If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Right-sided mechanical circulatory support for failing Fontan physiology has been largely unsuccessful due to inherent hemodynamic differences between these patients and the target populations for most assist devices. This study uses advanced benchtop modeling of Fontan physiology to examine the use of PediMag and CentriMag to improve failing Fontan hemodynamics.
Methods
Each device was attached to a compliance-matched, patient-specific total cavopulmonary connection in vitro model that used resistances, compliances, and programmable waveforms to establish “failing Fontan” baseline hemodynamics (cardiac output [CO] = 3.5 L/min and central venous pressure ∼17 mm Hg). The ability of the assist devices to improve failing Fontan hemodynamics (reduce inferior vena cava pressure and augment CO) was investigated.
Results
Requiring complete Fontan pathway restriction, PediMag reduced inferior vena cava pressure by ∼10 mm Hg and supported CO augmentation up to 5 L/min. This was accompanied by an increase in superior vena cava pressure of ∼6 mm Hg. CentriMag produced similar hemodynamic changes without the need for pathway restriction or an increase in superior vena cava pressure.
Conclusions
PediMag and CentriMag right-sided support led to a decrease in inferior vena cava pressure and augmentation of cardiac output. In the case of CentriMag, this is accomplished without an increase in superior vena cava pressure or the need for restrictive banding. This work provides further data to help with the optimal design of a Fontan assist device to ameliorate the growing need.
The ability to mechanically support the growing number of failing Fontan patients with minimal alterations to the Fontan anatomy or support device is needed. In this study, PediMag and CentriMag right-sided support reduced vena cava pressure and supported cardiac output augmentation in an in vitro failing Fontan model, offering a novel and clinically relevant application of these devices.
The use of mechanical circulatory support (MCS) for failing Fontan patients is an area of growing interest, as the increased life expectancy of Fontan patients continues to be accompanied by numerous end-organ complications.
Many of these complications are hypothesized to be hemodynamic driven, with hepatic congestion, elevated central venous pressure, and low cardiac output (CO) noted as some of the driving forces.
Progression of liver pathology in patients undergoing the Fontan procedure: chronic passive congestion, cardiac cirrhosis, hepatic adenoma, and hepatocellular carcinoma.
Fontan patients can experience different modes of circulatory failure, which may require the need for different types of support. Compromised ventricular function may benefit from ventricular assist devices (VADs), whereas poor ventricular filling due to high pulmonary vascular resistance (PVR) and/or low flow across the Fontan circuit may require systemic venous MCS positioned as Fontan assist devices (FAD). These 2 modes of failure are characterized by different anatomic and hemodynamic conditions and therefore may require different MCS performance profiles. In addition, most MCS devices are not designed specifically for Fontan use, adding to the complexity of device selection for these patients.
Motivated by the limited success seen in previous work, the purpose of this study is to investigate the use of the PediMag and CentriMag (both Abbott, Santa Clara, Calif) devices as right-sided support for failing Fontan patients. In this study, we use an in vitro testing platform to explore device efficacy, dye flow visualization techniques to understand the underlying fluid mechanics, and computational modeling to assess the applicability to multiple patient anatomies. We hypothesize that these devices will be able to provide successful failing Fontan support by decreasing central venous pressure and augmenting CO.
Methods
Fontan Circulatory Loop
An overall schematic of the in vitro Fontan circulatory loop is shown in Figure 1, A. A programmable piston pump powers the loop, representing the single systemic ventricle (70 bpm), with a bileaflet mechanical heart valve on either side representing the atrioventricular and aortic valves. Compliance chambers and ball valves were used to model aortic and systemic compliance and resistance, respectively. All pathways were designed as 3-element Windkessel models (a common modeling technique), employing a resistance−compliance−resistance circuit (only one resistance shown in Figure 1 for clarity).
Pressures and flows were recorded at each of the total cavopulmonary connection (TCPC) vessels.
Figure 1In vitro loop design. A, Schematic of in vitro Fontan circulatory loop with labeled components: (1) programmable piston pump (single ventricle), bileaflet mechanical heart valves for the (2) aortic valve and (3) atrioventricular valve, (4) aortic compliance chamber, (5) inferior and (6) superior systemic compliance chambers, (7) inferior and (8) superior systemic resistance, (9) patient-specific total cavopulmonary connection, (10) left and (11) right pulmonary resistance, (12) left and (13) right pulmonary compliance chambers, (14) single atrium compliance chamber. “P” and “F” indicate locations of pressure and flow measurements respectively. B, Orientation of total cavopulmonary connection and inflow/outflow cannula. SVC, Superior vena cava; RPA, right pulmonary artery; LPA, left pulmonary artery; IVC, inferior vena cava.
