Prognostic Value of Acute Vasoreactivity in Chronic Thromboembolic Pulmonary Hypertension

Authors: Justin A. G. Uphus, Jiun‐Ruey Hu, Shi Huang, Thomas M. Hofbauer, Adelheid Panzenboeck, Roela Sadushi‐Kolici, Inbal Shafran, Nika Skoro‐Sajer, Christian Gerges, Evan Brittain, Irene M. Lang

Abstract

Vasoreactivity, which refers to the reduction of mean pulmonary arterial pressure in response to inhaled vasodilators, is a well‐established metric for prognostication and treatment selection in patients with pulmonary arterial hypertension. However, the role of vasoreactivity in chronic thromboembolic pulmonary hypertension is less studied. We investigated whether vasoreactivity at time of diagnosis carries prognostic value in all patients with chronic thromboembolic pulmonary hypertension, and how it relates to the effect of long‐term pulmonary vasodilator treatments. Patients diagnosed with CTEPH were prospectively subjected to 40 ppm inhaled nitric oxide testing at the diagnostic right heart catheterization. Classic (acute decrease of mean pulmonary arterial pressure > 10 mmHg to a level below 40 mmHg, n = 25), absolute (acute decrease of by 10 mmHg, n = 47) and percent (acute decrease by 10%, n = 129, all with maintained cardiac output) definitions of vasoreactivity were examined in retrospect. The relationship between each definition and transplantation‐free survival was assessed with Cox regression models adjusted for baseline mPAP, age, sex, and WHO functional class. Patients (n = 325) were observed over a median of 5.5 years (interquartile range 2.6–9.2). Vasoreactivity, by the percent definition, was associated with significantly improved adjusted 5‐year transplantation‐free survival (hazard ratio = 0.61, 95% confidence interval 0.38–0.97, p = 0.036). Among not operated patients (n = 174, 53.5%), vasodilator PH medications were associated with improved survival in vasoresponders (hazard ratio = 0.46, 95% confidence interval 0.22–0.96, p = 0.04), but had no impact on survival in non‐vasoresponders (hazard ratio = 0.67, 95% confidence interval 0.36–1.24, p = 0.20). Vasoreactivity at baseline catheterization carries prognostic value in CTEPH.

Introduction

Vasoreactivity refers to the phenomenon in which pressure in the pulmonary arteries can be acutely lowered by reducing the tone of smooth muscle cells via administration of inhaled nitric oxide (iNO) or inhaled iloprost during right heart catheterization in pulmonary hypertension (PH) [1, 2]. Classically, vasoreactivity to iNO has been defined as a reduction of mean pulmonary arterial pressure (mPAP) > 10 mmHg to a level below 40 mmHg, without a decline in cardiac output (CO). In patients with idiopathic, heritable and drug induced pulmonary arterial hypertension (PAH), vasoreactivity plays a central role because vasoresponders usually respond well to treatment with high‐dose calcium channel blocker therapy, and have significantly better prognosis than non‐vasoresponders [3, 4, 5, 6].

However, acute vasoreactivity testing has not been a routine component of diagnostic assessment of chronic thromboembolic pulmonary hypertension (CTEPH) [7]. CTEPH is characterized by persistent organized thrombotic and fibrotic material within the major pulmonary arteries, presumably as a late sequela of acute pulmonary embolism [8, 9]. In addition to its characteristic intravascular lesions, CTEPH is defined by mPAP above 20 mmHg, and a pulmonary vascular resistance (PVR) > 2 Wood Units, after at least 3 months of effective anticoagulation [10]. Recent studies have documented the presence of a secondary microvasculopathy in at least 40% of patients with CTEPH undergoing pulmonary endarterectomy (PEA), lending support to the concept that vasoreactivity may indicate a larger functioning vascular bed [11]. Furthermore, we previously demonstrated that vasoreactivity to iNO is present in at least 8% of patients with CTEPH [12] and associates with better prognosis [13]. In the present study, we sought to confirm the prognostic role of vasoreactivity in a larger population of patients with CTEPH. Secondly, we investigated whether alternative definitions of vasoreactivity may outperform the classic definition used in idiopathic, heritable and drug induced PAH [4]. Lastly, we investigated if any type of vasoreactivity was able to predict the response to vasodilator therapies in CTEPH, as iNO shares a common mode of action with many PH medications, particularly with stimulators of soluble guanylate cyclase.

