Prevalence of pulmonary arterial hypertension in an Australian scleroderma population: screening allows for earlier diagnosis

Authors


  • Funding: This study was supported by a research grant from Actelion Pharmaceuticals Australia. E. G. has received travel support from Actelion Pharmaceuticals Australia, the manufacturers of bosentan as well as honoraria for speaking and consulting engagements, and as a member of the Actelion Advisory Board. The Heart and Lung Transplant Foundation of Western Australia, of which E. G. is chair has received educational grants from Actelion. E. G. has also received travel support from Bayer-Schering, the manufacturers of iloprost, Encysive Pharmaceuticals, the manufacturers of sitaxentan and GSK, the distributors of ambrisentan in Australia. D. P. and J. R. have received travel support from Actelion Pharmaceuticals. B. D. has received consulting fees from Actelion.

  • Conflict of interest: G. S. is an employee of Actelion Pharmaceuticals Australia.

Eli Gabbay, Advanced Lung Disease Programme, Pulmonary Hypertension Service and Lung Transplantation Unit, Royal Perth Hospital, GPO Box X2213, Perth, WA 6001, Australia.
Email: eli.gabbay@health.wa.gov.au

Abstract

Background:  We sought to determine the prevalence of pulmonary complications and especially pulmonary arterial hypertension (PAH) in an Australian scleroderma population.

Methods:  Between July 2005 and June 2007, physicians in Western Australia were asked to refer patients with scleroderma specifically for pulmonary hypertension screening. All patients were assessed for PAH and other respiratory conditions using echocardiography, lung function testing and clinical assessments. Right heart catheterization was carried out in patients with evidence of increased right ventricular systolic pressure.

Results:  Of the 184 patients analysed, 44 had possible PAH on echocardiography. Right heart catheterization confirmed the diagnosis in 24 (13%). Diffuse interstitial lung disease was found in 32 patients representing a point prevalence of 17.4%. The severity of PAH at diagnosis varied according to whether the patients were referred for screening (group A) or for diagnostic (group B) purposes. The 6-min-walk test distance and median pulmonary vascular resistance were significantly worse in group B versus group A (324 vs 402 m; P= 0.02 and 884 dynes/s per cm−5 vs 486 dynes/s per cm−5; P < 0.01, respectively).

Conclusion:  Screening may result in earlier diagnosis of PAH with, in general more mild disease. This is important, given that early treatment for PAH while patients are less symptomatic is associated with improved exercise tolerance and pulmonary haemodynamics: indices indicative of disease progression and clinical worsening.

Introduction

Scleroderma is a multisystem autoimmune connective tissue disease of unknown aetiology1 with an estimated prevalence of 19–75 cases per 100 000.2 Pulmonary complications are relatively common with reports of up to 26% of patients developing pulmonary arterial hypertension (PAH)3 and up to 40% developing ventilatory restriction, suggestive of possible interstitial lung disease (ILD).4 The prognosis is worse in these patients and with the improved management of hypertensive renal crisis, pulmonary complications have overtaken renal disease as the leading cause of death in scleroderma.5,6

In patients with scleroderma, pulmonary hypertension (PHT) can arise not only in association with PAH but also with left heart disease, hypoxic parenchymal lung disease and chronic thromboembolic pulmonary hypertension (CTEPH).2 It is important to differentiate PAH from other causes of PHT, because the treatment is different and the development of PAH per se affects outcome and survival.7–9 Patients with scleroderma-related PAH have a worse prognosis compared with patients with idiopathic PAH (iPAH).10 PAH may be asymptomatic in the early stages and when symptoms develop, they are usually non-specific (e.g. fatigue and diminished exercise tolerance), so that diagnosis is delayed until the disease is advanced and less responsive to therapy.8,11–13

With the advent of specific therapies that improve prognosis in patients with PAH, early detection and screening of at-risk patients may improve clinical outcome.6,14,15 Treatment while patients are mildly symptomatic is associated with improved exercise tolerance and pulmonary haemodynamics,16 indices that are strongly indicative of disease progression and clinical worsening in PAH.13,17,18

In Australia, there has been increasing interest in screening patients with scleroderma for PAH.19 Screening programmes for PAH have been established in other countries.20–22 However, the prevalence of PAH in scleroderma from these studies varies considerably from 4.98 to 26.7%3 and there is no universal assessment to allow accurate comparisons between studies.

