Article Text
Abstract
Background Patients with normal (mean pulmonary arterial pressure (mPAP) ≤20 mm Hg) and borderline mean pulmonary pressures (21–24 mm Hg) are “at risk” of developing pulmonary hypertension (PH). The objectives of this analysis were to examine the baseline characteristics in systemic sclerosis (SSc) with normal and borderline mPAP and to explore long-term outcomes in SSc patients with borderline mPAP versus normal haemodynamics.
Methods PHAROS is a multicentre prospective longitudinal cohort of patients with SSc “at risk” or recently diagnosed with resting PH on right heart catheterisation (RHC). Baseline clinical characteristics, pulmonary function tests, high-resolution CT, 2-dimensional echocardiogram and RHC results were analysed in normal and borderline mPAP groups.
Results 206 patients underwent RHC (results showed 35 normal, 28 borderline mPAP, 143 resting PH). There were no differences in the baseline demographics. Patients in the borderline mPAP group were more likely to have restrictive lung disease (67% vs 30%), fibrosis on high-resolution CT and a higher estimated right ventricular systolic pressure on echocardiogram (46.3 vs 36.2 mm Hg; p<0.05) than patients with normal haemodynamics. RHC revealed higher pulmonary vascular resistance and more elevated mPAP on exercise (≥30; 88% vs 56%) in the borderline mPAP group (p<0.05 for both). Patients were followed for a mean of 25.7 months and 24 patients had a repeat RHC during this period. During follow-up, 55% of the borderline mPAP group and 32% of the normal group developed resting PH (p=NS).
Conclusions Patients with borderline mPAP have a greater prevalence of abnormal lung physiology, pulmonary fibrosis and the presence of exercise mPAP ≥30 mm Hg.
Statistics from Altmetric.com
Pulmonary hypertension (PH) and interstitial lung disease (ILD) are the leading causes of mortality in patients with systemic sclerosis (SSc).1 The recently concluded 4th World Symposium on Pulmonary Hypertension reclassified patients with resting mean pulmonary arterial pressure (mPAP) of 20 mm Hg or less as ‘normal pulmonary haemodynamics’ and 21–24 mm Hg as the ‘borderline’ group2 on right heart catheterisation (RHC). This distinction was based on a systematic review of 47 studies that measured the resting mPAP of healthy volunteers. In this review, the normal mean (SD) resting mPAP was 14 mm Hg (3.3) and the upper limit of normal was 20.6 mm Hg.3 Although it is well established that resting PH (mPAP ≥25) in SSc is associated with poor prognosis,1 ,4,–,6 data regarding the natural history and outcomes in patients with normal haemodynamics and borderline mPAP are lacking.7 ,8
The Pulmonary Hypertension Assessment and Recognition of Outcomes in Scleroderma (PHAROS) is a prospective longitudinal study that includes patients with SSc “at risk” of developing PH and those who have newly diagnosed resting PH. The study is being conducted in the USA with a goal of discovering risk factors for PH and defining the course of disease progression in patients with established pulmonary vascular disease. The objectives of this analysis were to examine the baseline demographics and clinical features in patients with SSc with normal resting haemodynamics (mPAP ≤20 mm Hg) versus borderline mPAP and to explore long-term outcomes in SSc patients with normal versus borderline mPAP resting haemodynamics.
Methods and materials
Patients
PHAROS9 is a multicentre study enrolling SSc patients who met the American College of Rheumatology classification criteria for definite SSc10 or the LeRoy definition11 of limited cutaneous or diffuse cutaneous SSc.9 The study was approved by the review board at each institution and each patient signed the voluntary consent form before participating in the study. PHAROS included two patient groups. The first group was patients “at risk” of developing pulmonary arterial hypertension (PAH). Entry criteria for patients “at risk” of PH were any one of the following three criteria:
Carbon monoxide diffusing capacity (DLCO) less than 55% predicted without severe ILD (defined as forced vital capacity (FVC) <65% predicted and/or a thoracic high-resolution CT scan of the lungs with moderate to severe ILD); or
FVC%/DLCO% ratio of 1.6 or greater; or
Estimated right ventricular systolic pressure (RVSP) greater than 35 mm Hg on echocardiogram with Doppler.
