Exercise-induced pulmonary hypertension associated with systemic sclerosis: Four distinct entities

Authors

  • Rajeev Saggar,

    Corresponding author
    1. David Geffen School of Medicine at University of California, Los Angeles
    • David Geffen School of Medicine, University of California, Los Angeles, Division of Pulmonary and Critical Care, Box 951690, 37-131 CHS, Los Angeles, CA 90095-1690

    Search for more papers by this author
    • Drs. Rajeev Saggar and Dinesh Khanna contributed equally to this work.

    • Dr. Rajeev Saggar has received consulting fees, speaking fees, and/or honoraria from Actelion (less than $10,000) and from Gilead and United Therapeutics (more than $10,000 each).

  • Dinesh Khanna,

    1. David Geffen School of Medicine at University of California, Los Angeles
    Search for more papers by this author
    • Drs. Rajeev Saggar and Dinesh Khanna contributed equally to this work.

    • Dr. Khanna has received consulting fees, speaking fees, and/or honoraria from Actelion, Gilead, and United Therapeutics (less than $10,000 each).

  • Daniel E. Furst,

    1. David Geffen School of Medicine at University of California, Los Angeles
    Search for more papers by this author
    • Dr. Furst has served as a paid consultant with investment analysts on behalf of Actelion and Gilead (less than $10,000 each); he also has received research grants from Actelion and Gilead.

  • Shelley Shapiro,

    1. David Geffen School of Medicine at University of California, Los Angeles
    Search for more papers by this author
  • Paul Maranian,

    1. David Geffen School of Medicine at University of California, Los Angeles
    Search for more papers by this author
  • John A. Belperio,

    1. David Geffen School of Medicine at University of California, Los Angeles
    Search for more papers by this author
  • Neeraj Chauhan,

    1. University of Southern California, Los Angeles
    Search for more papers by this author
  • Philip Clements,

    1. David Geffen School of Medicine at University of California, Los Angeles
    Search for more papers by this author
    • Dr. Clements has received consulting fees, speaking fees, and/or honoraria from Gilead (less than $10,000).

  • Alan Gorn,

    1. David Geffen School of Medicine at University of California, Los Angeles
    Search for more papers by this author
  • S. Sam Weigt,

    1. David Geffen School of Medicine at University of California, Los Angeles
    Search for more papers by this author
  • David Ross,

    1. David Geffen School of Medicine at University of California, Los Angeles
    Search for more papers by this author
  • Joseph P. Lynch III,

    1. David Geffen School of Medicine at University of California, Los Angeles
    Search for more papers by this author
  • Rajan Saggar

    1. David Geffen School of Medicine at University of California, Los Angeles
    Search for more papers by this author
    • Dr. Rajan Saggar has received consulting fees, speaking fees, and/or honoraria from Actelion and Gilead (less than $10,000 each) and from United Therapeutics (more than $10,000).


Abstract

Objective

Exercise-induced pulmonary hypertension (PH) may represent an early but clinically relevant phase in the spectrum of pulmonary vascular disease. There are limited data on the prevalence of exercise-induced PH determined by right heart catheterization in scleroderma spectrum disorders. We undertook this study to describe the hemodynamic response to exercise in a homogeneous population of patients with scleroderma spectrum disorders at risk of developing pulmonary vascular disease.

Methods

Patients with normal resting hemodynamics underwent supine lower extremity exercise testing. A classification and regression tree (CART) analysis was used to assess combinations of variables collected during resting right heart catheterization that best predicted abnormal exercise physiology, applicable to each individual subject.

Results

Fifty-seven patients who had normal resting hemodynamics underwent subsequent exercise right heart catheterization. Four distinct hemodynamic groups were identified during exercise: a normal group, an exercise-induced pulmonary venous hypertension (ePVH) group, an exercise out of proportion PH (eoPH) group, and an exercise-induced PH (ePH) group. The eoPH and ePVH groups had higher pulmonary capillary wedge pressure (PCWP) than the ePH group (P < 0.05). The normal and ePH groups had exercise PCWP ≤18 mm Hg, which was lower than that in the ePVH and eoPH groups (P < 0.05). During submaximal exercise, the transpulmonary gradient and pulmonary vascular resistance (PVR) were elevated in the ePH and eoPH groups as compared with the normal and ePVH groups (P < 0.05). CART analysis suggested that resting mean pulmonary artery pressure (mPAP) ≥14 mm Hg and PVR ≥160 dynes/seconds/cm−5 were associated with eoPH and ePH (positive predictive value 89% for mPAP 14–20 mm Hg and 100% for mPAP >20 mm Hg).

