Impaired Pulmonary Vascular Adaptation to Exercise in Emphysema without Pulmonary Hypertension – A Proof-of-Concept Study


 Background: Chronic obstructive pulmonary disease with emphysema lead to respiratory disability beyond bronchial obstruction. The functional impact of pulmonary vascular lesions in emphysema remains unknown. We investigated pulmonary vascular adaptation to exercise in patients with extended emphysema.Methods: Chest magnetic resonance imaging was used to quantitatively assess right-heart function, pulmonary artery and distal pulmonary blood flow. This was performed at rest and during cycling exercise with a magnetic resonance imaging-compatible cyclo-ergometer. Seven emphysematous patients without pulmonary hypertension were compared to 7 healthy non-smokers matched in gender and age.Results: At rest, cardio-pulmonary hemodynamics and distal pulmonary vascular parameters were similar in both groups. Intrasubject adaptation to exercise in emphysematous patients was characterized by a higher increase in right-ventricular ejection fraction (ΔRVEF +8.1 vs. -2.4 %, P=0.046) though a lower right-cardiac output (4.41 vs. 5.79 L/min, P=0.04) at exercise. Accounting for right-cardiac output variation, the distal pulmonary vascular yield index trended to be decreased in patients (ΔPBF/ΔQf -0.78 vs. +18.83 %, P=0.18).Conclusions: Pulmonary vascular adaptation to exercise is impaired in emphysematous patients without identified pulmonary hypertension.Clinical trial registration NCT 04126616.


Background
Chronic obstructive pulmonary disease (COPD) is a frequent respiratory condition providing important morbidity and mortality. It was historically described by spirometric indices of bronchial obstruction. Recent large cohort studies among smokers and COPD patients highlighted the heterogeneity of this disease; the bronchial obstruction severity does not always re ect the respiratory disability. For instance, in a population of emphysematous patients without spirometric bronchial obstruction, a signi cant proportion of cases reported an alteration of the quality of life, exercise hypoxemia or bronchopulmonary exacerbations (1). As a result, the paradigm to guide therapeutic strategies changed in 2017 (2) and is now based on the patients' symptoms (mainly dyspnea and exacerbations).
Although the disease is sometimes accompanied by emphysema, corresponding to alveolo-capillary destruction (3), the pulmonary vascular component has not been extensively studied in COPD (4). Yet early vascular lesions have been shown in COPD patients (5,6). Furthermore, functional alterations of pulmonary circulation were evidenced during exacerbations, both at the microvascular level (7) and in terms of large pulmonary artery vessels (8). They may participate to the respiratory disability beyond bronchial obstruction (9).
These vascular lesions may result in endothelial dysfunction. Indeed, the vascular endothelium has the skill to adapt the vessel caliber in order to match blood ow to increased needs. Pulmonary endothelial dysfunction can express as suboptimal adaptation of vessels to increase blood ow, which has been initially observed in explanted lungs of patients with severe COPD (10). However this feature, assumed to re ect vascular disease, is di cult to access in vivo.
Magnetic resonance imaging (MRI) is a non-invasive and repeatable technique presenting good contrast for vascular structures. It is widely used in the evaluation of the heart and intrathoracic vessels. Dynamic contrast-enhanced sequences allow to assess quantitatively the pulmonary blood ow in distal capillary regions (11). This method has been used in COPD patients with emphysema, con rming a decreased pulmonary blood ow early in the disease (12,13). However the functional impact of these pulmonary vascular alterations is still poorly known.
In this work, we assumed that pulmonary endothelial dysfunction can be evaluated using dynamic contrast-enhanced MRI. We assessed cardio-pulmonary hemodynamics and distal pulmonary blood ow variations from rest to exercise in emphysematous patients without pulmonary hypertension (PH).