The second patient group was resting PH patients enrolled within 6 months of RHC. Patients with resting PH were excluded if the left ventricular ejection fraction was less than 50% or the PH was non-SSc related. Exercise RHC was performed per local institutional protocols.12 ,13 No patients were enrolled based on exercise RHC data alone.
Resting PH was divided into three groups: group 1 (PAH) was defined as a mPAP of 25 mm Hg or greater with a pulmonary capillary wedge pressure (PCWP) of 15 mm Hg or less on RHC, with no significant pulmonary interstitial fibrosis (FVC ≥65% predicted and none to mild ILD on high-resolution CT); group 2 (pulmonary venous hypertension; PVH) was defined as an mPAP of 25 mm Hg or greater with a PCWP greater than 15 mm Hg on RHC; and group 3 (PH–ILD) was defined as a mPAP of 25 mm Hg or greater on RHC, PCWP of 15 mm Hg or less and significant ILD (FVC <65% predicted and/or moderate to severe ILD on high-resolution CT). Patients with no resting PH were further divided into normal pulmonary haemodynamics (mPAP <20 mm Hg) and borderline mPAP (21–24 mm Hg).
The baseline demographics collected at the time of enrolment included clinical history, SSc subtype, disease duration from first non-Raynaud's symptom, medications and smoking history. Autoantibodies were measured at the local laboratory. The patients completed questionnaires including the scleroderma health assessment questionnaire, the University of California at San Diego dyspnoea index, and the short form 36 (SF-36) at baseline. Patients also had a baseline physician evaluation, which included the modified Rodnan skin score. Pulmonary function tests, echocardiograms, high-resolution CT and 6-min walk test were encouraged in all patients and were completed in the majority of patients. High-resolution CT fibrosis was graded as normal (no fibrosis), mild, moderate or severe fibrosis by the local radiologist. The RHC were performed based on the clinical judgement of the treating physician. Medical history, hospitalisations, medication information and outcome events were recorded, and individual investigators independently initiated PAH-specific therapy when indicated.
All data were collected using paper case report forms and were manually entered into a central computerised database. For quality control, the authors contacted sites to confirm any outlying data values.
Statistical analysis
We compared patients with normal versus borderline mPAP using the two-sample t test for normally distributed continuous data, the Wilcoxon rank sum test for non-parametric continuous data and the χ2 test for categorical data. As the focus of the paper is to compare borderline mPAP versus normal groups, no adjustment for multiple comparisons was done. The survival rates of the two groups were analysed using a Cox proportional hazards model. The proportionality assumption was tested by introducing the interaction of log time and the independent variable group as a time-varying covariate. Analyses were performed using STATA 10.
Results
The PHAROS registry enrolled 322 patients from July 2005 to April 2010, of whom 206 patients underwent haemodynamic assessment; 177 RHC were performed during the initial visit and 31 were performed during follow-up visits. Of these RHC, 143 had resting PH (mPAP ≥25), 35 (56%) had normal pulmonary haemodynamics (mPAP ≤20 mm Hg) and 28 (44%) had borderline mPAP (21–24 mm Hg; figure 1).
Baseline characteristics
Baseline characteristics of the whole group
The average age for the whole group was 57.2 years, mean (SD) disease duration from the onset of Raynaud's phenomenon was 11.6 years (10.3), 85% were women and 63% had limited cutaneous SSc. There were no statistically significant differences in the demographics of normal versus borderline mPAP groups (table1).
