Haemodynamic evaluation of pulmonary hypertension

Pulmonary artery pressure

The pressure drop across the pulmonary circulation (i.e. the driving pressure) is often referred to as the transpulmonary pressure gradient (TPG): In analogy with the electric Ohm's law, TPG equals PVR times cardiac output 10–13: The equation can be rewritten as follows: The normal pulmonary circulation is a low resistance circuit, with little or no resting vascular tone, and the most important factors influencing mean PAP are hydrostatic pressure, intra-alveolar pressure, left atrial pressure and alveolar gases.

PH is most often defined as a mean PAP >25 mmHg at rest. According to the mechanistic classification 12, increases in mean PAP may be passive (as a result of increased downstream pressure), hyperkinetic (as a result of increased cardiac output through the lungs) or due to increased PVR resulting from changes in the pulmonary circulation itself. Granted that various hypertensive mechanisms can work jointly, two forms of PH have been defined. Postcapillary PH (or pulmonary venous hypertension) is a passive form characterised by an increased downstream pressure ≥15 mmHg and a normal TPG. Precapillary PH (or pulmonary arterial hypertension) is characterised by a normal downstream pressure (<15 mmHg). TPG is increased because of increased cardiac output or increased PVR. Increases in PVR are due to a significant reduction in the area of the distal (mainly resistive) and/or proximal (mainly capacitive) PAs. This classification illustrates the crucial role of cardiac catheterisation in determining not only the mean PAP and cardiac output but also the downstream pressure.

Pulmonary vascular resistance

The PVR is used to characterise PH in a more restricted sense and to quantify abnormalities of the pulmonary vasculature according to the following equation: PVR is expressed in Wood units=(1 WU=1 mmHg·min·L⁻¹=80 dyne·s·cm⁻⁵).

The PVR is mainly related to the geometry of small distal resistive pulmonary arterioles. According to Poiseuille's law, PVR is inversely related to the fourth power of arterial radius. PVR is therefore considered to mainly reflect the functional status of pulmonary vascular endothelium/smooth muscle cell coupled system 10–13. PVR is also positively related to blood viscosity and may be influenced by changes in perivascular alveolar and pleural pressure.

Pressure is independent of the size of the system, and PAP from various subjects can therefore be compared without the need to take into account potential differences in their body size. Conversely, variables such as volume are proportional to the size of the system. For comparative purposes, the PVRI is therefore defined as the pressure drop across the circuit divided by cardiac index (in WU·m² or in mmHg·L⁻¹·min·m² or in dyne·s·cm⁻⁵·m²). In patients with high BMI, the use of PVR instead of PVRI has been responsible for significant underestimation of PH, and for the occurrence of haemodynamic and respiratory failure following heart transplantation. The use of PVRI must be recommended in studies evaluating new therapeutic strategies in PH.

Pulmonary arterial hypertension results from three main elements: vascular wall remodelling, thrombosis and vasoconstriction 8. The increases in PVRI may be fixed and/or potentially reversible. Arterial obstruction, obliteration and remodelling are responsible for the fixed component, while active increases in vascular tone are responsible for the reversible component, which may account for >50% PVRI. The pulmonary vascular tone results from a complex interplay between the pulmonary endothelium, smooth muscle cells, extracellular matrix, and circulating blood cells and blood components. In PH, the dysfunction of pulmonary arterial endothelium plays a key role, whether due to external stimulus (e.g. shear stress, shear rate, hypoxia, acidosis) or to the disease process itself (e.g. primary PH).

Hypoxia, acidosis, endothelin, nitric oxide, thrombosis, neurohormones

Alveolar hypoxia is a major stimulus leading to pulmonary vasoconstriction either via a direct pressor effect or by causing mediators to discharge. Acidosis leads to pulmonary vasoconstriction as well as acting synergistically with hypoxia. Hypoxic vasoconstriction is the main mechanism explaining mild and moderate degrees of PH in patients with chronic obstructive pulmonary disease (COPD), in whom severe chronic long-standing hypoxia is observed. An oxygen tension in arterial blood (P_a,O2) <7.98 kPa (<60 mmHg) and a carbon dioxide tension in arterial blood >5.32 kPa (>40 mmHg) are thought to be accurate thresholds for the development of PH in COPD 14. A recent study has shown that mean PAP inversely correlated with arterial P_a,O2, forced expiratory volume in one second (%) and single-breath carbon monoxide diffusion capacity (%) and directly correlated with pulmonary wedge pressure in patients with severe emphysema 15. Surprisingly, all factors but P_a,O2 remained significant determinants of mean PAP when multiple regression analysis was used. Although methodological explanations may be discussed, this result suggests that factors other than hypoxia are involved in PH of patients with severe emphysema 15. In COPD patients with mild-to-moderate hypoxia, the progression of PAP is very slow (+0.4 mmHg·yr⁻¹) and only initial values of resting and exercise mean PAP are independently related to the subsequent development of PH 16. In patients with primary PH or chronic pulmonary thromboembolism, hypocapnia is commonly observed; the P_a,O2 may be within normal limits or only slightly decreased at rest, while hypoxaemia is observed with exercise 17, 18. Right-to-left shunt is the main mechanism of hypoxia in the Eisenmenger syndrome. Severe hypoventilation is associated with hypoxia and may lead to PH, particularly if there is associated acidosis. This may explain the PH observed in the setting of the obesity-hypoventilation Pickwickian syndrome and in a number of muscular disorders 6. In patients with PH, high resting PAP and PVRI values increase further following exercise or acute hypoxaemia (e.g. rapid eye movement (REM) sleep or respiratory failure in patients with COPD). It is important to prevent, diagnose and treat pulmonary infections in PH patients. Altitudes >1500 m must be avoided without supplemental oxygen, and altitudes >3000 m must be discouraged.