Methods
This prospective interventional study (NCT 04126616) was conducted following the declaration of Helsinki and our national ethics committee (CPP #2019-05-045). All subjects gave written informed consent. During the period from October 2019 to June 2020, we prospectively enrolled in this pilot study 7 COPD patients with emphysema and 7 healthy non-smokers matched by gender and age. Inclusion criteria for patients were: age between 40 and 70, diagnosed with COPD, diffuse emphysema on chest computed tomography (CT), forced expiratory volume at 1 s (FEV1) between 35 and 80% of predicted values, negative echocardiographic screening for PH. Healthy subjects were recruited following criteria: age between 40 and 70, non-smokers, no background for respiratory disease. Exclusion criteria for both groups were: contra-indication to exercise testing, MRI or Gadolinium-based contrast injection, history of heart disease, background of thoracic surgery with parenchymal resection, intra-thoracic cancer, pregnancy or breast feeding, refusal to be informed in case of incidental nding during the examinations.
Healthy subjects were pre-screened with electrocardiogram (ECG), 6 minute walk distance (6MWD) test and blood sample in order to con rm inclusion in the absence of contra-indications. Patients' data were collected from routine medical follow-up. Pulmonary emphysema extent was assessed by an experienced radiologist using Slicer3D® and a threshold at -950 UH.
Chest MRI was performed using a commercial imager at 3 T (Verio®, Siemens) and two surface array coils, under continuous monitoring of vital signs. An amagnetic cyclo-ergometer (Lode) was xed to the table of the MRI system, allowing for cycling exercise. After few minutes of warming-up, the mechanical workload was progressively increased to reach the patient's pre-established target heart rate. This latter was determined by the maximum heart rate measured during 6MWD test, capped at 70% of the subject's age-predicted maximum heart rate. If oxygen support was indicated from 6MWD test, the titrated oxygen ow was delivered to the patient to ensure pulse oximetry > 90% during MRI exercise. For technical reasons, imaging was performed while stopping cycling; exercise sessions were thus repeated to reach target heart rate before each sequence. At rest and after exercise, routine sequences were used to evaluate left and right heart function, but also the pulmonary artery output from velocity-encoded contrast imaging placed perpendicularly to the pulmonary artery trunk.
We dynamically followed gadolinium-based contrast agent passage through the pulmonary vasculature with repeated images in a coronal plane with a time resolution of one heart cycle ( Figure 1). Following the method initially proposed by Oechsner et al. (14), the pulmonary blood ow (PBF) could be mathematically derived from the signal variation in the distal parenchyma and using the simultaneous signal variation measured in the pulmonary artery. Half-dose contrast agent was injected at rest and exercise, allowing for deriving distal PBF variation from rest to exercise, that is distal PBF reserve.
Potential differences in right cardiac output variation between the two groups were accounted for by using a distal pulmonary vascular yield index derived as the ratio of PBF variation by right cardiac output variation ΔPBF/ΔQ f .
Statistical analysis was performed with Excel® (Microsoft). Quantitative variables were expressed as mean ± standard deviation. Comparisons were assessed with a bilateral paired Student t test and applying Welch's correction when variances were unequal. A P value < 0.05 was considered as statistically signi cant.

Results
The characteristics of the 7 patients and 7 healthy controls included are presented in Table 1. Mean age was 57.4 ± 6.6 years in patients' group and 5 were men (71%). Controls were matched in age and sex.
Blood in ammatory syndrome was signi cantly higher in patients (2.56 ± 1.38 vs. 0.63 ± 0.38, P=0.009). Disease characteristics of patients are presented in Table 2. They showed moderate bronchial obstruction (mean FEV1 55 ± 16% predicted), important thoracic distension (mean residual volume 174 ± 47% predicted) and moderate alteration of alveolar carbon monoxide transfer (mean DLCO 46 ± 11% predicted). The mean emphysema extent was 36 ± 17% of the total lung parenchyma. Mean room-air resting blood gas oxygen was 71 ± 8.5 mmHg. Two patients (28.5%) required oxygen support during exercise. Mean parameters for cardio-pulmonary hemodynamics at rest are presented in Table 3. Left and right heart function parameters, and right cardiac output measured in the pulmonary artery, were not signi cantly different between patients and controls (right cardiac output 3.53 ± 0.98 L/min vs. 3.53 ± 1.26 L/min, P=0.99). Table 3 Cardio-pulmonary hemodynamics parameters at rest a .  Figure 2).  Figure 3). Table 5 Distal pulmonary blood ow parameters at rest and exercise a .