Conclusion

We characterized the exercise hemodynamic response in at-risk patients with scleroderma spectrum disorders who did not have resting PH. Four distinct hemodynamic groups were identified during exercise. These groups may have potentially different prognoses and treatment options.

Systemic sclerosis (SSc) is a connective tissue disease (CTD) of unknown etiology characterized by microvascular injury, immunologic activation, skin fibrosis, and visceral involvement (1, 2). Importantly, pulmonary complications, including pulmonary arterial hypertension (PAH) and interstitial lung disease (ILD), are the leading causes of SSc-related death (3). PAH is a progressive disease leading to increased pulmonary vascular resistance and eventual right heart failure and death (4).

The prevalence of resting PAH in CTD is ∼2.3–10 cases per million (5), typically mostly SSc patients (5–50%) (6–12), patients with mixed CTD (21–29%) (13), and patients with systemic lupus erythematosus (5–43%) (14–16). Based on right heart catheterization, the prevalence of resting PAH in SSc is likely 7.9–20% (7, 17). However, histopathologic evidence of pulmonary arteriopathy has been documented in up to 72% of patients with limited scleroderma, suggesting discordance between clinical findings and pathologic diagnosis (18).

SSc with PAH tends to have a poorer prognosis than other World Health Organization Group I diagnoses. Isolated survival of SSc patients with PAH while receiving pulmonary vasodilator therapy was recently reported to be 78% (1-year rate), 68% (2-year rate), and 47% (3-year rate) (7, 19, 20). Importantly, Tolle and coworkers have suggested that exercise-induced pulmonary hypertension (PH) may represent an intermediate phenotype “between” normal and resting PH (21). In addition, Condliffe and coworkers found that 19% of SSc patients with PAH who had exercise-induced PH developed resting PAH after ∼2.3 years (19). Others, however, have argued that a certain degree of exercise-induced PH may be explained by increasing age, persistent hypoxia, and/or systemic hypertension (22, 23). Nonetheless, an exercise evaluation of SSc patients with PAH may allow for earlier diagnosis, initiation of therapy, and perhaps a more favorable outcome (24). The theoretical evolution from exercise-induced to resting PAH in SSc patients requires further study.

We retrospectively assessed clinical and hemodynamic data collected on consecutive patients with scleroderma spectrum disorders (limited cutaneous SSc, diffuse cutaneous SSc, or overlap syndrome) referred for right heart catheterization at a university hospital. Our objective was to delineate patterns of these patients undergoing exercise right heart catheterization.

PATIENTS AND METHODS

This is a retrospective, descriptive analysis of 80 consecutive patients with scleroderma spectrum disorders who underwent exercise right heart catheterization at a single university hospital. Collection of these data was approved by the local institutional review board. Based on prior reports, patients underwent right heart catheterization if they had dyspnea and at least 1 of the following: resting Doppler echocardiogram–estimated right ventricular systolic pressure ≥30 mm Hg (25), decline by 15% of predicted diffusing capacity for carbon monoxide (DLCO) during the preceding 2 years, DLCO ≤60% predicted, or forced vital capacity (FVC):DLCO (% predicted) ratio >1.4 (6, 10). Importantly, a Doppler echocardiogram–derived right ventricular systolic pressure threshold of 30 mm Hg was chosen to better capture “early” PAH in SSc, given a reported positive predictive value (PPV) of 73% in this population (25).

Using Doppler echocardiogram, we excluded patients who had evidence of significant resting systolic (ejection fraction <50%) and/or diastolic (more than mild) left heart dysfunction. We identified patients with scleroderma spectrum disorders based on the American College of Rheumatology (formerly, the American Rheumatism Association) classification (26). Disease duration was defined as the time since the first non–Raynaud's phenomenon sign or symptom attributable to SSc. We also explored the impact of older age (>50 years) (22) and significant ILD (FVC <60% predicted) (19) on exercise right heart catheterization hemodynamics.

Definition of PH.

PH is defined as a resting mean pulmonary artery pressure (mPAP) ≥25 mm Hg at right heart catheterization; PAH is defined as PH in addition to pulmonary capillary wedge pressure (PCWP) ≤15 mm Hg (27). A hemodynamic evaluation was performed during maximal exercise for patients without resting PH. Four distinct groups of patients were identified based on their exercise hemodynamics: a normal group, an exercise-induced pulmonary venous hypertension (ePVH) group, an exercise out of proportion PH (eoPH) group, and an exercise-induced PH (ePH) group. We defined normal as mPAP <30 mm Hg. We defined ePH as mPAP >30 mm Hg, PCWP ≤18 mm Hg (28), and transpulmonary gradient (TPG) ≥15 mm Hg (29–31), where TPG = mPAP − PCWP (32). We defined ePVH as mPAP >30 mm Hg, PCWP >18 mm Hg, and TPG <15 mm Hg. We defined eoPH as mPAP >30 mm Hg, PCWP >18 mm Hg, and TPG ≥15 mm Hg.

