The prevalence of heart failure keeps rising due to population aging and increased survival in coronary and valvular heart diseases.1,2 About 5% of heart failure patients have advanced heart failure with a poor short-term prognosis without heart transplant or a left ventricular assist device.3 Pulmonary hypertension (PH) secondary to left-heart disease (PH-LHD) constitutes the most common form of PH, affecting over 50% patients with left-sided heart failure and is associated with more severe symptoms, lower exercise tolerance, and higher mortality.4,5 A thorough characterization of PH is essential in the evaluation of candidacy for heart transplant, as the presence of significant PH is considered a contraindication.6

Right heart catheterization remains the gold standard for the analysis of pulmonary circulation hemodynamics.4 However, this procedure has several limitations. First, hemodynamic parameters are indirect surrogates of pulmonary vasculopathy and are highly influenced by many determinants, such as heart rate or volume status or volemia. Second, there is a poor correlation with histopathological findings and the prognostic implication of right heart catheterization data is inconsistent.7–12 Finally, measurement challenges and variability in interpretation limit the reliability of the results.4,13 Therefore, there is an unmet need for a better assessment of pulmonary vascular remodeling. Morphological evaluation of PH was, until recently, limited to biopsy and autopsy. Optical coherence tomography (OCT) is rapidly evolving and is used in some types of PH, providing relevant information on diagnosis and management.14–17 However, evidence about its feasibility and clinical usefulness in advanced heart failure is scarce. Previous data suggest that pulmonary vascular remodeling could translate into greater wall thickness (WT).14 Our aim was to study the correlation between OCT results and hemodynamics in patients with advanced heart failure.

Among 100 recruited patients, OCT could be performed and properly analyzed in 90 patients. OCT could not be achieved in 4 patients for technical reasons (ie, wire support and vessel tortuosity); 3 OCT recordings did not have sufficient quality for analysis, and 3 patients had complications leading to a premature ending of the procedure (2 catheter-induced ventricular arrhythmias requiring cardioversion without further relevance and 1 acute pulmonary edema). Complete echocardiographic assessment could be performed in 78 out of those 90 patients.

Demographic, hemodynamic, echocardiographic, and morphological data are depicted in table 1. Median age was 57.5 years. Interestingly, age was significantly associated with pulmonary vascular resistances but not with morphological data. The most frequent underlying heart disease was nonischemic dilated cardiomyopathy followed by ischemic heart disease.

INTERMACS status was not associated with pulmonary vascular resistances or morphological assessment (table 2 and table 3), whereas daily severe symptoms assessed by the NYHA classification were linked to higher pulmonary vascular resistances (table 2).

There was a good correlation between the OCT measurements obtained by the 2 operators. The mean difference was not statistically relevant for any variable (WT, WT index, WT area and iWTA) (figure 1 of the supplementary data).

WT was significantly associated with a higher pulmonary artery pressure, transpulmonary gradient, pulmonary arterial compliance, and pulmonary vascular resistance (figure 2). In addition, it also correlated with echocardiographic parameters such as pulmonary artery acceleration time, pulmonary systolic pressure, and right ventricular-pulmonary arterial coupling (figure 3 and figure 4). However, there was no association with the severity of tricuspid regurgitation or any measurement of right ventricle systolic function.

Figure 2.

Wall thickness and its association with hemodynamic parameters. A: linear Pearson correlation, top, and B: box-plot, bottom. mPAP, mean pulmonary artery pressure; PVR, pulmonary vascular resistance; TPG, transpulmonary gradient.

(0.35MB).

Figure 3.

Pearson correlation between wall thickness and echocardiographic findings of pulmonary hypertension and right ventricular function. A: right ventricular-pulmonary arterial coupling measured by the ratio of tricuspid annular plane systolic excursion to noninvasively estimated pulmonary artery systolic pressure; the correlation is nonlinear, and the best fit is polynomic. B: pulmonary artery acceleration time. C: noninvasively estimated pulmonary artery systolic pressure.

(0.15MB).

Figure 4.

Box-plot showing significant differences in WT regarding noninvasively estimated RV-PA coupling. RV-PA, right ventricular-pulmonary artery; WT, wall thickness.

(0.08MB).

Patients with a WT over 0.25mm had a higher mean pulmonary pressure, transpulmonary gradient and pulmonary vascular resistance than those in the lower percentiles (table 3). These findings were independent of the underlying disease and time from diagnosis, the presence of atrial fibrillation, or mitral regurgitation. Regarding the use of neurohormonal blockade drugs, patients receiving sacubitril-valsartan had significantly lower pulmonary vascular resistance but similar WT (table 2). Beta-blockers were associated with thicker vessels.