We sought to determine the prevalence of pulmonary complications, especially PAH, in an Australian scleroderma population. Furthermore, we suggested that a screening programme would allow for diagnosis of PAH before it would otherwise have been recognized.

Methods

Setting

This prospective study was conducted at Royal Perth Hospital (the state referral centre for adult PHT) between 1 July 2005 and 30 June 2007. Physicians in the state of Western Australia (WA) were asked to refer patients with scleroderma specifically for PHT screening irrespective of whether they believed that the patient had symptoms suggestive of cardiopulmonary disease. If the referring clinician was specifically concerned about the possibility of PAH, they were asked to state this on the referral form. To ensure that this study reflected a statewide population, rheumatologists who attend regional centres in WA were specifically asked to refer patients to the study. Cost of travel for regional patients was reimbursed as part of a research grant.

Ethical approval was obtained from the Royal Perth Hospital Ethics Committee and informed consent given by all patients.

Patient selection and evaluation

All patients attending the clinic as a result of external referrals were reviewed independently by a senior rheumatologist. This included a thorough evaluation of skin involvement. Only patients with scleroderma, as confirmed by American Rheumatism Association (ARA) criteria, were enrolled.23 Eight patients were deemed not to meet the ARA criteria for scleroderma, including six with mixed connective tissue disease. These eight patients are not discussed further in this manuscript.

Eligible patients were divided into two groups: group A (screening referrals) were those referred specifically for PHT screening, without the referring physician believing that PAH existed. Group B (diagnostic referrals) consisted of patients referred during the same time period because the physician was specifically concerned about the possibility of PAH.

After informed consent was obtained, all patients completed a clinical assessment, including a questionnaire relating to the type and duration of scleroderma, presence of Raynaud's, smoking history and relevant respiratory history. Evidence of cardiopulmonary disease was examined. Blood was collected for autoantibodies and full blood picture. Pulmonary function tests, echocardiography and 6-min-walk tests (6MWT) were carried out on all patients, usually on the same day. Based on these results, some patients underwent high-resolution computed tomography of the thorax (HRCT-thorax), and/or right heart catheterization (RHC).

To determine a true point prevalence of scleroderma-associated PAH in our sample, we also included our existing data of living patients who were diagnosed by our service before the commencement of the study period. Furthermore, we included patients known to have ILD or other pulmonary disease before referral so that we could determine the overall prevalence of pulmonary disease in the study group.

Pulmonary function

Pulmonary function was assessed using a standardized protocol as per the American Thoracic Society Guidelines.24 Spirometry was measured using a PK Morgan Rolling Seal Spirometer (PK Morgan, Kent, UK) and lung volumes measured by body plethysmography. Transfer factor was measured by single breath diffusing capacity using a PK Morgan autolink breath system.

6MWT

Exercise capacity was assessed using the standardized 6MWT25 with a course length of 20 m. Patients were instructed to walk at their own pace while attempting to cover as much distance as possible during the allotted time. The test was carried out by one of two experienced physiotherapists.

Echocardiography

All echocardiograms were carried out using a GE Vivid 3 by a senior technician and validated by a senior cardiologist. A standard complete 2-D echocardiogram was obtained in all patients. Quantification of right ventricular dimensions was carried out according to published criteria.26 Pulmonary artery pressure (PAP) was assessed by imaging the peak velocity (v) of the tricuspid regurgitant jet (TRV) and applying the simplified Bernoulli equation (Right ventricular systolic pressure (RVSP) = 4v2+ right atrial pressure (RAP)).27 For standardization, an RAP of 10 mmHg was assumed for all patients unless features that suggested otherwise were present. Where a complete tricuspid regurgitation velocity envelope could not be seen because of insufficient tricuspid regurgitation, Levovist (Boyce Australia, Pymble, NSW, Australia) opacification contrast was used and/or 250 mL of i.v. normal saline was infused to enhance the tricuspid regurgitation envelope. Pulmonary acceleration times were measured as a sign of PHT, but were not used to estimate the pulmonary pressures.

