While post-operative complications of the Norwood procedure may be linked to pre-operative patient characteristics (e.g., presence of non-cardiac/genetic abnormalities, weight <2.5 kg, right dominant single ventricle) (61), the specific hemodynamics and flow resistance of the systemic-pulmonary arterial shunt has also been suggested as an important prognostic factor for patient survival. As demonstrated in an early in vitro study, which evaluated pressure-flow relationships in systemic-to-pulmonary Blalock-Taussig shunts of different diameters, an increase in shunt diameter resulted in reduced pressure gradients at both the proximal and distal anastomoses (62). More recently, a computational parametric investigation conducted in an idealized Blalock-Taussig shunt geometry revealed that not only shunt diameter but also anastomosis placement was critical to oxygen delivery to both systemic and coronary circulations (63). Lastly, multiscale computational flow modeling has also been used to compare the hemodynamic performance of different shunt configurations. The simulations suggested that the Sano shunt consistently generated more favorable hemodynamics (i.e., lower RV systolic and diastolic pressures, lower pulmonary-to-systemic flow ratios, and higher coronary perfusion pressure).

The surgical management of HLHS consists of redirecting the venous deoxygenated blood toward the lungs while bypassing the RV via a staged surgical approach. A challenge raised by the construction of this bypass is the requirement to achieve a balanced flow split between both lungs while minimizing flow energy loss. The dominant objective in single ventricle management is to achieve the Fontan circulation, which depends upon kinetic energy to propel blood through the lungs without a sub-pulmonary ventricle. The concept of energy is critical to the success of this staged palliative surgery. In fact, the single RV must eject blood with sufficient kinetic energy to overcome energy losses caused by the viscous friction along the entire vasculature and the hemodynamic disturbances (e.g., mixing, collision, separation, and recirculation) generated at each stage of the surgical cardiac and vascular reconstructions (64). Therefore, optimization of the staged surgery is critical to produce efficient hemodynamics necessary for positive long-term clinical outcomes (65).

The hemodynamics of the Glenn and Fontan surgical reconstructions have been documented in many in vitro and CFD studies. In vitro measurements in realistic glass models mimicking a bidirectional Glenn cavopulmonary connection suggested the benefits of this connection over a dilated atriopulmonary connection by demonstrating its ability to reduce fluid energy dissipation, achieve physiologic distribution of total flow, and maintain some hepatic venous flow to each lung (66). Pulsatile flow simulations in hemi-Fontan and bidirectional Glenn geometries reconstructed from magnetic resonance, angiocardiogram, and echocardiogram anatomic data complemented the previous experimental findings by indicating no substantial difference in power loss and flow distribution to each lung (67).

However, the bulk of the literature published to date on Glenn and Fontan hemodynamics has focused on the identification of key geometrical parameters affecting the hydraulics of the Fontan total cavopulmonary connection (TCPC). Experimental particle image velocimetry measurements in realistic stereolithography intra-atrial connection models have revealed the existence of complex, unsteady, and highly three-dimensional flow structures, suggesting a substantial degree of energy loss in this type of connection (68). Those results are supported by 4D-flow MRI and patient-specific CFD studies, which have evidenced the existence of high vorticity magnitudes and atrial recirculation in the intra-atrial connection (69), and the hydraulic superiority of the lateral tunnel Fontan operation relative to any other method (67).

Another important finding suggested by the literature on TCPC hemodynamics is the stronger dependence of energy loss on the TCPC topology and geometrical features than on the Fontan surgical option. In fact, a retrospective analysis of CFD and in vitro flow data revealed that the minimum cross-sectional area of the pulmonary arteries at the TCPC outlets was a stronger predictor of energy loss characteristics than the surgical procedure (extra- vs. intra-cardiac conduit) (70). Another CFD analysis using realistic pulsatile flow boundary conditions and TCPC geometries featuring extra- and intra-cardiac conduits demonstrated that power dissipation was primarily influenced by the actual cross-sectional area of the inferior vena cava anastomosis (71). Lastly, steady flow pressure measurements and flow visualization conducted in idealized TCPC glass models suggested that caval offsets and anastomotic flaring could reduce the hydraulic power loss by half relative to no offset and by at least 45% relative to no flaring, respectively (72, 73). Altogether, those studies suggest that patient-specific flow modeling could be an effective surgical planning tool toward the improvement of HLHS patient outcome and the reduction of surgical risks (74).