Regarding right ventricular-pulmonary arterial coupling, the best cutoff value for WT was 0.20mm; patients with WT over this value showed significantly lower right ventricle-pulmonary artery coupling (figure 4). Patients with a pulmonary vascular resistance over 3 Wood Units had higher indexed measurements of WT (table 2). WT area index was also significantly associated with pulmonary hemodynamics (figure 2 of the supplementary data). Figure 3 of the supplementary data shows the receiver-operator curves of WT and iWTA regarding the prediction of high pulmonary vascular resistance.

The OCTOPUS-CHF study shows that OCT of pulmonary arteries is feasible in patients with advanced heart failure. In addition, intimal thickness is associated with classic hemodynamic measurements of PH. OCT has been associated with pulmonary vascular resistances in other forms of pulmonary hypertension, but information on PH-LHD is scarce. A small study in patients with mitral valve disease represents the only previous data in this setting so that the OCTOPUS-CHF study is the largest sample analyzed to date.

PH can be produced by any valvular, coronary, or myocardial disease that increases filling pressures. It affects most patients with heart failure and its severity is among the stronger prognosis predictors.4 However, specific PH-LHD therapies are lacking.4,21 Although right heart catheterization is still the gold standard for pulmonary vascular remodeling assessment, it has several drawbacks. Firstly, it has technical limitations and waveform analysis variability. In addition, the current steady-flow view of PH is simplistic since it does not consider relevant determinants such as the strong influence of load conditions (volume status, heart rate, and the treatment received),4 ventriculoarterial coupling, left atrial stiffness, vascular compliance, or upstream and downstream wave transmission.22–24 Diastolic gradient seemed promising but failed to demonstrate usefulness, whereas the transpulmonary gradient is even more influenced by left preload.25,26 Thus, pulmonary vascular resistance remains the cornerstone in the evaluation of PH-LHD since it has consistently proven to have linear correlation with mortality, even though it is also highly dependent on pulmonary wedge pressure.8,10,27 Last, there is uncertainty regarding the prognostic value of hemodynamic data together with its suitability to guide heart failure therapy.7–10,12,28

The mechanisms producing PH-LHD remain poorly understood since most knowledge comes from the field of primary PH. There are probably different phenotypes of PH-LHD, and it is not clear why patients with the same left-heart disease display such a wide range of PH-LHD severity. It is postulated that pressure overload may induce endothelial dysfunction, activation of inflammatory pathways, and a proliferative response of the extracellular matrix through a complex signaling interaction between post- and precapillary beds in a subset of genetically predisposed patients.29,30 Current data argue against the classic paradigm that pulmonary vascular disease protects from left-heart overload. In fact, not only alveolar-capillary interface remodeling, lung diffusion impairment, uncoupling of right ventricle-pulmonary artery, and ventilation-perfusion matching alterations become more severely affected but also lung congestion is greater in left-heart disease with combined pulmonary hypertension.31 Recent morphological studies found a larger increase in intimal thickening, muscular hypertrophy, and fibrosis in pulmonary veins than in arterioles in PH-LHD. These findings challenge the classic hemodynamic compartmentation of the vascular tree.7 A deeper understanding of these biological processes might serve as future therapeutic targets.

For all these reasons, morphological parameters might provide valuable additional information. However, morphological assessment was limited to pulmonary biopsy until the recent rise of intravascular imaging techniques. Histopathological findings in congestive pulmonary vasculopathy have been mainly described in the setting of rheumatic mitral valve disease.32–35 Greater pulmonary artery pressure seems to be associated with fibrosis, intimal thickness, and medial hypertrophy of the pulmonary tree in PH-LHD.7,36 In the last few years, there has been increasing interest for “in vivo” evaluation of pulmonary vascular remodeling with different intravascular imaging techniques, among which OCT seems to be the most promising. OCT is a near-infrared light source-based imaging technique with a resolution of 10 to 20mm, which is particularly useful for tissue characterization.37 In previous OCT studies, the pulmonary arterial vessel wall has been described as a homogeneous, single layer, signal-rich structure, without a clear distinction between intima and media layers.14,38 Thickening of the wall by a structure with high reflectivity and low attenuation corresponds to fibrosis in pathological studies.38 Therefore, the distance between the luminal border of the vessel wall and the vessel wall-lung tissue interface defines the WT in pulmonary OCT studies with the same calculations described in the present project. A study comparing pulmonary OCT imaging of 79 patients with different forms of PH and 35 non-PH participants reported an association between the severity of intimal fibrosis and right ventricular afterload. Additionally, the authors were able to identify not only early stages of intimal thickening in borderline PH patients but also reverse remodeling after vasodilator therapy in a significant proportion of those with primary PH, suggesting the high sensitivity of OCT for early diagnosis.17 Regarding primary PH, intimal fibrosis predicts clinical outcomes and correlates with histological findings.38 OCT is emerging as a diagnostic tool in different forms of PH, such as chronic thromboembolic PH where it has successfully demonstrated the presence of thrombi and flaps, and guided angioplasty.14,16,39 Moreover, a larger study of 64 patients with primary PH and 23 with chronic thromboembolic PH also proved good correlation between intimal thickness obtained by OCT and hemodynamic data.40