Methods

Detailed methods are available in the online data supplement. Patients diagnosed with CTEPH [8, 14] between 1995 and 2021 provided informed consent. PH vasodilator medical therapy included riociguat, treprostinil/epoprostenol/iloprost, sildenafil/tadalafil, bosentan/macitentan/ambrisentan. The study was approved by an institutional review board.

Statistical Analysis

Continuous variables were summarized as means with standard deviations (SD) if normally distributed, and medians with interquartile ranges (IQR) if non‐normally distributed. Categorical variables were summarized as counts with proportions. p values for comparisons of baseline characteristics between survivors and non‐survivors are omitted. Kaplan‐Meier curves were constructed, with log‐rank tests to assess differences in transplantation‐free survival between vasoresponders and non‐vasoresponders. Cox proportional hazard regression models were constructed to assess the relationship between each definition of vasoreactivity with transplantation‐free survival, adjusting for baseline mPAP, age, sex, and WHO‐FC. Non‐linearity was assessed for each term in the regression model. Because the association between the variables and the outcome are non‐linear, the hazard ratio is expressed at two points along the curve. Hazard ratio (HR) was expressed as a function of the hazard at the 75% compared to the 50% in continuous variables. Percent change in mPAP under nitric oxide was fixed as a linear term in the model. HR was expressed as a function of a more severe relative to a less severe class. Four patients with WHO‐FC I were analyzed together with patients who presented with WHO‐FC II (n = 60) to avoid overfitting. Two‐tailed p values below 0.05 were considered significant. Statistical analyses were conducted in R (R Statistical Foundation, Vienna, Austria).

Results

Baseline Characteristics

In 325 treatment‐naive patients with CTEPH median follow‐up was 5.5 years (2.6–9.2, medians with IQRs) (Figure 1). At baseline diagnostic right heart catheterization (Table 1) patients were 61 years old (50–71); 161 (49.5%) were female. 80.3% of the patients were in WHO‐FC 3 or 4, Borg dyspnea score was 5 (2–7) and 6‐min walking distance was 347 m (234–451). 82.5% of patients had a history of acute pulmonary embolism, while 41.8% had a history of deep venous thrombosis. Cardiovascular comorbidities such as arterial hypertension (48.3%) and coronary artery disease (21.8%) were common. Metabolic disorders including hyperlipidaemia (29.2%), diabetes mellitus (11.7%) or hypothyroidism requiring thyroid supplementation (10.4%) were common, while inflammatory comorbidities such as inflammatory bowel disease (2.5%) or osteomyelitis (0.6%) were rare. While 25 (7.7%) patients met the criteria of the classical definition, 47 (14.5%) met the criteria of the absolute definition and 129 (39.7%) met the criteria of the percent definition of vasoreactivity to NO (Figure E1, online data supplement).

Hemodynamic Measurements at Baseline

At baseline right heart catheterization, mPAP was 48.0 mmHg (39.0–55.2), mean pulmonary arterial wedge pressure (mPAWP) was 10.0 mmHg (7.0–13.0), CO was 4.5 l/min (3.9–5.3) and cardiac index (CI) was 2.4 l/min/m^2^ (2.1–2.8). Baseline systolic arterial pressure was 127.5 mmHg (113.0–149.0). Mean right atrial pressure (mRAP) was 9.0 mmHg (5.0–13.0), mixed venous oxygen saturation (MVS) at baseline was 58.9% (52.0–64.0), oxygen saturation in the aorta was 90.0% (86.0–93.5), and pulmonary vascular resistance (PVR) was 8,3 WU (5.9–10.5) (Table E1, all medians and IQRs).

Hemodynamic Measurements Under iNO

Upon acute vasoreactivity testing with iNO, mPAP dropped to 44.0 mmHg (36–52), MVS rose to 64.0% (58–68) and aortic oxygen saturation to 95.0% (92–97). Mean PAWP was 10 mmHg (6–14), CO 4.7 l/min (4.0–5.5), CI 2.5 l/min/m^2^ (2.1–2.9), and PVR 6.7 WU (4.9–9.3) (Table E1, medians with IQRs). During NO inhalation, mPAP decreased significantly (p = 0.0131) by 4 mmHg (7–1) (n = 325).

