Avoiding use of total circulatory arrest in the practice of congenital heart surgery

Abstract

Deep hypothermic circulatory arrest (DHCA) technique has been an important armamentarium in the correction of congenital heart diseases. There have been many controversies and concerns associated with DHCA, particularly neurological damage. Selective ante grade cerebral perfusion (SACP) was introduced as an adjunct to DHCA with the objective of limiting the neurologic injury during aortic arch repairs. Over the past two decades, various aspects of cardiopulmonary bypass and DHCA have been studied and modified such as optimisation of flows, anti-inflammatory interventions, haematocrit, and temperature to improve neurologic outcomes. With the changes in practice of DHCA, outcomes have significantly improved but SACP intuitively appears attractive to offer better neuroprotection. The strategy of conduct of SACP is evolving and needs to be standardised for comparing outcomes. In this review we have discussed the various physiological and technical factors involved in conduct of SACP in paediatric cardiac surgery and outcomes with SACP.

Introduction

The earliest research work on hypothermia for cardiac surgery is attributed to the team led by Wifred G Bigelow. They hypothesised that a reduced oxygen requirement could be achieved with hypothermia, which in turn enables exclusion of the heart from the circulation for cardiac surgical interventions. It was John F Lewis in 1952 who did the first human cardiac surgery utilising inflow occlusion technique under hypothermia. Christian Barnard and Schrire in 1963 were the first to use cardiopulmonary bypass (CPB) and deep hypothermic circulatory arrest (DHCA) on patients with aortic aneurysm. DHCA offered a bloodless field with no cannulas for surgeons to operate efficiently on complex congenital heart defects. This resulted in significant improvements in the surgical results. Soon the deleterious effects of DHCA on neurodevelopment emerged raising a concern regarding its use [1, 2].

Pathophysiology of neurologic injury

The brain unlike other organs is more susceptible to hypoxic injury because of high resting total metabolic oxygen consumption. It receives almost 20% of the cardiac output although accounting for only 2% of the body weight [3]. Adenosine triphosphate (ATP) generated from oxygen-dependent glucose metabolism is the primary source of energy for neurons. The brain needs a constant supply of these substrates as there is no storage of glucose in the brain and therefore cerebral blood flow is finely autoregulated in health for adequate delivery of energy substrates [4]. These energy dynamics make the brain most susceptible to injury during total circulatory arrest, and hence, effect of DHCA on brain injury is a matter of significant concern.

Anaerobic metabolism during hypoxic states results in insufficient ATP generation and lactic acid accumulation which results in significant decrease in cellular pH [5]. The ATP depletion results in Na+/K+ pump failure leading to osmotic diffusion of Na+, Cl− and water leading to cellular swelling. Accumulation of Na+ results in depolarisation of the neuron and activation of the Ca2+ channels. The high cytoplasmic Ca2+ activates the lipase and proteases causing the neuronal injury [6] High calcium concentration also causes cell death by excessive production of nitric oxide (NO) by activating nitric oxide synthetase [7]. Nitric oxide can be neurodestructive by reacting with superoxide anion to form peroxynitrite which causes lipid membrane peroxidation. Hypoxia and ischemia compromise re-uptake of excitatory neurotransmitters such as glutamate which activates the N-methyl-d-aspartarte (NMDA) channels, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic (AMPA) and kainite receptors causing excitotoxic effects by elevating intracellular concentration of Na+ and Ca2+ [8, 9].

The apoptotic machinery is triggered through intracellular or extracellular stimuli. The intrinsic pathway is activated by excessive intracellular calcium, NO and reactive oxygen species (ROS) accumulation resulting in mitochondrial outer membrane permeabilisation. The hypoxic ischaemic injury on other cellular organelle further increases the pro-apoptotic B cell lymphoma 2 (BCL 2) protein family balance, c-Jun N-terminal kinase (JNK) and caspase-2 activity at the level of the mitochondrial outer membrane. These molecular changes cause release of pro-apoptic proteins such as Cytochrome C (Cyt C) into the cytosol which activates executioner Caspase 3 leading to deoxyribonucleic acid (DNA) cleavage. Hypoxic ischaemia promotes apoptosis through the extrinsic pathway through increase of circulating death receptor ligands such as tumour necrotic factor (TNF) alpha Fas and TNF-related apoptosis-inducing ligand (TRAIL) [10].