The pathogenic role of endogenous endothelin-1 has been stressed, together with impaired synthesis of vasorelaxant nitric oxide, and this may have therapeutic implications 19, 20. PH is associated with activation of the endothelin system, which has potent vasoconstrictive and mitogen properties. Thrombosis can play a part in the pathophysiology of the disease, as attested to by the platelet activation, disturbances of various steps of the coagulation cascade and abnormal thrombolysis described in PH 21. Primary or secondary endothelial dysfunction increase the risk of thrombotic events. Dilated right heart chambers, sluggish pulmonary blood flow and sedentary lifestyle also increase this risk. Adrenergic overdrive may precipitate right ventricle (RV) failure and high plasma noradrenaline is associated with increased mortality in patients with primary PH, suggesting that the level of sympathetic activation relates to the severity of the disease 22.

Other precipitating factors

Changes in the rheological blood properties may aggravate PH. Erythrocytosis may be secondary to hypoxia, and is potentially responsible for increased blood viscosity and changes in erythrocyte deformability. PH may be aggravated by increases in cardiac output in the setting of hyperadrenergic states, anaemia and hyperthyroidism. Decreased diastolic time (e.g. tachycardia) or the loss of atrial contribution to left ventricle (LV) filling (e.g. atrial fibrillation) tend to increase left atrial pressure and thus may also aggravate PH. Finally, permanent PH is self-aggravating, as it favours several local pathological processes, including the remodelling of small distal arteries and the loss of the elastic properties of proximal arteries, thus leading to a vicious circle.

Right ventricular function

RV has a complex geometry and is characterised by a crescentic shape and a thin wall 23. The crista supraventricularis divides the RV into inflow and outflow regions. Inflow (sinus), which is located posterior and inferior, has a greater fibre shortening and is the effective flow generator, pumping >85% of the stroke volume. Outflow (conus), which is located anterior and superior, is a resistive and pulsatile conduit with a limited ejection capacity. RV contraction proceeds from the sinus to the conus according to a peristaltic movement. Right ventricular output equals heart rate times stroke volume. For a given stroke volume, RV output increases when heart rate increases (chronotropic reserve). For a given heart rate, RV output increases when RV end-diastolic volume increases (preload reserve) or when RV end-systolic volume decreases. The RV end-systolic volume decrease can be due to increased inotropy (inotropic reserve) or to decreased RV end-systolic pressure. The latter mechanism may be observed in some healthy subjects during exercise (distensibility, recruitment) and in a subgroup of PH patients following vasodilators (pulmonary vasodilatory reserve).

The thin-walled, highly compliant RV can accommodate with high volumes at physiological pressures (e.g. exercise), thanks to its marked preload-dependence and RV-LV interdependence 23. However, the RV is unable to face an acute increase in PVRI, and works inefficiently when confronted with PH 24. Thus, unlike the LV, the RV performance is markedly afterload-dependent. In PH patients, the use of preload reserve (Frank-Starling's mechanism) helps preserve RV function. Furthermore, according to Laplace's law, an increase in afterload can be offset by an increase in RV wall thickness (RV hypertrophy), and this normalises RV wall stress and myocardial oxygen consumption. There is a linear correlation between RV mass and free wall area, indicating that an increase in afterload causes RV enlargement with both dilatation and hypertrophy 25. Chronic cor pulmonale is defined as dilatation and hypertrophy of the RV secondary to PH caused by diseases of the pulmonary parenchyma and/or pulmonary vascular system between the origins of the main PA and the entry of the pulmonary veins into the left atrium 26, 27.

Right ventricular ejection fraction (RVEF) is much more sensitive to changes in ventricular afterload than LV ejection fraction 28. The RVEF has been found to augment with exercise in healthy subjects given the decreased PVRI and increased contractility of the RV free wall 29, 30. In patients with PH, the decreases in RVEF are not indicative of decreased RV contractility but mainly reflect increased afterload 31. In primary PH studied at rest, cardiac function is characterised by RV systolic overload due to PH and diastolic overload with tricuspid regurgitation (TR), whereas the LV is subject to diastolic underloading and reduced compliance 32, 33. With exercise, RV systolic performance further declines with a reduction in stroke volume and ejection fraction, and consequently heart rate becomes the mechanism by which cardiac output increases.

Right ventricular failure

RV failure may be related to the natural evolution of the disease or to acute exacerbation of PH (e.g. following acute hypoxaemia). In both primary PH and thromboembolic PH, the main cause of death is RV failure 1–8. COPD is the most common pulmonary disease that culminates in RV dysfunction 24, 34. An autopsy study suggests that cor pulmonale is found in 40% COPD patients 35. Cor pulmonale is mainly observed in “blue bloaters” who develop a low cardiac output profile and severe hypoxaemia and erythrocytosis resulting in PH. Conversely, “pink puffers” exhibit less severe hypoxaemia, decreased cardiac output and increased arteriovenous oxygen content difference, and cor pulmonale is less likely to develop despite increased PVRI 24. In a number of COPD patients, peripheral oedema may not always be explained by RV failure, but rather relate to hypercapnia, acidosis and increased sympathetic and renin-angiotensin-aldosterone activities, and the related changes in renal haemodynamics and redistribution of body water 36.

When pulmonary arterial impedance is chronically increased, RV function and output are preserved thanks to RV dilation and hypertrophy, increased inotropy and faster heart rate. Both increased impedance and increases in RV size and annulus diameter lead to TR, which further compromise RV function. Decreased RV function is observed when the limits of cardiac reserves are reached or when significant RV-LV interdependence or chronic RV ischaemia are present (fig. 2⇓). The mechanisms for RV ischaemia include increased pressure work leading to increased myocardial oxygen consumption; reduced systemic pressure leading to decreases in the coronary perfusion driving pressure; and higher sensitivity to hypoxia and reduced vasodilatory reserve of the hypertrophied RV 23, 34. In patients with end-stage RV failure, the maintenance of a sufficiently high systemic pressure is of major importance to prevent RV ischaemia, and this may require high doses of vasoconstrictive agents (e.g. noradrenaline).