Functional class

Functional class was determined according to the standardized World Health Organization (WHO) classification.28

RHC

Before the study commencement, it was determined that a standard RHC (at rest, supine) would be carried out in all patients with probable PHT on echocardiography. Probable PHT was defined as (i) an estimated RVSP > 40 mmHg, (ii) in symptomatic patients with an RVSP of 35–40 mmHg or (iii) where RVSP could not be determined, in the presence of echocardiographic-derived features suggestive of increased PAP (e.g. reduced pulmonary acceleration time, increased right ventricle or right atria size and/or reduced right ventricle function). In most patients where RVSP could be determined, this equated to a maximum TRV of at least 2.7 m/s or at least 2.5 m/s in the presence of symptoms.

Based on current guidelines, precapillary PHT on RHC was defined as a mean PAP (mPAP) ≥25 mmHg at rest or ≥30 mmHg during exercise, with a concomitant pulmonary capillary wedge pressure (PCWP) <15 mmHg and pulmonary vascular resistance (PVR) of >240 dynes/s per cm−5. When the mPAP was between 18 and 24 mmHg, straight leg raising was carried out to determine whether PAH could be defined on exercise alone. When PCWP exceeded 15 mmHg at rest, left heart disease was diagnosed. If the PCWP was less than 5 mmHg, then an i.v. fluid bolus was given up to a maximum of 500 mL or until the PCWP reached 5 mmHg or higher and the PAP remeasured.

Where a patient was confirmed to have precapillary PHT on RHC, subsequent investigations were carried out (e.g. V/Q scanning, HRCT-thorax, and sleep study) as per our usual protocol for patients newly diagnosed.29 Patients were confirmed to have PAH (group 1 of the WHO revised classification of PHT) once other potential causes (e.g. CTEPH) were excluded.

HRCT-thorax

To assess for ILD, HRCT-thorax was obtained in patients with restrictive physiology (haemoglobin (Hb)-adjusted DLCO≤ 75% predicted and reduction in forced vital capacity or total lung capacity to below 80% predicted), clinical signs of ILD (e.g. chest crackles), and in patients in whom precapillary PHT was confirmed on RHC. Where a patient was known to have ILD or other parenchymal lung disease and a HRCT-thorax had been obtained within 24 months, this was not repeated unless clinically indicated.

ILD was defined by the presence of characteristic appearances on HRCT-thorax in combination with either restrictive lung physiology or clinical signs.

Statistical analysis

The prevalence of PAH (confirmed by RHC) was calculated as the ratio of patients with either known or newly diagnosed PAH over the eligible sample group. For two group comparisons, a Student's t-test (parametric) or Wilcoxon rank sum test (non-parametric) for continuous variables and χ2-analysis for categorical variables was used. A Bonferroni correction was made for measurements with multiple variables and a p value of less than 0.05 was considered statistically significant.

Results

One hundred and eighty-four patients with scleroderma were assessed between July 2005 and June 2007. Figure 1 shows the screening profile and Table 1 shows patient characteristics at the time of enrolment. One hundred and seventy patients were referred for screening (group A) and 14 referred specifically because of possible PAH (group B). The reasons for referral in group B were abnormal echocardiogram in six, unexplained dyspnoea in eight, palpitations in four and syncope in one. In five patients, more than one reason was given for referral. No significant differences were observed for age, ratio of female to male patients or for the duration of scleroderma symptoms between groups. Limited cutaneous disease was diagnosed in 139 (75.5%) patients and diffuse systemic disease in 45.

Figure 1.

Screening profile of subjects with scleroderma. Includes seven patients investigated for both interstitial lung disease (ILD) and pulmonary arterial pressure (PAH). Includes nine patients known to have emphysema and six patients known to have asthma. ARA, American Rheumatism Association; HRCT, high-resolution computed tomography; PFT, pulmonary function test; 6MWT, 6-min-walk test; PHT, pulmonary hypertension; RHC, right heart catheterization.