These hypoxic ischaemic injury and apoptotic mechanisms have been reported in various animal species exposed to DHCA. In an animal study on 32 dogs who were subjected to DHCA for 2 h at 18 °C by Tseng et al., dogs treated with 17477AR (Astra Arcus, USA), a neuronal nitric oxide synthase (NOS) inhibitor, had superior neurologic function and significantly reduced neuronal necrosis by histopathology [11]. In another animal model of DHCA, it was demonstrated that neuronal death involved activation for apoptotic cascade in spite of adequate ATP concentrations. Cyt C release was seen as early as 1 h after reperfusion, Fas expression was seen at 4 h and Caspase 8 was activated at 8 h in this study. Caspase 3 activity coincided with cell death at 8–72 h [12].

Hypothermia bestows neuroprotection by inhibiting or suppressing various pathways of neural injury. The metabolic rate at 18 °C is only 12–25% of the metabolic rate at normothermia [13]. Hypothermia decreases the oxygen and glucose demand in the neurons and therefore reduces ATP breakdown making the ATP energetics favourable for neuronal survival [5]. Hypothermia decreases the release of the excitatory neurotransmitters which provides protection against excitotoxicity induced injury. Hypothermia has also shown to inhibit the pro-apoptotic activity and reduce the production of free radicals and inflammatory cytokines [14].

Is DHCA unsafe?

The Boston circulatory arrest trial is the largest prospective trial to assess the safety and effects of DHCA. This study enrolled 171 patients with transposition of great arteries with or without ventricular septal defect (VSD) between 1988 and 1992 at a single centre. They were randomised to DHCA or low-flow CPB group and followed up for neurologic impairment at various time intervals. In the immediate postoperative period, there was higher incidence of seizure (odds ratio 11.4) and significantly higher release of brain isoenzyme creatine kinase indicating clinically significant neuronal injury in the DHCA arm of the study. On continuous electroencephalographic monitoring during the first 48 h in the postoperative period, there was higher incidence of post ictal activity and longer recovery time to first reappearance of electroencephalography (EEG) activity [1].

The same cohort of 155 patients underwent developmental and neurologic evaluation by two independent observers using the Bayley Scales of Infant development at 1 year of age. A score of less than 80 was considered 2 standard deviations (SD) below the mean score. Patients also underwent magnetic resonance imaging (MRI) of brain which was also assessed by two independent neuroradiologists for structural and intensity abnormalities. The psychomotor development index (PDI) was significantly lower among children assigned to circulatory arrest, a mean difference of 6.5 points. Similarly the mental development index was lower by 4.1 points in the circulatory arrest group compared with low-flow CPB group. Children with longer DHCA time had lower scores on neurological assessment. Perioperative seizure activity was associated with lower PDI scores and increased likelihood of abnormalities on the MRI scans of the brain. The occurrence of seizure was associated with poorer outcomes attributable to severe hypoxic ischaemic injury [2].

At preschool age, development and neurologic evaluation of the whole cohort showed a below expectation intelligence quotient (IQ), expressive language skill, motor function, oromotor control and visual motor control. Further, the DHCA group had lower motor function, both fine and gross, and had more oromotor apraxia compared with the low-flow CPB group. Perioperative seizure was associated with lower IQ and increased risk of neurologic abnormalities [15]. At 8-year follow-up, the cohort had below expectation academic achievement, fine motor function, working memory, visual spatial skills and higher order language. The DHCA arm had lower motor function, visual motor tracking and phonologic awareness compared with the low-flow CPB group. The children assigned to the low-flow CPB had more impulsive response and worse behaviour as rated by their teachers [16]. Even after 16 years, the DHCA group had worse visual spatial function and social cognition but academic achievement, memory, executive function, and attention were comparable between the DHCA and low-flow CPB groups [17]. In an ongoing study this cohort will be assessed again between 24 and 28 years to compare the impact of DHCA versus low-flow CPB.