TPG is the pressure gradient across the pulmonary vascular bed (mPAP − PCWP) and increases linearly with cardiac output in healthy volunteers. Importantly, TPG accounts for the normal increase in left-sided filling pressures during exercise and allows differentiation between precapillary and postcapillary (pulmonary venous) pulmonary vascular disease (33). As such, we hypothesized that TPG would be better able to distinguish pulmonary vascular disease from pulmonary venous disease.

Resting and exercise right heart catheterization protocol.

A pulmonary artery catheter (Edwards Scientific) was floated by internal jugular approach under ultrasound guidance. End-expiratory values were transduced for right atrial, pulmonary artery systolic, pulmonary artery diastolic, and pulmonary capillary wedge pressures. Thermodilution cardiac output was analyzed after averaging the sum of triplicate measurements at rest and single measurements during each exercise interval.

All exercise right heart catheterizations were conducted with patients receiving 6 liters of oxygen via nasal cannula to prevent significant hypoxemia. A supine, lower extremity cycle ergometer (Medical Positioning) test was performed using a 3-minute interval ramp protocol with increments of 10–15W, starting at 25W or 30W until exhaustion or 75% maximum predicted heart rate. In normal subjects, there is a linear relationship between heart rate and cardiac output versus V̇O2 during exercise. Since we did not directly measure V̇O2, the heart rate response was used as a surrogate for maximizing cardiac output during exercise (21, 32). End-expiratory mPAP was measured continuously, and end-expiratory PCWP was obtained during the final minute of exercise.

Statistical analysis.

Descriptive statistics of baseline characteristics and resting hemodynamic variables were computed for each group and are reported in Table 1 (mean and SD for continuous variables; frequencies and percentages for categorical variables). Analysis of variance (ANOVA) was used to detect between-group differences for continuous variables, and data not normally distributed were log-transformed. In the case of categorical variables, the chi-square test and Fisher's exact test were used. Where significant between-group differences were detected by ANOVA, pairwise comparisons were made for each possible pair of groups using the Tukey-Kramer method to adjust for multiple comparisons. These analyses were repeated for the hemodynamic values during peak exercise.

Table 1. Baseline demographic and clinical characteristics of the 4 groups of patients*
 TotalNormalePVHeoPHePHP
  • *

    Four groups of patients were identified based on their exercise hemodynamics: the normal group, the exercise-induced pulmonary venous hypertension (ePVH) group, the exercise out of proportion pulmonary hypertension (eoPH) group, and the exercise-induced PH (ePH) group. SSc = systemic sclerosis; MCTD = mixed connective tissue disease.

  • Adjusted for multiple comparisons using the Tukey-Kramer method.

  • P < 0.05 versus eoPH group.

No. (%) of patients57 (100)15 (26)12 (21)9 (16)21 (37)
No. (%) female45 (78.95)10 (66.7)11 (91.7)8 (88.9)16 (76.2)0.452
Type of SSc, no. (%)     0.212
 Diffuse25 (43.9)9 (60.0)7 (58.3)3 (33.3)6 (28.6)
 Limited22 (38.6)5 (33.3)4 (33.3)4 (44.4)9 (42.9)
 MCTD/overlap10 (17.5)1 (6.7)1 (8.3)2 (22.2)6 (28.6)
Disease duration, mean ± SD years5.6 ± 5.83.2 ± 3.55.0 ± 4.47.9 ± 7.26.8 ± 7.00.35
Age, mean ± SD years50.9 ± 12.842 ± 10.253.3 ± 11.560.1 ± 13.252.1 ± 11.90.007
Ethnicity, no.     0.024
 Caucasian81268
 Hispanic4014
 African American0005
 Asian/Pacific Islander3024

To systematically determine the presence or absence of exercise abnormality in a given patient, a classification and regression tree analysis was performed using the statistical software package R, version 2.9.1. All analyses were done using Stata 10.1 software (Stata Corporation). P values less than 0.05 were considered significant.

These decision tree models are technically known as binary recursive partitioning. A parent node is always split into 2 child nodes (binary); the process is repeated for each child node (recursive), and each split results in portioning into mutually exclusive subsets. The model aims to recursively partition input variables in order to maximize the purity in a terminal node. The decision to make a partitioning split is done after searching each possible threshold for each variable included in order to find the split that leads to the greatest improvement in the purity score of the resultant node; cutoff points are sought for continuous variables (rather than deciding on an arbitrary dichotomous point). Each node is split on just 1 variable.

RESULTS

Patient demographics.