Numerous factors may precipitate RV failure 23, 24, 34. Factors related to preload include impaired venous return (e.g. mechanical ventilation), decreased preload reserve (e.g. hypovolaemia) and mechanical limitation (e.g. effusive pericarditis). Factors related to afterload include acute load increases (although pulmonary embolism-related RV failure is observed only for acute systolic pressure >60–70 mmHg), increased pulsatile load (e.g. proximal obstruction with markedly increased pulse wave reflections), and factors related to transpulmonary pressure, lung compliance and mechanical ventilation. Major tricuspid valve insufficiency may precipitate RV failure, and the anatomical and functional status of the tricuspid valve has a key role in preserved RV function. Other factors potentially precipitating RV failure are decreased inotropy, RV ischaemia, REM sleep and obstructive sleep apnoea syndrome. The deleterious role of hypoxia and neurohormonal factors has also been demonstrated.

Therapeutic implications

The main goals in the prevention and treatment of RV failure include: 1) reducing RV afterload by decreasing PVRI; 2) limiting pulmonary vasoconstriction (e.g. prevention of hypoxaemia); 3) optimising RV preload; and 4) preserving coronary perfusion through maintenance of systemic blood pressure. Other strategies may depend upon the cause of PH (e.g. recipient selection for preventing acute RV failure following cardiac transplantation) 37.

Pulmonary artery pressure

Cardiovascular textbooks indicate a 9–19 mmHg range for normal mean PAP at sea level 47, 48. In resting healthy subjects aged 6–45 yrs, mean PAP remains constant at 14±3 mmHg 2, 4. Although mean PAP increases slightly to 16±3 mmHg between 60–83 yrs, it is generally agreed that mean PAP is little influenced by age in healthy subjects 2, 4, 49, 50. There is no clear consensus as to what resting mean PAP level constitutes PH. Pressure values >18 mmHg 4, 51, 20 mmHg 5, 6, 38, 52 and 25 mmHg 5 have been proposed, the latter cut-off value being the one most often used in recent clinical trials. Other definitions of PH include a systolic PAP >30 mmHg at rest and a mean PAP >30 mmHg on exercise 5, 51, 52. The cause and type of PH also influence the haemodynamic profile, which is the major determinant of PH prognosis. For example, PH of mild (mean PAP=25–35 mmHg) and moderate (mean PAP=35–45 mmHg) intensity are more common in PH secondary to cardiac diseases or COPD, whereas severe PH (mean PAP ≥45 mmHg) is generally found in primary PH and in chronic pulmonary thromboembolism 1–8.

Pulmonary occlusion pressure, pulmonary vascular resistance index, total peripheral resistance index and pulmonary hypertension classification

Downstream pressure is approximated on the basis of the pulmonary occlusion pressure (P_A,op). P_A,op provides a more accurate estimate of left atrial pressure than mean PA capillary wedge pressure (P_A,wp), which overestimates left atrial pressure when pulmonary venous resistance is increased 53, 54. In normoxic healthy subjects, the difference between mean PAP and P_A,wp is 5–9 mmHg 4. The P_A,op value allows diagnosis of the type of PH, namely pulmonary venous (P_A,op ≥15 mmHg) or arterial (P_A,op <15 mmHg) hypertension. Measurements of mean PAP, P_A,op and cardiac index allow the calculation of PVRI according to standard formula. Unlike mean PAP, PVRI increases significantly with age 4, 49, 50. The upper limit for PVRI (in mmHg·L⁻¹·min·m²) in normal subjects increases from ∼2.8 (6–10 yrs) to 3.2 (32–45 yrs) to 4.6 (60–83 yrs) 2, 4. There is no consensus as to what resting PVRI level constitutes PH, and threshold values ranging from 3–6 mmHg·L⁻¹·min·m² have been used previously. Although it has often been reported that an accurate P_A,op may be unobtainable in patients with severe PH, this is rarely the case in experienced teams. Total peripheral resistance (TPR; equal to mean PAP/cardiac output) and total peripheral resistance index are calculated in cases where reliable P_A,op recordings are not obtained.

A PVR threshold value may be included in the diagnostic criteria of some forms of PH, e.g. portopulmonary hypertension (POPH) 55–57. POPH can be defined as a pulmonary arterial hypertension associated with portal hypertension with or without hepatic disease. A moderate increase in mean PAP (25–35 mmHg) is observed in up to 20% of patients with cirrhosis and portal hypertension. This passive increase in mean PAP, which relates to increased cardiac output and/or blood volume, is associated with near normal PVRI (i.e. minimum pulmonary vascular remodelling) and normal or increased P_A,op (pulmonary venous hypertension). A severe pulmonary arterial hypertension with extensive pulmonary vascular remodelling and elevated PVR is more rarely observed in patients with portal hypertension. This latter condition represents the entity of POPH and is associated with poor outcome. To distinguish between these two forms of PH, the following POPH criteria have been proposed 55, 56 :1) mean PAP >25 mmHg at rest; 2) PVR >120 dyne·s·cm⁻⁵; and 3) P_A,op <15 mmHg.

Cardiac index, evaluation of intracardiac shunt, venous oxygen saturation, uric acid

The determination of cardiac index is essential for the diagnostic work-up and follow-up programme in patients with PH. The thermodilution technique is used in the vast majority of PH centres. In patients with severe PH, a recent study indicates that thermodilution was equally accurate to the Fick method (the gold standard technique) over a broad range of cardiac output values (1.7–7.8 L·min⁻¹) 58. Although conflicting results have been published, the agreement between the two methods was not affected by the severity of tricuspid regurgitation 58. In most cases of PH, the accuracy of the thermodilution technique is acceptable. However, in case of intracardiac shunt, as in the patients with patent foramen ovale, the thermodilution technique is inappropriate. Only the Fick method, with measurement of pulmonary venous O₂ content is able to measure pulmonary blood flow.

