The Impact of Homogeneous Versus Heterogeneous Emphysema on Dynamic Hyperinflation in Patients With Severe COPD Assessed for Lung Volume Reduction.

Dynamic hyperinflation (DH) is a pathophysiologic hallmark of Chronic Obstructive Pulmonary Disease (COPD). The aim of this study was to investigate the impact of emphysema distribution on DH during a maximal cardiopulmonary exercise test (CPET) in patients with severe COPD. This was a retrospective analysis of prospectively collected data among severe COPD patients who underwent thoracic high-resolution computed tomography, full lung function measurements and maximal CPET with inspiratory manouvers as assessment for a lung volume reduction procedure. ΔIC was calculated by subtracting the end-exercise inspiratory capacity (eIC) from resting IC (rIC) and expressed as a percentage of rIC (ΔIC%). Emphysema quantification was conducted at 3 predefined levels using the syngo PULMO-CT (Siemens AG); a difference >25% between best and worse slice was defined as heterogeneous emphysema. Fifty patients with heterogeneous (62.7% male; 60.9 ± 7.5 years old; FEV1% = 32.4 ± 11.4) and 14 with homogeneous emphysema (61.5% male; 62.5 ± 5.9 years old; FEV1% = 28.1 ± 10.3) fulfilled the enrolment criteria. The groups were matched for all baseline variables. ΔIC% was significantly higher in homogeneous emphysema (39.8% ± 9.8% vs.31.2% ± 13%, p = 0.031), while no other CPET parameter differed between the groups. Upper lobe predominance of emphysema correlated positively with peak oxygen pulse, peak oxygen uptake and peak respiratory rate, and negatively with ΔIC%. Homogeneous emphysema is associated with more DH during maximum exercise in COPD patients.

Introduction

Expiratory flow limitation is the pathophysiologic hallmark of Chronic Obstructive Pulmonary Disease (an class="Disease">COPD). One of the consequences of airflow limitation and permanent parenchymal destruction is lung hyperinflation (1,2). Static hyperinflation is due to the reduced elastic recoil of the lung. Dynamic hyperinflation (DH), which occurs independently or in addition to static hyperinflation, refers to the increase of lung volumes above their resting value that occurs with increases in respiratory rate due to expiratory flow limitation. It is associated both with excessive loading and mechanical disadvantage of inspiratory muscles and with restriction of normal tidal volume expansion during exercise (2). Previous studies have indicated that lung hyperinflation contributes significantly to exercise limitation, exertional dyspnoea and exercise desaturation in COPD patients (3,4).

Thoracic computed tomography (CT) allows precise assessment of emphysema extent and distribution (5,6) and correlation of these parameters with functional and pathologic data (7,8). The distribution of parenchymal damage varies widely between individuals (9), and its categorization in homogeneous and heterogeneous an class="Disease">emphysema is of clinical significance, since it guides selection for surgical and bronchoscopic lung volume reduction procedures (LVR) (10–13). Although there are some data regarding the association between emphysema distribution, clinical features and resting lung function values in COPD patients of various severities (9,14), no study has yet investigated the potential impact of emphysema distribution on DH among COPD patients.

Thus, we conducted a retrospective study to identify: (a) differences in the degree of DH occurring during maximum cardiopulmonary exercise testing (CPET) between COPD an class="Species">patients with heterogeneous and matched COPD patients with homogeneous emphysema and (b) associations between emphysema heterogeneity and CPET parameters in the two patient groups. The primary hypothesis was that patients with homogeneous emphysema hyperinflate more due to the diffuse distribution of the emphysematous lung tissue destruction.

Material and Methods

Study population

Data were collected for COPD outan class="Species">patients who had been assessed at the Respiratory Biomedical Research Unit of Royal Brompton Hospital between June 2009, and August, 2013 for potential eligibility to undergo a bronchoscopic lung volume reduction procedure. These included full lung function measurements, thoracic high resolution computed tomography (HRCT) and breath by breath data from maximal cycling CPET with inspiratory capacity manoeuvers every minute. The COPD patients were in stable clinical condition and optimally treated with combinations of β2-agonists, anticholinergic drugs and inhaled corticosteroids, according to guidelines. All patients had provided informed consent for their initial participation in the studies (15–18).

