Assessing pulmonary hypertension severity in lung disease is a key step to improving outcomes: embrace resistance and don't be pressurised to go with the flow

Identifying patients with PH-CLD with a pulmonary vascular phenotype

Severe PH occurs in a minority of patients with lung disease [1, 5]; however, estimates of its prevalence suggest it is several-fold more common than group 1 PH, pulmonary arterial hypertension (PAH) [6]. Studies that improve our understanding of PH-CLD and aid classification and patient selection for therapeutic trials are therefore urgently needed. In this edition of the European Respiratory Journal, studies by Zeder et al. [7] and Olsson et al. [8] take us a step further by providing evidence-based pulmonary haemodynamic thresholds that more accurately identify patients with severe PH who are most likely to have a pulmonary vascular phenotype [9]. In doing so they challenge current haemodynamic thresholds for severe PH and highlight that a measure of resistance and not flow best predicts mortality in both COPD and ILD. They also remind us that PH is a haemodynamic diagnosis and can occur in the absence of pulmonary vascular involvement. In lung disease in particular, elevations in intrathoracic pressure and cardiac output commonly cause mild to moderate PH without significant pulmonary vascular disease. Excluding such patients from trials of PAH therapies would seem logical as would enriching studies for populations with pulmonary vascular remodelling where such therapies would be expected to have most benefit.

Is it time to update the definition for haemodynamic severity in PH-CLD?

The 2015 European Society of Cardiology/European Respiratory pulmonary hypertension guidelines defined severe PH as a mean pulmonary arterial pressure (mPAP) ≥35 mmHg or a mPAP ≥25 mmHg with a cardiac index (CI) of <2.5 L·min⁻¹·m⁻², not explained by other causes [10]. Reflecting data that a mPAP >20 mmHg is abnormal, the 6th World Symposium [11] defined PH-CLD as mPAP 21–24 mmHg and pulmonary vascular resistance (PVR) ≥3 WU or mPAP 25–34 mmHg and severe PH-CLD as mPAP ≥35 mmHg or mPAP ≥25 mmHg with a CI <2.0 L·min⁻¹·m⁻². As both Zeder et al. [7] and Olsson et al. [8] highlight, the choice of cut-off values of haemodynamic severity, including measures of flow, namely CI, were arbitrary and based on expert opinion.

Zeder et al. [7] noted that both mPAP and PVR, but not CI, were associated with survival in 139 patients with COPD, after adjustment for age, sex and forced expiratory volume in 1 s. Using an unbiased approach, of all haemodynamic measurements, a PVR >5 WU was the best prognostic cut-off. Two equivalent mPAP thresholds were identified: one of 33 mmHg (similar to current definitions) and a second of 40 mmHg (previously identified by the ASPIRE Registry [12]). Whereas a PVR >5 WU was rare in the absence of mPAP ≥33 mmHg (4%), an mPAP ≥33 mmHg with a PVR ≤5 WU was associated with more severe pulmonary disease and left heart disease.

Olsson et al. [8] adopted a similar approach in patients with ILD. In a large, well-characterised cohort from the COMPERA registry of 449 patients with ILD, age, sex and higher PVR, but not mPAP or CI, predicted survival [8]. A PVR >5 WU was associated with a significantly worse survival compared to a PVR ≤5 WU, although a PVR >8 WU had the lowest p-value and was a better discriminator of survival. Importantly, the currently proposed haemodynamic definition of severe PH did not distinguish between survivors and non-survivors. Both these studies illustrate that it is a measure of resistance and not flow that is key in defining disease severity in PH-CLD. Figure 1 incorporates the findings of the studies by Zeder et al. [7] and Olssen et al. [8] in a proposed new haemodynamic definition for PH and severe PH in CLD, and how it could be used to enrich PAH therapy trials for patients most likely to have a pulmonary vascular phenotype and a circulatory rather than ventilatory limit to exercise.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1

Evidence based haemodynamic definition for pulmonary hypertension (PH) severity in lung disease to enrich for a pulmonary vascular phenotype in pulmonary arterial hypertension (PAH) therapy trials. D_LCO: diffusing capacity of the lung for carbon monoxide; ABG: arterial blood gas; NT-proBNP: N-terminal-pro-B-type-natriuretic peptide; CTPA: computed tomography pulmonary angiography; mPAP: mean pulmonary arterial pressure; PVR: pulmonary vascular resistance; RCT: randomised controlled trial.