Eighty consecutive patients with scleroderma spectrum disorders underwent resting and exercise right heart catheterization between July 1, 2007 and July 15, 2009. Of those 80 patients, 23 had evidence of resting PH and were excluded from further analysis. The remaining 57 patients had evidence of normal resting pulmonary hemodynamics, with New York Heart Association functional class I or II, and subsequently underwent exercise testing. The clinical and hemodynamic characteristics of these 57 patients at baseline are outlined in Tables 1 and 2. The majority of the patients were female (79%) and Caucasian (60%), and their mean ± SD age was 50.9 ± 12.8 years. Twenty-five (43.9%) had diffuse cutaneous SSc, 22 (38.6%) had limited cutaneous SSc, and 10 (17.5%) had overlap syndrome. For the cohort, the mean FVC (% predicted) was 76.6%, the mean DLCO (% predicted) was 59.4%, and the mean FVC:DLCO ratio was 1.4; 9 patients (16.1%) had FVC (% predicted) <60%.

Table 2. Baseline hemodynamic characteristics of the 4 groups of patients*
 TotalNormalePVHeoPHePHP
  • *

    Except where indicated otherwise, values are the mean ± SD. FVC = forced vital capacity; DLCO = diffusing capacity for carbon monoxide; BNP = brain natriuretic peptide; 6MW = 6-minute walk (see Table 1 for other definitions).

  • Adjusted for multiple comparisons using the Tukey-Kramer method.

  • P < 0.05 versus eoPH group.

  • §

    A total of 39 patients were tested.

No. (%) of patients56/57 (100)14/15 (26)12 (21)9 (16)21 (37)
FVC, % predicted76.6 ± 21.278.6 ± 19.883.3 ± 1765.6 ± 26.976.1 ± 21.30.4798
FVC category, no. (%)     0.373
 <60% predicted9 (16.1)1 (7.1)1 (8.3)3 (33.3)4 (19.0)
 ≥60% predicted47 (83.9)13 (92.9)11 (91.7)6 (66.7)17 (81.0)
DLCO, % predicted59.4 ± 20.463 ± 16.472.2 ± 2044 ± 15.756.4 ± 20.70.018
FVC:DLCO ratio1.4 ± 0.71.3 ± 0.41.2 ± 0.31.5 ± 0.51.6 ± 1.10.4084
BNP, pg/ml48.1 ± 46.733.2 ± 36.450.9 ± 41.365 ± 59.350.3 ± 51.20.09
6MW, meters§387.7 ± 101.5405.4 ± 111.2427.6 ± 59.4270.4 ± 94.8392.7 ± 98.30.0796

Resting Doppler echocardiogram data were available for all patients. The estimated tricuspid regurgitant jet velocity was detectable for 36 patients (63%) with an estimated mean ± SD right ventricular systolic pressure of 36 ± 11.3 mm Hg. Five of the 57 patients had evidence of mild diastolic dysfunction on Doppler echocardiogram; however, none had elevated PCWP (>18 mm Hg) at rest, and only 1 patient developed elevated PCWP during exercise.

Eighty-seven percent of the cohort was positive for antinucleolar antibodies. Of those patients, 20% were positive for anticentromere antibodies; however, there were no correlations with hemodynamic parameters at rest or during exercise.

Resting hemodynamics.

Four distinct hemodynamic groups were identified on exercise: normal (n = 15), ePVH (n = 12), eoPH (n = 9), and ePH (n = 21) (Table 3). In addition, the eoPH and ePH groups had higher pulmonary vascular resistance (PVR) and TPG compared with the normal and ePVH groups. Patients in the normal group were younger (mean age 42 years), had lower mean mPAP (14.4 mm Hg) compared with patients in the other groups, and had lower mean PCWP (8 mm Hg) compared with patients in the ePVH and eoPH groups (Table 3). Compared with the other 3 groups, patients in the normal group had a greater percentage of males, a higher frequency of diffuse cutaneous SSc, and a shorter disease duration. Finally, the eoPH group had the lowest DLCO (44% predicted) and the highest resting mPAP (20.7 mm Hg) values compared with the other 3 groups. There were no other significant differences between groups in resting brain natriuretic peptide levels or Doppler echocardiogram parameters.

Table 3. Resting hemodynamics in the 4 groups of patients*
 TotalNormalePVHeoPHePHP
  • *

    Values are the mean ± SD. RA = right atrial; PAsys = pulmonary artery systolic; PAdia = pulmonary artery diastolic; mPAP = mean pulmonary artery pressure; PCWP = pulmonary capillary wedge pressure; CO = cardiac output; CI = cardiac index; PVR = pulmonary vascular resistance; TPG = transpulmonary gradient (mPAP − PCWP) (see Table 1 for other definitions).

