Article Text

Download PDFPDF

Extended report
Prevalence and incidence of systemic sclerosis in southern Sweden: population-based data with case ascertainment using the 1980 ARA criteria and the proposed ACR-EULAR classification criteria
  1. K Andréasson1,
  2. T Saxne1,
  3. C Bergknut2,3,
  4. R Hesselstrand1,
  5. M Englund2,3,4
  1. 1Department of Clinical Sciences Lund, Section of Rheumatology, Lund University, Lund, Sweden
  2. 2Epi-centre Skåne, Skåne University Hospital, Lund, Sweden
  3. 3Department of Orthopedics, Clinical Sciences Lund, Lund University, Lund, Sweden
  4. 4Clinical Epidemiology Research and Training Unit, Boston University, Boston, Massachusetts, USA
  1. Correspondence to Dr Kristofer Andréasson, Department of Rheumatology, Skåne University Hospital, Lund S-221 85, Sweden; kristofer.andreasson{at}med.lu.se

Abstract

Objectives To estimate the prevalence and incidence of systemic sclerosis (SSc) in southern Sweden.

Methods In Skåne, the southernmost region of Sweden (total population 1.2 million), healthcare provided is registered in the Skåne Healthcare Register. We identified all Skåne residents who had received an International Classification of Diseases 10 diagnosis of SSc (M34) or Raynaud's phenomenon (I73.0) between 1998 and 2010. Every single case was ascertained by review of medical records in reference to the 1980 American Rheumatism Association preliminary classification criteria for SSc and the proposed American College of Rheumatology (ACR)-European League Against Rheumatism (EULAR) classification criteria presented at the ACR/Association of Rheumatology Health Professionals Annual Meeting 2012. We calculated the point prevalence by the end of 2010 by linkage with the population register to exclude deceased persons and we also estimated the mean annual cumulative incidence for 2006–2010.

Results Using the 1980 ARA criteria, the adult prevalence and annual incidence of SSc in the Skåne region were 235 and 14 per 1 million inhabitants respectively. Applying the proposed ACR-EULAR criteria, the corresponding figures were 305 and 19 per 1 million inhabitants. A majority (82%) of the prevalent cases had the limited cutaneous SSc subtype.

Conclusions The prevalence and incidence of SSc in southern Sweden, based on the 1980 ARA criteria, are higher than previously reported in northern Europe and do not support the concept of a north-south gradient of SSc occurrence in Europe. Application of the proposed ACR-EULAR classification criteria in this population results in about 30–40% higher estimates of SSc prevalence and incidence compared to the 1980 ARA criteria.

  • Systemic Sclerosis
  • Epidemiology
  • Autoimmune Diseases

Statistics from Altmetric.com

Introduction

Systemic sclerosis (SSc) is a rare chronic connective tissue disease with significant mortality and morbidity. Much of the aetiology of SSc is unknown, but both environmental and genetic factors have been proposed to contribute to the pathogenesis of the disease.1

Previous studies have reported divergent prevalence and incidence estimates of SSc in different parts of the world, with at least a 10-fold variation between various regions. Lower estimates of incidence (<10 per 1 million per year) and prevalence (<150 per 1 million) have been reported from northern Europe2–7 and Japan8 compared to the USA,9 ,10 Canada,11 southern Europe12–15 and Australia16 (an incidence of >10 per 1 million per year and prevalence estimates of >150 per 1 million). Based on these data, geographical factors have been suggested to influence the prevalence of SSc.17 ,18 A north-to-south gradient of disease occurrence has been proposed to exist in Europe.17 Population registers have been used in only a minority of the above mentioned reports and there are methodological discrepancies between the studies that may possibly have contributed to the conflicting estimates. One challenge is the lack of sensitive diagnostic criteria for SSc.19

The 1980 American Rheumatism Association (ARA) preliminary criteria for the classification of SSc are the most widely used criteria in epidemiological studies of SSc.17 ,20 An important disadvantage of this set of criteria is its inability to identify SSc subjects with limited fibrotic manifestations.19 ,21 The proposed American College of Rheumatology (ACR)- European League Against Rheumatism (EULAR) classification criteria for SSc, presented at the 2012 ACR/ Association of Rheumatology Health Professionals Annual Meeting, has been developed by a joint EULAR and ACR committee. It has been constructed to identify the complete spectrum of SSc disease, including patients with minimal or even absent skin fibrosis, and to be feasible in clinical practice. The criteria have also been developed to include vascular, immunological and fibrotic manifestations of the disease.22 In contrast to the 1980 ARA criteria, the proposed ACR-EULAR criteria have been developed in reference to a comparison population having SSc-like features.23 Preliminary reports show that these criteria have higher sensitivity and specificity compared to the 1980 ARA criteria.24

The use of population-based healthcare registers containing doctors’ diagnostic codes available in Sweden, offers the opportunity to identify patients seeking inpatient or outpatient healthcare for a particular disease.25 ,26 Complete case ascertainment by individual review of medical records is expected to provide a highly accurate and reliable estimate of the true disease occurrence.

