Loss of alveolar membrane diffusing capacity and pulmonary capillary blood volume in pulmonary arterial hypertension

Background Reduced gas transfer in patients with pulmonary arterial hypertension (PAH) is traditionally attributed to remodeling and progressive loss of pulmonary arterial vasculature that results in decreased capillary blood volume available for gas exchange. Methods We tested this hypothesis by determination of lung diffusing capacity (DL) and its components, the alveolar capillary membrane diffusing capacity (Dm) and lung capillary blood volume (Vc) in 28 individuals with PAH in comparison to 41 healthy individuals, and in 19 PAH patients over time. Using single breath simultaneous measure of diffusion of carbon monoxide (DLCO) and nitric oxide (DLNO), DL and Dm were respectively determined, and Vc calculated. Dm and Vc were evaluated over time in relation to standard clinical indicators of disease severity, including brain natriuretic peptide (BNP), 6-minute walk distance (6MWD) and right ventricular systolic pressure (RVSP) by echocardiography. Results Both DLCO and DLNO were reduced in PAH as compared to controls and the lower DL in PAH was due to loss of both Dm and Vc (all p < 0.01). While DLCO of PAH patients did not change over time, DLNO decreased by 24 ml/min/mmHg/year (p = 0.01). Consequently, Dm decreased and Vc tended to increase over time, which led to deterioration of the Dm/Vc ratio, a measure of alveolar-capillary membrane functional efficiency without changes in clinical markers. Conclusions The findings indicate that lower than normal gas transfer in PAH is due to loss of both Dm and Vc, but that deterioration of Dm/Vc over time is related to worsening membrane diffusion.


Background
Lung diffusing capacity for carbon monoxide (DL CO ) is a valuable clinical tool in the assessment of pulmonary diseases. It measures the ability of the lungs to transfer gas from the alveolar space to the red blood cells in the pulmonary vessels. Measurement of DL is informative for pathophysiologic diagnoses of lung diseases and serial measurements are used to follow the course of disease [1]. The components of DL can provide more detailed knowledge of the mechanisms of loss of gas exchange capacity. Based on the Roughton and Forster model [2], DL CO is described as a series of resistances: the diffusion of the gas across the alveolar-capillary membrane, the transfer into the plasma and across the red blood cell membrane, and the chemical reaction of the gas with hemoglobin (Hgb). The following equation summarizes these concepts: Where DL represents total lung diffusing capacity; D m alveolar-capillary membrane diffusing capacity; V c pulmonary capillary bed available for gas transfer; and θ is the rate of reaction of the gas with the red cell. Traditionally, measure of the two components of DL CO relied upon repeated performance of the test at different oxygen concentrations and solving the equation for D m and V c . These technical challenges resulted in an arduous method and inaccuracies in measures that limited utility. However, nitric oxide (NO) has a greater affinity to Hgb compared to CO or O 2 , and thus NO diffusion is mainly limited by the transfer of gas across the alveolar capillary membrane [3]. Based on this, DL NO can be used as a direct measure of D m . Recent advances allow measure of DL NO and DL CO simultaneously and direct determination of D m and V c from single breath maneuver at one oxygen tension [3,4]. Few studies have evaluated the components of DL in pulmonary hypertension (PH) in an attempt to understand the pathophysiology of the decreased DL. Cross-sectional data from studies are inconsistent and there are no longitudinal data. In 1968, Nadel et al. showed that DL CO was reduced in PH as a result of low V c [5]. Borland et al. were the first to apply a single breath technique to measure DL NO and DL CO and identified that both were reduced in PH [6]. However, recent studies have attributed the low DL to a loss of D m rather than V c [7,8]. Oppenhiemer et al. measured D m and V c in different groups of PH patients and found that D m was reduced out of proportion to V c as shown by a reduction in the ratio of D m /V c [9]. Measuring the components of DL in PAH would give insight into the disease, especially that recent advances have unveiled the intricacy of the pathogenesis of PAH. We hypothesized that both D m and V c are decreased in PAH patients and that there is progressive decline of DL and its components over time in parallel to disease progression. To test this, DL NO and DL CO were measured by single breath technique in PAH patients in comparison to healthy controls and over time in order to evaluate changes in D m and V c in relation to clinical parameters including echocardiography, 6MWD and BNP levels.