Study Measurements

Pulmonary Function Testing

Spirometry, gas transfer and lung volumes measurements by body plethysmography were conducted using a Compact Lab System (Jaeger, Hoechberg, Germany). Arterialized capillary blood samples were used to measure arterial blood gases (19). The European Coal and Steel Community predicted values were used for lung function measurements (20) and values of carbon monoxide diffusion capacity and transfer coefficient were adjusted for haemoglobin concentration (an class="Chemical">TLcoc and Kcoc, respectively) (21). All pulmonary function testing (PFT) values used were measured prior to any bronchoscopic intervention and within a 6-month interval of both the HRCT and the maximal CPET.

HRCT acquisition and interpretation

Imaging was performed for clinical indications on 4-slice multidetector CT (Volume Zoom, Siemens, Erlangen, Germany), or 64-slice CT (Somatom Sensation 64, Siemens, Erlangen, Germany). Images were either acquired at 10-mm intervals (4-slice CT) or using a volumetric acquisition (64-slice CT) in a supine position from the lung apices to the bases at full inspiration without the use of intravenous contrast. Images were reconstructed at thin section width (1.0 mm to 1.5 mm) using a high spatial resolution algorithm and reviewed on a workstation at appropriate window settings for viewing the an class="Disease">lung parenchyma (window centre = −500HU; window width = 1500HU).

Images were transferred to a post-processing workstation (Leonardo, Siemens) and quantitative lung density analysis was performed using the Pulmo CT program (Siemens AG), which automatically segments the lung and calculates pixel attenuation coefficients as previously described (22) with a minimum segmentation threshold of -1024HU (Figure 1). Three representative slices of the lungs were analysed: (a) at the level where the superior border of the aortic arch appears; (b) at the level of the main carina, where clear separation of the right and left main bronchi becomes visible; and (c) at the lowest level where neither diaphragm nor any abdominal viscera are visible.

Figure 1.

(A) Upper slice (at the superior border of the aortic arch) transferred to the post-processing workstation, where quantitative lung density analysis was performed using the Pulmo CT program. (B) Histogram of distribution of lung attenuation values, measured in HU in the upper slice (at the superior border of the aortic arch) for several potential emphysema thresholds.

At each level, the program provides a total emphysema index, defined as the percentage of whole lung with attenuation values below a threshold of -900 HU, as well as a severe an class="Disease">emphysema index, defined as percentage of whole lung with attenuation values below a threshold of -950 HU.(23)(24)(25) An alternative threshold for emphysema definition of -960 HU has also previously been reported and we included this threshold as an additional quantitative parameter (26).

To enable us to classify subjects as having either homogeneous or heterogeneous emphysema, we defined heterogeneous an class="Disease">emphysema as a difference of >25% between the highest and lowest quantitative emphysema scores obtained, based on previous precedents of visual analysis (27).

In addition, emphysema was treated as a continuous variable and specifically as a ratio of average ES of the upper and middle slice versus ES of the lower slice (UM/L) for both the 950 and 960 an class="Disease">emphysema thresholds. A high ratio therefore represented upper lobe predominance.

Maximal CPET

Patients performed incremental symptom-limited CPET on a cycle ergometer (Jaeger an class="Chemical">Ergoline 800), with continuous monitoring of pulse oximetry (SaO2), heart rate (HR), and a 12-lead electrocardiogram. The test consisted of three minutes of rest, 1 minute of unloaded pedaling at 50–60 rpm, and then a ramp protocol with work rate (WR) increasing either 5 or 10 watts/minute, followed by 2 minutes of recovery (28).