At baseline, patients with resting PH had numerically higher (worse) symptoms compared with the normal or borderline mPAP group measured by both health-related quality of life and dyspnoea instruments (health assessment questionnaire – disease index, visual analogue scale breathing, visual analogue scale overall health, University of California at San Diego dyspnoea and SF-36 physical component summary; table 1) and by the New York Heart Association classification (table 2). No difference in disease duration, modified Rodnan skin score, autoantibody profile, or degree of fibrosis on high-resolution CT was detected between resting PH versus normal or borderline mPAP. However, the resting PH group had numerically worse DLCO% predicted, 6-min walk test and higher RVSP on 2-dimensional echocardiogram compared with patients with mPAP less than 25 (normal and borderline mPAP; table 2). Resting RHC showed numerically higher PCWP, cardiac output, pulmonary vascular resistance (PVR) and transpulmonary gradient (TPG = mPAP − PCWP) in patients with mPAP of 25 or greater compared with mPAP less than 25.
Borderline versus normal haemodynamics
The borderline mPAP group had a lower mean FVC% compared with the normal group (69.3% vs 81.9%, p<0.05; table 2). Although more patients in the borderline mPAP group had fibrosis on the high-resolution CT (80.0% vs 62.5%), this was not statistically significant (p>0.05). Patients with borderline mPAP had statistically higher estimated RVSP on 2-dimensional echocardiogram (46.3 vs 36.2 mm Hg, p<0.001). On RHC, the mean mPAP was higher in the borderline mPAP group (23.6 vs 16.5 mm Hg). PCWP was also significantly higher in the borderline mPAP group (10.2 vs 7.9 mm Hg) including one patient in each group with a PCWP of 16 mm Hg or greater (16 and 17 mm Hg, respectively). None of them had echocardiographic evidence of systolic or diastolic dysfunction. The PVR and TPG were significantly higher in the borderline mPAP group (p<0.05 for both). Other results are presented in table 2.
Baseline characteristics of patients undergoing exercise RHC
Thirty-four of 63 (53.9%) of the mPAP less than 25 group underwent exercise RHC performed on the basis of local standards.12,–,14 On exercise, borderline mPAP patients were more likely to have a mPAP of 30 or greater than the normal group; 14/16 (88%) versus 10/18 (56%); p=0.04. The mPAP values on exercise were significantly elevated in the borderline mPAP group compared with normal (39 vs 30 mm Hg, p<0.05). No significant differences were found between the two groups in PCWP, cardiac output or PVR (table 2). Thirty-three per cent of the normal group versus 44% with borderline mPAP had PCWP of 18 mm Hg or greater during exercise (p=NS).
Baseline characteristics excluding patients with moderate to severe ILD
We explored further the impact of moderate to severe ILD on normal versus borderline mPAP groups. We excluded 26 patients with either moderate to severe fibrosis on high-resolution CT or an FVC less than 65% predicted (figure 1). Twenty-nine patients, 21 (72%) of the normal, and 8 (28%) of the borderline mPAP groups were included (table 3); others were excluded because of missing FVC and high-resolution CT data. On echocardiogram, the mean (SD) RVSP in the borderline mPAP group was 45.0 mm Hg (10.8) compared with 37.0 mm Hg (9.5) in the normal group (p=0.14). Resting haemodynamics showed significantly higher mPAP and PVR in the borderline mPAP versus the normal group (p<0.05; table 3).
Eighteen of these 21 patients underwent exercise RHC (table 3). The mean (SD) value of mPAP for the borderline mPAP group was higher than the normal group; 42 mm Hg (5) versus 33 mm Hg (7; p=0.02). No other exercise haemodynamics reached a significant difference (table 3).
Follow-up data on patients with borderline and normal haemodynamics
As of January 2011, 24 (38%) patients (13 from the normal group, 11 from the borderline mPAP group) underwent repeat haemodynamic evaluation (figure 2). Patients had repeat RHC predominantly due to progressive unexplained dyspnoea and was left at the discretion of the investigators. There were no predefined criteria for repeat RHC. The mean (SD) follow-up period was 25.7 months (16.4), and the mean time between the initial and follow-up RHC was 13.67 months (8.16; borderline mPAP vs normal NS; table 4).