While the re-engineering of the right side of the heart is a critical requirement to support both pulmonary and systematic circulations in HLHS patients (7), RV function is a key determinant of HLHS patient survival (60). Mathematical and computational approaches with various degrees of sophistication have been proposed to estimate RV hemodynamics and function in HLHS. Flow simulations in a post-Norwood RV geometry confirmed the hemodynamic superiority of the Sano shunt over the Blalock-Taussig shunt, as suggested by the predicted reduction in RV workload and no substantial difference in systemic blood flow (75). Another computational model based on patient-specific RV anatomies reconstructed from magnetic resonance imaging revealed the strong dependence of interventricular pressure gradients, filling dynamics and capacity on RV shape and temporal deformation patterns (76). To date, the only quantification of the native pre-surgical HLHS RV hemodynamics was performed numerically in a fetal heart in the context of cardiac development (77). This investigation, which assessed fetal blood flow using 4D spatiotemporal image correlation ultrasound and numerical modeling, revealed that despite a larger right-ventricular cavity size and a greater cardiac output, HLHS fetal hearts generated essentially the same global interventricular hemodynamics as a normal heart.

The synergies suggested by previous RV studies between interventricular RV flow dynamics, cardiac function, and TCPC power loss suggest that the knowledge of pre- and post-surgical RV hemodynamics on a patient-specific basis could guide clinical decision making and promote patient survival (78). However, determining the exact role played by RV hemodynamics in the long-term outcome of the right heart surgical reconstruction requires the assessment of the hemodynamic alterations experienced by the RV during the course of the staged Fontan palliation. While this knowledge gap has not been addressed yet, computational strategies have been designed and successfully implemented to assess the potential mechanical changes of other cardiac defects such as discrete subaortic stenosis (DSS) (79–81). Cine-magnetic resonance imaging data was used in tandem with state-of-the-art fluid-structure interaction modeling to predict native blood flow patterns in patient-specific left-ventricular models featuring normal and DSS-prone outflow tract anatomies, and to characterize the resulting myocardial mechanical stresses. A similar modeling strategy could be deployed to elucidate RV hemodynamics in patient-specific post-Norwood RV anatomies, and to identify particular blood flow patterns and mechanical stresses associated with good clinical outcomes.

Stem cell therapy for myocardial infarction in adult hearts is now a large and expansive field, with numerous lessons being learnt (82). However, our goal is to focus this discussion on the potential mode of action of injected cells. How are the injected cells acting on the host tissue to provide a functional benefit? Initially, it was hypothesized that injected cells would integrate with host myocardium, transform to become contracting cardiomyocytes, electromechanically couple with host cardiomyocytes, and provide direct functional improvement to an otherwise failing heart (82). Conceptually, this was the basis of the field and if realized, would truly be revolutionary in the field of adult cardiac regeneration. However, this was not the case for many reasons, perhaps the most significant of which was the low rate of local cell retention: <5% of injected cells were retained at the site of injury (90). Irrespective of the challenges in the field, many lessons have been learnt regarding the potential mode of action of injected stem cells. It is now hypothesized that injected stem cells act through one of several mechanisms, including the release of paracrine signaling factors, promoting neovascularization and/or recruitment of resident or circulating stem cells, all of which serve to either increase the number of functional cardiomyocytes or promote neovascularization (90). With such a rich literature in the adult cardiac stem cell transplantation space, it provides a strong background to initiate similar studies in the pediatric congenital cardiac space, particularly related to HLHS.

Many of the factors discussed earlier apply to stem cell therapy for HLHS patients, to include the source and number of cells and mode of delivery. However, specific to HLHS patients is the timing of stem cell delivery. The surgical palliation of HLHS is a complex three-stage process, progressing through the Norwood, Glenn, and Fontan surgeries. Therefore, the timing of stem cell delivery becomes crucial and most studies making use of stem cell therapy have been at the time of stage two palliation.