Gas exchange values and exercise parameters were collected breath-by-breath using a Jaeger Oxycon system (29,30), allowing measurement of: tidal volume (Vt), respiratory rate (RR), oxygen uptake (an class="Chemical">VO2), carbon dioxide production (VCO2), end-tidal oxygen (PETO2) and end-tidal carbon dioxide (PETCO2). The anaerobic threshold (AT), oxygen pulse (VO2/HR), respiratory exchange ratio (RER), minute ventilation (VE) and the ventilatory equivalent for carbon dioxide at peak VO2 (VE/VCO2@VO2 peak) were also calculated, as previously described (29).

Each patient performed a total of four inspiratory capacity (IC) mical">aneuvers at the beginning and after 1, 2, and 3 minutes of rest as previously described (31); resting IC (an class="Gene">rIC) was calculated by averaging these values. End-exercise IC (eIC) was calculated by another inspiratory maneuver which was conducted during the last 30 seconds of peak exercise. ΔIC was utilized as a measure of DH and was calculated as rIC-eIC. For every patient, ΔIC was expressed as a percentage of rIC, that is: ΔIC% = (ΔIC/rIC) × 100%. All exercise tests were performed before any bronchoscopic intervention took place and without oxygen supplementation.

Statistical analysis

Statistical analysis was conducted using the PASW (Predictive Analytics Software by SPSS Inc®) version 19 for Windows 2008. Distribution of values was assessed using the Shapiro-Wilk test of normality; continuous variables are described as mean±1 standard deviation or as median (minimum-maximum), accordingly. Group comparisons in PFTs, DH and exercise parameters were conducted utilizing either the independent samples t-test or the Mann–Whitney test, depending on the normality of their distribution. Pearson r or Spearman rho were used to describe parametric and non-parametan class="Gene">ric correlations between emphysema distribution indices, DH measures, and exercise parameters. Level of p < 0.05 was considered significant.

Results

Baseline characteristics

Sixty-four COPD an class="Species">patients (61.3 ± 7.3 years old; FEV1%predicted = 31.5 ± 11.2%, 61.8 male); fulfilled the enrolment criteria and constituted the final study population. Fifty patients (78.1%) presented with heterogeneous and 14 patients (21.9%) with homogeneous emphysema (group Het and group Hom correspondingly). An initial attempt for patients in Hom group to match the same number of patients in Het group (1:1 matching) for age, FEV1 and TLcoc was made. However, the two patient groups (N1 = 14 patients in Hom group and N2 = 50 patients in Het group) were found to be already matched not only regarding these three selected variables, but also regarding gender, body mass index (BMI), rest pulmonary function testing variables (PFTs) and gas transfer parameters. Results were identical with the use of either emphysema threshold (-950 HU or -960 HU). The baseline demographic characteristics of the two groups are presented in Table 1.

Table 1.

Baseline characteristics of the study population

	Heterogeneous (n = 50)	Homogenous (n = 14)	P
Age	60.9 ± 7.5	62.5 ± 5.9	0.471
Gender (%)			0.936
-Male	62.7	61.5
-Female	37.3	38.5
BMI (kg/m2)	24 ± 4.5	25.6 ± 3.6	0.243
FEV1%predicted	32.4 ± 11.4	28.1 ± 10.3	0.213
FEV1/FVC	29.1 ± 7	26.5 ± 6.4	0.216
rIC (lit)	2.2 ± 0.7	2.2 ± 0.5	0.932
rIC /TLC	0.18 ± 0.05	0.17 ± 0.05	0.660
rEELV (lit)	6.1 ± 1.4	6 ± 1.5	0.840
rEELV/TLC	0.73 ± 0.07	0.72 ± 0.06	0.660
TLC %predicted	137 ± 16.4	132.4 ± 9.4	0.198
RV %predicted	232.8 ± 44.6	220.7 ± 32	0.365
RV/TLC (%)	61.6 ± 8.5	60.8 ± 5.2	0.757
FRC %predicted	189.9 ± 31.5	168.8 ± 47.4	0.059
TLCOc %predicted	34.1 ± 10	33.3 ± 11.5	0.867
KCOc %predicted	45.8 ± 12	40 ± 12.1	0.300
PaO2 (kPa)	9.6 ± 1.3	9.2 ± 1	0.527
PaCO2 (kPa)	4.8 ± 0.7	5.1 ± 1.5	0.398

BMI: Body Mass Index; FEV1: Forced Expiratory Volume in 1 second; FVC: Forced Vital Capacity; TLC: Total Lung Capacity; EELV: End-Expiratory Lung Volume; RV: Residual Volume; FRC: Functional Residual Capacity; TLcoc: Carbon monoxide transfer factor corrected for haemoglobin; Kcoc: Carbon monoxide transfer coefficient corrected for haemoglobin; PaO2: arterial Oxygen Partial Pressure; PaCO2: arterial Carbon Dioxide Partial Pressure.