Table 1.  Patient characteristics
 Group A (screening, n= 170)Group B (diagnostic, n= 14)Limited cutaneous disease (n= 139)Diffuse systemic disease (n= 45)
Female, n (%)148 (87.1)12 (85.7)122 (87.8)38 (84.4)
Median age, years (range)58 (30–90)64 (42–78)56 (30–90)58 (30–78)
Duration of scleroderma, months (range)124 (3–512)136 (8–492)120 (3–492)126 (7–512)

In the 184 patients, RVSP could be estimated on echocardiography in 160 (87%). After subsequent investigations, including RHC and HRCT-thorax, 24 were confirmed to have PAH, 32 ILD, 6 bronchiectasis, 6 asthma and 9 emphysema. Six patients had both mild ILD and PAH. In all the patients found to have asthma or emphysema, this diagnosis was known before the study. An additional nine patients had an increased PCWP in keeping with left heart disease. There were 104 patients without evidence of cardiac or respiratory disease.

Forty-four patients (24%) had possible PAH based on echocardiography results and were subsequently referred for RHC assessment. This consisted of 20 with an RVSP of 41–50 mmHg, 21 with an RVSP greater than 50 mmHg and 3 patients where the RSVP could not be estimated, but in whom there were additional features suggestive of increased right-sided pressures. This included enlarged right atria in two patients, additional RV dysfunction in one and reduced pulmonary acceleration time in another. Seven patients in this group were also investigated for ILD.

The screening protocol identified eight asymptomatic patients with an RVSP 35–40 mmHg, but without any other echocardiographic features suggestive of raised pulmonary pressures. None of these patients underwent RHC.

RHC findings and prevalence of PAH

PAH was confirmed on RHC in 24 subjects (13.0% of the study group). This included four patients in whom the diagnosis was made on exercise alone and three patients in whom an i.v. bolus was given. One additional patient with mild increase in pulmonary pressure (mPAP = 29 mmHg) had severe ILD (total lung capacity (TLC) = 48% predicted, PaO2= 49 mmHg on air). This patient was felt to have PHT secondary to hypoxaemic lung disease and was not considered to have PAH. Six patients with PAH had additional mild ILD (TLC > 70%, PaO2 > 72 mmHg on air), which was not considered to be a causative factor in their PHT. All other investigations, including V/Q scanning, excluded another causes of increased pulmonary pressures. Therefore, for these 24 patients, a diagnosis of scleroderma-associated PAH (WHO group 1) was made.30

A further 7 living patients were known to our service with PAH diagnosed before the commencement of the study. If these patients are included, the overall point prevalence of PAH in the scleroderma population at our centre is 16.2%.

Accuracy of echocardiography predictions

The RVSP estimates of PAH at echocardiography are presented in Table 2. In the total cohort, RVSP could not be obtained in 24 patients. Three of these patients underwent RHC and one was confirmed to have PAH. Increased PAP was not present in the other two patients.

Table 2.  RVSP estimates at echocardiography (n= 184)
 RVSP could not be determined, n (%)RVSP ≤ 40 mmHgRVSP 41–50 mmHg, n (%)RVSP > 50 mmHg, n (%)
  • Including four patients with evidence of left heart disease on echocardiography. mPAP, mean pulmonary artery pressure; NA, not applicable; PAH, pulmonary arterial hypertension (defined by mPAP > 25 mmHg with PCWP < 15 mmHg); PCWP, pulmonary capillary wedge pressure; RHC, right heart catheterization; RVSP, right ventricular systolic pressure.

n24 (13)11920 (11)21 (11)
Undergoing RHC3Nil2021
With RHC who had PAH1 (33)NA8 (40)15 (71.4)
With increased PCWPNilNA4 (20)5 (23.8)
With mPAP < 25 mmHg2 (67)NA8 (40)1 (4.8)

In individual patients an increased estimated RVSP did not always equate to increase in mPAP at RHC, which is the gold standard for the diagnosis of PAH. Eight of 20 patients with an estimated RVSP of 41–50 mmHg had PAH confirmed on RHC. Eight did not have an increased mPAP and four had increased mPAP but in association with increased PCWP compatible with left heart disease. Fifteen of 21 patients with an RVSP greater than 50 mmHg on echocardiography had PAH confirmed on RHC. Five also had an increased mPAP, but in association with increased PCWP, suggesting left heart disease. The one remaining patient did not have an increased mPAP.