Oximetry (i.e. diagnostic saturation run and computation of the pulmonary to systemic flow ratio) is required to exclude intracardiac shunting. Further investigations (dye dilution curves and angiography) may also be required although they do not intrinsically belong to the standard evaluation of PH. The high pulmonary blood flow-related hyperkinetic component of PH is often associated with vasoconstriction (reactive component) and arterial remodelling (obstructive/obliterative component). The prevalence of patent foramen ovale in PH is similar to the normal population, and this finding provides no detectable influence on resting haemodynamics or exercise tolerance in patients with PH 59.

Assuming normal haemoglobin content and normal S_a,O2, the fall in mixed venous oxygen saturation (S_v,O2) is a reliable indicator of the fall in cardiac index, as predicted from the Fick equation. It is important to compare S_v,O2 values between subjects at similar, optimal S_a,O2, and this may require supplemental oxygen. Decreased S_v,O2 has been found to be a strong predictor of poor outcome in some studies 7, 60 while others do not document such a link 40. PH patients with an S_v,O2 <60% are more likely to respond to prostaglandin treatment 61. An increase in S_v,O2 under treatment may indicate an improved cardiac index and a decreased RV afterload both during vasodilator challenge 62 and long-term therapy 63.

Tissue hypoperfusion and hypoxia resulting from reduced cardiac output induce both overproduction and impaired excretion of uric acid, leading to increased serum levels of uric acid in patients with severe primary PH 64, 65. Serum uric acid negatively correlates with cardiac output and positively correlates with TPR 64.

Right heart function

In the setting of PH, decreased cardiac index and elevated mean RAP and RV end-diastolic pressure indicate RV dysfunction. Although threshold values may depend upon each laboratory and upon the type of catheter used, mean RAP <6 mmHg and RV end-diastolic pressure <8 mmHg are generally considered normal values 47, 48. A mean RAP ≥8 mmHg and RV end-diastolic pressure ≥12 mmHg are abnormally high values highly suggestive of RV failure. The decreased compliance of the hypertrophied RV also contributes to increase RV end-diastolic pressure. In the case of typical TR, RAP rises throughout RV systole, with pressure regurgitation waves. In early stages of PH, the RV contractility is increased and the peak rate of RV pressure rise may be increased by up to 10 times the normal value 66. Analysis of RV waveform in more severe forms of the disease reveals RV systolic dysfunction with depressed upstroke and delayed relaxation.

Pulmonary arterial hypertension

Isolated arterial hypertension is characterised by a P_A,op <15 mmHg and a diastolic PAP minus P_A,op gradient >10 mmHg. Major increases in PAP and PVRI are found in thromboembolic and primary PH. Vascular obstruction/obliteration is typically observed in thromboembolic 67 and primary PH 8, 68. In chronic pulmonary thromboembolism, the thrombi follow an aberrant path of organisation and recanalisation, leaving endothelialised residua that narrow and stiffen major (main, lobar or segmental) PAs, i.e. elastic arteries. In primary PH, pulmonary obstructive vasculopathy involves distal, medium-to-small sized muscular (resistive) arteries. In the latter case, both a fixed obstructive component and a reactive component (increased vascular tone) contribute to PH. In COPD patients, hypoxia-induced vasoconstriction plays a major role in producing PH, and destruction of the pulmonary vascular bed is also involved. The increased resistance to flow through the pulmonary vascular bed may be due to several other diseases (table 1⇓), including the Eisenmenger syndrome. In this syndrome, the left-to-right shunts, mainly post-tricuspid shunts, lead to irreversible PH at systemic level which subsequently cause the shunt to be reversed 6.

Pulmonary venous hypertension

Increased resistance to pulmonary blood flow downstream leads to passive PH, with elevation of PAP but no significant elevation of PVRI. In patients with increased left heart filling pressure, pulmonary venous hypertension is characterised by a P_A,op ≥15 mmHg. Although the diastolic PAP minus P_A,op gradient is <10 mmHg, diastolic PAP must not be used as a surrogate of downstream pressure. In a recent study, diastolic PAP was >5 mmHg higher than P_A,wp in 21% and 57% of cardiac patients with mean PAP of 21–40 mmHg and >40 mmHg, respectively 69. Acute increases in pulmonary flow (e.g. exercise) worsens pulmonary venous hypertension when blood flow through the left atria or LV filling are impeded. Additional reactive vasoconstriction and vascular remodelling may occur, especially in cases where left heart filling pressure is markedly and durably increased. A decrease in left heart filling pressure usually reverses both the passive and vasoreactive components of PH and therefore tends to normalise PAP.

A P_A,op >15 mmHg may be observed together with a diastolic PAP minus P_A,op gradient >10 mmHg. In such combined venous and arterial PH, PAP does not normalise following the alleviation of the high downstream pressure, given the fixed arterial component of PH (e.g. evolved mitral stenosis). In these patients, high levels of PVRI are sometimes referred to as the “second stenosis” and may have both deleterious effects (by increasing the strain put on the RV), and protective effects (by limiting RV output and thus the occurrence of episodes of pulmonary oedema) 23. Finally, some catheterisation findings may be evocative of pulmonary veno-occlusive disease (e.g. the failure to obtain a P_A,wp tracing) 70.

Vasodilator testing

The potentially treatable vasoconstrictive component may be found in numerous forms of PH and thus must be sought during catheterisation by acute reversibility testing 71. The vasodilator challenge response is the first step of the algorithm currently used for the management of primary PH 8. Vasodilator challenge must be performed by an experienced team and requires intensive care equipment. Indeed, drug-induced systemic hypotension (e.g. with prostacyclin) may reduce RV coronary blood flow and result in acute RV ischaemia. There is no advantage in testing more than one drug. The most specific short-acting vasodilator for pulmonary vessels is inhaled nitric oxide (iNO), the maximal effect being obtained after 5–10 min at the dose of 10 parts per million (ppm). Since iNO has no systemic effects, the resulting drop in PVRI reliably indicates the amount of pulmonary vasodilatation in reserve 72. Other vasodilators have been proposed, including i.v. iloprost (epoprostenol sodium), a prostacyclin analogue with fewer systemic effects, and i.v. adenosine.