BMI: Body Mass Index; FEV1: Forced Expiratory Volume in 1 second; FVC: Forced Vital Capacity; TLC: Total Lung Capacity; EELV: End-Expiratory Lung Volume; RV: Residual Volume; FRC: Functional Residual Capacity; TLcoc: an class="Chemical">Carbon monoxide transfer factor corrected for haemoglobin; Kcoc: Carbon monoxide transfer coefficient corrected for haemoglobin; PaO2: arterial Oxygen Partial Pressure; PaCO2: arterial Carbon Dioxide Partial Pressure.

BMI: Body Mass Index; FEV1: Forced Expiratory Volume in 1 second; FVC: Forced Vital Capacity; TLC: Total Lung Capacity; EELV: End-Expiratory Lung Volume; RV: Residual Volume; FRC: Functional Residual Capacity; TLcoc: an class="Chemical">Carbon monoxide transfer factor corrected for haemoglobin; Kcoc: Carbon monoxide transfer coefficient corrected for haemoglobin; PaO2: arterial Oxygen Partial Pressure; PaCO2: arterial Carbon Dioxide Partial Pressure.

Exercise parameters

Exercise parameters for both groups are presented in Table 2. Dyspnoea was the reason for CPET termination for all an class="Species">patients, and the degree of breathlessness did not differ between groups. Only 15 patients (11 from Het group and 4 from Hom group) reached their AT during maximal CPET (data not shown); the rest terminated the exercise before reaching AT, due to respiratory reserve depletion.

Table 1.

Baseline characteristics of the study population

	Heterogeneous (n = 50)	Homogenous (n = 14)	P
Age	60.9 ± 7.5	62.5 ± 5.9	0.471
Gender (%)			0.936
-Male	62.7	61.5
-Female	37.3	38.5
BMI (kg/m2)	24 ± 4.5	25.6 ± 3.6	0.243
FEV1%predicted	32.4 ± 11.4	28.1 ± 10.3	0.213
FEV1/FVC	29.1 ± 7	26.5 ± 6.4	0.216
rIC (lit)	2.2 ± 0.7	2.2 ± 0.5	0.932
rIC /TLC	0.18 ± 0.05	0.17 ± 0.05	0.660
rEELV (lit)	6.1 ± 1.4	6 ± 1.5	0.840
rEELV/TLC	0.73 ± 0.07	0.72 ± 0.06	0.660
TLC %predicted	137 ± 16.4	132.4 ± 9.4	0.198
RV %predicted	232.8 ± 44.6	220.7 ± 32	0.365
RV/TLC (%)	61.6 ± 8.5	60.8 ± 5.2	0.757
FRC %predicted	189.9 ± 31.5	168.8 ± 47.4	0.059
TLCOc %predicted	34.1 ± 10	33.3 ± 11.5	0.867
KCOc %predicted	45.8 ± 12	40 ± 12.1	0.300
PaO2 (kPa)	9.6 ± 1.3	9.2 ± 1	0.527
PaCO2 (kPa)	4.8 ± 0.7	5.1 ± 1.5	0.398

BMI: Body Mass Index; FEV1: Forced Expiratory Volume in 1 second; FVC: Forced Vital Capacity; TLC: Total Lung Capacity; EELV: End-Expiratory Lung Volume; RV: Residual Volume; FRC: Functional Residual Capacity; TLcoc: Carbon monoxide transfer factor corrected for haemoglobin; Kcoc: Carbon monoxide transfer coefficient corrected for haemoglobin; PaO2: arterial Oxygen Partial Pressure; PaCO2: arterial Carbon Dioxide Partial Pressure.

