Potential conflict of interest: Nothing to report.
The hyperdynamic circulatory syndrome observed in chronic liver diseases is a great example of research that originated from clinical observations and progressed in the last 50 years from the patient to the experimental laboratory. Our knowledge has evolved from the patient to the molecule, using experimental models that serve as a source for understanding the complex pathophysiological mechanisms that govern this complex syndrome. We now know that progressive vasodilatation is central to the detrimental effects observed in multiple organs. Although nitric oxide has been shown to be the primary vasodilator molecule in these effects, other molecules also participate in the complex mechanisms of vasodilatation. This review summarizes three major areas: first, clinical observation in patients; second, experimental models used to study the hyperdynamic circulatory syndrome; and third, the vasodilator molecules that play roles in vascular abnormalities observed in portal hypertension. (Hepatology 2006;43:S121–S131.)
“Today's catch-phrase is “ ‘from bench to bedside’” and university medical centres are struggling with the promotion of translational research… [There are examples] where astute observations of small group of patients have been the starting point of voyages of scientific discovery—from, as it were, the bedside to the bench—but in doing… so illustrate[s] the synergy between clinical and basic biomedical research and that now provides medical scientists with investigative tools of hitherto unimagined power.”1
—Sir Keith Peters, FRCP
The hyperdynamic circulatory syndrome observed in chronic liver diseases is a great example of research that originated from clinical observations by astute bedside investigators2 and progressed in the last 50 years from the patient to the experimental laboratory. In this fascinating journey, investigators have gone from the patient to the molecule, and en route established a strong intermediate relay station in experimental models that have served and still serve as the source readily available to understand the complex pathophysiological mechanisms that govern this complex syndrome.
The hyperdynamic syndrome should be better called “progressive vasodilatory syndrome,” because vasodilatation is the factor that brings about all the vascular changes and finally leads to the multiorgan involvement observed as a consequence of this hemodynamic change. Early in the course, investigators recognized the primary importance of vasodilatation in the initiation of this syndrome, and considerable effort and resources were invested in finding the vasodilatory molecule(s). Finally, at the beginning of the 1990s it became evident that a recently discovered molecule involved in a multiple biological mechanisms was the main culprit.3 As a rule in biology, a single agent cannot explain a whole syndrome. However, it became evident that nitric oxide, a biologically active gas, is the main molecule responsible for vasodilatation and the multiple organ malfunctions that characterize hyperdynamic circulation.4 In the last 15 years, we have learned a lot about the patients, the experimental model, and the molecules that are the “prime movers” of the hyperdynamic state. We here review our current knowledge of this cardinal syndrome.
In 1953, based on the clinical observation that patients with cirrhosis frequently showed “warm extremities, cutaneous vascular spiders, wide pulse pressure, and capillary pulsations in the nail beds,” Kowalski and Abelmann2 first demonstrated that cirrhosis is associated with a hyperdynamic circulatory syndrome. This study demonstrated an increase in cardiac output and a decrease in peripheral vascular resistance in patients with alcohol-induced cirrhosis. These findings were reproduced in subsequent studies5; however, the recognition of the harmful effect of this syndrome on multiple organs was only recognized years later.4 We know now that vasodilatation is central to the detrimental effect observed in several vital organs4 and that the multi-organ failure observed in chronic liver diseases is in large part attributable to this progressive vasodilatation. The harmful effects observed in the systemic circulation and several other vital organs always originate via the vasodilatory state. Whereas in the heart, the splanchnic, the pulmonary, and the cerebral circulation, these deleterious effects are mediated by the hyperdynamic circulation itself, in other organs such as the kidney and the brain (such as in chronic encephalopathy); it is a response to vasodilatation in the other circulatory beds (Fig. 1).
The Systemic Circulation, the Heart.