Blood was modeled with a 36%/74% glycerin-saline solution (by volume) with 3.47 cSt kinematic viscosity. The in vitro loop employed a patient-specific (21-year-old female patient, body surface area 1.69 m2), flexible TCPC anatomy reconstructed from magnetic resonance imaging with institutional review board approval (H09279; approved August 4, 2016). The TCPC model was constructed of transparent silicone with a patient-specific matched bulk compliance (1.36 mL/mm Hg).
Baseline failing Fontan conditions were defined as a CO of 3.5 L/min with Fontan pressures near 15 to 20 mm Hg. Various levels of vessel stenosis were present in this patient model. The inferior and superior vena cava (IVC and SVC) and left and right pulmonary artery (LPA and RPA) flow splits were set to 60/40 and 50/50, respectively.
Fontan Assist Devices
In this study, the PediMag and CentriMag devices were tested as FADs. This right-sided support was attempted by placing the inflow cannula in the IVC near the diaphragm and positioning the egress at the LPA/RPA confluence as shown in Figure 1, B. Both devices are extracorporeal devices, which use a magnetically levitated/driven pump impeller. PediMag is currently approved for clinical use up to 6 hours as a short-term solution (eg, cardiopulmonary bypass), whereas surgical/long-term circulatory support decisions can be made, and CentriMag is currently approved for use as a right ventricle assist device for up to 30 days in patients in cardiogenic shock due to acute right ventricular failure. Both pumps run off the same console and motor but are different sizes. The priming volumes for PediMag and CentriMag are 14 and 31 mL, respectively. Cannula sizing also differs, with 1/4- and 3/8-inch tubing for the PediMag and CentriMag, respectively. As both pumps run off the same console and motor, each has a pump rotational range of 0 to 5500 rpm. The instructions for use state that the outflow capacity for the PediMag and CentriMag devices is 0 to 1.5 L/min and 0 to 9.99 L/min, respectively.
Experimental Protocol
In this study, successful Fontan support was defined as the ability to simultaneously decrease IVC pressure (Fontan pressure) by 5 mm Hg and increase CO to 4.25 L/min (a 0.75 L/min increase) without increasing SVC pressure.
If these criteria were met, the FAD configuration was deemed successful. Identical failing Fontan baseline conditions were first established in each scenario. The FAD was then turned on and a change in Fontan pressures was observed. If the IVC pressure decreased (reduced venous congestion) and PA pressures increased (improved ventricular filling), ventricle adaptation was modeled by increasing the output of the “single-ventricle” pump until the pulmonary artery pressures returned to their baseline values. This is based on the assumption that the increased flow-related pressure rise in pulmonary arteries represents the increased cardiac preload to be delivered to the ventricle under equivalent conditions. The resulting hemodynamic conditions were then compared with the criteria for success. If unsuccessful, the FAD configuration/parameters were adjusted and the process was repeated with the goal of identifying optimal FAD settings for the given configuration. Hemodynamics at each support scenario were recorded for approximately 35 cardiac cycles. The average flow/pressure across this period is reported in this study. Variability between cardiac cycles was less than 5% for these metrics. Measurement error for the both the pressures and flows was less than 2%.
Dye Flow Visualization
Dye flow visualization techniques were used to better understand the fluid dynamics inside the TCPC. A 20% (by volume) solution of red dye was injected into the SVC at baseline failing Fontan conditions at various FAD rpm. A Phantom VEO 340L (AMETEK Vision Research, Wayne, NJ) high-speed camera was used to record video at 1000 frames per second to allow for a qualitative understanding of the flow fields within the TCPC.
