Systemic sclerosis (SSc, scleroderma) is an autoimmune connective tissue disease characterised by vascular lesions, induration and thickening of the skin, and fibrotic changes in other organs, predominantly in the lungs, heart, intestinal tract, kidneys, muscles and joints.1 2 The leading cause of death in SSc is pulmonary involvement consisting of pulmonary fibrosis with or without pulmonary hypertension (PH) or pulmonary arterial hypertension (PAH).3^–5 Published estimates of the prevalence of SSc vary widely, depending on the observation period, methods used and geographical area and it is reported to be from 0.4 to 469 per 100 000 people.2 6^–13 Furthermore, the prevalence of PAH in SSc is unclear, but is estimated to be 8% to 12% in studies based on right-heart catheterisation, the gold standard for diagnosing PAH.3 14 Several treatments that successfully delay the process of PAH, such as epoprostenol,15 trepostinil,16 bosentan,17 18 sitaxentan19 and sildenafil,20 have recently become available. In addition, two studies have shown that cyclophosphamide has a significant although clinically minor effect on pulmonary fibrosis in SSc patients.5 21 In light of these treatment possibilities and given the late occurrence of symptoms of eventually life-threatening pulmonary involvement in SSc, screening for both PAH and pulmonary fibrosis has become a major goal in scleroderma referral centres.

To date, the optimal screening method for PAH has not been determined. A combination of decreased carbon monoxide diffusion capacity and elevated systolic pulmonary arterial pressure (PAP) estimated by echocardiography is the method used in most clinics.3 22 23 A decreased decreased carbon monoxide diffusion capacity is, however, also present in other diseases limiting oxygen uptake and echocardiography is flawed by both false-positive and false-negative results. Right-heart catheterisation is the gold standard for establishing the diagnosis of PAH.24 25 In order prospectively to investigate the optimal screening method for PAH in SSc the POEMAS study was started in 2005. POEMAS, an investigator initiated study, is a Dutch acronym for “Pulmonale hypertensie Opsporing Een Multidisciplinaire Aanpak bij Sclerodermie” translated as “Pulmonary hypertension screening, a multidisciplinary approach in scleroderma”, and intends to establish the optimal screening method for PAH in SSc. For this purpose, patients with SSc, without PAH or interstitial lung disease at inclusion, will prospectively be studied for 5 years. In order to find eligible patients for the POEMAS study, all patients with SSc treated in the participating hospitals were screened and documented according to our protocol. Their data were entered into the POEMAS registry.

The objectives of this study are to determine the point prevalence and incidence of SSc in The Netherlands and to establish the prevalence of PAH and interstitial lung disease in the same population.

PATIENTS AND METHODS

Prevalence of patients with SSc in The Netherlands

In The Netherlands, patients with SSc are referred by the family practitioner to either rheumatologists or internists, mostly clinical immunologists. All rheumatologists and internists use a computerised diagnosis system for all patients, which is designed to calculate costs for health insurance. Occasionally, referral is to dermatologists, who in their turn refer patients with SSc to rheumatologists. To determine the prevalence and incidence of SSc and the associated PAH and interstitial lung disease in The Netherlands, cases were found using two sources: first, the POEMAS registry, and second, a nationwide questionnaire among rheumatologists and clinical immunologists in The Netherlands. As a denominator the total population in The Netherlands aged 18 years or older on 1 January 2007 was used, as provided by the Statistics Netherlands registry (www.cbs.nl).

Fifteen hospitals currently participate in the POEMAS study, including both academic (n  =  5) and non-academic hospitals (n  =  10). All patients 18 years and older referred to these hospitals and diagnosed with SSc according to the preliminary classification criteria of the American College of Rheumatology26 and the LeRoy criteria for early SSc27 in the period 1 January 1 2005 to 1 August 2007 were included in the registry. Demographic and clinical characteristics of all patients were obtained from their clinical records.

To obtain information on the number of patients treated in other hospitals in The Netherlands, a questionnaire was sent to all registered rheumatologists and clinical immunologists, excluding the hospitals participating in the POEMAS study. If the questionnaire was not returned, a second was sent and the doctors were personally requested by telephone to return the survey. In the questionnaire they were asked to provide the number of all alive SSc patients aged 18 years and older and the number of SSc patients with PAH, as diagnosed by right-heart catheterisation, currently treated in their practice. The source of the information for the questionnaire was the computerised diagnosis lists available in all Dutch hospitals.

