flows were significantly elevated above the normal range (55–111 ml/kg/min at term, 43–140 ml/kg/min for preterm infants),5 with a median VGAM SVC flow of 374 (IQR 324–453) ml/kg/min (p<0.001). Elevations in SVC flow were related to increases in all components of the flow calculation, that is median heart rate 165 (IQR 137–181) per minute (normal range 120–160), median velocity time integral (VTI) 0.180 (IQR 0.130–0.240) (normal range 0.099–0.1777) mm and median SVC diameter 7.7 (IQR 6.6–8.7) mm (normal range 3.1–5.8). SVC flows decreased sequentially with each embolisation, with a reduction to less than twice the normal range in all survivors, allowing safe discharge home. Five cases (cases 1–5) with SVC flows >400 ml/kg/min at presentation had a poor outcome, with four deaths. The only survivor in this group (case 1) had extensive cerebral ischaemia on MRI only evident after the final embolisation. There were three antenatal diagnoses, one with fetal hydrops. The median age at presentation was 1 day (IQR 1–3). Associated echocardiographic findings on admission in all five cases were continuous retrograde flow in the aorta throughout diastole and extending into systole (figure 2), suprasystemic pulmonary artery pressures and a patent ductus arteriosus with right to left flow. All neonates in this group required ventilatory support at the time of referral and had multiorgan impairment. Bicêtre scores in this group were 8–12 (urgent embolisation) on admission, but deteriorated rapidly to <8 (not for embolisation), in cases 3, 4 and 5, despite intensive care. Embolisation was performed in cases 1 and 2, with the first embolisation on days 3 and 1, respectively. However, despite large reductions in SVC flow, there was clinical deterioration with left ventricular failure, systemic hypertension and renal impairment in both patients.
The remaining 10 cases (cases 6–15) had SVC flows that were elevated but <400 ml/kg/min at presentation, mostly in the range 300–400 ml/kg/min (cases 6–13). There were five antenatal diagnoses and the median age at presentation was 3.5 days (IQR 1–14). One case had fetal hydrops that resolved spontaneously following delivery. All cases had cardiac failure, with three requiring respiratory support. Bicêtre scores in this group were 8–17 at the time of referral. All cases required embolisation (median age at first embolisation, 6 days). Retrograde flow in the aorta was continuous in diastole in three neonates (cases 7, 10 and 12), all of whom had a patent ductus arteriosus. Retrograde flow limited to early diastole was recorded twice in the two neonates with the lowest SVC flows (cases 14 and 15). Pulmonary pressures were elevated, but were only suprasystemic in two cases (cases 6 and 9). Subsequent clinical deterioration occurred in two cases (cases 11 and 12) between days 3 and 5, with multiorgan impairment requiring emergency intervention in case 11. Two deaths were related to extensive intracranial haemorrhage following embolisation.
Complications following embolisation included acute left ventricular failure (three cases), continuing multiorgan failure (two cases), systemic hypertension (three cases) and intracranial haemorrhage (five cases), with death in three cases. There were no deaths following discharge from hospital during this study period. Developmentally, one child requires learning support in mainstream education, one child has visual field impairment and one has profound neurological damage.
In this study, substantially elevated SVC flows were demonstrated in all 15 neonates with VGAM. All five patients with SVC flows >400 ml/kg had multiorgan failure and an adverse outcome (died or severe neurological injury), even when very early emergency embolisation was technically successful. Ten patients with SVC flows <400 ml/kg presented with cardiac failure but rarely multiorgan failure. The relationship between clinical presentation and SVC flow <400 ml/kg/min was variable, although the case with lowest SVC flows had the latest presentation. Clinical improvement in survivors mirrored decreases in SVC flow. This relationship between clinical course and SVC flow is in contrast to the finding that ventricular outputs, although elevated in VGAM, do not provide useful prognostic information.4
Adverse outcomes associated with extremely high SVC flows were caused by multiorgan failure and intracranial haemorrhage as a consequence of excessive blood flow through a large intracranial shunt and ‘steal’ from the systemic circulation. Retrograde flow in diastole has been described previously as a poor prognostic factor,4 but we found that this was common and non-specific. However, retrograde flow extending into systole was associated with an adverse outcome. This pattern coexisted with pulmonary hypertension and continuous right to left ductal flow, compounding the problem of redistribution of blood flow from the systemic circulation in addition to high cardiac output.6 This scenario may be further aggravated by the use of conventional inotropes increasing peripheral vasoconstriction.7
Extremely low Bicêtre scores (<8) indicated a poor prognosis and these children were not treated. Those with a Bicêtre score of around 8 were more problematic, since the scores often were unstable over the first few days of life. This fluctuation made borderline scores difficult to use as an early prognostic indicator. This potential for rapid deterioration should reinforce the need for urgent referral and frequent reassessment when VGAM is diagnosed.
The mortality in this series is consistent with other published reports of VGAM presenting exclusively in the neonatal period.7 ,8 ,9 Frustratingly, effective reduction in SVC flow to levels tolerated in survivors did not result in clinical improvement in the cases presenting with the highest flow rates. Acute left ventricular failure developed following embolisation in three cases with the highest SVC flows and postembolisation hypertension, tissue oedema and intracranial haemorrhage were common features. This may reflect sudden increases in peripheral vascular resistance and fluid overload after rapid occlusion of the low resistance vein of Galen shunt, presenting a management challenge in patients where there is often pre-existing renal impairment and limited potential for diuresis. Therefore, we recommend meticulous attention to fluid balance, ventricular function, blood pressure and urine output during and following embolisation. Prompt treatment with diuretics, vasodilators and antihypertensive drugs may be required if there has been a rapid reduction in VGAM flow. We would caution against the suggestion that postintervention hypertension is a helpful adaptive circulatory response.10
The findings in this case series suggest that SVC flow measures in VGAM provide an indirect measurement of the size of the shunt and systemic haemodynamics. There is variability in presentation at flows <400 ml/kg/min, which may reflect differing biological tolerance or the need for refinement in SVC flow measurement technique. However, in conjunction with Doppler flow patterns in the aorta, this could be an additional useful prognostic indicator in determining the need and suitability for early intervention in VGAM. Additionally, this technique offers a non-invasive way of measuring shunt reduction, which could be used to guide interventional procedures. Further assessment of the impact of rapid shunt reduction on ventricular function and cerebral blood flow is required, but careful echocardiographic monitoring and intraoperative and postoperative use of vasodilators, inotropes and diuretics may facilitate successful adaptation to changes in the circulation.
The authors wish to acknowledge the support of the Departments of Cardiology and Neonatal Medicine at the Royal Hospital for Sick Children, Glasgow, UK.
Funding There was no NHS or external funding for this study.
Competing interests None.
Provenance and peer review Non-commissioned, externally peer reviewed.