Computational Modeling
Computational methods (validated with experimental data) were used to model blood flow through 5 additional patient anatomies (average age and body surface area of 14.2 years and 3.47 m2, respectively) to investigate the generalizability of experimental results across various patient anatomies. The 5 patients were deliberately selected to represent a range of anatomies including variations in vessel sizes and IVC-SVC offsets, a variety of diagnoses, and both intracardiac and extracardiac Fontan connections. All patient data were received from Children's Healthcare of Atlanta (institutional review board no. H09279). All computational fluid dynamic simulations were performed using ANSYS Fluent (Release 19.0; ANSYS Inc, Canonsburg, Pa) with patient-specific blood flow waveforms derived from magnetic resonance imaging and anatomy as the boundary conditions and domain, respectively. Blood was modeled as a single-phase Newtonian fluid (μ = 0.04 g/[cm·s], ρ = 1.06 g/cm3) and rigid vessel walls were assumed. The CentriMag device was modeled using the performance curve provided in the instructions for use.
Results
Physiologically accurate failing Fontan baseline conditions were achieved with the in vitro circulatory loop. IVC, SVC, and pulmonary artery pressures were near 15 mm Hg, with flow rates meeting the defined 60/40 and 50/50 IVC/SVC and LPA/RPA flow ratios, respectively.
PediMag
As with previous studies testing other devices, significant amounts of recirculation through the FP were observed with the PediMag device.
To reduce recirculation, various levels of FP banding were explored. Eighty percent FP restriction (by area) was required before any noticeable hemodynamic changes were observed (Figure 2, A). On the basis of these findings, the impact of complete restriction of the FP was assessed. With complete FP restriction, PediMag was able to decrease IVC pressure by >10 mm Hg while increasing PA pressures by approximately 2 mm Hg at the baseline flow conditions (Table 1). Following the established experimental protocol, these favorable changes allow for ventricle adaptation, which was modeled by increasing CO. PediMag easily surpassed the initial criteria for success, effectively maintaining decreased IVC pressure and increased PA pressures even up to a CO of 5 L/min (Figure 3, Table 1). However, these improvements in hemodynamics were accompanied by a concerning increase in SVC pressure of approximately 6 mm Hg.
Figure 2Effect of banding design on hemodynamics. A, Effect of banding on pressure. B, Theoretical relationship between banding length (L) and area to maintain resistance (r = radius). C, Top and side views of the 5 custom banding designs. D, Effect of banding design on SVC pressure (Fontan assist device rpm labeled above each bar). Banding is calculated as percent constriction by area. All points on the curve in (B) theoretically offer the same resistance. The asterisk in (D) indicates that SVC pressure increase was measured at the point when IVC pressure was reduced by 10 mm Hg. IVC, Inferior vena cava; SVC, superior vena cava; LPA, left pulmonary artery; RPA, right pulmonary artery.
Figure 3Hemodynamic changes achieved using PediMag. A, Average pressures and B, average flow rates. PediMag rpm are indicated below each cardiac output. LPA and RPA flow rates are nearly identical in (B) and therefore appear as a single line. IVC, Inferior vena cava; SVC, superior vena cava; LPA, left pulmonary artery; RPA, right pulmonary artery.
Various banding designs were then explored in an attempt to avoid complete FP restriction (a major factor in SVC pressure increase) while still effectively reducing recirculation. As mentioned earlier, 80% restriction was needed to produce any changes in hemodynamics (Figure 2, A). Using the assumption that flow through the FP can be roughly approximated as pipe flow, Poiseuille's law was used to determine the theoretical relationship between banding dimensions (radius and length) that would produce identical resistances (Figure 2, B). Five custom bands were prototyped for testing (Figure 2, C), corresponding to the 5 points in Figure 2, B. Band #3 was the only design able to decrease IVC pressure by >10 mm Hg while not increasing SVC pressure more than the completely clamped scenario (Figure 2, D). However, this required PediMag to run at 4000 rpm (approaching its upper limit), which did not allow for successful support at greater COs. Bands #4 and #5 did not produce enough resistance to prevent recirculation and reduce IVC pressure (Figure 2, D).
CentriMag
The need for greater output to overcome recirculation encouraged the testing of the CentriMag device. As shown in Figure 4, A, CentriMag was able to produce the desired hemodynamic changes with no FP banding. IVC pressure decreased ∼10 mm Hg whereas the average PA pressure increased ∼1.5 mm Hg (Table 1). In addition, SVC pressure decreased by ∼4 mm Hg, overcoming a major limitation of the PediMag device. Following these favorable changes, ventricle adaption was modeled (Figure 4, B and C). As with PediMag, CentriMag could easily support a CO up to 5 L/min while maintaining decreased IVC pressure and increased PA pressures (Table 1). Dye flow visualization techniques revealed that at high enough FAD output, SVC flow was pulled down through the FP and into the inflow cannula (Videos 1 and 2). This explains the decrease in SVC pressure seen with the CentriMag device while at the same time achieving increased forward flow through the pulmonary arteries.