In all, nine patients had increase in their PCWP compatible with left heart disease. Of these, the echocardiogram was not suggestive of left heart disease (systolic or diastolic dysfunction) in five. Subsequent left heart catheterization showed coronary artery disease in three, of whom one underwent angioplasty and stent insertion and one underwent coronary artery bypass surgery.

HRCT-thorax findings

HRCT-thorax was obtained in 44 patients who had restrictive lung physiology or crackles, including 7 patients who were also being investigated for possible PHT by RHC. All patients who were subsequently found to have precapillary PHT (i.e. PHT in the absence of raised PCWP) proceeded to HRCT-thorax as part of our protocol. Therefore, a total of 61 patients in the study underwent HRCT-thorax. Diffuse ILD was found in 32 (a point prevalence of 17.4% of the study group) and bronchiectasis (not including traction bronchiectasis as a result of ILD) in a further 6 (3.3%). Of the 32 patients with ILD, this diagnosis was already known in 21 and newly diagnosed in 11. In the 11 newly diagnosed patients, 7 were commenced on therapy with prednisolone and monthly i.v. cyclophosphamide according to a well-defined protocol.31 The decision to commence therapy was made by one or both the chief investigators.

Comparisons between patients with and without PAH

The demographics and functional characteristics of patients with and without PAH are presented in Table 3 and Table 4, respectively. Patients with PAH were significantly older than patients without PAH (67 vs 55 years, P= 0.03). There were no significant differences either in the proportion of women, or in the duration of disease in patients with or without PAH. As expected, the median 6MWT distance was greater for patients without PAH compared with those with PAH (485 vs 385 m, P < 0.05) and a higher percentage of patients with PAH were classified as having class II and III functional disease compared with patients without (67 vs 42% and 25 vs 6%, respectively). All the 104 patients without any diagnosed cardiorespiratory disease were in functional class I or II.

Table 3.  Demographics of patients with or without PAH
 Study group (n= 184)With PAH (n= 24)Without PAH (n= 160)P value
  1. NS, not significant; PAH, pulmonary arterial hypertension.

Median age, years (range)60 (30–90)67 (32–88)55 (30–90)0.03
Female, n (%)160 (87.1)21 (87.5)139 (86.8)NS
Durations of scleroderma, months (range)126 (3–512)134 (12–492)120 (3–512)NS
Table 4.  Clinical and DLCO characteristics of patients with or without pulmonary arterial hypertension (PAH)
 With PAH (n= 24)Without PAH (n= 160)Without any cardiorespiratory problem (n= 104)
  • *

    P= 0.03 versus without PAH group, P= 0.02 versus subjects without any cardiorespiratory problems.

  • **

    P= 0.05 versus without PAH group, P < 0.01 versus subjects without any cardiorespiratory problems.

  • ***

    P= 0.05 versus without PAH group, P= 0.05 versus subjects without any cardiorespiratory problems.

  • Eighty patients had one or more of the following: PAH, ILD, bronchiectasis, emphysema, asthma or left heart disease. DLCO, carbon monoxide diffusing capacity corrected for Hb; 6MWT, 6-min-walk test; PAH, pulmonary arterial hypertension.

Median 6MWT, m (range)345 (15–550)*445 (245–750)522 (265–750)
Functional class, n (%)   
 I1 (4)82 (51)76 (73)
 II**16 (67)67 (42)28 (27)
 III**6 (25)10 (6)0
 IV1 (4)1 (0.6)0
% predicted DLCO, median (range)59 (33–91)***72.4 (38–114)89.4 (73–114)
 >80% predicted2 (8.5)64 (40)59 (57)
 60–80% predicted10 (41.5)61 (38)45 (43)
 <60% predicted12 (50)35 (22)0

There was a trend (P= 0.08) towards a higher percentage of patients with the limited cutaneous scleroderma developing PAH compared with the diffuse systemic sclerosis subtype (Table 5). There was also a trend (P= 0.05), suggesting that a greater proportion of patients with diffuse systemic sclerosis subtype were at risk for the development of ILD. Further, the presence of anticentromere antibodies and anti-Scl-70 antibodies were identified as risk factors for the development of PAH and ILD, respectively.