Though the Boston circulatory arrest study was a well-planned randomised prospective study, there are important drawbacks in the conduct of CPB because of the era (1988–1992) effect. The typical strategy used was alpha stat cooling, haemodilution to haematocrit of 20% and no arterial filter which could have contributed to poorer outcome in the DHCA group. It should also be noted that, in both the treatment arms of the Boston circulatory arrest study, significant number of children needed remedial academic support compared with normal children and hence neurologic injury because of cardiopulmonary bypass, socioeconomic status, preoperative hypoxia, presence of VSD and postoperative haemodynamics cannot be completely discounted. The current understanding of mechanism of neurological injury has modified the DHCA strategies which include use of preoperative steroids to decrease inflammation [18, 19], ensuring adequate time for reaching the target temperature, target temperature often less than 18 °C [20] and using pH stat strategy to achieve cerebrovasodilatation to achieve adequate cooling of the brain [21–23]. Use of hyper-oxygenation prior to initiating DHCA [24] and modified ultrafiltration have also shown to improve cerebral metabolic recovery [25]. Low-flow CPB intuitively appears better for neuronal integrity but extends the inflammatory process on the brain during CPB which might be counterproductive.

Some non-modifiable factors have also been identified for poor neurodevelopment outcomes after paediatric cardiac surgery such as Apolipoprotein E e2 allele carriers. Carriers of this allele have lower PDI of approximately 7 points at 1 year after biventricular repair. This decrease in PDI was independent of ethnicity, socioeconomic status, cardiac defects and use of DHCA [26].

In a much later study from the Boston Childrens hospital, Menache et al. reported an incidence of neurologic complication as 1.6% after DHCA in the immediate postoperative period compared with 11.4% reported in the earlier Boston Circulatory arrest study cohort. The common causes of neurologic injury were Cyclosporin A neurotoxicity and perioperative cardiac arrest both unrelated to DHCA [27]. This clearly demonstrates the decrease in the neurologic damage caused by current changes in management strategies of CPB and DHCA.

Safe limit of DHCA

DHCA offers advantages of decreased exposure to cardiopulmonary bypass, improved surgical exposure, avoidance of multiple cannulas and reduced oedema. At the same time DHCA imposes the time pressure on the surgical team for favourable neurological outcomes. The need to set a safe limit appears to be justifiable to compare outcomes, but neuronal injury is a continuous process; therefore, it is unclear if any truly safe period of DHCA exists. The safe periods have been calculated retrospectively by looking at measurable end points of neurological damage, but the occurrence of these end points indicates severe brain injury. Therefore any assessment using these end points will overlook minor or moderate brain injury.

Various animal studies have been performed to predict the safe limit of DHCA. The safe limit of DHCA using a rabbit spine model using pH stat strategy at 15 °C was predicted to be 150 min by Miyamoto et al. [28]. Sakamoto et al. in a piglet model using various combinations of haematocrits and temperatures measured the time to reach the nadir tissue saturations after a cooling time of 40 min. In the group with haematocrit 20% at 15 °C and haematocrit 20% at 25 °C, this was 28.7 ± 2.4 min and 23.1 ± 1.7 min, whereas in haematocrit 30% at 15 °C and haematocrit 30% at 25 °C, this time was 51.3 ± 4.1 min and 41.4 ± 2.9 min [29].

Langley et al. studied the impairment of cerebral metabolic requirement of oxygen (CMRO2) as a marker of neurological injury during DHCA and benefit of intermittent perfusion in neonatal piglets. The strategy of intermittent reperfusion during DHCA resulted in normal cerebral metabolic recovery which was comparable to piglets instituted only CPB without DHCA. On electron microscopy, the uninterrupted DHCA group had collapsed vessels with swollen astrocyte foot process, thickened basement membrane, lobulated nucleus and swollen mitochondria and lysosome. In contrast the intermittent perfusion DHCA had normal looking endothelium and surrounding neuropil [30].