  • Adjusted for multiple comparisons using the Tukey-Kramer method.

  • P < 0.05 versus ePVH group.

  • §

    P < 0.05 versus eoPH and ePH groups.

  • P < 0.05 versus ePVH, eoPH, and ePH groups.

  • #

    P < 0.05 versus ePVH and eoPH groups.

  • **

    P < 0.05 versus ePH group.

  • ††

    P < 0.05 versus ePVH and ePH groups.

RA pressure, mm Hg6.3 ± 3.14.9 ± 3.38.1 ± 2.66 ± 2.66.4 ± 30.0347
PAsys pressure, mm Hg28.9 ± 6.124.3 ± 6.7§29.4 ± 3.731.6 ± 630.6 ± 5.30.0173
PAdia pressure, mm Hg12.3 ± 4.18.3 ± 2.613.7 ± 4.115.8 ± 2.612.8 ± 3.40.0001
mPAP, mm Hg18.2 ± 4.114.4 ± 3.919.2 ± 3.320.7 ± 3.419.3 ± 3.10.0007
PCWP, mm Hg10.7 ± 4.18 ± 3.6#14.6 ± 2.7**11.9 ± 4.39.9 ± 3.20.0003
CO, liters/minute5.2 ± 1.45.7 ± 1.65.8 ± 1.64.7 ± 1.64.7 ± 0.80.0491
CI, liters/minute/m23 ± 0.83.3 ± 0.83.3 ± 0.72.8 ± 0.82.7 ± 0.60.0517
PVR, dynes/seconds/cm−5127 ± 73.595.9 ± 50.4††67.7 ± 40.4§168.3 ± 86.7165.5 ± 65.60.0002
TPG, mm Hg7.5 ± 3.56.4 ± 3.1**4.6 ± 2.3§8.8 ± 3.49.4 ± 3.10.0005
Heart rate, beats per minute75.6 ± 13.972.7 ± 10.871.8 ± 13.879.2 ± 17.178.8 ± 15.30.5892

Exercise hemodynamics.

The majority of patients in all groups demonstrated acceptable effort, defined by peak heart rate ≥75% of predicted maximum; the mean ± SD peak heart rate was 128.4 ± 18.4 beats per minute. The ePVH, eoPH, and ePH groups had higher mPAP compared with the normal group (P < 0.05) (Table 4), while the ePVH (25.3 mm Hg) and eoPH (22.8 mm Hg) groups had higher PCWP compared with the normal (12.7 mm Hg) and ePH (14.6 mm Hg) groups (P < 0.05). PVR and TPG were higher in the eoPH and ePH groups than in the normal and ePVH groups (P < 0.05); there were no differences in PVR and TPG between the eoPH and ePH groups. In addition, cardiac output (liters per minute) was significantly higher in the normal group than in the ePH group, but only showed a higher trend in the ePVH group compared with the ePH group. PVR increased significantly between rest and submaximal exercise only in patients with eoPH and ePH (P < 0.01) (Figure 1).

Table 4. Exercise hemodynamics in the 4 groups of patients*
 TotalNormalePVHeoPHePHP
  • *

    Values are the mean ± SD. PAsys = pulmonary artery systolic; PAdia = pulmonary artery diastolic; mPAP = mean pulmonary artery pressure; PCWP = pulmonary capillary wedge pressure; CO = cardiac output; CI = cardiac index; PVR = pulmonary vascular resistance; TPG = transpulmonary gradient (mPAP − PCWP) (see Table 1 for other definitions).

  • Adjusted for multiple comparisons using the Tukey-Kramer method.

  • P < 0.05 versus ePVH, eoPH, and ePH groups.

  • §

    P < 0.05 versus ePVH and eoPH groups.

  • P < 0.05 versus ePH group.

  • #

    P < 0.05 versus eoPH and ePH groups.

  • **

    P < 0.05 versus eoPH group.

PAsys pressure, mm Hg52.9 ± 12.2740.1 ± 8.353.3 ± 8.961.9 ± 8.458.1 ± 10.6<0.0001
PAdia pressure, mm Hg24.0 ± 7.015.9 ± 3.925.8 ± 3.929.1 ± 7.226.57 ± 5.2<0.0001
mPAP, mm Hg34.53 ± 7.924.2 ± 3.435.9 ± 3.941.8 ± 4.837.9 ± 5.7<0.0001
PCWP, mm Hg17.7 ± 6.712.7 ± 6.2§25.3 ± 4.022.8 ± 3.014.6 ± 3.6<0.0001
CO, liters/minute9.4 ± 3.111.6 ± 3.0#10.3 ± 4.0**7.3 ± 1.898.2 ± 1.80.0007
CI, liters/minute/m25.4 ± 1.66.6 ± 1.4#5.8 ± 1.84.4 ± 1.04.7 ± 1.10.0006
PVR, dynes/seconds/cm−5165.1 ± 94.091.1 ± 61.6#90.2 ± 49.8#223.6 ± 91232.1 ± 61.0<0.0001
TPG, mm Hg16.96 ± 7.8411.5 ± 6.3#10.6 ± 4.7#19.0 ± 5.323.3 ± 5.4<0.0001
Heart rate, beats per minute128.4 ± 18.4133.0 ± 20.6124.1 ± 11.0120.1 ± 12.8129.5 ± 20.80.3634
Figure 1.