Discussion
This proof-of-concept study highlights the functional pulmonary vascular de cit of exercise adaptation in emphysematous patients without clinical and echocardiographic signs of PH. Though resting hemodynamics were similar with healthy non-smokers, our results emphasize a signi cantly different adaptation of the right heart at exercise, characterized by a larger increase in RVEF and a lower right cardiac output.
In this work, patients were recruited based on the presence of extended emphysema on chest CT (mean 36% of parenchymal volume) and showed moderate bronchial obstruction (mean FEV1 55% predicted), severe thoracic distension (mean residual volume 174% predicted) and moderate alteration of alveolar gas transfer (mean DLCO 46% predicted), compatible with a "pulmonary vascular phenotype" for which the respiratory disability exceeds the bronchial obstruction level (15). Blood in ammatory syndrome was signi cantly higher in patients, according to the known chronic in ammation in COPD (16).
The exercise performance was lower in patients (6MWD test or maximum workload), in relation with the expected respiratory disability. However the hemodynamics exercise intensity was similar in both groups ; the maximum heart rate at exercise (expressed as a percentage of age-predicted maximum heart rate) was not different between patients and controls.
Hemodynamic adaptation to exercise was characterized by a larger increase in RVEF in patients. Such variation was not observed for the left-sided heart. However the resulting right cardiac output measured in the pulmonary artery trunk was lower in patients at exercise. This adaptation pro le suggests an impaired pulmonary circulation downstream compatible with elevated pulmonary vascular resistances. This anomaly is revealed only at exercise and reminds an exercise-induced pulmonary artery hypertension pro le (17). Moreover, it may not be related with hypoxia-induced lesions as mean room-air resting blood gas oxygen was 71 mmHg in patients.
Regarding distal PBF, rest measurements found good right-left homogeneity at the group level. Importantly, resting distal PBF was similar between the two groups although rarefaction of the pulmonary parenchyma could have created a difference in emphysematous patients. The PBF measurement method was intrinsically sensitive to the parenchymal vascular density ( ow given for 100 mL of lung) and avoided such pitfall.
The intrasubject distal vascular yield index re ected the ability to perfuse pulmonary distal regions by increasing right cardiac output ( Figure 3). Though statistical signi cance was not reached, mean index showed striking difference, being < 1% in patients and 19% in controls. To the best of our knowledge, this is the rst study reporting functional pulmonary vascular impairment at exercise in vivo in emphysematous patients. This may explain one of the mechanisms causing the respiratory disability in patients with a "pulmonary vascular phenotype".
Besides the proof-of-concept nature of this study, main limitations of this protocol are technical. Assessment of PBF requires manual delineation of regions of interest on series of time frames and subsequent adjustable tting procedure. This is time consuming and could be improved with automatized and standardized detection processes in the future. Inclusion of patients with diagnosed pulmonary hypertension may bring better understanding of right-heart hemodynamics at exercise, however MRI supine exercise may be challenging in such population.

Conclusions
In conclusion, this study reports an impaired pulmonary vascular adaptation to exercise in emphysematous patients without PH, as measured with chest MRI. These vascular alterations may participate to respiratory disability. Furthermore, the associated right-heart adaptation pro le suggests isolated exercise pulmonary vascular impairment which could be a therapeutic target to prevent emphysema evolution towards pulmonary hypertension. Figure 1 Chest magnetic resonance images obtained at pulmonary artery (left) and pulmonary veins (right) times.

Abbreviations
Regions of interest placed in the pulmonary artery (green) and in the distal lung (dashed red) allow reconstruction of the signal time course during contrast agent passage. First-pass signal measured in the pulmonary artery region (green curve frame) is used to mathematically derive tting curve (red curve frame) of the distal lung signal (open dots) and assess the pulmonary blood ow.

Figure 2
A: Exercise right cardiac output in patients (left bar) and controls (right bar). B: Intrasubject rightventricular ejection fraction variation from rest to exercise, in patients (left bar) and controls (right bar).
Abbreviations: ΔRVEF, right-ventricular ejection fraction variation from rest to exercise.

Figure 3
Intrasubject distal vascular yield index in patients (left bar) and controls (right box).
Abbreviations: ΔPBF, pulmonary blood ow variation from rest to exercise; ΔQ f , right cardiac output variation from rest to exercise.