In this circulatory bed the evidence of this syndrome is usually first observed. “Warm extremities, cutaneous vascular spiders, wide pulse pressure, and capillary pulsations in the nail beds” are common findings in patients with overt liver diseases. However, over the years we have learned that these patients are hyperdynamic before the syndrome becomes clinically evident. The systemic circulatory syndrome seems secondary to changes occurring in regional vascular beds. Any change in peripheral vascular resistance is rapidly compensated by changes in cardiac output.6 This point is clearly seen in experiments and in patients with arteriovenous fistulae wherein opening of the fistula is enough to increase the cardiac output. However, in portal hypertension, the progressive reduction in peripheral resistance is slower, and compensatory mechanisms, such as sodium and water retention with expansion of the plasma volume, play a fundamental role in perpetuating and aggravating the hyperdynamic syndrome. We believe that the initial vasodilatation occurs in the splanchnic circulation and that the heart response is directly related to a combination of splanchnic vasodilatation and expansion of the plasma volume together with an increased venous return to the heart, in large part, through portal-systemic shunts (Fig. 2). Although vasodilatation is essential as the initiating factor, no hyperdynamic circulation occurs without expansion of the plasma volume and portal-systemic shunting.7, 8 In the long run, the heart behaves as in other forms of high cardiac output syndrome: Initial compensation according to the degree of individual cardiac reserve, followed sooner or later by some degree of cardiac insufficiency. The cardiac index is usually higher than normal (>4 L/min/m2) but insufficient to maintain arterial pressure on the face of progressive vasodilatation.9 Interestingly, high cardiac output failure is reversible once the initial cause leading to the high cardiac output is treated. This reversal has been observed also in patients with cirrhosis after liver transplantation.10, 11
The Hyperdynamic Splanchnic Circulation.
The hyperdynamic splanchnic circulation is central to the development of the syndrome, and although is commonly recognized as a complication of cirrhosis, it should be better conceptualized as a complication of portal hypertension. It has been observed in all forms of portal hypertension caused by a condition other than cirrhosis and confirmed in different experimental models of portal hypertension (Table 1); therefore, it cannot be considered solely a complication of cirrhosis.
Table 1. Experimental Models of Portal Hypertension and Hyperdynamic Circulation*
Representative publications for each model.
Portal Vein Constriction
Groszmann et al. Am J Physiol 1982;242:G156–G160
Vorobioff et al. Gastroenterology 1984;87:1120–1126
Hori et al. Dig Dis Sci 1993;38:2195–2202
Bile duct ligation
Lee et al. Am J Physiol 1986;251:G176–G180
Portal vein constriction
Iwakiri et al. Am J Physiol GI Liver Physiol 2002;283:G1074–G1081
Sarin et al. Am J Physiol 1990;258:G365–G369
Portal vein constriction
Cahill P.A. et al. Hepatology. 1995; 22(2):598–606
Bile duct ligation
Levy M. Am J Physiol. 1977;233:F572–F585
Bosch et al. Hepatology 1983;3:1002–1007
For many years the dominant theory explaining portal hypertension in cirrhosis was the “backward flow” theory, which postulated that increased portal vascular resistance was the only cause for the increase in portal pressure. This theory predicted a splanchnic hypodynamic circulation with increased mesenteric vascular resistance. Observations of a decreased portal blood flow at the hepatic hilum supported this theory.12 This hypodynamic or a normodynamic situation is observed in early cirrhosis when portal hypertension is mainly attributable to an increase in intrahepatic vascular resistance, and portal-systemic collaterals have not yet developed. In moderate to severe portal hypertension, however, when an extensive collateral circulation is present, the observation of a decreased portal blood flow entering the liver is misleading because does not take into account portal flow diverted through the collateral circulation. In the late 1960s and early 1970s, a series of studies in patients with well-established cirrhosis suggested that the splanchnic circulation was hyperdynamic.13–15 However, in the early 1980s, when a methodology to evaluate regional hemodynamics and portal-systemic shunting in rodent models of portal hypertension was developed, the hemodynamic events that follow the induction of portal hypertension were unequivocally demonstrated.16–18 An increase in splanchnic blood flow together with an increase in portal vascular resistance were shown to contribute to portal hypertension.17, 18 This process is called the “forward flow” theory, and it provides a rationale for the use of vasoconstrictors in patients with portal hypertension.
At that time, we coined the name “portal venous inflow” for the splanchnic blood flow entering into the portal system to distinguish it from the portal blood flow perfusing the liver.16 Portal hypertension is the only known pathophysiological situation in which the portal blood flow entering into the portal system is different from portal blood flow perfusing the liver.
In several studies, the hyperdynamic splanchnic circulation of the portal hypertensive patient has been demonstrated by using indirect techniques.19 We believe the initial signal that triggers the sequence vasodilatation-hyperdynamic circulation is located in this bed and that the signal is the initial increase in portal pressure itself (see Fig. 2).20, 21
The Hyperdynamic Pulmonary Circulation.