Treatments

Of 325 patients, 151 (46.5%) patients underwent PEA, 79 (24.3%) underwent ≥ 3 balloon pulmonary angioplasty (BPA) sessions, 156 (48%) received at least 6 months of PH medication, and 94 (28.9%) underwent various combinations of PEA, BPA, and PH medications (Figure 1). Comparing 237 survivors with 88 non‐survivors at 5 years, 132 (55.7%) survivors had undergone PEA as initial treatment compared to 19 (21.6%) non‐survivors. A total of 64 (27%) survivors had received ≥ 3 BPA interventions compared to 15 (17%) non‐survivors. A similar proportion of survivors (117, 49.4%) and non‐survivors (41, 46.6%) had received ≥ 6 months of PH medication. Regarding treatment with two modalities, 33 (13.9%) survivors received both PEA and PH medication, compared with 3 (3.4%) non‐survivors. 40 (16.9%) survivors had received both BPA and PH medication, compared with 5 (5.7%) non‐survivors. Regarding triple‐modality treatment, 9 (3.8%) survivors had received PEA, BPA, and PH medication, compared with 1 (1.1%) non‐survivor.

The Impact of Vasoreactivity on Cox‐Adjusted 5‐Year Transplantation‐Free Survival

Of 325 patients, 88 patients experienced death or transplantation at 5‐years, and 237 patients were alive. To assess whether vasoreactivity to iNO was associated with 5‐year transplantation‐free survival, we constructed Cox models adjusted for baseline mPAP, age at initial catheterization, sex, and WHO‐FC for each definition of vasoreactivity. Vasoreactivity, by the classic definition (HR = 0.82, 95% confidence 0.31–2.17, p = 0.69) and the absolute definition (HR = 0.73, 95% confidence 0.37–1.45, p = 0.37) showed no association with transplantation‐free survival, compared with no vasoreactivity. However, vasoreactivity by the percent definition, which included every patient meeting the classic and the absolute definition (HR = 0.61, 95% confidence 0.38–0.97, p = 0.036) was associated with greater transplantation‐free survival (Figure 2). In multivariable Cox regression analysis, smaller percent drop in mPAP under iNO, greater age at right heart catheterization, poorer WHO‐FC (III or IV) and male sex were associated with higher risk of mortality or lung transplantation (Table 2).

Vasoreactivity and Vasodilator Treatments

To determine whether acute vasoreactivity testing had any prognostic value for the response to PH medications in patients with CTEPH, we investigated whether vasoresponders survive better under PH medication regardless of mechanical treatments (PEA, BPA). In the full cohort, survival during the 5‐year observation period did not differ between patients who did or did not receive ≥ 6 months of PH medication by definition of vasoresponse (Figure E2). Because PEA is known to lead to significantly improved outcomes, we examined the subset of patients who did not undergo PEA. In not operated patients, treatment with PH medication for ≥ 6 months was associated with improved 5‐year survival compared to < 6 months or no PH medication (HR = 0.35, 95% confidence interval 0.15–0.82, p = 0.02, Figure 3) in percent vasoresponders. In addition, percent vasoresponders had a better 5‐year survival on PH medication than percent non‐vasoresponders (HR = 0.46, 95% confidence interval 0.22–0.96, p = 0.04, Figure 3). However, in not operated patients who were percent non‐vasoresponders, ≥ 6 months of PH medication had no impact on 5‐year survival compared to < 6 months of or no PH medication (HR = 0.67, 95% confidence 0.36–1.24, p = 0.20, Figure 3). Percent vasoreactivity in not operated patients who did not receive PH medication had no impact on 5‐year survival (HR = 0.89, 95% confidence 0.41–1.89, p = 0.76, Figure 3).

Discussion

Acute pulmonary vasoreactivity testing is a well‐established tool for prognostication and treatment selection in patients with idiopathic, heritable, and drug‐and toxin induced PAH, as recommended by both the European and American guidelines for pulmonary hypertension [4, 8, 15]. Because a pulmonary microvasculopathy resembling the classical pulmonary arteriopathy [16] is common in CTEPH [11], we examined the prognostic significance of acute pulmonary vasoreactivity in CTEPH for which acute vasoreactivity testing has traditionally not been performed [16, 17, 18]. In the present analysis of a large group of patients who underwent prospective vasoreactivity testing, we made two important observations. First, we found that vasoreactivity when defined as a reduction in mPAP by ≥ 10% in response to iNO inhalation, was associated with significantly better 5‐year transplantation‐free survival than non‐vasoreactivity. This implies that there is prognostic value to acute vasoreactivity testing in CTEPH, and that value is not captured by the classic definition of vasoreactivity. Second, we observed that in patients not treated with PEA, use of PH medication was associated with improved transplantation‐free survival in percent vasoresponders, but use of PH medication did not impact survival in non‐vasoresponders, within the limitations of a retrospective study. This implies that vasoreactivity testing may have a role in the selection of patients for PH medication, but this needs prospective validation.