There is no clear consensus as to what criteria may indicate beneficial response to vasodilator challenge. Because mean PAP and PVRI may vary spontaneously by ≤20–25%, measurements must be performed after stabilisation, and only changes greater than spontaneous haemodynamic variability must be taken into account. A significant response to vasodilator testing is currently defined as a drop in mean PAP or PVRI by >20% 60, 73–76. Other definitions include a >30% threshold 62, 77, 78 or a reduction in mean PAP >10 mmHg 1, or an increase in cardiac index by ≥30% 61.

In most studies on primary PH, the proportion of responders to vasodilators ranges from 12–40% depending on the patient's age, RV function, the cause of hypertension and the protocol used 79. The absence of pulmonary vasodilatory reserve has been associated with a poor outcome 60, 71, 80–82. A decrease in PVRI by ≥20% predicts good response to long-term vasodilator therapy with calcium channel blockers 83. Barst et al 74 found it safer to prescribe long-term chronic calcium channel blocker therapy only in patients exhibiting a decrease in PVR of ≥50%.

In PH patients evaluated for cardiac transplantation, some authors have defined irreversible (or fixed) PH as those patients who, after provocative vasodilator therapy, have either a PVR ≥4 mmHg·L⁻¹·min, a PVRI ≥6 mmHg·L⁻¹·min·m², a systolic PAP ≥60 mmHg or a tranpulmonary pressure gradient ≥15 mmHg 37. Patients with reversible (or reactive) haemodynamics have better prognosis than patients with irreversible PH. Serial aggressive sequences of long-term inotropic support followed by provocative vasodilatory testings may convert fixed PH into the reactive category and condition the pulmonary vasculature into a maximally dilated state (“vasodilator conditioning”) 37.

Pulmonary resistance partitioning

More insight into the pulmonary circulation can be obtained by estimating the effective pulmonary capillary pressure, in order to discriminate between increases in arterial resistance and in venous resistance. Clinically, capillary pressure is computed by analysing the PAP decay after single arterial occlusion by an intravascular balloon 53, 84, 85 (fig. 3⇓). Recent studies have confirmed a predominant increase in arterial resistance in thrombo-embolic disease, a significant increase in arterial but also in venous resistance in primary PH, and the expected major increase in venous resistance in pulmonary veno-occlusive hypertension 38, 53.

Pulmonary arterial pressure/flow relationship

The transposition of Poiseuille's law to pulmonary circulation rests on the main assumptions that the PAP-flow relationship is linear, and that the PAP is equal to downstream pressure at zero flow. The first assumption appears valid in both normal and diseased pulmonary circulation over physiological ranges of flows 38, 53, 86, 87 (fig. 4⇓). Initial work by Janicki et al. 38 suggested that the second assumption was also valid in normal subjects and in the majority of PH patients. However, in one fourth of the PH patients in this study, the average closure pressure exceeded P_A,wp 38. More recent experiments have shown that the pressure axis intercept of the mean PA pressure/flow relationship (i.e. the closing pressure) is higher than P_A,op or left atrial pressure in primary PH 53, 88 (fig. 3⇓). In such conditions, a change in PVRI can be interpreted unequivocally only when cardiac output remains unchanged, or when driving pressure and cardiac output change in opposite direction.

Increased driving pressure with decreased pulmonary flow attests to vasoconstriction, while decreased driving pressure with increased pulmonary flow attests to vasodilatation (fig. 5⇓). Other haemodynamic patterns are not easy to interpret 86 (fig. 6⇓). For example, increases in PAP and decreases in PVRI (e.g. after drug administration) can result from increases in cardiac output without any modification in pulmonary vascular tone. Therefore, the determination of multipoint mean PAP-flow plots provides more accurate insight into the nature of increased PVRI 38, 53, 86–89.

Thus, in PH, both closing pressure and the capillary pressure determined from the analysis of the PAP decay curve after balloon occlusion are increased 86, 88. The main site of increased resistance may be at the periphery of the pulmonary arterial tree, and this is possibly a cause of increased closing pressure 86. Greater venous involvement than previously assumed has also been suggested recently 88. In patients with PH, recent data using two-point curves have suggested that exercise is associated with some pulmonary vasoconstrictive effect, while dobutamine at doses ≤10 mcg·kg⁻¹·min⁻¹ allows cardiac output to increase without affecting pulmonary vascular tone 86, 89. Finally, in patients with primary PH, mean PAP-flow plots have indicated that the improvement in exercise tolerance seen after 6 weeks of prostacyclin therapy may be ascribable to a decrease in incremental PVR during exercise through a vasodilator effect 87.

Pulmonary vascular impedance

The pulmonary vascular impedance (PVZ) is a frequency-dependent function encompassing information about resistive, capacitive and inertial components of vascular hydraulic load, as well as the extent of pulse wave reflection 10, 43. Few studies have been performed in patients, due to a requirement for simultaneous recordings of high-fidelity pressure and flow signals. Results are reported as spectra of PVZ amplitude and PVZ phase versus frequency. PVZ spectra in patients with PH demonstrate increases of both the steady component (increased resistance) and oscillatory component (elevated PA characteristic impedance and increased pulse wave reflection) of hydraulic load 43. Decomposition of pressure and flow into forward and backward components show an increased amplitude and early arrival of the reflected pressure wave, contributing to unfavourable loading of the still-ejecting RV 43. A preliminary report has suggested that high characteristic impedance could predict better than PVR the occurrence of severe right ventricular failure after cardiac transplantation 90.