In the normal group, 4 patients (32% of the 13 who had a repeat RHC) developed resting PH during follow-up (table 4, figure 2). In contrast, repeat RHC in the borderline mPAP group revealed that 6 patients (55% of 11 with repeat RHC) developed resting PH. The mean (SD) time to developing resting PAH was 20.0 months (10.5) in the normal group and 17.4 months (17.0) in the borderline mPAP group (N=3 in both groups, p=0.83). The mean time to developing resting PH was 17.10 months (10.59) in the normal group and 18.85 months (10.95) in the borderline mPAP group (p=NS). Four of the 10 patients with exercise mPAP of 30 mm Hg or greater in the normal group had a repeat exercise haemodynamic evaluation; none developed resting PH during follow-up, 2 had exercise mPAP less than 30 mm Hg, and 2 continued to achieve exercise mPAP of 30 mm Hg or greater. Of the 14 borderline mPAP patients with exercise mPAP of 30 mm Hg or greater at baseline, 7 had repeat RHC during follow-up and 2 developed resting PH.
Complications including death
Follow-up data on patients with borderline and normal haemodynamics
PH-related complications were also investigated during the follow-up period (figure 2). In the normal group, 2 patients underwent lung transplantation for severe underlying ILD and 2 patients developed right-sided heart failure, including one patient who died (the 2 patients with right-sided heart failure both showed resting PAH on repeat RHC). Non PH-related major complications in the normal group included left-sided heart failure, myelodysplastic syndrome and metastatic lung cancer (the patient who developed lung cancer eventually died). In the borderline mPAP group, there was one death due to resting PH that was associated with severe ILD. An additional death was due to progressive pulmonary fibrosis, and another due to bacterial infection after lung transplantation.
Treatment
Follow-up data on patients with borderline and normal haemodynamics
Eighteen patients in the borderline mPAP group and 11 in the normal group were treated with PH-specific medications (64% vs 31%; p=0.009). Among the 28 patients initially classified as borderline mPAP, 6 were receiving treatment for newly developed resting PH, 11 for either previous or newly diagnosed exercise PH and one for severe Raynaud's phenomenon. Of the 11 patients in the normal group who were receiving PH-specific treatments, 3 were for newly developed PH and 5 were treated for previous or newly diagnosed exercise PH. Other indications for initiating PH medications in the normal group included severe Raynaud's phenomenon (n=2) and plexiform lesions seen on open lung biopsy (n=1).
Patients in the borderline mPAP group had a trend towards a worse survival (HR 1.80, 95% CI 0.40 to 8.05) compared with the normal group (p=0.44) but this was not statistically significant. The 3-year survival was 87% for the normal group versus 83% for the borderline mPAP group.
Discussion
PAH is now the leading cause of SSc-related mortality.1 ,4,–,6 Observational studies in SSc have focused on patients with definite PH, and the natural course and prognosis is well documented. Risk factors for developing PAH in SSc include limited cutaneous SSc, older age at disease onset, severity and duration of Raynaud's phenomenon, elevated estimated RVSP on echocardiogram, a decreased DLCO or a progressive decline of DLCO, and an increased FVC%/DLCO% predicted ratio greater than 1.6.15,–,19
Historical survival in SSc–PAH has ranged between 40% and 50% at 2 years after diagnosis of PAH.15 ,20 Survival has improved somewhat during the past decade and ranges between 47% and 56% at 3 years, probably due to early screening and the availability of PAH-specific therapies.8 ,21 ,22 This recent modest improvement in survival raises the question of whether early diagnosis/treatment of PAH or treatment of pre-PAH will improve long-term outcomes. However, we do not know if borderline elevations in pulmonary artery pressures on RHC are predictive of future PAH. In this regard, a borderline mPAP group on RHC may predict clinically relevant PAH.3
We assessed the baseline characteristics, morbidity and mortality in patients with normal and borderline mPAP in a large observational cohort of patients “at risk” of developing SSc–PH. When we compared patients who had resting PH with patients with mPAP less than 25 mm Hg, we found patients with resting PH had worse symptoms, lower DLCO% predicted, higher estimated RVSP on echocardiogram, and higher PCWP, TPG, PVR on baseline RHC. Among patients with mPAP less than 25 mm Hg we found that patients with borderline mPAP had greater evidence of restrictive lung disease, higher estimated RSVP on echocardiogram, and a higher PVR and TPG on resting RHC than patients with normal haemodynamics. In those patients who underwent exercise haemodynamic assessments, 88% in the borderline mPAP group versus 56% in the normal group had evidence of elevated exercise mPAP (≥30 mm Hg; p=0.04). Baseline demographics did not differentiate the normal versus the borderline mPAP groups. When we excluded patients with moderate to severe ILD (FVC ≤65% predicted and/or high-resolution CT chest with moderate to severe fibrosis), we still found higher RVSP on echocardiogram and elevated PVR and TPG on RHC in borderline mPAP versus normal group.