Table 2 provides an overview of several studies describing stem cell therapy in HLHS patients, and these are discussed in subsequent sections. Table 2 is designed to serve as a survey of the recent literature on stem cell therapy in HLHS, rather than an exhaustive list of all published studies. Earlier studies in 2010 (89) and 2015 (87) were designed to provide safety of injected cells in single HLHS patients at the time of the Glenn surgery, followed by the Phase I TICAP trial in 2015 (87) and Phase II PERSEUS trial in 2017 (86), all of which are discussed below.

One of the earlier case studies was described in 2010, in which intracoronary injection of bone marrow mesenchymal stem cells (BMMSCs) during stage II procedure in a single pediatric HLHS patient proved to be safe (89). This was an initial proof of concept study to demonstrate the feasibility of BMMSCs at the time of Glenn, without much detail to any potential functional benefit and/or mode of action. While the results of this study demonstrated safety of BMMSCs delivery in a single HLHS patient at the time of Glenn, many unanswered questions remained. A second case study was published in 2015, using mononuclear cells from umbilical cord blood, again injected during stage II surgery, increasing RV ejection fraction from 30–35 to 50% (84).

Phase I results of the TICAP trial were published in 2015, making use of intracoronary infusion of cardiac progenitor cells (CPCs) in 14 patients, seven undergoing staged palliation with CPCs infusion and an equal amount with only staged palliation (87). Patients treated with CPCs showed RV ejection fraction improvement from baseline to 3-month follow-up (46.9 ± 4.6% to 52.1 ± 2.4%; P = 0.008) (87). Compared with controls at 18 months, cardiac MRI analysis of CPC-treated patients showed a higher right ventricular ejection fraction (31.5 ± 6.8% vs. 40.4 ± 7.6%; P = 0.049) (87). This was followed up by a Phase II clinical trial, PERSEUS with 41 patients; at 3 months, the absolute changes in ventricular function were significantly greater in the CPC-treated group than in the controls (35.3%+/-9.2% to 41.7%+/-7.4%; P = 0.0002) (86).

Another recent study made use of autologous bone marrow derived mononuclear cells administered via cardiac catheterization to the coronary circulation in a single patient 23 years after the Fontan Surgery, showing decrease in ventricular size 3 months after cell injection (88).

A more recent study published in 2019 showed the feasibility of intramyocardial injections of mononuclear cells derived from umbilical cord bloods in a Phase I clinical trial with 10 patients, though there were no functional benefits resulting from this intervention (85).

Stem cell therapy in HLHS patients is at a stage of infancy, and very few studies have been conducted in a very small patient population. The mode of delivery has been either intracoronary or intramyocardial, and the number and source of cells has varied. The time of delivery has been consistent, with most studies conducted at the time of the second stage palliation. Source and number of cells, along with the timing of delivery, must be optimized, and the modes of action must be elucidated. Furthermore, a combinational therapy consisting of stem cells coupled with growth factors to increase the contractile function of the RV may also prove to be beneficial.

As one example, there is a high incidence of progressive myocardial dysfunction after the Norwood surgery, as discussed in an earlier section. The goal is to identify molecular indicators responsible for myocardial dysfunction. To accomplish this, blood samples from patients at the time of Norwood can be used to generate iPSCs and cellular difference can be used to correlate with observed clinical outcomes. These iPSCs can further be used to bioengineer 3D patches and that can be used to model tissue level contractility in post-surgical response of HLHS patients. Bioengineered right ventricles (93) can be used to study the effect of volume overload on heart muscle function and subsequent myocardial dysfunction. iPSCs-CMs, patches and ventricles can be used as powerful tools to provide mechanistic information at the cell, tissue, and organ levels.

Bioengineered models may someday be used as potential therapies to support HLHS hearts. As an example, biological contractile pumps (99) can be used in the Fontan circuit to provide pulsatile support for blood, as opposed to the inert grafts currently in use which only serve a passive role. These biological pumps can support the flow of blood through the Fontan circuit. Another example is the use of bioengineered heart muscle, which can be used to add musculature to the right ventricle, thereby augmenting contractile function to support the increased blood volume after the Norwood surgery. Heart transplantation is another important therapeutic approach, but the current scarcity of donor hearts precludes heart transplant as routine management strategy. However, in the future, if bioengineered hearts become available, heart transplantation may be expanded as a treatment option for all HLHS patients, or at minimum, expanded to patients considered high risk for palliative surgeries.