Video 1Slow-motion visualization of SVC flow interactions with CentriMag turned off. Dye is injected at the SVC upstream of the total cavopulmonary connection. Arrows indicate the direction of flow in the cannula. SVC, Superior vena cava. Video available at: https://www.jtcvs.org/article/S0022-5223(19)30933-X/fulltext.
Video 2Slow-motion visualization of SVC flow interactions with CentriMag running at 2700 rpm. Dye is injected at the SVC upstream of the total cavopulmonary connection. Arrows indicate the direction of flow in the cannula. SVC, Superior vena cava. Video available at: https://www.jtcvs.org/article/S0022-5223(19)30933-X/fulltext.
Figure 4Hemodynamic changes achieved using CentriMag. A, Effect of rotational speed on Fontan pressures. B, Average pressures and C, average flow rates during ventricle adaptation. This scenario required no banding of the Fontan pathway. CentriMag rpm are indicated below each cardiac output. IVC, Inferior vena cava; SVC, superior vena cava; LPA, left pulmonary artery; RPA, right pulmonary artery.
After validating computational results with the patient used for the in vitro experiments, 5 additional patients were computationally modeled to determine whether the phenomena of SVC flow moving through the FP and into the inflow cannula would occur for various Fontan anatomies. Streamlines originating from the SVC are shown in Figure 5. All patients showed similar results, suggesting that the CentriMag results achieved experimentally are generalizable across various patient anatomies.
Figure 5Superior vena cava (SVC) streamlines for 6 Fontan geometries. The upper left patient was the patient used for all experimental testing. These images show streamlines only from the SVC, which shows SVC flow is being pulled down the Fontan pathway and into the inflow cannula. FAD, Fontan assist device.
This study underscores the challenges of supporting a failing Fontan physiology. Similar to our previous studies, significant amounts of recirculation through the FP limited the effectiveness of the PediMag device and required complete clamping of the FP. In this scenario, PediMag was able to outperform any device previously tested and supported the failing Fontan circulation up to a CO of 5 L/min. However, this scenario (pulling from the IVC and ejecting into the SVC/PA junction with restrictive FP banding) resulted in increased SVC pressure. Use of CentriMag as an FAD overcame these shortcomings by providing enough output to overcome the recirculating flow and therefore not require the need for restrictive banding. Importantly, the decrease in IVC pressure and augmentation of CO achieved using CentriMag was accompanied by a decrease in SVC pressure as well. This is possible because even with recirculation through the FP, CentriMag is powerful enough to still provide sufficient forward flow to the pulmonary arteries, allowing for successful support. In addition, the recirculating flow is able to entrain SVC flow, effectively “pulling” SVC flow and therefore decrease SVC pressure. These observations provide an important insight into the pump parameters needed for suitable FADs.
The number of patients living with complex congenital heart disease continues to increase, and patients with single ventricle physiologies palliated with the Fontan TCPC continue to survive longer. This, however, sets them up for the inevitable Fontan failure; currently with very limited options for rescue, including heart transplant. The essential subtypes of Fontan failure have been described as (1) primary ventricular failure and (2) failure of the Fontan circulation in the setting of preserved ventricular function and pressure. Failure of the systemic ventricle can manifest similar to heart failure in anatomically normal hearts with increased end-diastolic pressure, increased left atrial pressure, and decreased CO with or without decreased systolic function. In this situation, at failure of medical therapy, VADs can be implanted in the standard fashion (systemic ventricle to aorta). Assistance in this manner has been shown to provide relief from such types of Fontan failure as a bridge to transplantation. Both pulsatile devices (Berlin Excor, Thoratec PVAD) as well as continuous-flow devices (HeartWare HVAD, CentriMag) have been used to support the systemic ventricle with varying success. Arnaoutakis and colleagues,
Outcomes of children with congenital heart disease implanted with ventricular assist devices: an analysis of the pediatric interagency registry for mechanical circulatory support (PEDIMACS).
(analyzing Pedimacs data, n = 19) shows 95% survival at 6 months for Fontan patients supported by an assist device. Therefore, supporting failing Fontan with ventricular dysfunction has become quite successful.