PAH therapies in ILD

A number of randomised controlled trials of PAH therapies in patients with ILD and COPD have been performed (table 1). The endothelin receptor antagonists bosentan, ambrisentan and macitentan were found to have no effect on idiopathic pulmonary fibrosis (IPF) progression in patients at low risk of having PH and ambrisentan was associated with worse outcomes [13–15]. The BPHIT trial subsequently studied the effects of bosentan in patients with IPF or fibrotic idiopathic non-specific interstitial pneumonia and significant PH (mPAP 37 mmHg and PVR 7.4 WU) and observed no significant effect on pulmonary haemodynamics, exercise capacity or quality of life [16]. A post hoc analysis of patients from the STEP-IPF trial suggested a stabilisation of 6-min walk distance (6MWD) and improvement in quality of life in those patients with right ventricular (RV) dysfunction who received sildenafil [17]. Two subsequent studies of sildenafil in patients with IPF receiving background antifibrotic therapy were therefore performed but failed to reach their primary endpoints [18, 19]. The INSTAGE study [18] in patients receiving nintedanib included 45% of patients with RV dysfunction on echocardiography, while in the study of Behr et al. [19] in patients receiving pirfenidone, the mean systolic pulmonary arterial pressure (sPAP) at echocardiography was 56 mmHg and mPAP 27 mmHg in 18% of patients undergoing right heart catheterisation (RHC). Of note, a subgroup analysis of the INSTAGE study demonstrated beneficial effects of sildenafil on N-terminal-pro-B-type-natriuretic peptide (NT-proBNP) levels in patients with RV dysfunction on echocardiography [20]. The RISE-IIP study assessed riociguat in patients with idiopathic interstitial pneumonias and RHC-proven PH (mPAP 33 mmHg and PVR 4.9 WU) [21]. The study was terminated early due to an increased incidence of severe adverse events, including deaths, in patients receiving riociguat. The INCREASE study assessed the benefits of inhaled treprostinil, a prostacyclin analogue, in patients with fibrotic lung disease and PH demonstrated at RHC (mPAP 37 mmHg and PVR 6.4 WU) [22]. Significant effects on 6MWD, NT-proBNP and clinical worsening were observed in patients receiving treprostinil. Interestingly, subgroup analysis revealed that beneficial effects were seen in patients with a PVR ≥4 WU but not in those with a PVR <4 WU.

PAH therapies in COPD

The effects of PAH therapies have also been assessed in a number of randomised controlled trials (RCTs) in patients with COPD (table 1) [23–26]. Studies of bosentan [23], sildenafil [24] and tadalafil [25] in populations enriched for PH but with echocardiographic evidence of modest PH have failed to meet their primary endpoints. The SPHERIC-1 trial [26] enrolled a small number of patients with more severe RHC-proven PH (mPAP 39 mmHg and PVR 7 WU), and demonstrated significant improvements in PVR and quality of life.

Improving outcomes for patients with PH-CLD: refining patient phenotyping

What, then, can we learn from the published RCT data of PAH therapies in PH-CLD? Can improvements in how we phenotype and evaluate patients translate into improved outcomes? First, some therapies have been associated with adverse outcomes, reinforcing the need for careful assessment of the risk–benefit ratio for treatment in individual patients. Second, published studies highlight the need for further RCTs to identify patient phenotypes who may benefit from PAH therapies. A significant number of studies have relied on diffusing capacity of the lung for carbon monoxide and echocardiography to enrich studied populations for the presence of PH. Patients in these trials have, on the whole, had evidence of modest PH and all studies failed to reach their primary endpoints. The studies by Zeder et al. [7] and Olssen et al. [8] highlight the importance of comprehensive haemodynamic characterisation in the assessment of patients with PH-CLD and support future RCTs of PAH therapies in lung disease enrolling patients who have undergone haemodynamic assessment using RHC. Third, RCTs published to date suggest that certain therapies (endothelin receptor antagonists in ILD) and certain phenotypes (mild PH where exercise limitation is more likely to be due to respiratory rather than pulmonary vascular disease) [27] are not associated with beneficial treatment responses in PH-CLD. Previous open-label studies have reported haemodynamic and functional response to sildenafil and prostacyclin analogues in patients with ILD and severe elevations in mPAP and PVR [28, 29], while clinical response to predominantly phosphodiesterase-5 inhibitor-based therapy has been associated with improved survival in patients with severe PH associated with COPD [12, 30]. Finally, in addition to refining haemodynamic severity of PH-CLD, advances in imaging technology using quantitative computed tomography and the application of artificial intelligence technologies may allow more detailed assessment of the extent of parenchymal and vascular involvement in CLD [31]. Magnetic resonance imaging, the gold standard for cardiac assessment, predicts mortality and clinical worsening in PAH [32], can identify patients with severe PH-COPD [33] and is highly sensitive to treatment effect in patients with PAH [34, 35], and may be a future endpoint for clinical trials of PAH therapies in CLD.

Conclusion

In summary, these data suggest that future RCTs of PAH therapies are warranted in PH-CLD but should enrol well-characterised patients with more severe PH. This issue of the European Respiratory Journal suggests that the optimal threshold for defining this haemodynamic state is a PVR >5 WU.

Shareable PDF