POEMAS study population

Demographic and clinical information on the patients are shown in table 1. Interstitial lung disease was determined by pulmonary function tests and high-resolution computer tomography (HRCT) scan. For the assessment of signs of interstitial lung disease on the HRCT scan the same method was used as in a recent, large idiopathic pulmonary fibrosis trial.5 In short, this scoring method consists of the evaluation of the HRCT scan at three levels: the midarch of the aorta, the main carina and 1 cm above the diaphragm. Both fibrosis and ground glass opacity (GGO) were graded from 0 to more than 75% of the pulmonary surface, providing a score between 0 and 5. Left and right lung scores were averaged, resulting in a total GGO score and a fibrosis score, both between 0 and 15.28 29 The patients were divided into two groups: one with HRCT scan scores of 3 or less for GGO and fibrosis, representing minimal or absent interstitial pulmonary involvement and another with more extensive interstitial involvement. Furthermore, all patients were classified as having a total lung capacity (TLC) of 70% or greater or less than 70% of predicted.22 30 Systolic PAP was calculated by applying the simplified Bernoulli equation on maximum tricuspid regurgitation velocity and adding the estimated right atrial pressure. Right atrial pressure was estimated using inferior caval vein diameter and the effects of respiratory variation.31 PAH was suspected when on echocardiography the estimated systolic PAP was 35 mm Hg or greater at rest in patients without a history of pulmonary embolism or other possible causes of PH and with a normal TLC. In all cases with an elevated systolic PAP on echocardiography right-heart catheterisation was performed, and when the mean PAP was greater than 25 mm Hg at rest with a normal wedge pressure, a diagnosis of PAH was established.22 25 32

Some patients were registered twice in the database due to referral to an academic centre or moving house. Those patients were recognised using date of birth and postal code and the duplicates were removed.

Statistical analysis

All statistical analyses were performed using SPSS version 14.0. To calculate the prevalence of SSc, the number of patients from the POEMAS registry was combined with the results from the survey. The response to the survey was 70%; thus for calculations the number of patients not included in the POEMAS registry was estimated to be 100/70 times the result of the questionnaires. Adding the POEMAS registry number to the estimated number in the other hospitals provided the total number of patients in The Netherlands. To provide an estimate of the prevalence, this number was divided by the number of inhabitants aged 18 years or older in The Netherlands. The percentage of patients in The Netherlands with SSc included in the POEMAS registry was calculated. To establish the incidence, the number of new cases in 2005 and 2006 in the POEMAS database was determined, generalised to the total population, using the calculated percentage of patients registered in the POEMAS database, and then divided by the total Dutch population. For determination of the prevalence of PAH, both the POEMAS registry and the questionnaires were used. For the presence of interstitial lung disease, information from the POEMAS registry was the sole source. The prevalence of the latter complication was calculated by generalising the POEMAS registry to the total population in The Netherlands. Quantitative variables were summarised as median and full range and were compared using the t test. Groups were compared using the Mann–Witney U test. All statistical analyses were two-tailed and p values less than 0.05 were considered statistically significant.

RESULTS

In the POEMAS registry 654 SSc patients have been included up to 1 August 2007. Eighty-two and 18% of the patients were included in the academic and non-academic hospitals, respectively. The study population consisted of 493 (75.4%) women and 161 (24.6%) men. The median (range) age of the patients with SSc was 59.1 years (19–90). The median age (range) at diagnosis of SSc was 50.0 years (3–86). Antinuclear antibodies were detected in 91.7% of the SSc patients, and the frequency of other disease-specific autoantibodies as well as the demographic description of the patients is shown in table 1.

Prevalence of SSc

The response to the questionnaire was 70%. The response rates of doctors working in academic and non-academic hospitals were similar. Therefore the questionnaire was considered to be representative. Through the questionnaire, another 346 SSc cases were recognised. When combining the results of the survey and the POEMAS study, the estimated number of SSc patients in The Netherlands in 2005 and 2006 is 1148. The total Dutch population aged 18 years and older on 1 January 2007 was 12 793 440. Using this as the denominator population, the estimated prevalence of SSc is 8.9 per 100 000 inhabitants (table 2).

Incidence of SSc

In the hospitals participating in the POEMAS study, 58 new SSc patients were diagnosed in 2005 and 53 in 2006. As 57.0% of the SSc patients in The Netherlands were calculated to be registered in the POEMAS study (654/(100/70*346 + 654)*100), the estimated annual number of new scleroderma patients is 103 in 2005 and 94 in 2006. Generalising this number to the denominator population, the estimated incidence of SSc in The Netherlands is 0.77 per 100 000 inhabitants aged 18 years or older per year (table 2).

Pulmonary hypertension

The presence of PH, either PAH or PH associated with pulmonary fibrosis, was reported in 11.6% of the SSc patients in the survey. In the POEMAS registry PH was prevalent in 56 (8.6%) SSc patients. Combining these results, the estimated prevalence of PH in SSc in The Netherlands is 9.9% (table 2). SSc patients with PH had a significantly lower TLC compared with patients without PH, namely 70.5% (35–132%) of predicted and 91.0% (32–137%) of predicted, respectively (p<0.01). Also, the median (range) age of patients with SSc and PH was significantly higher than the patients without PH: 67.2 years (31.6–84.7) versus 58.5 years (19.6–90.0). No differences regarding PH were found in gender or autoantibodies. When using a TLC less than 70% of predicted as a diagnosis of interstitial lung disease, this cohort consisted of 35 (61.4%) PAH and 22 (38.6%) PH patients.