Pressure pulsatility

For a given stroke volume, the TPG and thus mean PAP reflect the steady component of potential energy, i.e. the heat dissipated in the microcirculation 10. Mean PAP and PVRI are useful measurements reflecting the opposition to the mean component of flow and the functional status of the distal (resistive) pulmonary vasculature. However, they do not take into account the pulsatile component of hydraulic load, which is mainly related to the proximal (elastic) PAs. The precise measurements of instantaneous pulsatile pressure and wave reflections require the use of expensive, high-fidelity pressure catheters instead of the commonly used fluid-filled catheters.

In the normal pulmonary vasculature, proximal PAs are highly distendable. The normal pulmonary circulation exhibits very little wave reflection, and the composite reflected pressure wave arrives during the early diastolic period 10. Earlier, increased wave reflection may impede the ongoing RV ejection and thus impair right ventricular-arterial coupling in both primary PH and chronic pulmonary thromboembolism. In patients with primary PH, the extent of wave reflection is large at rest and the composite reflected pressure wave arrives during the mid-portion of RV ejection 43. Patients with chronic pulmonary thromboembolism have increased anticipated wave reflection compared with patients with primary PH, probably because the functional reflection site is more proximal in the former patient group 91, 92. In PH, increased PVR, decreased PA compliance and complex changes in the pulse wave reflection site lead to increased pulsatile load, as attested to by increased PA pulse pressure (PP; equal to systolic minus diastolic PAP) and increased wave reflection 43, 91. Some authors have suggested that patients with chronic pulmonary thromboembolism have a higher PP than patients with primary PH 93. Others have shown that no cut-off PP value could help distinguish these two populations when the mean PAP level was accounted for 91, 94. In hyperkinetic/high-flow states, increased pulsatile load may be explained by increased stroke volume.

Systolic pulmonary artery pressure: tricuspid regurgitation peak jet velocity

Continuous-wave Doppler transthoracic echocardiography can be used in subjects with TR to determine the RV systolic pressure minus RAP gradient with the simplified Bernouilli equation: where Vtr is the TR peak jet velocity. Assuming that systolic PAP equates RV systolic pressure in the absence of pulmonic stenosis and outflow tract obstruction the following is obtained: Patients must be in sinus rhythm to ensure stable Vtr recordings. According to the WHO recommendations, mild PH is defined as a resting systolic PAP of 40–50 mmHg, which corresponds to a tricuspid regurgitant velocity of 3.0–3.5 m·s⁻¹ 1, assuming a fixed 5 mmHg RAP. In a recent large-scale, echocardiographic study of 3,790 healthy subjects (2,432 females) from 1–89 yrs of age, a 37.9 mmHg 95% upper limit was documented for systolic PAP at rest, which corresponds to a Vtr of 2.64 m·s⁻¹ assuming a fixed 10 mmHg RAP 95. This study supports the use of age- and BMI-corrected values in establishing the normal pressure range, since a systolic PAP >40 mmHg was found in 6% of healthy subjects >50 yrs of age and 5% of healthy subjects with BMI >30 kg·m⁻² 95. Male sex and LV septal and posterior wall thickness are also predictive of systolic PAP 95. The level of physical training is also important to consider. In resting healthy males <23 yrs, the 95% upper limit of mean Vtr is 1.93 and 2.41 m·s⁻¹ in nonathletes (n=14) and athletes (n=26), respectively 96.

Tricuspid regurgitation is common in healthy subjects, and increases in prevalence and severity as PAP increases. It is widely accepted that continuous wave Doppler echocardiography is a valuable tool for diagnosing PH, with interobserver variability of <5%. However, some investigators were unable to document Vtr in 40–70% of patients with proven PH, mainly in cases of moderately elevated PAP or because of poor acoustic windows 97. Finally, as Doppler estimates of systolic PAP are operator-dependent, inter- and intra-observer reproducibility must be checked.

Doppler-derived systolic PAP values are strongly correlated with pressure measured with catheterisation in numerous studies 98, 99. In other studies, Doppler values markedly underestimated systolic PAP 45, 96, 100, 101. Systolic PAP values measured by catheterisation and estimated by Doppler were poorly related during dynamic changes 45. Catheterisation therefore remains the gold standard in establishing the diagnosis of PH, while Doppler allows a serial, noninvasive follow-up of PA haemodynamics.

Diastolic pulmonary artery pressure

The simplified Bernouilli equation also allows determination of diastolic PAP in normal subjects and subjects with PH: where VEDipis the end-diastolic velocity of the pulmonic insufficiency jet and RAP is a surrogate for the RV end-diastolic pressure 102. Satisfactory pulmonary flow recordings are obtained in a majority of normal subjects and in >90% of subjects with PH.

Mean pulmonary artery pressure

Mean PAP can be approximated either by measuring the early-diastolic velocity of the pulmonic insufficiency jet (VeDip) or by measuring time intervals. The first method 102 is based on the following equation: The correlation with catheterisation is weak, although it may improve in cases where only patients with heart rates of 60–100 beats·min⁻¹ are considered 103.

The second method relies on the observation that the pulmonary flow velocity signal is dome shaped in normal subjects while it is more triangular, with an earlier peak flow, in patients with PH. In these patients, time-to-peak flow (i.e. acceleration time (AcT)) is decreased and Kitakabe et al. 104 have documented a high correlation (r=0.88) between AcT and log₁₀ mean PAP. In some patients, premature deceleration of pulmonary systolic flow may result in mid-systolic notching, especially in severe PH or when there is an early return of the reflected wave (e.g. proximal obstruction) 105. Consistently, patients with chronic pulmonary thromboembolism (i.e. proximal obstruction) have a shorter time-to-peak pressure than patients with primary PH (distal obstruction) 91, 92. Finally, AcT may also be prolonged in patients with decreased RV function and/or increased preload. Several studies have confirmed the predictive value of a decreased AcT for diagnosing PH, with a 90–100 ms threshold 106, 107. The pre-ejection time /AcT ratio is also increased and correlates with PAP in patients with PH.