Nevertheless, there is still little information on OCT in PH-LHD to date. The only relevant evidence available about pulmonary OCT in PH-LHD comes from a sample of 20 patients with mitral valve disease whose WT index was significantly associated with pulmonary vascular resistance, transpulmonary gradient and pulmonary artery pressure.41

This association between OCT assessment and hemodynamic data was also reproduced in our study. We found a significant correlation between OCT assessment of vascular remodeling and functional measures, such as pulmonary pressures and resistances. However, agreement for this association was not strong. Indeed, an area under the curve of 0.622 for WT to predict high pulmonary vascular resistance can be considered poor. This is not necessarily unexpected, as it is widely known that hemodynamic data do not entirely explain structural changes. Firstly, many patients may have elevated pressures and pulmonary vascular resistances that quickly normalize after heart transplant36 or left ventricle unloading through mechanical support, suggesting the absence of significant pulmonary vascular remodeling.6,42 Secondly, the correlation between morphological and functional assessment in our study was noteworthy, despite the heterogeneity present in the sample, with a broad range of underlying diseases and load conditions. Furthermore, it is difficult to demonstrate a strong correlation between morphological and functional data since pulmonary vascular resistance was normal or just slightly increased in most of our patients. This was because our hemodynamic results were used to list patients for heart transplant.

Regarding clinical data, NYHA class was not associated with WT, nor was the INTERMACS status. Furthermore, although some subset of underlying myocardiopathies, long-term disease, the presence of atrial fibrillation or mitral regurgitation might be associated with pulmonary vascular remodeling, we were unable to prove these associations in our study. With respect to heart failure drugs, neurohormonal blocker tolerance may select low-risk patients. Therefore, interpretation of our results should be cautious. In addition, this is the first study that evaluates the relationship between pulmonary vascular remodeling measured by OCT and noninvasive assessment of PH. Both pulmonary artery systolic pressure and pulmonary acceleration time, widely known as the best echocardiographic parameters for estimating PH,19,43 were significantly correlated with OCT results. On the other hand, the lack of association with right ventricle systolic function measurements and the severity of tricuspid regurgitation, are not surprising given the broad spectrum of underlying diseases affecting the right-side irrespective of PH-LHD development. In addition, in our study, right ventricle-pulmonary artery coupling, represented through the ratio of tricuspid annular plane systolic excursion to noninvasively estimated pulmonary systolic pressure, displayed a significant nonlinear correlation with WT. Recent studies have found that this parameter reflects the adequacy of right ventricle adaptation to afterload and it is an early and independent predictor of outcomes in patients with PH and advanced heart failure.44–46

Quantification of pulmonary artery fibrosis is feasible and can be safely performed with OCT. It has the potential to emerge as a new diagnostic tool for PH-LHD as it is significantly correlated with hemodynamic parameters, providing direct morphological assessment of vascular remodeling.

This work was funded by the Instituto de Salud Carlos III (PI18/00254) and cofounded by the European Regional Development Fund, through the CIBERCV (CB16/11/00502).

Conception and design of the study; patient recruitment and data collection: all authors. OCT measurements: J. Martínez-Solano and E. Gutiérrez-Ibañes. Presentation to the ethics committee: E. Gutiérrez-Ibañes and M. Martínez-Sellés. Article drafting: J. Martínez-Solano, C. Ortiz-Bautista, E. Gutiérrez-Ibañes and M. Martínez-Sellés. Statistical analysis: J. Martínez-Solano and E. Gutiérrez-Ibañes.

None declared.