Interstitial pulmonary disease

Pulmonary function tests were performed in 388 (77.9%) SSc patients in the last 6 months before registration in the POEMAS database. The median (range) TLC in SSc was 90.0% (32–152%) of predicted. Figure 1 shows the TLC values as a percentage of predicted in SSc patients. The TLC as a percentage of predicted was 70% or greater in 81.0% of patients. To date, HRCT scanning has been performed and scored according to the described protocol in 292 SSc patients. A score of 3 or less for GGO and fibrosis, representing minimal or absent interstitial pulmonary involvement, was found in 53.0% of the HRCT scans of the SSc patients. In SSc patients without pulmonary involvement according to the HRCT scan, anticentromere antibodies were present more frequently than in patients with a GGO or fibrosis score greater than 3 (p<0.01). No differences were found between patients in the normal and abnormal HRCT scan groups regarding age, disease duration, other autoantibodies or sex. The median (range) TLC in patients with an HRCT scan score of 3 or less was 97.0% (46–137%) of predicted, which differed significantly from patients with higher HRCT scan scores, namely 71.5% (35–134%) of predicted (p = 0.01). When an HRCT scan score of more than 3 is used to define pulmonary interstitial disease, this complication is present in 47.0% (139/296) of the patients. If a TLC less than 70% of predicted is used to define interstitial lung disease, this complication is present in 18.8% (91/485) of the patients. The HRCT scores correlated significantly with the TLC as a percentage of predicted, r  =  0.527 (p<0.01).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Total lung capacity as a percentage of predicted in systemic sclerosis. TLR, total lung capacity.

DISCUSSION

This study on the prevalence and incidence of SSc is based on the largest European population of SSc patients described to date. We found an estimated prevalence of SSc in The Netherlands of 8.9 per 100 000 adult persons based on 1000 cases. This finding is comparable with recent findings in the United Kingdom, where an estimated prevalence of 8.2 per 100 000 has been published.7 The prevalence in France and Greece is reported to be 15.8 and 15.4 per 100 000 persons, respectively,6 10 whereas the prevalence in the United States and Australia is reported to be higher, 24.2 and 23 per 100 000 persons, respectively.12 33 The differences in prevalence in different countries can be explained by several factors, such as the criteria used to diagnose patients, the preliminary classification criteria of the American College of Rheumatology26 or the LeRoy criteria for early SSc,27 or the group size.11 12 Furthermore, differences in ethnic background may lead to a higher prevalence of SSc in the United States and Australia.8 12 33 For the identification of SSc patients in the POEMAS registry we used both classification criteria,26 27 which allowed us to include patients with early disease and limited cutaneous involvement without interstitial lung disease.

The POEMAS study is performed by doctors with a special interest in SSc and PH. To avoid an overestimation of the prevalence of SSc, we included a questionnaire to all other rheumatologists and clinical immunologists in The Netherlands. The responses are considered to be a reflection of the total population of The Netherlands. A proportion of patients with SSc is not yet referred to the rheumatologist but treated only in primary care or by other specialists such as dermatologists or vascular surgeons, but in our opinion this is a small minority of the total number of patients. By combining the results of the questionnaire and the POEMAS registry we were able to identify 87% of the SSc patients, and therefore the estimated prevalence presented in this study is robust.

The incidence of SSc was calculated using only the POEMAS study population, of which the date of diagnosis is known in every case. The calculated incidence of 0.77 per 100 000 is between the reported incidence in the United States (1.9 per 100 000), Greece (1.1 per 100 000) and the United Kingdom (0.4 per 100 000).6 12 34 As our estimated incidence is based on a study period of only 2 years and the number of incident cases is somewhat smaller than the US and Greek cohorts, this figure should be interpreted with care and should be confirmed in the second part of the POEMAS study, with an observational period of 5 years.

The prevalence of PH in this study is a combined prevalence of PAH and PH associated with pulmonary fibrosis. As PH associated with fibrosis is a hypoxia-driven disease rarely occurring in SSc, probably all patients with PH in this study can be categorised as PAH. We found a prevalence of PH in SSc of 9.9%, which is lower than expected.35^–37 This may be that patients with PH associated with SSc are treated by pulmonologist or cardiologists, sometimes without recognising the systemic disease, as can be the case in every country. Another reason could be that screening for PH by annual echocardiography has not yet been incorporated into standard care in all hospitals. In the POEMAS study this will be performed twice a year, providing us with the opportunity to study further the prevalence and incidence of PH in the scleroderma spectrum of diseases.

The POEMAS registry provides detailed information on 654 SSc patients. The percentage of patients with a TLC of 70% or greater in SSc was 81.2%. The expected percentage of interstitial lung disease in this connective tissue disease is much higher, but may differ depending on the definition and methods used. Earlier studies reported pulmonary fibrosis in 35–53% of SSc patients.37 Apparently, a TLC of 70% or greater does not exclude interstitial lung disease. Combining TLC with HRCT scan findings is, however, expected to be a more sensitive method for this purpose; the POEMAS study will provide further information.

In summary, we have described the epidemiology of SSc in The Netherlands. Our findings reveal that the prevalence and incidence of SSc in The Netherlands is similar to the data from the United Kingdom. The prevalence of PH and interstitial pulmonary disease is, however, lower than expected. With the POEMAS study in progress, additional information on epidemiology, PAH and interstitial lung disease is pending.