The hyperdynamic circulation also affects the lungs. Pulmonary vasodilatation is associated with the hepatopulmonary syndrome, one of the most severe complications of chronic liver diseases. Although the intrinsic mechanism that triggers this syndrome into its full expression is not fully known, local vasodilatation mediated by several endothelial vasodilators, including nitric oxide and carbon monoxide, plays a major role.22 This review deals with the consequences of the systemic hyperdynamic syndrome on the different organs affected by the progression of chronic liver diseases. Local factors in the pulmonary circulation may determine why only some patients develop the hepatopulmonary syndrome. High cardiac output, by increasing shear stress in the pulmonary vascular endothelium and by shortening pulmonary and tissue transit time of red cells,23, 24 may contribute as well to the severity of the hepatopulmonary syndrome.
The Renal and Cerebral Circulation.
The renal circulatory bed appears to be indirectly affected by the consequences of the hyperdynamic state (Fig. 1). The kidney responds to a perceived hypovolemia, in the only way it knows, by retaining sodium and water. In reality, progressive vasodilatation in these patients induces a state of “relative hypovolemia.” Rather than a decrease in intravascular volume, the relative hypovolemia results from an increase of the vascular compartment caused by vasodilatation, leading to a reduction in central blood volume, and activation of vasoconstrictive and volume-retaining neurohumoral mechanisms that perpetuate the sodium and water retentive state.25–27 In compensated cirrhosis, in the face of progressive vasodilatation, the intravascular volume and the cardiac output increase to maintain arterial perfusion pressure. With the progression of the disease, vasodilatation is accentuated, and the cardiac output continues to increase. Eventually the cardiac response is not enough to maintain perfusion pressure, the renal blood flow drops, and renal failure develops. This chain of events was recognized in the mid 1970s by Dr. J. N. Cohn, who suggested that this syndrome of progressive vasodilatation, increased cardiac output, and renal failure should be called cardiocirculatory hepatorenal failure.28 Recently, Ruiz-del-Arbol and others29 have shown that the hepatorenal syndrome seems to develop when the heart cannot compensate any longer for the progressive decrease in peripheral resistance. This limitation can be attributable to different causes,30 including the known effect of chronic high cardiac output, sepsis, or a decrease in the heart pre-load caused by the continuously increasing vasodilatation. Interestingly, Cohn et al.31 also suggested that treatment of the systemic vascular dilatation could result in an improvement in renal function. Again he has proved to be correct; chronic infusions of vasoconstrictors, especially vasopressin analogs, improve renal function in patients with hepatorenal syndrome.32 Infusions of these powerful vasoconstrictors not only improve systemic peripheral resistance but also, by constricting mainly the splanchnic bed, reduce portal pressure and increase renal perfusion pressure.
The relationship between the systemic vasodilatation of liver diseases and the cerebral circulation is perhaps the most difficult to define. Decreases as well as increases in cerebral blood flow have been described.33 The increase in blood flow has been mainly associated with acute liver failure; portal hypertension and systemic vasodilatation have also been encountered.33 The brain (as well as the kidney and the splanchnic circulatory bed) autoregulates blood flow within a wide margin of arterial pressure and, as Guyton6 had demonstrated, the heart output increases in response to the demands of the regional circulations. Therefore, if an increase occurs in cerebral blood flow, the driving force must be determined by local factors. Conversely, in chronic liver diseases, cerebral blood flow is decreased, and this decrease runs in parallel with the reduction in renal blood flow, suggesting that the mechanisms of this reduction are similar in these two vital organs.34 Local disruptions of the autoregulatory mechanisms, a decrease in arterial perfusion pressure beyond the autoregulatory margins, and organ vasoconstriction induced by the activation of the neurohormonal compensatory responses are probably the causative factors. Although increasing systemic perfusion pressure is beneficial to the renal circulation, such an increase in systemic perfusion pressure may be detrimental to the brain. In acute liver failure, increases in cerebral blood flow may lead to the development of brain edema.35 However, whether in chronic encephalopathy an increase in systemic perfusion pressure may be beneficial is of interest.