Q10 is the ratio of metabolic rates at temperatures 10 °C apart. This index indicates the protective effect of hypothermia. Greeley et al. measured Q10 in paediatric patients undergoing hypothermic CPB and found it to be 3.65 which is much higher than the adult value of 2.3–2.6. Therefore, hypothermia offers greater suppression of cerebral metabolism in paediatric population than adult patients. Based on this index they predicted safe duration of DHCA at 18 °C as 39–65 min [31, 32]. The neurodevelopment outcomes of the Boston circulatory arrest trial cohort at 8 years showed a non-linear effect of duration of total circulatory arrest. There was little influence on neurologic outcomes with DHCA times less than 41 min but worsening outcomes with longer duration [16].

Considering the uncertainty of safe time period of DHCA and the proven benefit of even intermittent perfusion during DHCA, continuous flow during all stages of CPB to both hemispheres of the brain during CPB intuitively seems a safe option to prevent adverse neurological outcome.

Techniques for avoiding DHCA

Presently the choice of technique of perfusion for surgeries not involving arch repair involves arterial cannulation and bicaval cannulation with venting of left heart and temperature appropriate CPB flows. This approach is presently a challenge only in preterms requiring intracardiac surgery not involving the aortic arch. With surgeries involving the arch of aorta, significant modification to arterial cannulation and perfusion techniques are necessitated to establish continuous cerebral perfusion. The first report of arch repair with ante grade cerebral perfusion without use of DHCA was by Sakamoto et al. [32]. Asou and associates perfused the brain through a polytetrafluoroethylene (PTFE) graft sutured to the innominate artery [33]. Various modifications of this approach with the principle of ante grade cerebral perfusion have been described for aortic arch surgeries to avoid DHCA.

• Pigula et al. achieved cerebral perfusion by cannulating the pulmonary artery and perfusing and cooling the brain through the patent ductus arteriosus. During the arch repair, the brain was perfused through the proximally sutured Blalock Tausig (BT) shunt. The distal end of BT shunt was sutured with arterial cannula placed in the neo-aorta [34].

• Tchervenkov technique involves a completed modified Blalock Taussig shunt as a perfusion conduit with arterial cannula in the pulmonary artery. Innominate artery is perfused by blood flowing retrogradely through the BT shunt from right pulmonary artery during the arch repair. The disadvantage of this method is that the cannula is in the surgical field [35].

• Imoto et al. established perfusion for arch repair by cannulating both the PTFE graft anastomosed to innominate artery and the descending aorta. Initial CPB was established through the proximally sutured PTFE graft. On CPB the additional descending aortic cannulation was done. Perfusion through the descending aortic cannula maintains cerebral flow during completion of the BT shunt [36].

• Direct innominate artery cannulation and selective clamping of the proximal innominate artery [37]. This has been the technique for establishing SACP (selective ante grade cerebra perfusion) in our centre [38].We have used this technique in 52 patients with weight range of 2.0 to 8.1 kg requiring single-stage repair of aortic arch and associated cardiac defects.

Continuous cerebral perfusion

As SACP involves perfusing only through the right inominate artery, the functional integrity of the circle of Willis is essential for adequate oxygen delivery to the left cerebral hemisphere. The other minor collateral system which contributes to blood supply to the left hemisphere includes the leptomeningeal collaterals, ophthalmic artery and perforating arteries of cortical arteries. The circle of Willis is complete in only 24–27% in adult population. Though cerebral circulation is assumed to be intact in neonates, 10% of healthy full-term infants do demonstrate abnormal flow pattern by transcranial Doppler (TCD) [39]. van Kooij et al. showed variations in circle of Willis in 74% term and preterm infants by MRI, but the integrity of the circle was maintained in all the infants [40].

Significant knowledge and experience has been gained regarding the physiology and conduct of continuous cerebral perfusion over the last 2 decades. In spite of this, still certain controversies related to the use of selective ante grade cerebral perfusion persist such as the optimal perfusate temperature, the minimum flow rate required, pH strategy, the perfusion pressure range and the role of neuro-monitoring.