Pulmonary Vascular Lesions in CTEPH

Two different types of pulmonary vascular lesions contribute to an increase in PVR in CTEPH (1): intra‐vascular fibrotic obstructions of pulmonary arteries [19], and (2) a microvasculopathy characterized by obstructive remodeling of pulmonary arterioles which are thought to account for persistent or recurrent PH after PEA [11, 17, 20]. While the former does not have “dilatable” properties, the latter has the potential to respond to cGMP‐based stimulation. In addition, a systemic collateral circulation arising from bronchial, intercostal, or coronary arteries feeds into large vascular obstructions, and connects to pulmonary veins, maintaining perfusion to the capillary bed, thus precluding lung ischemia in these areas [20]. Principally, iNO can act on any of these vessels if they are carrying a functioning endothelium and smooth muscle cells.

Despite the importance of pulmonary microvascular changes, there exists no modality for direct imaging of microvascular changes. Non‐invasive methods to assess microvasculopathy in CTEPH include measurement of mPAP, PVR or diastolic pulmonary vascular gradient out of proportion to the degree of mechanical major vessel obstruction. Invasively, the pulmonary artery occlusion technique has been used to assess the presence of small artery and capillary‐venous disease. The degree of microvasculopathy by upstream resistance measurements at the baseline hemodynamic evaluation of patients with CTEPH correlated well with the results of intraoperative lung biopsies [11]. The percent definition represents a novel application of vasoreactivity testing at novel thresholds to investigate the biologically plausible hypothesis that open vascular territories may be relevant for response to vasodilator treatments and for survival in CTEPH. We believe that acute vasoreactivity testing with iNO [13] may serve as an important technique to estimate the severity of secondary microvasculopathy in CTEPH because its prognostic value was equal across all subsets of CTEPH, regardless of operability.

Defining Vasoreactivity to iNO in Patients With CTEPH

Our study is not the first to document a vasodilatory response to iNO in CTEPH [13, 21, 22]. However, previous studies made use of the classic definition of vasoreactivity. We investigated if alternative definitions would be able to disclose more subtle degrees of vasoreactivity and demonstrate an association with CTEPH survival and response to PH medication. We have previously demonstrated that a reduction in mPAP > 10.4% during acute vasoreactivity testing was associated with improved survival after PEA. This was the basis for the “percent definition” in the present study [13]. Work by Halliday et al. in patients with PAH examined a classic definition of vasoreactivity as well as what they termed a non‐classic response, defined as a decline in mPAP > 10 mmHg without the requirement of reaching a value below 40 mmHg. This was the basis for the “absolute definition” in the present study. Halliday et al. noted that in PAH, the classic definition selects patients with milder disease at baseline, and that non‐classic vasoresponders showed a similar reduction in mPAP but did not have improved survival [23].

Ulrich et al. and Xu et al. found similar rates of acute vasoreactivity to iNO or iloprost in patients with idiopathic PAH and CTEPH according to the classic definition [21, 22]. Frantz et al. found that 20% of their study population of 49 patients reacted to inhaled O2 plus NO with a decrease of 20% in mPAP and PVR. 8% of patients were classic vasoresponders. In the entire cohort, mPAP was reduced by 11.7% without a significant change in CO [12]. In our study, non‐survivors were 10 years older at baseline than survivors, suggesting that early CTEPH diagnosis and subsequently an earlier targeted treatment may be relevant [13].

Prognostication in CTEPH

Observations in patients with not operated and persistent/recurrent CTEPH have suggested a potential utility of PAH risk scores in this patient population [24, 25]. However, the consensus is that there are risk factors that are unique for CTEPH such as lesion characteristics, associated medical conditions [26], the degree of venous involvement, and left ventricular filling pressures [14] that are not depicted by scores developed for PAH. Based on the presented data, vasoreactivity to iNO is another risk factor for CTEPH that should be considered in the future design of CTEPH‐specific risk scores.

Treatment Response to PH Medication

In idiopathic, hereditary, and drug/toxin‐induced PAH, calcium channel blockers are well‐established single‐drug treatments for the rare classic vasoresponders [8]. In contrast, vasoreactivity has had no relevance in the use of other PH medication. In our study, PH medication was given independent of the vasoreactivity status, and response to these medical treatments was assessed retrospectively. Patients without PEA and BPA treatment, but with PH medication only were investigated, as their survival rate has been shown to be significantly worse than after mechanical treatment.