Pulmonary vascular resistance (PVR) at rest and during submaximal exercise in the 4 groups of patients identified based on their exercise hemodynamics: the normal (NL) group, the exercise-induced pulmonary venous hypertension (ePVH) group, the exercise out of proportion pulmonary hypertension (eoPH) group, and the exercise-induced PH (ePH) group. Data are shown as box plots. Each box represents the 25th to 75th percentiles. Lines outside the boxes represent the 10th and the 90th percentiles. Lines inside the boxes represent the median.

Hemodynamics in patients with FVC ≥60% predicted versus FVC <60% predicted.

Of the 9 patients with FVC <60% predicted, 3 were in the eoPH group, 4 were in the ePH group, 1 was in the normal group, and 1 was in the ePVH group. These 9 patients had significantly higher mean ± SD baseline mPAP (22.1 ± 2.1 mm Hg) compared with patients with FVC ≥60% predicted (17.6 ± 3.7 mm Hg) (P < 0.001). Otherwise, there were no significant differences in the baseline characteristics between these groups (data not shown). During exercise, patients with FVC <60% predicted had higher mean ± SD mPAP (40.6 ± 7.4 mm Hg versus 33.5 ± 7.5 mm Hg; P = 0.02) and higher mean ± SD PVR (238.3 ± 93.5 dynes/seconds/cm−5 versus 150.4 ± 89.2 dynes/seconds/cm−5; P = 0.02) compared with patients with FVC ≥60% predicted. When patients with FVC <60% predicted were excluded from the analysis, the 4 distinct clinical entities remained discernible (data not shown).

Hemodynamics in patients age ≤50 years versus age >50 years.

Patients in the normal group were younger than those in the other 3 groups (Table 1). We analyzed all patients who achieved mPAP >30 mm Hg during exercise (patients in the ePVH, eoPH, and ePH groups). Fifteen patients (36%) were age ≤50 years, while 27 (64%) were age >50 years. There were no differences between those age ≤50 years and those age >50 years in mean exercise mPAP (38.0 mm Hg versus 38.3 mm Hg, respectively), PVR (168.0 dynes/seconds/cm−5 versus 201.8 dynes/seconds/cm−5, respectively), or PCWP (19.0 mm Hg versus 19.7 mm Hg, respectively). Similarly, both groups of patients (age ≤50 years and age >50 years) achieved at least 75% of their predicted maximum heart rate (134.8 beats per minute versus 120.8 beats per minute, respectively).

Decision tree.

Figure 2 demonstrates the decision tree that is applicable to an individual patient with available resting right heart catheterization hemodynamics data. In this analysis, baseline mPAP is the parent or primary node. A baseline mPAP <14 mm Hg excludes ePH and eoPH (negative predictive value [NPV] 100%). While a resting mPAP >20 mm Hg had a PPV of 90% for abnormal physiology (ePH, eoPH, or ePVH), a concurrent PVR ≥160 dynes/seconds/cm−5 had a PPV of 100%. Similarly, for patients with mPAP 14–20 mm Hg (inclusive), PVR <67 dynes/seconds/cm−5 had an NPV of 100%, and PVR ≥160 dynes/seconds/cm−5 had a PPV of 89% for predicting ePH and eoPH.

Figure 2.

Decision tree applicable to an individual patient with available data on resting right heart catheterization (RHC) hemodynamics. The baseline mean pulmonary artery pressure (mPA; in mm Hg) is the primary mode. PVR values are in dynes/seconds/cm−5. NPV = negative predictive value; PPV = positive predictive value (see Figure 1 for other definitions).

DISCUSSION

In this study, we demonstrated a variable exercise response in a “real-life” cohort of patients with scleroderma spectrum disorders who had normal resting hemodynamics. There were 4 distinct groups of patients characterized by the hemodynamic response to exercise: a normal group, an exercise-induced pulmonary venous hypertension group, an exercise out of proportion PH group, and an exercise-induced PH group. The mPAP and PVR are hemodynamic variables at rest that best distinguished the ePH and eoPH groups from the ePVH and normal groups. In addition, we provide a preliminary algorithm based on resting hemodynamics that may help researchers specifically predict eoPH and ePH for an individual patient with a scleroderma spectrum disorder.