Temperature and cerebral blood flow

The optimal flow on SACP is an unresolved issue with various practices reported in literature. The flow on SACP cannot be considered in isolation but would have to be integrated with other variables such as the temperature, perfusion pressure and extra cranial shunts. There has also been a change in the SACP flow rate strategy over time influenced by the monitoring used. Optimising flow on SACP is essential to avoid neuronal damage due to decreased flow causing hypoxic ischaemic damage or excessive flow resulting in cerebral oedema.

Animal studies

Most of our practices of SACP are an extrapolation of data from animal studies on whole-body hypothermia. In SACP though the blood flow in the cerebral arteries can be controlled as compared with whole-body perfusion, it still does not ensure delivery of this blood into the cerebral capillaries. The flow of blood into the cerebral capillaries further depends on the cerebrovascular resistance and the extra cranial shunts. The effect of hypothermia on vascular resistance seems to vary in different vascular beds resulting in variable blood flow. In a canine study of differential blood flow, the cerebral blood flow decreased to 20% of the flow at 37 °C after cooling to 20 °C. This decrease in cerebral flow was attributed to increase in cerebrovascular resistance by 339% ± 48% on cooling. It was also noted that the cerebral blood flow remained constant from 90 mmHg to 40 mmHg but abruptly decreased below 30 mmHg suggesting definite autoregulatory response even in hypothermia [41]. Another factor affecting the cerebral flow which needs to be compensated for during SACP is the extra cranial shunt. It has been demonstrated in an animal study that the extra cranial shunt is higher at 15 °C than at 22 °C. Therefore, the flow required to achieve a perfusion pressure of 35 mmHg was higher at 15 °C than at 22 °C. These authors suggest pressure-based management of perfusion rather than flow-based management during SACP [42].

The interplay of hypothermia, pump flow rate, cerebral blood flow (CBF), haematocrit and oxygen extraction was investigated by Kern et al. in neonates and children undergoing cardiac surgery under moderate and deep hypothermic CPB. At 18–20 °C they reduced the pump flow by 45–70% of the baseline pump flow. The corresponding CBF decreased from 16.0 ± 6.6. to as low as 7.3 ± 3.4 ml/100 g/min, but the oxygen extraction did not increase significantly .This protective effect was not seen at 26–29 °C thus emphasising the importance of deep hypothermia–induced decrease in CMRO2 during low-flow CPB [43].

De Campli et al. studied neonatal piglets with an SACP model at 20 vs. 40 ml/kg/min, and concluded that the higher SACP flow rates maintained cortical oxygen tension at pre-CPB levels and this was not achieved with lower flow rates [44]. Sasaki et al. concluded that at deep hypothermia in neonatal piglet model, SACP flow rate of 50 ml/kg/min matched the CBF at baseline pump flows of 100 ml/kg/min. At 30 ml/kg/min of SACP, the CBF was 60% of the baseline flow but the regional cerebral oxygen saturation was equal to the baseline. They recommended a SACP flow of 30–50 ml/kg/min based on these observations [45].