Materials and methods
41 healthy controls and 28 subjects with PAH were enrolled in the study. The Institutional Review Board at the Cleveland Clinic approved the study and all subjects gave written informed consent (IRB 7853). Healthy controls were recruited from the community, had no medical problems and were on no medications. PAH subjects were recruited from the PH clinic. They had an established diagnosis of PAH based on right heart catheterization. Their medical records were reviewed to verify diagnosis. Clinical data including 6MWD, BNP, complete cell count and metabolic panel were collected from the medical records when available. Lung function testing including spirometry, lung volumes, DL CO and DL NO as well as exhaled NO and echocardiogram were measured as part of the study. These tests were repeated longitudinally for a subgroup of PAH patients.
Exhaled NO, lung function, volumes and diffusing capacity for carbon monoxide and nitric oxide Single-breath on-line measurement of fractional NO concentration in expired breath (FE NO ) was measured at the beginning of each visit using the NIOX (Aerocrine, NJ) [10].
The forced expiratory volume at 1 second (FEV 1 ) and vital capacity (FVC) were measured by spirometry (MasterLab; Viasys/Jaeger; Höchberg, Germany) following established guidelines [11]. Total lung capacity (TLC) was measured using the single breath measurement method [12]. Helium was used as a tracer gas following the ATS/ERS guidelines.
Single breath measurements of DL CO and DL NO were performed using a modified Masterscreen PFT (Viasys/ Jaeger; Höchberg, Germany) adapted to measure nitric oxide. The measurement was performed with the patient rested in a seated position using a nose clip. The inspired gas contained 70 ppm nitric oxide (Ikara, NJ) and a blend of ultra high purity 21% oxygen, 0.28% carbon monoxide, 9.5% helium and balance nitrogen. The single breath DL CO method was performed in duplicate to a maximum of 4 measurements to obtain 2 measurements within 5%,~4 minutes apart using a washout volume of 750 ml and an alveolar volume of 750 ml per standard guidelines [1,13]. The breathhold time was~6 seconds [13]. The instrument was calibrated daily. D mCO was calculated as DL NO divided by 1.97 based on solubility factors for CO and NO respectively of 0.0183 and 0.0364 [14].
V c was calculated from the following formula [2]:

Echocardiographic analysis
All echocardiographic analysis were performed following the American Society of Echocardiography Guidelines and Standards [15]. From the parasternal long axis view the Interventricular septal (IVS) thickness in end-diastole, left ventricular end-diastolic dimension (LVEDD), left ventricular endsystolic dimension (LVESD) and posterior wall thickness in diastole were measured from the 2D parasternal long axis image following ASE guidelines. Left ventricular (LV) mass was determined from 2D measurements using the cubed formula: LV function: LV ejection fraction was determined by visual assessment, and/or apical biplane volumes. LV end-diastolic and end-systolic volumes were calculated from the apical 4 and 2 chamber views using the modified Simpson method. LV fractional shortening was determined from parasternal 2D analysis as [(LVEDD-LVESD)/LVEDD] × 100.
Right ventricular (RV) function: RV end-diastolic area (RVAD) and end-systolic area (RVAS) were measured in the apical 4-chamber view by tracing the endocardial border of the RV and the tricuspid annular plane. RV fractional area change was calculated as follows: Right atrial volume was measured in the apical 4 chamber view by using the single plane area length method.
The right ventricular systolic pressure (RVSP) was estimated from the systolic pressure gradient between the RV and the right atrium by the peak continuouswave Doppler velocity of the TR jet using the modified Bernoulli equation plus estimated right atrial pressure (RAP). RAP was estimated from the subcostal window approach measuring changes in inferior vena caval size and collapsibility as determined by the respiratory sniff test following ASE guidelines.
Echo-Doppler estimation of pulmonary vascular resistance (PVR): The highest Doppler continuous wave tricuspid valve peak velocity jet obtained from multiple views (parasternal long axis, parasternal short axis, apical 4 chamber, subcostal or apical off-axis imaging) was determined as the maximum tricuspid regurgitant velocity (TRV). Pulsed wave Doppler sample was placed in the right ventricular outflow tract (RVOT) at the level of the aortic valve in the parasternal short axis view just below the pulmonic valve so that pulmonic valve closure is identified. The Doppler spectrum was traced to determine the time velocity integral of the RVOT (RVOT-TVI). PVR calculation was determined by:  Table 1 shows the baseline characteristics of the two groups and Table 2 lists the clinical data of the PAH patients.
The study spanned 15 months with a mean time of 5 months between visit 1 and visit 2 and of 4 months between visit 2 and visit 3. Twenty-eight PAH patients were evaluated at baseline, of those 19 elected to continue with the longitudinal study and were followed over time. Fourteen patients completed all 3 visits. There was no difference in baseline characteristics between the patients who were followed longitudinally and those who were not.   Table 3]. Hgb did not vary over time [  Table 3]. The rest of the lung functions measured and exhaled NO did not vary (all Wilcoxon SR p > 0.05).

Changes in clinical parameters over time
Baseline clinical data for the PAH patients are summarized in Table 2. Right heart catheterization data were retrieved from the medical records and the most recent one was used. Some (N = 15) were performed within the same year, however others were performed in past years. Clinical parameters including echocardiogram, 6MWD and BNP, did not decline significantly over time of the study. At baseline, DL NO and D m did not correlate with clinical markers [ Table 4].  Table 4].

