The Johns Hopkins University Institutional Review Board approved this study. Consecutive patients with SSc diagnosed with PAH by RHC and evaluated at the Johns Hopkins Pulmonary Hypertension (PH) Program between January 1, 2000 and March 31, 2009 were included. The diagnosis of SSc was based on specific criteria as previously reported (7) and all diagnoses were confirmed by rheumatologists with expertise in SSc (F.M.W. and L.K.H.). Date of onset of scleroderma was defined as the date of first non-Raynaud symptom attributable to scleroderma.

The selection process is summarized in Figure 1. Out of the 242 patients with SSc referred to the PH Clinic for suspected PH, 210 patients underwent RHC. A diagnosis of PH (defined as a resting mean pulmonary artery pressure [mPAP] ≥ 25 mm Hg) (13) was confirmed in 185 patients. Twenty-eight patients were excluded due to the presence of high pulmonary capillary wedge pressure (>15 mm Hg) indicating pulmonary venous hypertension. Patients with other potential causes of PH, including significant chronic obstructive pulmonary disease and interstitial lung disease (ILD), were excluded. The criteria for significant chronic obstructive pulmonary disease or ILD were indicated in previous studies (7, 10, 14) and are detailed in the online supplement. Patients were also excluded if they had been previously treated with active drugs for PAH.

Figure 1. Algorithm of selection of patients. Highlighted box indicates patients included. Patients with overlap syndrome (n = 15) included 6 with polymyositis/dermatomyositis, 3 with systemic lupus erythematosus, 1 with rheumatoid arthritis, 1 with Sjögren syndrome, 3 with mixed connective tissue disease, and 1 with combined rheumatoid arthritis and Sjögren syndrome. Patients excluded with other overlapping diagnoses (n = 3) included one patient with acute pulmonary embolism, one with sarcoidosis, and one with severe restrictive disease due to obesity. * Refers to patients already on pulmonary arterial hypertension (PAH)-specific therapy before diagnostic catheterization. COPD = chronic obstructive pulmonary disease; ILD = interstitial lung disease; mPAP = mean pulmonary artery pressure; PCWP = pulmonary capillary wedge pressure; PH = pulmonary hypertension; PVH = pulmonary venous hypertension; PVR = pulmonary vascular resistance; SSc = systemic sclerosis.

Download Figure | Download Powerpoint

[More] [Minimize]

Hemodynamic measurements also included stroke volume index (SVI), pulmonary arterial capacitance (SV/PP) calculated from stroke volume (SV) divided by the pulmonary artery pulse pressure, and right ventricle stroke work index as calculated by SVI × (mPAP − RAP) × 0.0136 (15).

Echocardiogram reports, serum autoantibodies, pulmonary function tests results, computerized tomography, and World Health Organization functional classification (WHO FC), were obtained from clinical records. Renal function was assessed by the estimated glomerular filtration rate (eGFR) using the Modified Diet in Renal Disease equation (16). Renal dysfunction was defined as eGFR less than 60 ml/min/1.73 m².

Survival time was considered starting on the date of diagnosis of PAH by RHC. Death was determined from the clinical and hospital records as well as the Social Security Death Index. All causes of death were considered for survival analysis.

Treatment was established in accordance with the standards of practice of the Johns Hopkins PH Program and based on the availability of different treatments across the years. The effect of period-specific treatment on overall outcomes was analyzed.

Continuous variables are shown as mean ± standard deviation (SD) or median (quartile [Q]1–Q3 range). Group comparisons were made using Student t test or Wilcoxon rank test as appropriate for continuous variables, and χ² statistics or Fisher exact test, as appropriate for categorical variables. A P value less than 0.05 was considered significant.

Survival analysis was performed using the Kaplan-Meier method. Comparisons between groups were assessed by log-rank test. Univariable Cox and multivariable proportional hazards modeling were performed to determine the variables associated with mortality. All computations were performed using Stata statistical software (version 10.1; Stata, College Station, TX).

As described in Methods and Figure 1, 185 out of 210 patients with SSc who underwent RHC had PH. Thirty-three patients were excluded based on the presence of significant ILD as defined by a TLC <60% or TLC between 60 and 70% with the presence of significant radiological abnormalities as detailed in the online supplement.