Supporting a failing Fontan with preserved ventricular function (the purpose of this study) is even more challenging in our view. This is supported by the fact that there are a large number of Fontan patients that develop this pathophysiology with essentially only a few reports of attempts to support these patients mechanically.
In addition, we believe that there may be underreporting of other unsuccessful attempts at such support. In all of the published attempts to support the Fontan side failure, modifications have been required in the form of a Fontan take down (Prêtre and colleagues
). One of our interests here was to assess the feasibility of offering successful support using off-the-shelf devices with no or minimal modification to the Fontan anatomy or MCS device.
Patients with pulmonary hypertension and high PVR pose an additional challenge to this type of support. Patients with pulmonary hypertension would be expected to have elevated pulmonary artery pressures at baseline and perhaps even more pronounced as flow was increased with VAD support. If this type of VAD support were being used clinically, these individuals may require pulmonary vasodilator therapies to allow for increased flow and manageable pulmonary artery pressures. If PVR were too elevated, such a device may not be a viable therapy. We would also expect that such a device would require flow to be increased over a period of days to weeks to allow for accommodation to the changes in circulation. Very few Fontan patients will ever survive having severe pulmonary hypertension and would not be candidates for this intervention. Mild elevation in PVR is poorly tolerated by the Fontan patient, with greater venous filling pressures, congestion, low CO, etc. These patients may benefit the most from unloading the venous circulation and augmenting CO. Until clinical trials are completed, this would be unknown.
In general, Fontan patients with elevated PVR would be treated medically. In the setting of inadequate response to medical therapy, the options are very limited. Furthermore, the calculated PVR in Fontan patients is further confounded by the low cardiac index. Low CO leading to progressive hypoxia further increases PVR in these patients. Therefore, augmentation of CO using a FAD along with concomitant aggressive vasodilatory therapy could potentially reverse the incentive for elevated PVR. However, if on hemodynamic assessment before placement of FAD, the patient has nonresponsive PVR as well as high ventricular end-diastolic pressures, this strategy would not be relevant. Patient selection and medical management is important to avoid a scenario of forcing blood into a vascular bed with high resistance.
In light of these limitations and challenges associated with supporting failing single-ventricle patients, specifically failing Fontan patients, development of devices that allow for the adaptability necessary for practical use is critical. The current study provides data to support the notion that a centrifugal pump such as CentriMag may be able to assist the failing Fontan physiology (graphical abstract). The advantage of this pump design is its high output (designed for full cardiac support) as well as the cannulation options adaptable to the TCPC anatomy. Allowing for both an inflow and outflow cannula to be custom sized as needed depending on the patient anatomy permits optimal positioning of these cannula.
The devices investigated here offer paracorporeal support and are therefore not a long-term strategy. However, the study finding that CentriMag may be able to effectively support a failing Fontan circulation and potentially improve hemodynamics has important clinical implications. It provides an option for temporary relief of the failing Fontan scenario and the ability to optimize the patient before consideration for long-term support or transplantation. CentriMag and PediMag have both been used as bridge to transplant devices in the pediatric population as left VADs.
Their ability to support the failing Fontan circulation provides a novel and clinically relevant application.
For long-term support, currently available durable intracorporeal continuous flow devices however have fixed, short, inflow cannula (metal) that necessitate positioning of the device proximate to the ventricle. Such proximate positioning as well as the straight metal cannula is not well suited for attachment to the Fontan conduit. This therefore necessitates design modifications that would allow for flexibility in cannula length and positioning, as well as device positioning due to significant variations in FP anatomy and intrathoracic relationships. A conceptual design for such a pump is shown in Figure 6. Here the intracorporeal, durable centrifugal pump is anchored to the right hemidiaphragm with inbuilt anchoring hooks. In addition, the device can be anchored to the chest wall. The inflow and outflow cannula made of flexible material such as Dacron can be custom sized and appropriately sutured to the Fontan conduit. The driveline is externalized to be connected to the controller and power source. A design such as this may be able to provide long term support for this group of patients as a bridge to transplant or even as destination therapy.
Figure 6Conceptual design and placement of an intracorporeal durable centrifugal pump for failing Fontan support. Driveline not shown.