Previous studies have explored the association between elevated resting mPAP on RHC in SSc and elevated exercise mPAP. Baseline resting mPAP between 19 and 21 mm Hg were associated with exercise mPAP of 30 mm Hg or greater in three studies.12 ,13 ,23 Exercise physiology on RHC is a focus of intense debate in the PH literature. In a systematic review, Kovacs and colleagues3 showed that submaximal exercise during RHC in healthy volunteers (based on 10 studies, N=193) led to an increased mPAP of greater than 30 mm Hg in approximately 20% of subjects under 50 years of age, and in nearly half aged 50 years or older. This led to the removal of exercise PH from the definition of PAH.2 However, exercise mPAP of 30 mm Hg or greater may be an important intermediate step in “at risk” populations.12 ,13 ,23,–,26 This is supported by a longitudinal study that followed 42 patients with SSc-related elevated exercise mPAP. Nineteen per cent of these patients developed resting PAH after a mean (SD) time of 30 months (16), and of those, four (9.5%) patients died due to PH-related complications within 3 years.21 In the current study, a greater percentage of patients with borderline mPAP also had exercise mPAP of 30 mm Hg or greater compared with the normal group at baseline (88% vs 56%, p=0.04). Longitudinal follow-up on 14 borderline mPAP patients with elevated exercise mPAP disclosed the development of resting PH in 2 patients (18%). In contrast, none of the 4 patients in the normal group with exercise mPAP of 30 mm Hg or greater who had a repeat RHC demonstrated newly developed resting PH. Abnormal exercise haemodynamic profiles in SSc may represent an abnormal haemodynamic phenotype, which is part of a continuum from normal to resting PH.
The mean follow-up period for our patient group was 25.7 months, during which 13 and 11 patients in the borderline mPAP and normal groups, respectively underwent repeat RHC. Fifty-five per cent of the borderline group and 32% of the normal group developed resting PH (p=0.41), although more in the borderline mPAP group had PVH or PH–ILD than in the normal group. Another study by Schreiber and colleagues,27 presented as an abstract, also assessed their prospective cohort of patients with borderline mPAP and normal haemodynamics. During follow-up at 5 years, 58% of the borderline mPAP group versus 30% of the normal group had progressed to resting PH, and PVR greater than 200 dyn/s/cm−5 was an independent predictor of progression to resting PH. Our study also supports the significance of PVR as there is a statistically significant difference in PVR between the normal (137 dyn/s/cm−5 at rest) and borderline mPAP group (210 dyn/s/cm−5 at rest) at baseline resting RHC, and the significance is maintained after the exclusion of patients with moderate to severe ILD.
There were 7 deaths (4 in the borderline mPAP group and 3 in the normal group). Of these, 2 borderline mPAP and one normal death were due to PH-related complications (HR 1.80, p=NS). As the majority of the patients were recruited in an era when exercise PH was part of the definition of PH, these patients were treated with PAH-specific therapies. This is exemplified by our data in which a significantly higher proportion of patients in the borderline mPAP group (11/14 or 78.6%) with exercise mPAP of 30 mm Hg or greater were receiving PAH-specific treatments (4/10 for normal group). We noted an increased frequency of resting PH and increased mortality in the borderline mPAP group, but these numbers may have been underestimated due to concomitant PAH therapies.