Baseline characteristics are shown in Table 1. Most patients were white (86.8%) and women (84.2%). The median time of duration of SSc at diagnosis was 10.8 years (range, 0–37 yr), whereas symptoms of Raynaud phenomenon preceded the diagnosis of PAH for a median time of 15 years (range, 0.2–49 yr). The diagnoses of SSc and PAH were established in the same year in 17 patients.

SSc-related autoantibodies were found in 57 out of 63 patients who had completed studies, including detection of antinuclear, anticentromere, and antitopoisomerase antibodies. Anticentromere autoantibodies were present in 33 patients, 32 of them with limited disease, and were the predominant antibodies in female compared with male patients (57.4% vs. 22.2%; P = 0.07). Antitopoisomerase antibodies were present in four patients, all women, all with limited disease. Antinucleolar antibodies were present in 15 (23.8%) patients and were the predominant antibodies in African Americans compared with whites (50% vs. 19%; P = 0.07). There were also three patients with antiRNP antibodies, one patient with antiRNA polymerase III, one patient with both anticentromere and antinucleolar antibodies, and six patients with positive undefined antinuclear antibodies.

Echocardiographic and hemodynamic data are shown in Table 2. Results from baseline echocardiography were available in 65 patients (85%). Forty-four patients (71%) had evidence of RV dilation, and 23 (35%) had evidence of pericardial effusion. Fifteen out of 50 patients (30%) had evidence of nonsystolic dysfunction of the left ventricle. Estimated left ventricular systolic function was normal (mean left ventricular ejection fraction, 60 ± 6%). Traditional hemodynamic measurements indicated moderate-to-severe PAH (mean RAP, 8 ± 4 mm Hg; mPAP, 42 ± 11 mm Hg; cardiac index, 2.4 ± 0.7 L/min/m²; and PVR, 8.6 ± 5.6 Wood units). Mean stroke volume index (31 ± 10 ml/m²) and SV/PP (1.47 ± 0.84 ml/mm Hg) were similarly depressed.

WHO FC was I or II for 36 patients (47.4%) and III or IV for 40 patients (52.6%) at time of diagnosis (Table 1). The proportion of patients diagnosed in early stages of PAH varied over the years (Table 3). The proportion of patients in WHO FC I or II at diagnosis was 5.6%, 55.6%, and 38.9% during the years 2000–2002, 2003–2005, and 2006–2009, respectively (P = 0.02).

Serum creatinine levels within 3 months of RHC were available in 68 patients. Based on eGFR, 31 (45.6%) patients had renal dysfunction at the time of diagnosis. Patients with renal dysfunction were older compared with those with normal renal function (64.6 vs. 58.9 yr; P = 0.03) and tended to have diffuse SSc subtype (19.3% vs. 8.1%; P = 0.28) and systemic hypertension (45.2 vs. 24.3%; P = 0.08). Only 2 of the 31 patients with renal dysfunction (6.5%) had a documented previous history of renal crisis. The eGFR was significantly but weakly associated with several baseline hemodynamic parameters, with the following Spearman correlations: PVR, −0.27; P = 0.02; cardiac index, 0.28; P = 0.02; SVI, 0.39; P < 0.01; SV/PP, 0.33; P < 0.01; and mixed venous blood oxygen saturation (Sv_O2), 0.40; P < 0.01. The correlations with RAP and mPAP were not significant. The proportion of renal dysfunction was higher in patients with WHO FC III to IV (58 vs. 31.2%; P = 0.02). The proportion of patients taking diuretics at the time of the analysis was 62.1 and 44.4% for patients with and without renal dysfunction, respectively (P = 0.16).

Thirty-eight (64.4%) patients were already receiving or were started on calcium channel blockers at low doses (i.e., < 60 mg of nifedipine daily) during follow-up due to different conditions: systemic hypertension, cardiac arrhythmia, or Raynaud symptoms. At least 60 patients (88.2%) received diuretics during follow-up.

Sixty-nine (90.8%) patients received PAH-specific therapy after RHC. Initial treatment consisted of intravenous prostacyclin in 8 patients (11.6%), endothelin receptor antagonists in 26 (37.7%), phosphodiesterase 5 inhibitors (PDE5-I) in 34 (49.3%), and calcium channel blockers at high dose in 1 (1.4%), all as monotherapy. As shown in Table 3, initial therapy varied across the years (P < 0.01). At the end of follow-up, 5 patients were on prostanoids alone (7.2%), 10 were on endothelin receptor antagonists alone (14.5%), 19 on PDE5-I alone (27.5%), and 35 patients were receiving combined therapy (50.7%).