Discussion
This is the first longitudinal study to describe the changes in gas transfer in PAH over time. Lung diffusing capacity for NO, but not CO, dropped due to the disproportionate loss of membrane diffusion, Dm, as compared to changes in the vascular bed available for gas exchange, Vc. These findings suggest that the efficiency of the alveolar-capillary unit in PAH worsens over time independent of traditional clinical and echocardiographic measures. The results also put forward the potential utility of DL NO in tracking progression of disease in PAH. As in prior work, DL was reduced in PAH patients. Although PAH patients and controls were not well matched by age and gender, the decrease in DL is not likely explained by age and gender alone. Earlier studies measured DL CO at two different oxygen concentrations and calculated D m and V c except for the study by Borland et al. where the single breath test measuring both CO and NO uptake was used. All discovered a decrease in DL, but earlier studies showed predominant loss of V c while more recent ones showed that  D m is reduced out of proportion to V c . The difference between earlier studies and more recent ones could be attributed to differences in techniques or use of pulmonary vasodilators leading to higher measured V c . In this study, the single breath technique measuring DL NO and DL CO was used to assess D m and V c . D m was reduced proportionally to V c in patients with PAH, and thus D m /V c was not different from controls. This confirms the interdependence of D m and V c in pulmonary vascular diseases and any change in the vascular bed that reduces V c would lead to a reduction in D m through a reduction in the surface area available to gas exchange. In PAH, the reduction in V c is likely multifactorial: increased pulmonary vascular resistance, decreased cardiac output and local thrombosis of the vascular bed. Mechanisms underlying the reduced D m may include an increase in the alveolar-capillary membrane thickness caused by fibrotic or proliferative process, and/or interstitial edema. Of note, none of the patients recruited for the study had interstitial lung disease. The decreased DL in association with decreased D m and V c has been described in chronic heart failure patients in stable clinical condition [18]. There is growing evidence that as disease progresses; the left ventricular function may become independently compromised in PAH [19,20]. This may contribute to the reduced DL, D m and V c .
Over time, the increase in V c in association with a decrease in D m could be explained by the effect of Table 4 Correlations among clinical markers and lung diffusion capacity and its components at baseline vasodilator therapy on the diseased pulmonary vasculature. D m would not be expected to increase proportionally to V c with vasodilation alone as the thickened alveolar-capillary membrane is not affected and the fibroproliferative process is not responsive to pulmonary vasodilators. Another explanation could be worsening left side function with increase pulmonary capillary pressure and congestion associated with interstitial edema. In fact, similar findings are noted in heart failure, in which acute decompensation and increased wedge pressure cause a drop in D m paralleled by an increase in V c [21]. Worsening RV function in PAH can lead to LV dysfunction through ventricular interdependence. Our findings showed that V c increased with worsening RV function but there was no relation between V c and LV function. In view of the limitations of echocardiography and the absence of repeat right heart catheterization, the contribution of LV dysfunction to the increase of V c noted over time remains undetermined. Overall, the lack of perceptible changes in standard clinical markers highlights the limitations of available tests, and suggests the possibility that lung diffusing capacity as measured by NO, i.e. the lung D m , may be a potential marker for disease progression in PAH. A limitation of the study is the small number of patients followed longitudinally and studies with a large cohort of PAH patients over time in correlation with clinical parameters and outcomes are needed to confirm our findings. Evidence of airway obstruction in PAH is suggested in this study based on the decreased %FEV 1 and FEV 1 /FVC. This is in keeping with published data that identifies peripheral airway obstruction in PAH by lung functions and pathologic findings [22][23][24][25][26]. Inflammation typically surrounds plexiform lesions in PAH [27] and abundant mast cells have been noted in PAH lungs [28]. Lung biopsies show small airways narrowing with thickened walls infiltrated by lymphocytes, plasma cells and polymorphonuclear leukocytes [29]. Moreover, studies reveal shared mechanistic features of asthma and PAH in experimental models [30,31]. Thus, although PAH is defined as a pure vascular disease, impaired airflow occurs in patients on PAH therapies, suggesting a potential role for airway-directed therapies in the care of these patients.
There are several limitations to the study. Patients were recruited from our PH clinic and the research testing was done on the day of their scheduled clinical visits. Their primary physician scheduled follow up visits and adjusted PAH therapy. We had no control on the time interval between visits or on treatment. As such though most patients were stable on their PAH therapies, the effect of medications on the findings can not be excluded. Another limitation is the difference in age between the PAH and control groups. However, the novel finding here is the longitudinal changes in the PAH group. The drop in D m over time has not been described previously. Another limitation is the lack of concurrent hemodynamic data at the time of the DL measurements. Future studies evaluating D m and V c in relation to hemodynamic data are essential to better understand the significance of the changes noted here.

Conclusion
In summary, a reduction in both D m and V c was noted in PAH patients in association with highly efficient alveolarcapillary units. However, over time, there was a drop in D m /V c . Although the present study does not fully unveil the clinical importance of the decline in D m /V c , it opens the field to further investigate the single breath measurement of DL NO /DL CO and its components in PAH.  Association of brain natriuretic peptide (BNP) to lung capillary blood volume (V c ) and to the ratio of membrane diffusing capacity (D m ) and V c . D m /V c is inversely related to disease severity measured by BNP whereas V c correlates directly to BNP. The points represent measurements obtained at time of first visit.