Table 5.  Risk of developing pulmonary arterial hypertension (PAH) according to subtype of scleroderma and presence of autoantibodies
 With PAH (%)With ILD (%)
  • *

    P= 0.08 versus subjects with diffuse systemic disease with PAH,

  • **

    P= 0.05 versus subjects with diffuse systemic disease with ILD,

  • ***

    P= 0.07 versus subjects with anti-scl-70 with PAH,

  • ****

    P= 0.05 versus subjects with anti-scl-70 with ILD. Ab, antibody; ILD, interstitial lung disease.

Limited cutaneous disease (n= 139)20 (14.4)*11 (7.9)**
Diffuse systemic disease (n= 45)4 (9)21 (46.7)
Anticentromere Ab positive (n= 130)20 (15.4)***10 (7.7)****
Anti-scl-70 Ab positive (n= 38)3 (7.9)19 (50)

Comparison between subgroups with PAH

The severity of PAH at diagnosis was different in patients identified from the screening referrals (group A) compared with patients identified from the diagnostic referrals (group B), Table 6. The median 6MWT distance and median PVR was significantly worse in group B versus group A (324 vs 402 m; P= 0.02 and 884 vs 486 dynes/s per cm−5; P < 0.01, respectively). Five of 7 patients (71%) in group B were diagnosed with functional class III or IV disease compared with 2 of 17 (12%) in group A (P= 0.06). No differences were found at time of diagnosis in terms of median 6MWT distances, PVR or functional class between group B and our known cohort of patients diagnosed with PAH-related scleroderma before commencement of the study period.

Table 6.  Severity of pulmonary arterial hypertension (PAH) according to diagnostic group
 Group A (screening, n= 17)Group B (diagnostic, n= 7)
  • *

    P= 0.02 versus group A,

  • **

    P= 0.06 versus group A,

  • ***

    P < 0.01 versus group A. 6MWT, six-min-walk test; PVR, pulmonary vascular resistance.

Median 6MWT, n (range)402 (145–550)324 (15–480)*
Functional class, n (%)
 I**1 (6)0
 II**14 (82)2 (29)
 III2 (12)4 (57)
 IV01 (14)
PVR, median dynes/s per cm−5 (range)486 (325–820)884 (410–1407)***

Discussion

To our knowledge, this is the first study to examine prospectively the prevalence of PAH in an Australian scleroderma population. Of the 184 patients assessed in the study, 24 (13%) had PAH confirmed on RHC. Cox identified an 11% incidence of PAH after a systematic review of the South Australian Scleroderma Register.32 However, that was retrospective and included patients with PAH diagnosed on echocardiography alone.

The prevalence of 13% present at our study is similar to other catheter-based screening studies. Mukerjee et al. reported a prevalence of 12% on RHC in 722 scleroderma patients,20 and Ungerer et al. reported a prevalence of 16%.33 Mukerjee and Ungerer used a background population of asymptomatic patients with scleroderma, similar to our cohort. However, in a French cohort of 599 patients from 21 centres, Hachulla et al., estimated a point prevalence of systemic sclerosis-related PAH of 7.85% (95% confidence interval 5.70–10.00).21

By comparison, echocardiography-based studies may overestimate the true prevalence of PAH. Prevalence rates vary widely, ranging from 4.9 to 38.6%.3,7,8,22,34,35 This may be because of the fact that whereas echocardiography estimates RVSP, the diagnosis of PAH is based on an increase in mean PAP. Further, the diagnosis of PAH requires exclusion of other causes, some of which may not be evident on echocardiography alone. It is of note that in our study only 24 out of 44 (56%) of patients with probable PHT on echocardiography (defined predominantly by increased estimated RVSP) were confirmed to have PAH (defined by mPAP increase). Of these, an additional nine patients did have PHT, but in association with left heart disease, and another had PHT in association with severe ILD.