Although frequently used, in vitro modeling cannot necessarily replicate the dynamic nature of human pathophysiology. In addition, the Fontan circulatory loop employed in this study has not been validated for all scenarios of failing Fontan physiologies. Lack of chronic or failing Fontan animal models, however, make the in vitro modeling very necessary. Our model of failing Fontan is able to replicate acute effects of changes in Fontan physiology in response to mechanical support very well. However, such modeling cannot replicate long-term adaptations that may occur in the systemic and pulmonary vascular bed. We do believe that a positive correction such as the one obtained by this study will lead to improved oxygenation, CO, and therefore lead to positive remodeling of the pulmonary vascular bed. This is unsubstantiated at this time and can be only assessed with future clinical studies. The other limitation of the study comes from the fact that the immediate application of such a pump is limited to temporary support only due to its paracorporeal nature. As suggested previously, design modifications are necessary for an optimal FAD.
Conclusions
This in vitro study demonstrates that optimal support to correct the failing Fontan physiology can be achieved using PediMag and CentriMag devices, leading to a decrease in IVC pressure and augmentation of CO. In the case of CentriMag, this is accomplished without an increase in SVC pressure or the need for restrictive banding. This in vitro study provides further engineering data to help with the optimal design of a FAD to ameliorate the growing need.
Conflict of Interest Statement
Authors have nothing to disclose with regard to commercial support.
The devices under investigation were provided by Abbott. The authors acknowledge the use of ANSYS software, which was provided through an Academic Partnership between ANSYS, Inc, and the Cardiovascular Fluid Mechanics Lab at the Georgia Institute of Technology. We also acknowledge the use of EchoPixel, Inc, for visualization of the proposed, conceptual design.
Slow-motion visualization of SVC flow interactions with CentriMag turned off. Dye is injected at the SVC upstream of the total cavopulmonary connection. Arrows indicate the direction of flow in the cannula. SVC, Superior vena cava. Video available at: https://www.jtcvs.org/article/S0022-5223(19)30933-X/fulltext.
Slow-motion visualization of SVC flow interactions with CentriMag running at 2700 rpm. Dye is injected at the SVC upstream of the total cavopulmonary connection. Arrows indicate the direction of flow in the cannula. SVC, Superior vena cava. Video available at: https://www.jtcvs.org/article/S0022-5223(19)30933-X/fulltext.
References
Rychik J.
Goldberg D.
Rand E.
Semeao E.
Russo P.
Dori Y.
et al.
End-organ consequences of the Fontan operation: liver fibrosis, protein-losing enteropathy and plastic bronchitis.
Progression of liver pathology in patients undergoing the Fontan procedure: chronic passive congestion, cardiac cirrhosis, hepatic adenoma, and hepatocellular carcinoma.
Outcomes of children with congenital heart disease implanted with ventricular assist devices: an analysis of the pediatric interagency registry for mechanical circulatory support (PEDIMACS).
This work was partially funded by the Michael P. Fisher Cardiac ICU research fund (Children's Healthcare of Atlanta), as well as an American Heart Association Predoctoral Fellowship 17PRE33630117.
Long-term survival of the patients with a functional single-ventricle physiology has dramatically improved with the establishment of staged surgical palliation and subsequent Fontan operation. As the number of the patients who live with a Fontan circulation has been increasing exponentially, management of the failing Fontan circulation becomes one of the most critical issues in pediatric and adult congenital heart disease care. Dr Trusty and colleagues1 reported an in vitro analysis of the mechanical circulatory support (MCS) system for failing Fontan circulation using a PediMag (Thoratec Corp, Pleasanton, Calif) or CentriMag (Thoratec Corp) centrifugal pump.
The “un-natural” history of the Fontan circulation portends a rather grim long-term outlook for patients with single ventricle congenital heart disease. Decades of experience with the Fontan circulation have shown the advantages of the total cavopulmonary connection that minimizes power loss to improve subpulmonary flow efficiency; yet, the altered hemodynamics impart a chronic attrition to the organ systems of the body leading to a range of complications and limited life expectancy for this patient cohort.
30933-X&id=fx1.jpg){kind=link}
30933-X&id=fx2.jpg){kind=link}
30933-X&id=gr1.jpg){kind=link}
30933-X&id=gr2.jpg){kind=link}
30933-X&id=gr3.jpg){kind=link}
30933-X&id=fx3.jpg){kind=link}
30933-X&id=fx4.jpg){kind=link}
30933-X&id=gr4.jpg){kind=link}
30933-X&id=gr5.jpg){kind=link}
30933-X&id=gr6.jpg){kind=link}