Right atrial pressure

The enlargement of the right atria reflects increased RAP. A precise estimation of RAP is especially needed in PH patients because increased RAP reflects RV dysfunction and is a strong predictor of outcome, and also because Doppler estimates of PAP depend upon the RAP value. This may be of critical importance for Doppler-derived diastolic PAP since RAP is a much higher percentage of diastolic than systolic PAP. In practice, the RAP may be arbitrarily fixed 102, clinically evaluated by the jugular venous pulse 2, or echocardiographically measured by the inferior vena cava collapse with spontaneous respiration 108. Kircher et al. 108 have suggested that per cent collapse ≥50% or <50% reflect RAP values of <10 mmHg or ≥10 mmHg, respectively. Given that RAP increases with PAP, a fixed RAP value tends to overestimate or underestimate low and high PAP values, respectively 109. Recently, a new index of RV filling pressure has been described, namely the tricuspid E/Ea ratio, where E is Doppler tricuspid inflow velocity and Ea is the tissue-Doppler early diastolic velocity of the tricuspid annulus 110. Mean RAP was linearly related to E/Ea, and E/Ea ratio >6 has a sensitivity of 79% and a specificity of 73% for mean RAP ≥10 mmHg 110. Finally, Doppler-derived hepatic vein systolic filling and filling fraction are negatively related to mean RAP, but the specificity and sensitivity of these indices over a large RAP range are still under study.

Right ventricular function

Given the complex architecture of the RV, indices quantifying RV diameter, area or volume are not always reproducible and are highly operator- and model-dependent. Following pressure overload, RV enlargement is best evidenced on the apical four-chamber view, and is often associated with a rounded rather than triangular (physiological) shape. The RV fractional area change in the apical four-chamber view is decreased. An RV/LV area ratio ranging from 0.6–1 indicates mild RV dilation, while severe RV dilation is associated with a ratio >1 111. The RV function may be indirectly estimated by using the global RV function index and the velocity of the tricuspid annulus plane. The Doppler global RV function index is defined as the sum of RV isovolumic contraction and relaxation times divided by ejection time 112. The index is calculated by dividing the TR jet duration minus pulmonary jet duration difference by pulmonary jet duration. The global RV function index is a useful, independent predictor of adverse outcome in patients with PH, and patients with an index <0.83 have a more favourable survivorship 113. Doppler tissue imaging can be used to calculate the velocity of the tricuspid annulus plane which is positively correlated with RV systolic function 114.

Interventricular septum and left ventricle function

PH results in RV systolic overload characterised by leftward septal shift at end-systole and early diastole, the septum behaving has an intrinsic part of the RV 32, 111. RV geometric changes influence LV function, and prolonged LV isometric relaxation, reduction in LV early diastolic filling and paradoxic septal motion in systole are also observed 32, 111. Deceleration time of early transmitral flow is decreased and negatively correlates with LV deformity index and PVR 115. Interventricular septum kinetics are best evidenced on the parasternal short-axis view. Significant tricuspid insufficiency leads to RV diastolic overload with the RV filling at the expense of the LV, together with leftward septal shift at end-diastole, unchanged LV isometric relaxation and a reduction in the contribution of left atrial systole to LV diastolic filling 32. In PH, the underfilling of the LV may be due to reduced pulmonary venous return, left-to-right interatrial shunting and/or interventricular septal shift due to RV overload. In patients with chronic thromboembolic PH, the impairment of LV filling is potentially reversible, as attested to by the increases in the transmitral early to atrial filling velocity ratio (E/A) quickly after pulmonary thromboendarteriectomy, with postoperative E/A >1.5 correlating with the absence of severe, residual PH 116.

Right ventricle hypertrophy and other findings

RV free wall thickness is measured at end-diastole using the subcostal view, and normal values are ≤4 mm. Marked (>10 mm), heterogeneous RV hypertrophy is currently found in chronic cor pulmonale, while mild RV hypertrophy (6–8 mm) or subnormal values are found in acute cor pulmonale 111. Echocardiography may reveal effusive pericarditis. Transoesophageal echocardiography and contrast echocardiography may be especially valuable for diagnosing a patent foramen ovale. Colour Doppler indices relying on the jet length into the RA allow the severity of TR to be quantified. Measurements of PA diameter (normal value <2.35 cm) belong to the systematic echocardiographic analysis of PH patients. Finally, echocardiography may diagnose an underlying cardiac disease responsible for secondary PH.

In primary PH, higher mean PAP are associated with more pronounced abnormalities in the septal curvature and depressed RV contractile function 45. Patients with poor performance on the 6-min walk test exhibited a greater degree of RV dilation, more pronounced septal displacement in diastole, larger pericardial effusions, and more severe TR 45. Patients with Class IV symptoms are more likely to have pericardial effusions.

Imaging

Pulmonary magnetic resonance angiography (PMRA) is a promising imaging tool for the identification of patients with PH. In a recent study, gadolinium-enhanced PMRA showed that a right PA diameter >28 mm, along with the measurement of diameter of the proximal-to-distal tapering of the PAs and absence of intravascular filling defects, allow the diagnosis of PH with a high sensitivity and high negative predictive value 117. These results confirm earlier studies with computed tomography scanning 118 and magnetic resonance (MR) angiograms 119, 120. The diameter of the descending right PA on chest radiography has been used for a long time to screen for the prevalence of PH in populations 121, and others indices are currently used, namely the hilar width percent and hilar/thoracic index percent.

Classic pulmonary angiography and ventilation/perfusion scans help diagnose chronic pulmonary thromboemboli. In primary PH, quantitative analysis of the irregularities of lung perfusion scans could be related to the severity of the disease 122. Some authors have used thallium scintigraphy and cine MR imaging to study RV pressure overload in PH patients. Perfusion abnormalities suggestive of RV ischaemia are associated with RV dysfunction 123. Finally, first-pass radionuclide angiography and gated stress myocardial perfusion imaging 124 allow the routine evaluation of RVEF, bearing in mind that RVEF reflects afterload rather than RV contractility.