We believe that the prevalence figure in our study is broadly representative of the true prevalence in Australia at least of patients with scleroderma managed by specialist physicians. Although it is possible that physician referral bias resulted in a higher prevalence of PAH in our study group compared with the community as a whole, this is unlikely to be significant as physicians were encouraged to refer all their patients with scleroderma for screening, regardless of symptoms. Further, the prevalence figure as a percentage of total patients screened did not change significantly throughout the study period, suggesting that there was no bias to refer symptomatic patients earlier.

We carried out RHC on those patients who were more likely to have PAH based on their echocardiogram. It is possible that some patients who did not have a RHC may have had PAH. However, short of carrying out RHC on all patients (which we believe to be impractical and unnecessary), we used the echocardiogram as a screening tool, which is more reflective of real life. Similar protocols followed by Hachulla et al.21 and Mukerjee et al.20 suggest that this approach is unlikely to miss many, if any, patients with PAH.

Doppler echocardiography is recognized as the main screening tool for identification of patients with PAH, with subsequent confirmation required by RHC.36,37 However, the echocardiographer's experience is crucial for the accurate identification of the tricuspid regurgitation (TR) jet. In our study, adequate TR velocity was available to enable estimation of RVSP in 160 of 184 patients (87%). This concurs with Hachulla, who reported adequate TR jets in 81.5% of patients.21 However, we found one patient with PAH where RVSP could not be estimated, although other echocardiographic features of PHT confirm that an inadequate TRV does not exclude PAH.

With regard to optimal echocardiographic screening thresholds to detect PAH, a RVSP > 40 mmHg appears to be reasonable cut-off for RHC referral. Indeed, PAH was confirmed on RHC in 23 of 41 patients (56%) with a RVSP > 40 mmHg. This included 8 of 20 subjects (40%) with a RVSP between 40 and 50 mmHg. These findings are supported by previous research, which suggests that a Doppler echocardiographic threshold of <40 mmHg versus ≥40 mmHg has a positive predictive value of 92% and a negative predictive value of 44%.3,38

Nine patients were found to have left heart disease with increased PCWP despite this not being identified on echocardiography assessment in five patients. The clinical diagnosis of patients with heart failure symptoms despite normal ejection fraction remains challenging and mitral flow Doppler alone, with its 40–70% specificity, cannot reliably detect diastolic dysfunction in this cohort of patients.39–41

We found that a transfer factor (Hb corrected) of <60% is not specific for PAH, but is specific for lung or pulmonary vascular disease (e.g. all patients with <60% predicted had some form of pulmonary parenchymal or vascular disease (PAH, ILD, emphysema, bronchiectasis)). Approximately 50% of patients with vascular or parenchymal lung disease had a DLCO between 60 and 80%, and two patients with PAH had a DLCO > 80%. Therefore, a normal DLCO cannot be used to exclude PAH. These results are consistent with those of Mukerjee et al. who found DLCO < 55% not to be useful in identifying patients with early pulmonary vasculopathy associated with scleroderma.38

PAH can present in both sexes of all ages.42,43 For the patients identified in our study the median age was 67 years (range 32–88 years), significantly older than those without PAH (P= 0.03). These figures are similar to screening studies by Hachulla et al. (65 years ± 11.7),21 Mukerjee et al. (66 years ± 7),20 Wigley et al. (62.4 ± 13.0 and 55.5 ± 13.1 years in retrospective and prospective groups, respectively),3 and a body of clinical trial data.12,15,44 Other studies have reported associations between PAH and late age of disease onset.45 The older age of PAH onset in patients with scleroderma may promote complacency among clinicians. However, it indicates the need for a high index of suspicion and routine screening of all patients to permit early diagnosis.22 Our study also confirms the female predominance of scleroderma-related PAH: 87.5% of patients identified were women, similar to the 88.9% described in the French screening study.21 This reflects the female predominance among scleroderma patients rather than suggesting that female gender per se is a risk factor for developing PAH in patients with scleroderma.