Comparison of exercise parameters between patients with heterogeneous and homogeneous an class="Disease">emphysema

e: end-exercise values; IC: Inspiratory Capacity; EELV: End-Expiratory Lung Volume; TLC: Total Lung Capacity; IRV: Inspiratory Residual Volume; rBorg: rest Borg scale score; pBorg: peak Borg scale score; VE: Minute ventilation; Vt: Tidal Volume; VO2: an class="Chemical">Oxygen Consumption; WR: Work Rate; VCO2: Carbon Dioxide Production; RER: Respiratory Exchange Ratio; HR: Heart Rate; PETO2: End Expiratory Oxygen partial pressure; PETCO2: End Expiratory Carbon Dioxide partial pressure; VE/VCO2: Ventilatory Equivalent for Carbon Dioxide; VO2/HR: Oxygen Pulse; rSPO2: rest arterial Oxygen Saturation; pSPO2: peak arterial Oxygen Saturation.

Patients in the Hom group displayed more DH during exercise, as ΔIC% was significantly higher among Hom group compared to Het group; 39.8% vs 31.2% (p = 0.031). An additional analysis was undertaken, utilizing the change from rest to peak exercise values of End-Expiratory Lung Volume to Total lung capacity ratio (ΔEELV/TLC); Again, Hom group presented with significantly higher ratio, that is greater DH during exercise, compared to Het group (p = 0.035) (Table 2). No other differences in CPET parameters were noted between the two groups (Table 2 and Table 3).

Table 2.

Comparison of exercise parameters between patients with heterogeneous and homogeneous emphysema

	Heterogeneous (n = 50)	Homogenous (n = 14)	p
eIC (lit)	1.5 ± 0.5	1.3 ± 0.3	0.240
eIC/TLC	0.18 ± 0.05	0.17 ± 0.05	0.385
ΔIC %	31.2 ± 13	39.8 ± 9.8	0.031
eEELV (lit)	6.8 ± 1.4	6.9 ± 1.5	0.833
eEELV/TLC	0.82 ± 0.05	0.83 ± 0.05	0.385
ΔEELV/TLC	0.079 (0.01–0.21)	0.119 (0.06–0.15)	0.035
eIRV (lit)	0.4 ± 0.3	0.3 ± 0.3	0.352
rBorg	1(0–5)	2(0–2)	0.441
pBorg	7(4–10)	7(5–7)	0.807
ΔBorg	5(1–9)	5(3-7)	0.852
Peak VE (lit/min)	34(15–47)	32(18–84)	0.593
Peak Vt (lit/min)	1.1(0.5–2.1)	1(0.7–1.4)	0.739
Peak VO2 (ml/kg/min)	13.3(7.9–26.2)	12.5(7.8–26)	0.349
Peak VO2%predicted	49(13–99)	44(14–67)	0.246
Peak WR %predicted	26(5–73)	34(12–59)	0.361
Peak VCO2 (ml/min)	780.6(404–2023.4)	934.6(361–1025)	0.944
RER	0.9(0.4–1.3)	0.9(0.4–1.5)	0.662
Peak HR %predicted	75.5(42–106)	75.5(65–86)	0.802
PETO2 (kPa)	14.9(13–18)	14.9(11.4–16)	0.441
PETCO2 (kPa)	4.7(2.3–6.7)	4.6(3.6–8.3)	0.687
Peak VE/VCO2	39.4(26.9–72.2)	41.6(21–48.3)	0.834
Peak VO2/HR %predicted	66.6(14.4–112.3)	59(20.4–82)	0.131
RR (breaths/min)	32.9 ± 8.9	31.2 ± 9.4	0.545
rSPO2 (%)	94.5(88–99)	96(91–98)	0.323
pSPO2 (%)	89(79–98)	89(80–95)	0.869
ΔSPO2 (%)	6(0–12)	6.5(3–11)	0.674

