• Echocardiography-exercise
• pulmonary arterial hypertension (PAH)
• pulmonary vascular disease

Exercise stress tests have been used for the diagnosis of pulmonary hypertension since the introduction of the haemodynamic definition of the disease in the 1950s. Paul Wood reasoned on the pulmonary vascular resistance (PVR) equation to show that pulmonary artery pressure (PAP) can increase as a result of resistance but also because of flow and/or left atrial pressure.1 One of his patients with borderline pulmonary hypertension on mitral stenosis and severe exercise-induced pulmonary hypertension is illustrated in figure 1. Right heart catheterisation studies in normal subjects showed that the average slope of the PAP–cardiac output relationship is 1 mm Hg/l per minute in young adults and increases to 2.5 mm Hg/l per minute after six to eight decades of life, and that left atrial pressure is transmitted upstream to PAP in an approximately 1:1 ratio.2 Cardiac output-associated high PAP may thus be amplified by an increased PVR due to pulmonary arteriolar remodelling, reflecting lung disease, and/or by upstream transmission of an increased pulmonary outflow pressure, reflecting left ventricular diastolic dysfunction.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Pulmonary artery pressure (PA) and left atrial pressure (PCV) in a patient with mitral stenosis at rest and at exercise, reported by Paul Wood.1 Exercise was associated with an increase in cardiac output (not shown) and parallel increases in PA and PCV.

What are the limits of normal? Pulmonary hypertension was traditionally defined by a PAP higher than 25 mm Hg at rest and 30 mm Hg at exercise.3 Those who used this definition were well aware that the upper limit of normal of PAP at rest is 20 mm Hg1 and that PAP may increase to 40 mm Hg and sometimes more in healthy individuals at high levels of exercise.2 4 5 However, cut-off values were deliberately different from the limits of normal because of the need for a safety margin at rest, and the assumption that patients with dyspnoea and fatigue symptoms caused by pulmonary hypertension would not exercise heavily anyway. It thus came as a surprise when very recently, in 2008, an expert consensus conference held in Dana Point decided to withdraw exercise criteria from the definition of pulmonary hypertension because of insufficient certainty about the limits of normal.6 It is interesting that a few months later, exercise-induced pulmonary hypertension was characterised as a clinical entity.7

In their paper published in this issue of Heart, D'Alto and colleagues (see page 112)8 report on a high frequency of abnormal exercise pulmonary haemodynamics in 172 patients with systemic sclerosis and no pulmonary hypertension at rest compared with 88 age and gender-matched controls. High PAP responses to exercise were related to the presence of mild interstitial lung fibrosis, which was present in 25% of the patients, but also to subclinical echocardiographic evidence of left ventricular diastolic dysfunction. The authors calculated a systolic PAP from the maximum velocity of tricuspid regurgitation. This measurement has a bad reputation because of a reported high incidence of false positives and negatives for the diagnosis of pulmonary hypertension in daily clinical practice.9 However, when properly implemented by dedicated investigators, the method has actually allowed for the detection of abnormal pulmonary haemodynamics in relatives of patients with pulmonary arterial hypertension10 and a realistic description of pulmonary vascular function by pressure–flow relationships in healthy individuals.11 The distribution of PAP in the study of D'Alto et al8 was unimodal in controls and bimodal with an additional mode of increased PAP in patients with systemic sclerosis, in contrast to previous report of an additional mode of increased PAP also to be seen in a small percentage of normal subjects.10 This discrepancy is probably explained by the fact that PAP and cardiac output were measured by D'Alto et al8 immediately after rather than during the exercise test. Both PAP and cardiac output return rapidly back to normal after exercise, but this is slower for cardiac output, which probably accounted for relatively lower slopes PAP–cardiac output relationships, to values approximately half of those previously reported in normal individuals.11

D'Alto et al8 considered a systolic PAP of 48 mm Hg as an upper limit of normal at exercise. As mean PAP can be calculated as systolic PAP×0.61+2 mm Hg,12 this corresponds to a mean PAP of 30 mm Hg. Systolic PAP at exercise exceeded this upper limit of normal in 13% of the systemic sclerosis patients and none of the controls. This prevalence of pulmonary hypertension was slightly higher than the 8% reported in a large cohort of patients with systemic sclerosis screened by a tricuspid regurgitation measurement at rest and confirmed at right heart catheterisation.13 It may thus be that exercise stress echocardiography still overestimates the prevalence of pulmonary hypertension in this patient population. The finding by D'Alto et al8 that approximately half of the systemic sclerosis patients with pulmonary hypertension presented with diastolic dysfunction is in keeping with long-term follow-up invasive studies.13 The strength of non-invasive echocardiography is that it allows for sensitive tissue Doppler measurements of ventricular function, as well as the detection of early changes in pulmonary vascular function by measurements. D'Alto et al8 could have thought of adding an internal control such as the acceleration time of pulmonary flow to improve the diagnostic accuracy of exercise PAP–cardiac output relationships.14

D'Alto et al8 are well aware that their report does not provide the final word about exercise-induced pulmonary hypertension. Further work is needed to improve stress echocardiography for the diagnosis of pulmonary hypertension. A better definition of PVR by multipoint pulmonary vascular pressure–flow relationships should provide improved insight compared with isolated measurements of PAP, PAP at maximal workload, or PAP–workload relationships.2 11 14 Cardiac output is linearly correlated to workload, but differently from one patient to another as a result of unequal mechanical efficiencies of the work. Measurements of PAP, cardiac output and left atrial pressure should be performed during rather than after exercise because of variable and non-proportional rates of recovery. The type of exercise should always be specified, with preference for dynamic rather than resistive, to avoid the associated increase in systemic blood pressure and intrathoracic pressure. Limits of normal should be defined on larger reference populations of various ages, because of the known age-dependency of both PVR and left ventricular diastolic function. All this is needed for a renewed definition of exercise-induced pulmonary hypertension. There is thus still a long way to go. It is hoped that interest will be raised for it in forthcoming meetings on pulmonary hypertension, so that appropriate multicentre collaborations may be undertaken.

Recent guidelines for the diagnosis and the treatment of pulmonary hypertension express reticence about exercise stress tests and echocardiography.15 D'Alto et al8 provide a refreshing and clinically valuable counterpoint. Their work is an important step forward to an improved earlier diagnosis of the pulmonary vascular complications of systemic sclerosis.