The development of experimental models to study the hyperdynamic circulation has been of critical importance for our understanding of the syndrome. Although the models result in different forms of liver diseases, the common denominator is portal hypertension. Vasodilatation and the hyperdynamic circulation have been observed in all forms of portal hypertension (see Table 1). We recently showed20 that increased portal pressure is the initial signal that triggers the molecular mechanisms that initiate the vasodilatory stimulus leading to splanchnic vasodilatation (see Fig. 2). The portal vein constricted model in the rat has served as the “war horse,” in which the different stages of the portal hypertensive syndrome as well as the day-to-day changes leading to the hyperdynamic state have been studied.36, 37 It is a highly reproducible and predictable model. The model has shown that a decrease in peripheral resistance leads to a decrease in the central blood volume before the onset of sodium retention; retention of sodium ceases once the central blood volume returns to baseline values.38
Although this model of pre-hepatic portal hypertension mimics in many ways some of the most notable hemodynamic characteristics observed in models of cirrhosis, it still lacks the severity and complexity of the multiple mechanisms leading to vasodilatation and the hyperdynamic circulation observed in cirrhosis39, 40 (Fig. 2). Again, in all these models, splanchnic and systemic vasodilatation is the initial step leading to the hyperdynamic syndrome (Fig. 2). Whereas the involvement of the splanchnic and systemic circulation is observed in all experimental models, significant abnormalities in other regional beds are seen only in some certain specific models and sometimes only after provocative maneuvers.33, 41
Blood flow autoregulation is very prominent in the brain, kidneys, and splanchnic circulation within a large margin of blood pressures values.6 Within the autoregulatory perfusion pressure margins, local metabolic changes modulate blood flow in these circulatory beds. However, outside these margins, perfusion pressure becomes the determinant of organ blood flow. Disruption of the autoregulatory process, by derangement of the local milieu, allows inappropriate and detrimental end organ responses to the changes in systemic hemodynamics. Such disruption may occur even within the organs' autoregulatory margins.
Clinical research poses critical questions that cannot be completely answered with human investigation. Results of animal research carried out to resolve these questions must be reexamined clinically to confirm their relevance to the human condition. Technological advances are providing increasing opportunities to carry out investigations directly on portal hypertensive patients; however, inherent limitations make research with animal models essential to elucidate basic mechanisms, improve diagnosis, and test potential therapies.
Nitric oxide (NO), an endothelial derived relaxing factor, is a key player in arterial vasodilatation in the splanchnic and systemic circulation, which leads to the hyperdynamic circulatory syndrome in portal hypertension.42–46 NO causes vasodilatation by stimulating soluble guanylyl cyclase to generate cyclic guanosine monophosphate (cGMP) in vascular smooth muscle.47
NO is synthesized by a family of three nitric oxide synthases (NOS): constitutively expressed isoforms, endothelial NOS (eNOS),48 and neuronal NOS (nNOS),49, 50 and inducible NOS (iNOS).51 Among these isoforms, eNOS is the major enzymatic source of the vascular NO overproduction in the splanchnic arterial circulation.21, 43, 52 Recent evidence suggest that nNOS, found in the neuron and vascular smooth muscle cells, is also upregulated in the mesenteric artery53, 54 and in aorta55 and plays a role in the development/maintenance of the hyperdynamic splanchnic circulation in experimental cirrhosis.55 The iNOS is synthesized de novo in a variety of cell types, including macrophages and vascular smooth muscle cells, only after induction by endotoxin and inflammatory cytokines.56 Interestingly, despite the presence of endotoxemia in cirrhosis, iNOS is not detected in the splanchnic arterial vasculature in both portal vein ligated and CCl4 rats with cirrhosis.40, 42, 44, 57 However, increased iNOS expression was observed in aorta isolated from rats with biliary cirrhosis.58, 59
eNOS is Ca2+/calmodulin-dependent and requires co-factors such as tetrahydrobiopterin (BH4) for its activity.60 In rats with cirrhosis, an increase in circulating endotoxin activates guanosine triphosphate (GTP)–cyclohydrolase I, which generates BH4 production in mesenteric arteries. This increase in BH4 is associated with enhanced eNOS activity and eNOS-derived NO overproduction in the mesenteric arterial beds.61