The median follow-up time was 36 months (3 yr), with a maximum follow-up of 9.4 years. The overall median survival time was 4.02 years and there were 42 deaths observed. Four patients were lost to follow-up. The cause of death could be determined in 33 patients and included right heart failure (24 patients), sudden cardiac death (1), lung cancer (2), sepsis/multisystem failure (2), postoperative complications of bowel obstruction (1), motor vehicle accident (2), and lower gastrointestinal bleeding (1). There were no deaths in the group of patients with WHO FC I.

The Kaplan-Meier survival curves are shown in Figure 2. Overall survival was 85%, 72%, 67%, 50%, and 36% at 1, 2, 3, 4, and 5 years from diagnosis, respectively. As expected, survival was worse in patients with WHO FC III to IV compared with I to II (log rank test: P = 0.02), as shown in Figure 2. The median survival time varied by WHO FC, with a median survival for FC IV patients shorter than 10 months.

Figure 2. Survival curves for all patients and by World Health Organization functional class (WHO FC).

Download Figure | Download Powerpoint

[More] [Minimize]

Univariable and multivariable analysis were performed using the first 5 years of follow-up (with 36 deaths observed during that period), as survival curves show a relatively small group of patients at risk after that period. Hazard ratios (HRs) for the univariable analyses are shown in Table 4. WHO FC, cardiac index, PVR, SVI, SV/PP, pulmonary artery oxygen saturation, and eGFR were significant predictors in the univariable analysis. Neither the year of diagnosis nor the duration of scleroderma before PAH diagnosis was a significant predictor for mortality.

Several multivariable analysis models were built using a single hemodynamic parameter in each model and adjusting for age, WHO FC, and eGFR. These variables were included because they have been shown to be significant predictors of survival in other studies. Initial treatment was not included as this did not necessarily reflect the treatment received during the follow-up period. Echocardiographic measurements were not introduced in multivariate analysis due to statistical limitations with the sample size and number of missing values. Multivariable analysis models are shown in Table 5. Although PVR, SVI, SV/PP, and pulmonary artery oxygen saturation were significant predictors in the adjusted models, cardiac index, RAP, and mPAP were not. eGFR was a strong predictor across all the models. Neither WHO FC nor age was significant in the multivariable models, and the addition of these adjusting variables did not result in a higher prediction value after comparing the various models by the likelihood ratio test. A posterior comparison between the different models was performed using the Akaike Information Criterion. The models including SVI, PVR, or SVPP had a better prediction.

Although PVR and SV/PP did not follow a normal distribution, the analysis using log-transformed variables did not change the significance of the results. The proportional hazards assumption was confirmed for each of the models. The survival curves for SVI, PVR, and SV/PP, dichotomized using the respective median values, are shown in Figure 3 along with the corresponding HRs and significance. The dichotomization by median values shows the following HRs for mortality: HR, 3.13 (95% CI, 1.50–6.52; P < 0.01) for PVR greater than 7.2 Wood units; HR, 2.34 (95% CI, 1.11–4.96; P = 0.03) for SVI less than 30 ml; and HR, 3.06 (95% CI, 1.41–6.65; P < 0.01) for SV/PP less than 1.25 ml/min/mm Hg.

Figure 3. Survival curves by median values of pulmonary vascular resistance (PVR), stroke volume index (SVI), and pulmonary artery capacitance (SV/PP).

Download Figure | Download Powerpoint

[More] [Minimize]

In summary, we have shown that measures of RV function (SVI), pulmonary vascular compliance (SV/PP), and pulmonary vascular resistance (PVR), rather than traditional hemodynamic measures commonly used in PAH (e.g., cardiac index, mPAP, and RAP), were independently associated with survival in this cohort of patients with SSc-PAH. These findings suggesting that the pathophysiology of PAH and RV response in SSc may differ from that of other forms of PAH should be confirmed in larger prospective studies.

Furthermore, survival remains unacceptably poor, with 1-, 3-, and 5-year rates of 85%, 67%, and 36%, respectively, despite the use of PAH-specific therapy. Future studies should focus on identifying mechanisms of RV dysfunction to increasing afterload in SSc and on possible direct effects of SSc on the heart and pulmonary vasculature.