Electrocardiogram criteria

Although the echocardiogram has decreased dependency on the electrocardiogram (ECG) and catheterisation to diagnose PH and monitor RV function, certain ECG parameters nevertheless remain important in patient management. Those reflecting right atrial and ventricular overload have been associated with decreased survival 125. The p-wave amplitude in lead II, a p ≥0.25 mV in lead II, the presence of a qR pattern in V1, and WHO criteria for RV hypertrophy are highly predictive of an adverse outcome 125.

Natriuretic peptides

In patients with PH, a high level of plasma atrial natriuretic peptide (ANP) and brain natriuretic peptide (BNP) negatively correlates with parameters of RV function 126–128. Following prostacyclin treatment, the decreases in both ANP and BNP parallel pulmonary vasodilatation and haemodynamic improvement 127, 128. During follow-up, a further increase in plasma BNP is an independent predictor of mortality. Repeated plasma BNP measurement may well offer a noninvasive tool for identifying patients refractory to treatments 127, although it remains to be established how the predictive value of BNP compares to that of other indices, e.g. the walking distance.

Primary pulmonary hypertension

Several studies have focused on the haemodynamic predictors of adverse outcomes in primary PH. Although the heterogeneity of the clinical course of primary PH has been acknowledged, the prognosis appears mainly related to both RV function (static parameters) and to the amount of pulmonary vasodilatation in reserve (dynamic parameter). Adverse prognosis correlates with the NYHA functional classification 40, 61, the distance walked during the 6-min walk test 18, 39, the mean RAP 40, 60, 65, 129, 130, cardiac or stroke volume index 40, 60, 61, 129–131 and systemic arterial and mixed S_v,O2 60, 61, 64. Prognosis also relates to mean PAP and PVRI (or PVR) in most studies, although conflicting results have been reported in some studies. The lack of pulmonary vasodilatory reserve is also associated with poor prognosis 60, 71, 80–82, 132. Echocardiographic predictors of adverse outcomes in primary PH include pericardial effusion, right atrial enlargement, septal displacement and an increased Doppler global RV function index 113, 133, 134. Numerous other prognostic factors have been described, including the forced vital capacity 60, pulmonary S_a,O2 132, oxygen saturation at peak distance during the 6-min walk test 18, anticoagulant therapy 132 and increased serum uric acid 64, 65. Overall, determining the walking distance, stroke volume index, mean RAP, pulmonary vasodilatory reserve, serum uric acid and echocardiographic characteristics should help in predicting primary PH clinical course.

The results of two recent studies deserve particular attention. Sitbon et al. 135 have tested the predictive value of the responses to a 3-month continuous infusion of epoprostenol in primary PH patients in NYHA functional class III-IV. In the Cox model of multivariate analysis including both baseline and 3-month clinical and haemodynamic variables, the persistence of NYHA functional class III–IV and the absence of fall in PVRI >30% relative to baseline at 3 months, and a history of right heart failure, were associated with a poor 5-yr survival. This suggests that haemodynamic responses to 3-month prostaglandin I₂ infusion may help diagnose a subset of primary PH in whom lung transplantation could be proposed 135. Furthermore, Wensel et al. 65 have performed symptom-limited exercise testing, pulmonary function test, cardiac catheterisation and uric acid measurements in primary PH. They reported that low peak V′_O2 and low peak exercise systolic arterial pressure were the strongest predictors of impaired survival. The predictive value of the V′_E/V′_CO2 slope was not tested due to the high percentage of patients with a patent foramen ovale 65. Patients with two risk factors, namely peak V′_O2 ≤10.4 mL·kg⁻¹·min⁻¹ and peak exercise systolic arterial pressure ≤120 mmHg had poor survival rates at 12 months (23%), whereas patients with one or none of these risk factors had 79% and 97% survival rates, respectively 65. The prognostic consequence of a low systemic pressure could be explained by greater sensitivity and more severe dysfunction of the RV as a result of ischaemia 123. Finally, amongst the new therapeutic approaches for primary PH, several trials, either published or in progress, have suggested benefits for a patient assigned to a specific treatment group, and further studies are needed to confirm and summarise new therapeutic guidelines.

Chronic obstructive pulmonary disease

In COPD patients, factors related to survival include forced expiratory volume in one minute, P_a,O2 and mean PAP 24. In patients receiving long-term oxygen therapy, mean PAP is the best prognostic factor 136. Finally, the higher the mean PAP, the higher the risk of hospitalisation for acute exacerbation 137.

Heart failure

As far as right heart failure is concerned, there is a complex interplay between the severity and the cause of PH. Despite a tendency towards higher PAP, patients with Eisenmenger syndrome have lower mean RAP and greater systemic cardiac output than those with primary PH 138, and the preservation of biventricular function could explain the more favourable outcome in the former group 138, 139. The degree of hypoxaemia at baseline is a strong predictor of survival in the Eisenmenger syndrome.

The RVEF is an independent prognostic factor in patients with PH complicating chronic heart failure 140. In patients with advanced heart failure, RVEF ≥0.35 at rest and during exercise is a more powerful predictor of survival than V′_O2 46. A few recent studies have suggested that hyperventilatory response to exercise, as measured by increased V′_E/V′_CO2 slope, may be a prognostic factor of poor survival, but this point remains to be confirmed in large population studies. Mean PAP is particularly important in stratifying risk in cardiomyopathy caused by myocarditis 141. Patients with PH are a particularly high-risk group, probably because PH may develop more acutely, overwhelming compensatory mechanisms such as RV hypertrophy.

In PH patients evaluated for cardiac transplantation, patients with reversible (or reactive) haemodynamics following provocative vasodilatory therapy have better prognosis than patients with irreversible (or fixed) PH 37.