eNOS is regulated by complex protein–protein interactions and post-translational modification.60 Among positive regulator proteins, Hsp9062 and serine/threonine kinases Akt/protein kinase B contribute to the activation of eNOS in the splanchnic arterial circulation in portal hypertensive rats.63 Akt/protein kinase B directly phosphorylates eNOS at Ser1177 (human) or Ser1179 (bovine) and enhances its ability to generate NO.64–67 Various forms of stimuli, such as vascular endothelial growth factor, inflammatory cytokines, and mechanical forces by shear stress, stimulate the production of NO by this mechanism64–66 in portal hypertension.21, 58, 63 In the intestinal microcirculation, we observed that vascular endothelial growth factor upregulates eNOS protein expression,68 leading to the development of the hyperdynamic circulatory syndrome.69
In hepatopulmonary syndrome, eNOS-derived NO also plays a major role in intrapulmonary vasodilatation, but not iNOS-derived NO, similar to what is observed in the splanchnic circulations.70 Biliary cirrhosis is the only model in which the hepatopulmonary syndrome spontaneously develops. It is suggested that in this model the production and release of endothelin 1 (ET-1) by cholongiocytes is increased. ET-1 leads to activation of endothelin B (ETB) receptors in the pulmonary circulation inducing vasodilatation. ETB in the pulmonary circulation can be upregulated by shear stress, which is caused by the systemic hyperdynamic circulatory state. This outcome seems to be the mechanism for the local NO overproduction in the lung.22
Like NO, carbon monoxide (CO) is an endogenously produced gas molecule that activates soluble guanylyl cyclase, resulting in increased production of cGMP71 and regulating vascular tone in a manner similar to NO.72, 73 CO-induced vasodilatation is through the activation of Ca2+-activated potassium channels, independently from cGMP.74, 75 CO is produced from the breakdown of heme to biliverdin via the activity of the enzyme heme oxygenase (HO). Two isoforms of HO have been identified: HO-1 and HO-2. HO-1 is inducible by multiple agents, whereas HO-2 is constitutively expressed.76
A progressively increased expression of HO-1 was found in aorta and mesenteric arteries of rats with biliary cirrhosis, whereas no change was observed in HO-2 expression. The acute intraperitoneal injection of zinc protoporphyrin, a selective inhibitor for HO activity, to those rats at a dose that normalizes aortic HO activity, ameliorates the hyperdynamic circulatory syndrome, suggesting the role of CO in arterial vasodilatation in portal hypertension.77
CO also plays a role in intrapulmonary vasodilatation in hepatopulmonary syndrome both in humans78 and in the experimental model of biliary cirrhosis.79 CO circulates tightly bound to hemoglobin, resulting in the formation of carboxyhemoglobin,80, 81 used as an indicator of CO production. It has been observed that levels of blood carboxyhemoglobin were significantly increased in a cohort of patients with cirrhosis78, 80 and in experimental hepatopulmonary syndrome.79 The inhibition of pulmonary HO normalizes arterial carboxyhemoglobin levels and reduces vasodilatation.79 In the brain, activation of HO-1 and CO production, but not NO,82 could be a cause of cerebral hyperemia, leading to the development of brain edema in acute liver failure.83
Prostacyclin (PGI2) is synthesized by cyclooxygenase and released from the endothelium; however, PGI2 elicits smooth muscle relaxation by stimulating adenylyl cyclase and generation of cyclic adenosine monophosphate.84 An increase in circulating PGI2 has been observed in patients with cirrhosis85 and in portal hypertensive rabbit.86 A pathogenesis role for PGI2 in the hyperdynamic circulatory syndrome has been suggested.85–87
Endothelium-Derived Hyperpolarizing Factor.
Which molecule constitutes endothelium-derived hyperpolarizing factor (EDHF) is controversial. The EDHF-induced vasodilatation is characterized by an essential hyperpolarization of the vascular smooth muscle and could be blocked by inhibitors of potassium channels. The major molecules currently considered to explain EDHF-mediated dilations are: (1) arachidonic acid metabolites88–91; (2) the monovalent cation, K+92; (3) components of gap junctions93–95; and (4) hydrogen peroxide.96–100 EDHF seems to be more prominent in smaller arteries and arterioles than in larger arteries. This observation has been made in a number of vascular beds, including those from the mesenteric, cerebral, ear, and stomach.101–105 Presence of EDHF was suggested in mesenteric arteries isolated from rats with biliary cirrhosis.106 The role of EDHF becomes more significant when the production of NO is inhibited, because NO seems to inhibit the release of EDHF.107 This evidence may explain our previous finding that portal hypertension still causes the development of the hyperdynamic circulatory state in mice lacking eNOS.63 EDHF may be upregulated in mice lacking eNOS to compensate for the lack of NO production in endothelial cells.
