To the Editor:

Left heart disease (LHD) is the most common cause of pulmonary hypertension (PH) worldwide [1]. PH-LHD results from the upstream transmission of elevated left atrial pressure (LAP) and is associated with global pulmonary vascular remodelling of various degrees involving mainly pulmonary veins, but also capillaries and pulmonary arteries [2–4]. Morphological changes of the pulmonary vasculature in PH-LHD have been well described since the 1950s in patients with severe mitral stenosis. While pulmonary vascular remodelling was generally mild in the majority of cases, severe pulmonary vascular disease was observed in 13–28% of patients [3, 5]. Recent histological studies in PH-LHD cohorts confirm that pre-capillary pulmonary vascular remodelling, which shares similarities with pulmonary arterial hypertension (without plexiform lesions), is rare, occurring in approximately 15% of cases [2, 4]. In these cases, the increase in mean pulmonary artery pressure (mPAP) may be out of proportion to LAP, exceeding the 1:1 ratio as predicted by the PVR equation [6].

In the past 15 years we have witnessed dynamic shifts in the haemodynamic definition and nomenclature of PH, especially in the PH-LHD subgroup, reflecting evolving insights in pathophysiology gained from population science. Several haemodynamic metrics, such as transpulmonary gradient (TPG; mPAP − mean pulmonary arterial wedge pressure (mPAWP)), diastolic pressure gradient (DPG; diastolic pulmonary artery pressure − mPAWP), and pulmonary vascular resistance (PVR; TPG/cardiac output) with different thresholds have been proposed to distinguish between isolated pre-capillary PH (IpcPH) and combined post- and pre-capillary PH (CpcPH) [2, 7–22], but their prognostic significance remains controversial [2, 7–22]. Discordances between the respective data analyses can be attributed to demographic differences in cohorts, competing risk factors and technical limitations of haemodynamic metrics [23, 24]. In addition, the majority of these studies were performed in historic heart failure (HF) cohorts before the introduction of modern HF treatments. Furthermore, right ventricular adaptation to increased afterload may matter more than pulmonary vascular disease itself, and haemodynamic correlates of pulmonary vascular disease are not always prognostic indicators [25].

In the present study we sought to investigate the impact of different haemodynamic definitions on the prevalence of PH subsets in LHD. For this purpose, we queried our haemodynamic database and identified 1659 patients with a mPAWP >15 mmHg who underwent first diagnostic right and left heart catheterisation at the Medical University of Vienna, a tertiary care centre. A retrospective cohort was included between May 1996 and June 2006 and a prospective cohort was included between January 2012 and September 2013, as previously described [2, 14, 26]. Catheterisations were performed for various indications, mostly for the diagnosis of elevated systolic PAP on echocardiography in patients with chronic HF and/or suspected PH, but also before valve replacements, percutaneous interventions and surgical procedures. Referral patterns did not change throughout the study period. Diagnoses were validated on the grounds of patient symptoms, histories, comorbidities, imaging, clinical data and pathoanatomical evidence. Patients were on optimised HF treatments, but not on pulmonary arterial hypertension-specific treatments at the time of haemodynamic measurements. Diagnosis of HF and valvular heart disease was based on current guidelines [27, 28].

Haemodynamic assessment was performed using a 7F Swan-Ganz catheter (Edwards Lifesciences, Irvine, CA, USA). Mean right atrial pressure, right ventricular pressure, systolic, diastolic and mean PAP, mPAWP and respective oxygen saturations, including inferior and superior vena cava saturations, were measured. Left ventricular end-diastolic pressure was measured via femoral or radial arterial access with a 5 or 6F pigtail catheter (Cordis, Bridgewater, NY, USA). All pressures were recorded as averages of 8–12 time-pressure integral derivations during several respiratory cycles using Sensis (Siemens AG, Berlin and Munich, Germany). mPAWP was assessed at the pre-c-wave nadir averaged over the 8–12 cardiac cycles [29]. Zero reference was at midthoracic level. Cardiac output was assessed in at least triplicate by thermodilution.

Based on the same dataset, we analysed the prevalence of PH-LHD subsets in patients with mPAWP >15 mmHg based on six successive definitions by the European Society of Cardiology (ESC)/European Respiratory Society (ERS) guidelines and World Symposium on Pulmonary Hypertension (WSPH) proceedings:

 1) mPAP ≥25 mmHg, TPG >12 mmHg (ESC/ERS 2009 guidelines) [30]

 2) mPAP ≥25 mmHg, DPG ≥7 mmHg (5th WSPH 2013 proceedings) [31]

 3) mPAP ≥25 mmHg, DPG ≥7 mmHg and/or PVR >3 WU (ESC/ERS 2015 guidelines) [32]

 4) mPAP >20 mmHg, PVR >3 WU (6th WSPH 2018 proceedings) [33]

 5) mPAP >20 mmHg, PVR >2 WU (ESC/ERS 2022 guidelines) [1]

 6) mPAP >20 mmHg, PVR >5 WU (definition proposed by the ESC/ERS 2022 guidelines for the identification of PH-LHD patients with a severe pre-capillary component [1]

Finally, we analysed the impact of TPG, DPG, PVR and pulmonary arterial compliance (CPA; stroke volume/pulmonary pulse pressure) on survival in patients with mPAP >20 mmHg using flexible Cox proportional hazard models with restricted cubic splines.