Abnormal exercise physiology has been recognized in SSc, idiopathic PAH, idiopathic pulmonary fibrosis (IPF), and chronic obstructive pulmonary disease (21, 28, 34–40). Studies have described exercise-induced PH in 46–59% of patients with scleroderma spectrum disorders, characterized by Doppler echocardiogram pulmonary artery systolic pressure >40 mm Hg during exercise (23, 41, 42). Collectively, these studies suggest an abnormal phenotype in “at-risk” populations after assessing the response of the pulmonary vasculature to exercise.

There is a paucity of data regarding longitudinal followup in patients with exercise-induced PH. The UK group prospectively reported survival in 42 SSc patients with exercise-induced PH, of whom 8 (19%) progressed to having resting PAH within a mean ± SD 838 ± 477 days and of whom 4 (9.5%) died secondary to complications of pulmonary vascular disease within 3 years of diagnosis (19). PH in SSc may be an abnormal hemodynamic response and part of a continuum from normal to resting PH. We postulate that patients with ePH and eoPH may be at risk for developing progressive pulmonary vascular disease given the otherwise indistinguishable response to exercise, although probably with different frequencies and response to existing therapies.

Few studies have used right heart catheterization to address whether an mPAP >30 mm Hg during exercise represents an aberrant pulmonary vascular system in “at-risk” populations (19, 21, 28, 34, 38, 39, 43). Recently, Tolle and coworkers evaluated a large cohort of non-SSc symptomatic patients who underwent simultaneous cardiopulmonary exercise testing and right heart catheterization (21). At maximal exercise, mPAP and PVR were highest in patients with resting PAH, lowest in normal patients, and intermediate in patients with exercise-induced PH (21). Exercise-induced PH was defined as mPAP >30 mm Hg, PCWP <20 mm Hg, and PVR ≥80 dynes/seconds/cm−5 (21). Importantly, after repeat analysis of our ePH group using the definition described by Tolle et al, there was no reclassification of patients. In our study, both resting mPAP >20 mm Hg and PVR ≥160 dynes/seconds/cm−5 predicted the development of either eoPH or ePH. In contrast, the study by Tolle et al did not find an association between baseline hemodynamic parameters (including mPAP 21–24 mm Hg) and the development of exercise-induced PH. This difference may be due in part to our study of a more homogeneous, “at-risk” population.

More recently, Kovacs and coworkers described the association of resting and exercise mPAP with exercise capacity (6-minute walk [6MW] and peak V̇O2) in a well-characterized SSc cohort (38). Patients with baseline FVC ≤60% predicted and PCWP >15 mm Hg were excluded, representing an important difference from our study. A median mPAP >17 mm Hg at rest and a median mPAP >23 mm Hg during slight exercise were associated with both a significant decrease in 6MW and peak V̇O2 (38). The PVR from baseline to slight exercise was relatively unchanged, however, similar to our experience, suggesting that a PVR ≥160 dynes/seconds/cm−5 at rest was potentially abnormal.

In our study, normal patients were significantly younger than patients with eoPH and trended toward being younger than patients with ePVH and ePH; however, it remains unclear whether exercise-induced PH is an age-related phenomenon specifically in SSc. A recent systematic analysis suggested that during submaximal exercise, mPAP exceeds 30 mm Hg in 47% of all cases in normal, healthy adults age >50 years (22). However, measurements of PVR were not comprehensively reported. In our study, for individuals with mPAP >30 mm Hg during exercise (n = 42), mPAP, PVR, and PCWP values were independent of age. Moreover, after accounting for exercise PCWP and cardiac output, the PVR values in both the normal and ePVH groups were not significantly different between rest and exercise. In contrast, the PVR increased significantly from rest to exercise in the ePH and eoPH groups compared with the ePVH and normal groups, suggesting a blunted cardiac output response and vascular stiffness in ePH and eoPH. Importantly, another study showed that both exercise capacity (6MW) and effective PVR during exercise were significantly improved after 8 weeks of epoprostenol in a cohort of patients with idiopathic PAH, even prior to appreciable changes in resting hemodynamics, further supporting the strength of this variable (44, 45).

Similar to a recent study (28), we evaluated patients with an FVC <60% predicted that was related to significant ILD. This subgroup had significantly higher resting and exercise mPAP compared with the group with an FVC ≥60% predicted. Importantly, reanalysis after exclusion of these ILD patients (n = 9) did not change the hemodynamic response to exercise in each of the 4 categories. Interestingly, IPF is a particularly well studied ILD, with a spectrum of pulmonary complications similar to those of SSc, and in fact, PAH complicates IPF in ∼40% of patients during rest (40, 46, 47) and in 80% during exercise (40). We hypothesize that a pathophysiology similar to IPF exists within the scleroderma spectrum disorders, especially given the high prevalence of ILD in the latter; as such, earlier diagnosis and potential intervention may modify the natural history of this disease.