Endogenous cannabinoids (or endocannabinoids) is a collective term describing a novel class of endogenous lipid ligands, including anandamide (arachidonyl ethanolamide).108 Endocannabinoids, through their binding to CB1 receptor, cause hypotension. Anandamide is increased in monocytes in cirrhosis, and over-activation of CB1 receptors within the mesenteric vasculature may contribute to the development of splanchnic vasodilatation and portal hypertension.109 The blockade of CB1 receptor by the antagonist SR141716A not only increases mean arterial pressure39, 109 and peripheral resistance39 but also reduces mesenteric blood flow in rats with CCl4-induced cirrhosis109 and increases the vascular tone in mesenteric artery in biliary cirrhosis.110 CB1 receptors are present in vascular endothelial cells, and activation of endothelial CB1 receptors has been shown to lead to increased production of NO.109, 111, 112 In contrast, a study by Ros et al.39 showed anandamide-induced hypotension is NO-independent.
Tumor Necrosis Factor Alpha.
Tumor necrosis factor alpha (TNFα), produced by mononuclear cells on activation by bacterial endotoxins, is found in increased levels in portal hypertension113, 114 and is a well-known mediator of NO release.115 Antagonism of TNF-α with anti-TNF-α antibody or inhibition of TNF-α synthesis by thalidomide or pentoxifylline blunts the development of the hyperdynamic circulation in portal hypertensive rats114, 116 and rats with biliary cirrhosis.117 The mechanism of action of TNF-α in portal hypertension with cirrhosis remains to be elucidated. However, TNF-α stimulates gene expression and activity of the key enzyme for regulation of BH4biosynthesis, GTP-cyclohydrolase I, in endothelial cells,118, 119 and enhanced BH4 production directly increases eNOS-derived NO bioavailability.118, 120 Conversely, in biliary cirrhotic rats, TNF-α through the activation of iNOS in lung and aorta, was shown to play a role in the development of the hyperdynamic circulatory syndrome and hepatopulmonary syndrome.117
Adrenomedullin is a potent vasodilatory peptide with 52 amino acid residues in the human and 50 amino acid residues in the rat. In patients with cirrhosis, there is an increase in circulating adrenomedullin level,121, 122 which was associated with increased plasma nitrite (a stable NO metabolite) and plasma volume expansion, and inversely correlated with peripheral resistance.122 The administration of anti-adrenomedullin antibodies prevents the occurrence of the hyperdynamic response in the early sepsis123 and ameliorates blunted contractile response to phenylephrine in aorta isolated from rats with cirrhosis.124 Adrenomedullin phosphorylates and activates Akt and increases cGMP production in rat aorta, an indicator of NO production. It has been suggested that adrenomedullin-mediated vasorelaxation may occur through the production of NO.125
A growing body of recent evidence suggests that hydrogen sulfide (H2S) is a potent endogenous vasodilator in aorta126, 127 and mesenteric arteries.128 H2S is synthesized endogenously from L -cysteine mainly by the activity of two enzymes, cystathionine-γ-lyase and cystathionine-β-synthase.126, 129 An intravenous bolus injection of H2S was shown to transiently decrease blood pressure of rats by 12 to 20 mm Hg. H2S-mediated vasodilatation is through the opening of KATP channel. Unlike NO or CO, H2S relaxes vascular tissues independent of the activation of cGMP pathway.130 The role of H2S in the hyperdynamic circulatory syndrome has not been studied.
The hyperdynamic circulation is fascinating for those who work in this field and equally so for workers in other specialties or other branches of biology. We have made tremendous progress in the last 50 years. Over the last 25 years, Hepatology has been integral to the process through which we have gone from the patient to the molecule. We are now returning to the patient in the form of various clinical treatments that are becoming available for the complications of the hyperdynamic state. No doubt knowledge in this area will continue to grow in basic science as well as the clinical arena, including studies in the experimental models that have given us a unique opportunity to provide a molecular basis for pathophysiological findings. 3
We apologize to colleagues whose references were omitted for the sake of brevity or whose contributions were cited in reviews.