Aetiology of LHD was HF with reduced ejection fraction (n=566, 34%), HF with preserved ejection fraction (n=527, 32%) and left-sided valvular heart disease (n=561, 34%). As illustrated in figure 1a, the prevalence of CpcPH has demonstrated variability based on evolving haemodynamic definitions. Following the ESC/ERS 2009 guidelines definition with a TPG >12 mmHg threshold, the prevalence was 47%. However, with the introduction of the 5th WSPH proceedings definition in 2013, utilising a DPG ≥7 mmHg criterion (figure 1a, red font), the prevalence notably decreased to 15%. Subsequent definitions, such as ESC/ERS 2015 guidelines (DPG ≥7 mmHg and/or PVR >3 WU) and 6th WSPH proceedings in 2018 (PVR >3 WU combined with a lower mPAP threshold at >20 mmHg), led to prevalence rates of 44% and 38%, respectively. Since the 5th WSPH definition in 2013, the prevalence of CpcPH has increased to 61% with the ESC/ERS 2022 guidelines definition (PVR >2 WU threshold in combination with mPAP >20 mmHg). Last, the PVR >5 WU as recommended by ESC/ERS 2022 guidelines for the identification of a severe pre-capillary component in patients with PH-LHD leads to a prevalence of 14%. The change in mPAP threshold for defining of PH from ≥25 mmHg to >20 mmHg increased the overall number of patients with PH-LHD from 88% to 97% (figure 1a); however, the impact on new diagnoses of CpcPH was limited (PVR >3 WU: no new cases; PVR >2 WU: 6 new cases). TPG (hazard ratio (HR) 0.679, 95% CI 0.444 to 0.914), DPG (HR 0.523, 95% CI 0.280 to 0.767), PVR (HR 0.685, 95% CI 0.461 to 0.909) (figure 1b) and CPA (HR −0.475, 95% CI −0.748 to −0.203) were significant predictors of survival (all p<0.001).

a) Proportions of subsets of pulmonary hypertension (PH) associated with left heart disease (LHD). Haemodynamic subsets of PH-LHD based on 1659 patients with elevated mean pulmonary arterial wedge pressure (mPAWP >15 mmHg). The grey bars represent the proportion of patients without PH. The blue bars depict the percentage of isolated post-capillary PH, while the red bars indicate patients with superimposed pre-capillary pulmonary vascular disease (combined post- and pre-capillary PH; CpcPH) according to various definitions introduced over the past 15 years by European Society of Cardiology (ESC)/European Respiratory Society (ERS) guidelines and World Symposia on Pulmonary Hypertension (WSPH) proceedings. The red font emphasises the original definition of CpcPH, introduced at the 5th WSPH in 2013. The current ESC/ERS definition of CpcPH is highlighted in orange. b) Impact of pulmonary vascular resistance (PVR) on survival in patients with PH-LHD. The flexible hazard ratio function demonstrates a continuous increase in mortality risk. This increase is most pronounced up to a PVR of 5 WU. 95% confidence intervals are depicted in blue. HFrEF: heart failure with reduced ejection fraction; HFpEF: heart failure with preserved ejection fraction; VHD: valvular heart disease; DPG: diastolic pressure gradient; mPAP: mean pulmonary artery pressure; TPG: transpulmonary gradient; WU: Wood units.

Our study is limited by its retrospective study design and the experience of a single tertiary care centre and may, therefore, not be entirely generalisable to a community PH/HF centre. Nevertheless, the potential advantages of a single-centre approach are the enrolment of a homogenous patient population, the adherence to a consistent clinical routine, and a consistent quality and safety of diagnostic procedures and invasive haemodynamic assessments. In addition, the diversity in indications for catheterisation at our centre contributes to a broader understanding of PH in the context of LHD.

In this work, DPG ≥7 mmHg and PVR >5 WU performed similarly well in predicting pulmonary vascular disease (figure 1a) [2]. Guidelines recommend to send those patients to a PH centre for an individualised approach, including inclusion in randomised controlled trials [1]. This is a wise decision, because the 2022 booming of CpcPH diagnoses (figure 1a) is due to changing definitions, while the proportion of patients with a high likelihood of pulmonary vascular disease has remained unchanged. Further studies are required to better define phenotype (extending beyond haemodynamics and survival, e.g. advanced imaging) and potential responses to therapies targeting the pulmonary vascular bed in this specific population.