A common goal of clinical research studies is to develop reliable clinical decision rules that can be used to classify patients into clinically important categories. Our central hypothesis is that patients with normal exercise physiology and ePVH have a different pathophysiology from those with ePH and eoPH. We also postulate that exercise-induced PH in SSc is an abnormal hemodynamic response and part of a continuum from normal to resting PH. Since exercise right heart catheterization testing is not standardized, we developed a preliminary decision rule using resting hemodynamics that can be used by researchers to differentiate normal patients and patients with ePVH, ePH, or eoPH. We used mPAP and PVR to differentiate these groups. Our decision model shows that a resting mPAP <14 mm Hg has a 100% chance of ruling out groups with an abnormal exercise response (ePVH, ePH, and eoPH), whereas a resting mPAP >20 mm Hg has a 90% chance of predicting mPAP >30 mm Hg during exercise. Our decision rule supports the recent Dana Point classification that defined a resting mPAP of 21–24 mm Hg as borderline PAH, considering patients with this resting mPAP a population “at risk.” In addition, a PVR ≥160 dynes/seconds/cm−5 predicted eoPH and ePH. Since patients with ePH and eoPH are otherwise hemodynamically indistinguishable during exercise, they may be at risk of developing progressive pulmonary vascular disease. Importantly, this algorithm consistently separates ePVH, a condition with a potentially different therapeutic paradigm and natural history.

The pathologic mechanism of exercise-induced PH has not been well elucidated. Alveolar hypoxia can cause pulmonary vasoconstriction and is believed to play an important role in the increase in mPAP during exercise (48). Others have shown increased levels of catecholamines, vasopressin, and endothelin 1 during exercise (48–50). With regard to SSc, the vascular abnormality is dominated by the presence of both arteriolar and venular involvement, a pattern distinct from that of idiopathic PAH (51). These distinctive abnormalities may account for the different responses to exercise as well as to therapy seen in SSc patients with PAH as compared with those seen in patients with idiopathic PAH.

Systematic longitudinal followup of all of our patients will be important in determining the clinical significance of these 4 groups and the potential risk factors for progression to resting PAH. To date, 2 patients have died of acute respiratory failure, one had a baseline FVC ≤60% predicted and the other had a baseline FVC >60% predicted. Two additional patients who had ePH progressed to having resting PAH in less than 1 year.

Our study has several strengths. First, we comprehensively assessed exercise hemodynamics in a “real-life” cohort of consecutive patients with scleroderma spectrum disorders, and we included patients with ILD, a frequent finding in this population. To our knowledge, this study represents the largest experience of invasive exercise assessment in a population of patients with scleroderma spectrum disorders. Second, we developed a preliminary decision tree based on resting right heart catheterization hemodynamics that may predict eoPH or ePH.

Our study has some limitations apart from being a single-center experience. First, we did not assess exercise Doppler echocardiogram and cardiopulmonary exercise testing in our cohort, as recently done by other investigators. Second, although we developed a decision tree, it must be validated in an independent external cohort before it can be accepted in clinical practice. Third, some might argue that elevated mPAP >30 mm Hg during exercise may be a consequence of increasing age. However, we did not find any difference between younger and older age groups regarding exercise hemodynamics, providing some confidence in our results. Finally, we acknowledge the small sample size that was used to model the decision tree. Our decision tree should be considered preliminary until validated in another independent sample.

In conclusion, we have characterized 4 potential hemodynamic phenotypes of scleroderma spectrum disorders that emerge during exercise right heart catheterization. Further studies including SSc patients with borderline mPAP of 21–24 mm Hg and/or exercise physiology consistent with ePH and eoPH are warranted. To this end, we are actively enrolling SSc patients with isolated exercise-induced PH to receive open-label therapy with ambrisentan as part of a single-center clinical pilot study. Finally, our decision tree algorithm will require external validation incorporating a larger number of subjects and longitudinal followup before it can be used in clinical practice.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Rajeev Saggar had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Rajeev Saggar, Khanna, Furst, Shapiro, Chauhan, Rajan Saggar.

Acquisition of data. Rajeev Saggar, Khanna, Furst, Shapiro, Clements, Gorn, Rajan Saggar.

Analysis and interpretation of data. Rajeev Saggar, Khanna, Furst, Maranian, Belperio, Clements, Weigt, Ross, Lynch, Rajan Saggar.

Ancillary