Human studies

In an early series by Tweddell et al. in management of stage 1 Norwood, flows of 30–40 ml/kg/min were used. No specific monitoring was used to guide the SACP flows [46]. Golberg et al. in their randomised trial comparing outcomes of DHCA and SACP for Norwood stage 1 used 20 ml/kg/min of flows during SACP. Though near infrared spectroscopy (NIRS) was used it was not used to optimise the SACP flows [47]. Compared with these studies, most of the other studies have used higher flows during SACP. Tchervenkov and colleagues used mean SACP flows of 44 ml/kg/min (18–76 ml/kg/min). The measured right radial arterial pressures were 24 ± 10 mmHg and the left radial pressures were 15 ± 8 mmHg. Though these were monitored, they were not used to titrate the flows [48]. Using TCD jugular venous saturation and infrared spectroscopy in neonates undergoing Norwood stage 1 procedure, the SACP flows were varied from 30 ml/kg/min to 20 ml/kg/min to 10 ml/kg/min by Hofer and associates. The decrease in flows was associated with corresponding decrease in cerebral oxygen saturation from 78 ± .8 to 72 ± .9 and 66 ± 8; the jugular venous saturations decreased from 94 ± 6 to 89 ± 13 and 83 ± 15. On TCD monitoring there was loss of flow in 1 patient with SACP flow of 30 ml/kg/min, increasing to loss 3 patients with 20 ml/kg/min and the loss was seen in 4 patients with SACP of 10 ml/kg/min [48]. Hofer et al. conclude that SACP flows should be > 30 ml/kg/min [49]. The SACP strategy of Guo and associates involved systemic cooling to rectal temperature of 18 °C under pH stat and then to perfuse the brain with perfusate starting at 10 ml/kg/min increasing to 30–40 ml/kg/min with pressure of 30–40 mmHg [50]. Visconti and associates used SACP flows of 30–40 ml/kg/min guided by the distal left radial pressure which was kept at 20–25 mmHg [51]. Fraser et al. advocate NIRS and TCD monitoring guided strategy with a goal to keep regional cerebral saturation (rSO2) of 90–95% on both sides during SACP. The flow used for SACP was 57 ± 11 ml/kg/min with SACP time of 76 ± 24 min [52].

Rajasinghe and associates performed extended end to side anastomosis of descending aorta to proximal aortic arch for coarctation of aorta through left thoracotomy under mild hypothermia, with cerebral perfusion maintained only through unobstructed innominate arterial flow in 19 neonates [53]. This concept has been extended to perform SACP at moderate hypothermia for arch repair. Gupta et al. used SACP at 25 ± .0.9 °C in 61 neonates for arch reconstruction with univentricular and biventricular repair. SACP was performed at mean flow rate of 46 ± 6 ml/min/kg and with a median time of 48 min (range, 10–123 min) [54]. In a similar group of patients requiring aortic arch repair, SACP was initiated at 25 °C with SACP flows of 82 ± 20 ml/kg/min. The flows were adjusted to keep the radial arterial pressure below 50 mmHg. The rSO2 was78.8% ± 9.5% during high flow regional cerebral circulation with incidence of isolated seizure of 4.9% with no neurologic deficit or stroke in this cohort of 18 patients [55]. SACP for aortic arch repairs have been instituted even with mild hypothermia (30 °C) with SACP flows of 104 ml/kg/min. On comparison of mild hypothermia with moderate hypothermia strategy (SACP flow 85 ml/kg/min, temperature 26 °C), during SACP no statistical difference in end organ injury, ventilatory days or mortality was seen [56].

The postulated advantage of using higher SACP flows is higher extra cranial shunt which could prevent other organ damage. The conduct of SACP at high temperature requires higher flows. This imposes an additional risk of cerebral oedema if adequate vasodilators are not used to decrease the regional perfusion pressure. The strategy of moderate or mild hypothermia with higher flows needs to be evaluated with larger randomised trials with assessment of long-term neurological and other organ injury.

Cerebral monitoring during CPB

NIRS is a simple non-invasive and continuous method to assess the neurological status. The disadvantages of NIRS are inability to detect ischaemia in deep brain structures and wide inter patient variability. The other techniques which have been used for cerebral monitoring are EEG, bispectral index and TCD velocity. EEG monitoring is technically difficult to use and interpret and requires numerous electrodes. BIS is user-friendly processed EEG but not well tested in children younger than 2 years. TCD is good monitor to assess the embolic load, and it measures the velocity of flow in the cerebral arteries which is influenced by the vascular resistance. Each technique has its advantages and disadvantages, and hence, a multimodal monitoring should be employed to detect ongoing ischaemia.