Our study has significant strengths. First, this was a longitudinal study in SSc patients to describe the differences between patients with borderline mPAP and normal haemodynamics including both baseline and follow-up RHC. Second, our study is a multicentre study that involves 16 scleroderma centres. Finally, we followed a ‘real-life’ cohort of patients with SSc, and included patients with ILD, a frequent finding in this population. Our data were robust after excluding patients with moderate to severe ILD.
Our study is not without limitations. First, given the design of the study as an observational cohort, a heterogenous population and missing data were unavoidable. Second, longitudinal follow-up, diagnostic (such as repeat RHC) and treatment decisions were based on the discretion of the treating physician and thus limited homogeneity. Third, it was not feasible to study the natural history of borderline mPAP and normal because patients, particularly those with exercise mPAP of 30 mm Hg or greater were treated with PH-specific therapy based on a previous definition of PAH that included an exercise component. Although we do not endorse the use of these medications without the diagnosis of resting PAH, this reflects real-life practices. Future studies may include further subgroup analyses on patients with/without ILD or right ventricular dysfunction based on predictive parameters such as NT-proBNP. Furthermore, identifying the role of exercise pulmonary haemodynamics in the evaluation of pulmonary vascular disease and developing a standardised approach for when and how to perform exercise pulmonary haemodynamics, in the context of an evidence-based definition, is clearly needed.
In conclusion, this is a prospective observational study that separates SSc patients without resting PH into borderline mPAP and normal groups on the basis of RHC, describing the baseline characteristics and follow-up data. Patients with borderline mPAP have a greater prevalence of abnormal lung physiology, pulmonary fibrosis and presence of exercise mPAP of 30 mm Hg or greater compared with patients with mPAP of 20 mm Hg or less. Further longitudinal studies will be needed to confirm these findings and to validate the importance of identifying and prognosticating borderline mPAP patients.
Acknowledgments
The authors wish to thank the coorindators at each centre for their dedication to the registry.
References
Footnotes
Funding The PHAROS registry is funded by unrestricted grant from Actelion Inc and Gilead Inc and the Scleroderma Foundation. DK was supported by a National Institutes of Health award (NIAMS K23 AR053858-05).
Ethics approval The study received ethics approval from the review board at each institution.
Patient consent Obtained.
Competing interests SB, RS, MBB, RD, AG, MH, VH, JM, NN and PM have no conflicts of interest to declare; LC reports being involved in clinical trials with Gilead, United Therapeutics and Pfizer; MEC reports serving on the advisory board meeting on PAH and rheumatology by Gilead; CD has received research grants from Aspreva and Gilead; AF reports being a speaker at Actelion and Gilead, serving as a consultant, advisory board member at Actelion; LH reports research support from Actelion and Medimmune; MDM reports consulting, speakers bureaux and/or received grant/research support from Actelion, Gilead, United Therapeutics and Novartis; VMcL reports receiving research grants from Actelion, Novartis, United Therapeutics, acting as a consultant for Actelion, BMS, Mondo Biotech, Gilead, United Therapeutics and as a speaker's bureau for Actelion, Gilead and United Therapeutics; DEF reports honoraria, speaking, receiving research grants, and serving on the advisory board of Actelion and Gilead, grants and advisory board of Pfizer; VS reports research grants from United Therapeutics, Pfizer, Gilead, Actelion, Bristol Myers Squibb and Sibley foundation, acting as a consultant to United Therapeutics and Gilead, and as a speaker at Gilead and Actelion; DK reports serving on the advisory board of Actelion and Pfizer and received grants from Actelion and Gilead.
Provenance and peer review Not commissioned; externally peer reviewed.