e: end-exercise values; IC: Inspiratory Capacity; EELV: End-Expiratory Lung Volume; TLC: Total Lung Capacity; IRV: Inspiratory Residual Volume; rBorg: rest Borg scale score; pBorg: peak Borg scale score; VE: Minute ventilation; Vt: Tidal Volume; VO2: Oxygen Consumption; WR: Work Rate; VCO2: Carbon Dioxide Production; RER: Respiratory Exchange Ratio; HR: Heart Rate; PETO2: End Expiratory Oxygen partial pressure; PETCO2: End Expiratory Carbon Dioxide partial pressure; VE/VCO2: Ventilatory Equivalent for Carbon Dioxide; VO2/HR: Oxygen Pulse; rSPO2: rest arterial Oxygen Saturation; pSPO2: peak arterial Oxygen Saturation.

Correlations between emphysema distribution and exercise parameters, for both an class="Disease">emphysema thresholds

Emphysema distribution is expressed as the UM/L ratio of an class="Disease">emphysema score for both the 950 and 960 emphysema thresholds; a high ratio represents upper lobe predominance. IC: Inspiratory Capacity; rBorg: rest Borg scale score; pBorg: peak Borg scale score; VE: Minute ventilation; Vt: Tidal Volume; VO2: Oxygen Consumption; WR: Work Rate; VCO2: Carbon Dioxide Production; RER: Respiratory Exchange Ratio; HR: Heart Rate; PETO2: End Expiratory Oxygen partial pressure; PETCO2: End Expiratory Carbon Dioxide partial pressure; VE/VCO2: Ventilatory Equivalent for Carbon Dioxide; VO2/HR: Oxygen Pulse; rSPO2: rest arterial oxygen saturation; pSPO2: arterial oxygen saturation at peak exercise.

Effect of upper lobe predominance of emphysema distribution

The UM/L ratio of emphysema score for both the 950 and 960 an class="Disease">emphysema thresholds was evaluated; a high ratio represents upper lobe predominance. As presented in Table 3, the UM/L ratio established a weak, inverse but significant correlation to ΔIC% for both emphysema thresholds (Spearman rho = -0.264, p = 0.049; Spearman rho = -0.246, p = 0.049, correspondingly). Moreover, peak VO2%predicted, peak VO2/HR %predicted and peak RR were all positively correlated to UM/L ratio (Table 3).

Discussion

We aimed to investigate whether the distribution of emphysema has any impact on dynamic lung volumes during exercise in an class="Species">patients with COPD. The primary hypothesis was confirmed; we found that patients with homogenous emphysema hyperinflated more than those with heterogeneous disease, although no differences in other exercise parameters were noted. Furthermore, a measure of upper lobe predominance (UM/L ratio) correlated inversely with DH, and positively with peak oxygen consumption, peak oxygen pulse and peak RR.

Significance of findings

Emphysema distribution varies significantly among individuals and possibly represents different pathogenic patterns of disease development (9). Three different subtypes of an class="Disease">emphysema have been recognized: (a) centriacinar emphysema, which predominantly involves the upper lobes and is associated with long-standing cigarette smoking, (b) panacinar emphysema, which mainly involves the lower lobes and is frequently found in patients with alpha-1 antitrypsin deficiency and (c) distal acinar emphysema, which tends to occur adjacent to the pleura or the fibrous septa (32).

These pathologic lesions are found in various combinations in each patient, comprising a heterogeneous or homogeneous an class="Disease">emphysema pattern, as classified by CT imaging (32,33). Mair et al. indicated that emphysema distribution is associated with several clinical features, such as FEV1, BMI, BODE index and health status;(9) however, these associations were most evident among patients with core versus rind predominant, rather than upper versus lower predominant emphysema. Of note, the presence of heterogeneity has been repeatedly associated with improved outcomes and increased survival after a LVR procedure (12,13,34). Although the theoretical background of these interventions is the restoration of lung elastic recoil and the improvement of lung mechanics, the impact of emphysema distribution itself on DH has not previously been assessed.