The duration of scleroderma in our study was not indicative of PAH development: one patient developed PAH within 12 months of the diagnosis of scleroderma. The published work is divided on this topic with some reports suggesting a positive association between the duration of scleroderma32 and others finding no association.22 Our study suggests that any patient with scleroderma may present at any stage in the course of their disease with PAH, with or without ILD.19,46

This is in contrast to the previously held belief that patients with limited scleroderma develop PAH and patients with diffuse scleroderma develop ILD.9 Whereas patients with limited scleroderma may be at some increased risk of developing PAH compared with those with diffuse scleroderma, it is clear from our study and others, that all patients with scleroderma are at risk of developing PAH.20–22 Further, the presence of antibody subtype does not adequately discriminate for the risk of development of PAH or ILD.

We defined patients as having ILD if they had compatible HRCT-thorax features, as well as either restrictive physiology or clinical signs. This is similar to an approach that we have previously used for ILD complicating rheumatoid arthritis and is a more rigorous criterion for the definition of ILD to better reflect disease that is probably clinically significant to warrant close monitoring and/or intervention.47 The point prevalence of ILD was 17.4% and the screening programme identified 11 new patients, seven of whom have been initiated on therapy. These figures are in concordance with those of Chang et al., who, in a cohort of 619 patients with scleroderma, identified restrictive ventilatory defects in 139 (22.5%).46 Additionally, we identified six patients with bronchiectasis, perhaps reflecting aspiration in association with gastro-oesophageal reflux and gut dysmotility, although this was not formally assessed.

We separated patients in our study according to their mode of referral. We did this so that we could directly examine the effects of a screening programme compared with the existing referral pattern. The haemodynamics at diagnosis for patients in group B were not significantly different from the haemodynamics at diagnosis of patients with scleroderma-associated PAH already known to our centre before commencement of the study. This is consistent with our view that patients from group B probably would have been referred for investigation regardless of whether a screening programme existed.

However, the screening programme identified 17 out of 170 patients (group A; 10%) with PAH where the referring physician was not particularly concerned about a diagnosis of PAH. Although it was not specifically assessed, it is reasonable to conclude that in some of these patients, the reduction in exercise capacity was either not assessed or attributed to other causes. Therefore, we believe that without a screening programme these patients would not have been identified at that time. Those patients diagnosed with PAH from group A had, in general, better haemodynamics, functional class and exercise capacity than those in group B, suggesting that screening may lead to earlier diagnosis.

Treatment while patients are mildly symptomatic is associated with improved exercise tolerance and pulmonary haemodynamics,16 indices that are strongly indicative of disease progression and clinical worsening in PAH.13,17,18 Further, early detection of PAH and a multidisciplinary approach to diagnosis and treatment in specialized PAH centres improve clinical outcomes.6,14,15 For example, Williams et al. reported that earlier treatment was associated with better outcomes and that the risk of death increased by 11% for every 10 mmHg in mPAP at diagnosis.15

Conclusions

The present study has several limitations. This is an observational study and therefore has no comparative control. However, it was prospectively designed to differentiate subjects screened versus those referred for diagnostic assessment. Probably, not all patients in the state were screened. Some patients are managed solely by their family physician without referral to specialist centres. This is probably a small proportion because of the complexity and relative rarity of the disease. Finally, we did not examine the cost-effectiveness of our screening programme.

In summary, this is the first Australian study to assess the prevalence of PAH in a state-based scleroderma population. Using a prospective screening protocol, we found a prevalence of PAH to be 13% in the study group. Screening resulted in an earlier diagnosis of PAH. It is imperative that RHC is carried out to confirm PAH and exclude left heart disease as echocardiography alone may overestimate the true prevalence of PAH. Finally, whereas measures of lung function are a useful part of a screening protocol, measurement of DLCO on its own is not specific for PAH, but DLCO < 60% is highly specific for respiratory abnormalities.

Acknowledgements

The authors thank Ms Corina Jary and Ms Melanie Murrell (Royal Perth Hospital) for their assistance in coordinating the scleroderma assessment clinic.

Ancillary