The NIRS decreases at initiation of CPB because of lower flows and fall in haematocrit. During cooling the cerebral oxygenation increases secondary to decreased cerebral metabolic rate. With the cessation of pump flows for DHCA, the cerebral oxygen saturation reaches a nadir. This slow decline to nadir tissue saturation suggests cerebral tissue oxygen consumption even during DHCA [57, 58]. In contrast to DHCA the cerebral oxygen saturation increases during SACP and may reach above the baseline [59]. Post CPB it takes up to 24 h for the tissue oxygen saturation to reach normal cerebral oxygen saturations. The delay is attributed to restoration of the oxygen debt incurred preoperatively and on the CPB and because of increased cerebral vasoconstriction [60].

Bilateral cerebral oxygen saturation monitoring detects any differential flow during SACP because of abnormality in the circle of Willis which can exist in 5% of population. Usually the cerebral oxygen saturations are comparable between the two cerebral hemispheres during the pre CPB and at initiation of CPB. During SACP the correlation between the saturations of the two hemispheres of the brain diverges, and in more than 50% of patients, the left side saturations are 10% lower than the right hemispheric saturations even when simultaneously measured transcranial Doppler velocity in the cerebral vessels remains comparable. This difference in saturation between the two cerebral hemispheres is hypothesised to be because of cerebral vasoconstriction [61]. The mean total oxygen index measured during SACP or DHCA by NIRS has been shown to correlate with neurological injury detected by MRI scanning by Mc Quillen et al. In this study among the 16 patients with new onset MRI changes, 8 of them had undergone surgery under SACP [62].

With use of NIRS and TCD, 70–75% of cerebral blood flow distribution is monitored. A decrease in rSO2 index by more than 20% from baseline hypothermic full flows or an absolute value of < 30% during SACP requires intervention to increase the cerebral oxygen delivery or decrease the CMRO2 or adjustment of the cannula. A cerebral oxygen saturation of 95% is also of concern during SACP; this may indicate a high cerebral perfusion which can result in postoperative cerebral oedema. To avoid neurologic injury because of high flows during DHCA, flows are adjusted to achieve TCD velocity within 10% range of baseline hypothermic full flow TCD velocity. During DHCA absolute cerebral saturation < 30% indicates need for reperfusion [59]. When only TCD velocity is measured, a decrease of > 25% TCD velocity from baseline hypothermic full flow needs intervention [54].

Is SACP better THAN DHCA?

Various animal and human experiments have been done to compare the outcomes of surgeries with SACP and DHCA. The variable SACP and the DHCA strategies in these studies make comparison among these studies difficult. With improved understanding of each strategy, the outcomes have improved in both the groups.

Animal experiments

Comparing the effect of regional low-flow perfusion (RLFP) and DHCA in neonatal piglet, Myung et al. reported lower abnormal postoperative neurobehavioural scores (25% versus 88%) in RLFP group compared with DHCA group. The neurologic injury on postoperative day 1 was significantly less in RLFP compared with DHCA. The degree of apoptosis assessed by terminal deoxynucleotidyl tranferase (TdT)–mediated deoxyuridine triphospahte nick end labeling (TUNEL) was also lower in the RLFP group. All the piglets in this study were cooled to 18 °C for 20 min on CPB with alpha stat strategy and were ultrafiltered and perfused at 10 ml/kg/min through innominate artery [63]. Similarly fewer staining of cells was also reported by Chock et al. for activated caspase, which is an enzyme in the apoptotic pathway in neonatal piglets undergoing SACP compared with DHCA [64].

Human studies

In a large retrospective analysis of postoperative outcome of neonates and infants who underwent biventricular repair with or without arch repair under DHCA or SACP, the neurological outcome was poorer with DHCA than with SACP (6.23% vs 2.64%). There was no difference in the intensive care unit (ICU) length of stay, mortality, low cardiac output, pulmonary or renal complications between the two groups. The SACP strategy involved systemic cooling to rectal temperature of 18 °C under pH stat and then to perfuse the brain with perfusate starting at 10 ml/kg/min increasing to 30–40 ml/kg/min with pressure of 30–40 mmHg [49]. Similarly better neurological outcomes were also reported by Kornilov and colleagues in infants with SACP compared with DHCA (OR 0.19). They established SACP through innominate artery graft and cooled patients to 18–27 °C; the SACP was established with a lower flow rate of 30 ml/kg/min [65].