Both the pathological hallmarks of COPD, airway an class="Disease">inflammation and parenchymal destruction contribute to the development of DH.(35) During exercise, increased airway resistance and decreased elastic recoil result in increased time constants for alveolar units, so as RR and expiratory flow increase, the expiratory time available for exhalation becomes insufficient (36,37). In our study, patients with a heterogeneous pattern of emphysema, that is with unequally distributed parenchymal damage, experienced significantly less DH during maximal exercise.

In patients with heterogeneous an class="Disease">emphysema, lung areas with distinct destruction coexist with areas where lung parenchyma is relatively well preserved (32). We speculate that this coexistence poses a mechanical barrier to the further increase of end expiratory lung volume in emphysematous lung areas, resulting in less DH during maximal exercise. On the contrary, patients with homogeneous emphysema have, by definition, more widespread disease and diffuse floppy airways, meaning that mechanical restriction to whole lung hyperinflation is less.

Interestingly, no CPET parameter differed significantly between patients in the Het and Hom groups. It is well established that medical or other interventions targeting an class="Disease">lung hyperinflation improve exercise tolerance among COPD patients (38–40). Tzani et al. reported that DH was associated with increased exertional dyspnea and reduced maximum exercise capacity in a cohort of COPD patients (3); however, a relatively high value of end expiratory lung volume (≥75% TLC) was used as a threshold for group categorization, while resting TLC was significantly higher between patients who hyperinflated compared to those who did not.

In another study ΔIC was associated with exercise desaturation among male COPD an class="Species">patients with severe disease; nevertheless, this study utilized six-minute walking test, that is a submaximal exercise testing and not maximal CPET, so results are not easily comparable (4). In our study, the highly selective patient population, the relatively small absolute difference of ΔIC (approximately 8.5%), and the presence of similarly increased resting lung volumes between the groups are the most probable causes for this lack of difference in measures of dyspnea and exercise capacity between patients with homogeneous and those with heterogeneous emphysema.

When emphysema distribution was treated as a continuous variable, weak but significant correlations were established with several CPET parameters. The higher the an class="Disease">emphysema heterogeneity with upper lobe predominance, as manifested by increased UM/L ratio, the lower the ΔIC%, and the higher the peak O2 consumption and the peak oxygen pulse. Lung hyperinflation is an important cause of circulatory impairment during exercise, through several mechanisms (3,41,42).

The development of an intrinsic positive end-expiratory pressure (PEEP) during active expiration and the high intrathoracic pressure swings that have to be generated to overcome the increased high elastic and resistive loads result in functional hypovolemia during exercise among an class="Disease">COPD patients, which may have an impact on stroke volume. (35)(41)(36)(43) Whether the unequal distribution of parenchymal destruction itself, as seen in heterogeneous emphysema, compromises cardiac functions less, irrespectively from the degree of DH, due to the potential compensatory effect of preserved lung areas, is a hypothesis that remains to be tested.

Methodological issues

Although data were analyzed retrospectively, they were collected prospectively minimizing recall bias. COPD an class="Species">patients who were included in the study were followed-up in a single tertiary hospital and were pre-screened for assessment as potential patients for a LVR procedure. Thus, they presented with severe disease and with a higher proportion of heterogeneous emphysema than expected. Testing was performed by an experienced and highly trained team of physiologists and researchers with strict quality control measures in place to minimize testing variability.

Moreover, patients were matched for other parameters which could potentially affect DH, such as resting lung function. Although the categorization of an class="Disease">emphysema distribution was performed using a well-established technique, the use of UM/L ratio as a continuous variable to describe upper-lobe predominance in sub analysis has not been reported previously and needs to be further validated; Nevertheless its correlation with parameters of exercise capacity and DH is in accordance with published literature on emphysema heteroneneity (3,41), which strengthens the rationale of its use.

Conclusions

In conclusion, we found that patients with homogeneous an class="Disease">emphysema hyperinflate significantly more during maximum exercise than to those with heterogeneous emphysema. Upper lobe predominance was also associated with less DH and with a higher peak VO2, peak VO2/HR and peak RR. Therefore, CT assessed emphysema pattern and degree of heterogeneity may have a role to play when phenotyping COPD; however, further prospective studies are needed to address this.