PH medications such as riociguat and prostacyclins are routinely given to patients lacking classical vasoreactivity. We observed that being on ≥ 6 months on PH medication was associated with improved transplantation‐free survival in patients who were percent vasoresponders, but not in patients who were not. One possible explanation is that in CTEPH, 10% percent vasoreactivity may indicate a threshold of vascular beds responsive to iNO that mediate mPAP change, and thus predict response to chronic vasodilators. In contrast to PAH, secondary microvasculopathy in CTEPH is clearly in a territorial distribution that depends on the major vessel lesions [11]. It is of practical interest to predict a potential benefit of PH medication in patients with not operated CTEPH. The implications of this finding are that percent vasoresponders may benefit from PH medication upfront. However, this concept needs to be tested in a prospective setting.

Limitations

Our study has several limitations. First, missing data were limiting the number of patients eligible for statistical analysis (Table E2). Patients from abroad were lost to follow‐up, potentially leading to underestimation of the number of events because they had to be excluded. 82 patients did not have baseline NT‐proBNP measurements. The definitions of classical and absolute vasoreactivity, being stricter than the definition of percent vasoreactivity, necessitate a smaller sample size of patients meeting criteria for vasoreactivity, and thus may conceal a survival signal that may emerge with further testing of these definitions in a large population (Figure E1). Second, because this was a retrospective analysis of a prospectively‐enrolling registry, medications were not standardized. The decision of whether a patient received BPA, PEA, or medications was influenced by many factors, including clinician and patient preference, that were not controlled for. This is an inherent limitation of registry‐based investigations. Any inference regarding potential treatments with calcium channel blockers in classic responders cannot be made. Third, secular trends may have caused temporal bias. Over the course of the observation period, BPA was introduced in Europe in 2014, and riociguat (Adempas) and treprostinil (TrePulmix) received market authorization in the European Union in 2014 and 2020, respectively. The wide time window for enrollment accounts for the overall poorer survival of the cohorts compared with that of contemporary CTEPH registries [27]. Fourth, NO‐testing was performed with supplemental oxygen, giving rise to the possibility that the degree of vasoreactivity may have been influenced. Fifth, lung transplantation in CTEPH has become an extremely rare event over the years. Lastly, we acknowledge that the p values which could be construed as marginal, suggest associations that should be tested with more robust, prospective cohorts.

Conclusions

We found that percent vasoreactivity, as defined by a ≥ 10% reduction in mPAP with nitric oxide inhalation, is associated with improved transplantation‐free survival. We also found that in not operated patients who were percent vasoresponders at baseline, ≥ 6 months of PH medication was associated with better survival than in patients with no vasoreactivity to NO. These findings suggest that vasoreactivity in CTEPH is a prognostic indicator that warrants further investigation in the evaluation of patients with CTEPH.

Author Contributions

Justin A. G. Uphus and Jiun‐Ruey Hu contributed equally to the conception, data acquisition, data analysis and interpretation of the work. Irene M. Lang conceived the study, treated the patients, and supervised the implementation of the study. Shi Huang performed statistical data analysis. Thomas M. Hofbauer, Adelheid Panzenboeck, Roela Sadushi‐Kolici, Inbal Shafran, Nika Skoro‐Sajer, Christian Gerges and Evan Brittain critically reviewed this work and especially contributed by improving the conception, data quality and interpretation of the results in a broader context.

Funding

The authors received no specific funding for this work.

Ethics Statement

Ethics Committee of the Medical University of Vienna (EK#1041/2019).

Conflicts of Interest

J.A.G.U. has received compensation for scientific symposia from AOP‐Health. T.M.H. has received compensation for scientific symposia from Sanofi, AOP‐Health and MSD, and speaker fees from Sanofi. N.S.S has relationships with AOP‐Health, Actelion‐Janssen/J&J, MSD, Cordis, Medtronic, GlaxoSmithKline, and United Therapeutics. C.G. has received compensation for scientific symposia from AstraZeneca, AOP‐Health, Cordis, Actelion‐Janssen/J&J and MSD, speaker fees from AOP‐Health, AstraZeneca, Actelion‐Janssen/J&J and Ferrer, and an educational grant from OrphaCare. I.M.L. has relationships with drug companies including AOP‐Health, Actelion‐Janssen/J&J, MSD, United Therapeutics, Medtronic, Novo Nordisk, Neutrolis and Pulnovo. In addition to being investigator in trials involving these companies, relationships include consultancy service, research grants, and membership of scientific advisory boards. J.R.H., S.H., A.P., R.S.K. I.S. and E.B. have nothing to declare.

Guarantor Statement

I.M.L. takes the responsibility for the content of the manuscript, including the data and analysis.

Supporting information