However the superiority of SACP has not been consistently reported from various other trials. Goldberg and colleagues randomised 77 patients undergoing Norwood stage 1 to DHCA or SACP groups [47]. Psychomotor and mental developmental score using the Bayley Scales of Infant Developmental scale was assessed prior to Norwood stage 2 and at 1 year of age. They found no statistical difference in mental or psychomotor developmental scores between SACP or DHCA groups at both the assessment periods. The SACP flows used were only 20 ml/kg/min unlike many other studies with higher flows. In a retrospective non-randomised study of 29 patients, Visconti et al. similarly showed comparable psychomotor and mental developmental index measured by Bayley Scales of Infant Development scale [51] at 1 year among patients undergoing stage 1 Norwood using SACP or DHCA. Selder et al. reported no immediate neurological complications in infants undergoing biventricular repair with arch repair neither with SACP nor with DHCA but found significantly prolonged length of stay in the DHCA group [66]. A more comprehensive randomised study was done at the same centre to compare SACP and DHCA in neonates and infants undergoing univentricular and biventricular repair for congenital heart diseases. MRI scans were done pre- and postoperatively, NIRS and amplitude-integrated EEG was monitored for 48 h and biomarker of neurologic injury S100 B was measured at 4 h and at 24 h. These children were evaluated at 2 years for neurodevelopment outcomes. New white matter injury (WMI) was seen in 72%and 78%respectively in SACP and DHCA groups with relative risk of 0.86. The severity of WMI was also comparable between the groups [67].

However centrally located ischaemic infarcts involving the thalamus, basal ganglia and internal capsule were seen only in patients in the SACP arm of the study. The infarcts are probably due to microparticles or air emboli travelling through there middle cerebral artery during the SACP. Though the mean cerebral saturations intraoperatively were lower in the DHCA group 52% compared with 76% in the SACP group, the seizure activity was comparable between the two groups on EEG. Neurobehavioural score was also not statistically significant at 2 years between the two groups. Additionally no differences in motor function markers for fine and gross motor and walking age were seen between the two groups [67].

Although the optimal flow temperature and pH strategy have not be identified during SACP, data from the Society of Thoracic Surgeons Congenital Heart Surgery database points towards preference of SACP (43%) by surgeons for neonatal arch repair. DHCA was still used by 32% surgeons and a mixed technique by 16% of surgeons [68].

Conclusions

With the present technical knowledge, it is possible to maintain continuous cerebral blood flow during all the stages of paediatric cardiac surgery including aortic arch repairs thus avoiding DHCA. Lower DHCA times have shown to have similar outcomes to SACP, but DHCA imposes time pressure which can be relaxed to some extent with the use of SACP. Optimal flow rate during SACP has individual variability, and hence, a multimodal neuromonitoring–guided flow strategy would be ideal. In case these monitoring devices are not available, a flow rate > 40 ml/kg/min should be used. The issues of acid base strategy and use of moderate and mild hypothermia during SACP need further evaluation. We still lack adequate monitoring capability to prevent and predict neurological outcomes in real time. It is not only the conduct of DHCA or SACP which determines the neurological outcome but other factors such preoperative status, genetic polymorphism, genetic syndromes, inflammatory response, the conduct of CPB and postoperative haemodynamics have overbearing influence on the neurological outcome. An integrated approach addressing all these issues will have to be evolved during conduct paediatric cardiac surgery for better outcomes. Though there is a concern of cerebral infarctions in children undergoing SACP, presently there is no evidence to suggest inferiority of SACP compared with DHCA.

Funding

There is no funding from any source involved in this review article.

Compliance with ethical standards

Conflict of interests

The authors declare that they have no conflicts of interests.

Research involving human participation or animals

In this review article, the authors were not involved with any research related to human participants or animals.

Ethical approval

Not applicable as review article.

Informed consent

Since this is a review article, the manuscript contained purely material related to review of literature. There is no material in the manuscript related to our own institutional experience related to the subject. We only mentioned the reference of the previous publication from our unit. For this reason, the informed consent is not applicable.