Abbreviations

BMIBody Mass IndexCOPDChronic an class="Disease">obstructive pulmonary diseaseCPETCardio pulmonary exercise testDHDynamic hyperinflationESEmphysema scoreEELVEnd Expiratory Lung VolumeFEV1Forced Expiratory Volume in 1 secondFRCFunctional Residual CapacityFVCForced Vital CapacityHRHeart RateHUHounsfield unitHRCTHigh resolution computed tomographyeICEnd-exercise Inspiratory CapacityKcocCarbon Dioxide transfer coefficient corrected for haemoglobinpBorgpeak Borg scale scorePaO2arterial Oxygen Partial PressurePaCO2arterial Carbon Dioxide Partial PressurePETCO2End Expiratory Carbon Dioxide partial pressurePETO2End Expiratory Oxygen partial pressureppSPO2peak arterial Oxygen SaturationrBorgrest Borg scale scoreRERRespiratory Exchange RatiorICrest Inspiratory CapacityrSPO2rest arterial Oxygen SaturationRVResidual VolumeTLCTotal Lung CapacityTLcocCarbon Dioxide transfer factor corrected for haemoglobinUM/LRatio of upper and middle lobe emphysema score to lower lobeVCO2Carbon Dioxide ProductionVEMinute ventilationVE/VCO2Ventilatory Equivalent for Carbon DioxideVO2Oxygen ConsumptionVO2/HROxygen PulseVtTidal VolumeWRWork Rate

Authors’ contribution

AKB contributed to study design, data acquisition, analysis and interpretation and manuscript drafting. ZZ, AN, CD and DMH contributed to data acquisition and interpretation and they critically revised the mical">anuscript. AJ contributed to study design and critically revised the mical">anuscript. an class="Disease">MIP and NSH contributed to study conception and design, data acquisition and interpretation and critically revised the manuscript. All authors approved the final version of the manuscript to be published and agree to be accountable for all aspects of the work.

Table 3.

Correlations between emphysema distribution and exercise parameters, for both emphysema thresholds

Exercise parameter	950 UM/L	960 UM/L
ΔIC %	-0.264*	-0.246*
rBorg	-0.018	0.016
pBorg	0.096	0.068
ΔBorg	0.078	0.030
Peak VE (lit/min)	0.186	0.040
Peak Vt (lit/min)	0.025	0.053
Peak VO2 (ml/kg/min)	0.050	0.033
Peak VO2%predicted	0.340**	0.341**
Peak WR %predicted	0.087	0.040
Peak VCO2 (ml/min)	0.130	0.128
RER	0.075	0.110
Peak HR %predicted	-0.074	-0.086
Peak PETO2 (kPa)	0.193	0.197
Peak PETCO2 (kPa)	-0.097	-0.109
Peak VE/VCO2	-0.006	0.010
Peak VO2/HR %predicted	0.398**	0.390**
Peak RR (breaths/min)	0.300*	0.266*
rSPO2 (%)	0.046	0.030
pSPO2 (%)	0.079	0.069
ΔSPO2 (%)	-0.099	-0.091

Emphysema distribution is expressed as the UM/L ratio of emphysema score for both the 950 and 960 emphysema thresholds; a high ratio represents upper lobe predominance. IC: Inspiratory Capacity; rBorg: rest Borg scale score; pBorg: peak Borg scale score; VE: Minute ventilation; Vt: Tidal Volume; VO2: Oxygen Consumption; WR: Work Rate; VCO2: Carbon Dioxide Production; RER: Respiratory Exchange Ratio; HR: Heart Rate; PETO2: End Expiratory Oxygen partial pressure; PETCO2: End Expiratory Carbon Dioxide partial pressure; VE/VCO2: Ventilatory Equivalent for Carbon Dioxide; VO2/HR: Oxygen Pulse; rSPO2: rest arterial oxygen saturation; pSPO2: arterial oxygen saturation at peak exercise.

* Significant correlation at 0.05 level.

**Significant correlation at 0.01 level.