In this study, we aimed to determine the current point prevalence and provide an estimate of the cumulative incidence of SSc in southern Sweden by applying the 1980 ARA and the proposed ACR-EULAR classification criteria.

Methods

The Swedish population register

All residents in Sweden are registered in the mandatory population register by a 10-digit personal identification number. This register includes information on sex, date of birth, residential address and, if applicable, date of death. It is continuously updated by the Swedish Tax Agency.

The Swedish healthcare system

In Sweden, healthcare is provided mainly by public and, to a certain degree, private healthcare providers to all inhabitants at a standardised low co-pay through a tax-based financing system. Access to public healthcare is high, irrespective of socioeconomic factors.27

The Skåne region

The Skåne region is located in the southernmost part of Sweden. It contains urban and rural areas. Of a total population of 1 243 329 inhabitants in December 2010, 15% were born outside the Nordic countries and 7% outside Europe, and 990 464 were above 18 years of age.28

The SSc unit at the Department of Rheumatology of the Skåne University Hospital in Lund is a referral centre for SSc and SSc-like rheumatic diseases. It was established in 1983 and has a longstanding tradition of clinical cooperation with the six other public rheumatology units in the Skåne region. For all inhabitants in the region, a rheumatology clinic is within 1 h of travel using public transportation.

The Skåne Healthcare Register

All public and private healthcare visits in this region are registered in the Skåne Healthcare Register (SHR), together with the patient's unique personal identification number, date of visit/discharge from hospital and name of healthcare provider. For public healthcare, the register also includes the patients’ confirmed and occasionally tentative diagnoses according to the International Classification of Diseases (ICD) 10 system.

Identification and case ascertainment

From the SHR database, we identified all subjects with an ICD-10 diagnosis of SSc or probable SSc (M34) as well as Raynaud's phenomenon (RP) (I73.0) between 1 January 1998 and 31 December 2010. Cases where the M34 diagnosis had been made at least once at any specialised unit of rheumatology, internal medicine, dermatology or vascular surgery in Skåne were considered eligible, together with every individual diagnosed with RP who had visited the rheumatology clinic in Lund. Patients were subject for further review by the first author (KA). This case ascertainment included but was not limited to analysis of the patient's medical records at the different rheumatology clinics in the Skåne region, as well as the clinic at which the patient had received the M34 diagnosis.

All potential SSc cases were validated in reference to the 1980 ARA and the proposed ACR-EULAR criteria. In accordance with both these criteria sets, patients with localised scleroderma were excluded, while possible overlap patients were included.20 ,24 Subjects younger than 18 years by the end of December 2010 were excluded.

The three private rheumatologists in the Skåne region were all personally contacted and asked about their experience of patients with probable SSc. All of them had received training in Lund.

Point prevalence by 31 December 2010

Estimations of point SSc prevalence were calculated from the number of subjects satisfying the two criteria sets alive and resident in Skåne in reference to a denominator population comprising all residents in Skåne aged 18 years and above.

Incidence estimates for 2006–2010

Using data from the SHR, we identified all patients who between 1 January 2006 and 31 December 2010 satisfied any SSc criteria for the first time.

Characterisation of subjects

In this study, cases were categorised into diffuse cutaneous SSc or limited cutaneous SSc (lSSc) according to LeRoy et al.29 ,30 Date of diagnosis was noted for every patient and defined as the first date when the patient satisfied any SSc criteria used. Date of disease onset was defined as the onset of the first non-RP symptom of SSc.

Any history of positive testing for antinuclear antibodies (ANA), any antibodies against extractable nuclear antigens or autoantibodies against anti-RNA polymerase III was noted.

Statistical analysis

We calculated sex-specific prevalence and incidence estimates using data from Statistics Sweden.28

Results

Point prevalence by 31 December 2010

We identified 233 patients who satisfied the 1980 ARA criteria. Application of the proposed ACR-EULAR criteria identified all these subjects and an additional 69 cases, totalling 302 SSc patients.

We recognised 434 adult subjects in the SHR with at least one clinic visit with an ICD-10 code of M34 (figure 1). Of these 434 cases, 23 patients were without further review and are considered to have been coded erroneously by the doctor; these were typically single diagnostic codes registered at an emergency clinic (n=11) or orthopaedic clinic (n=6). Of the remaining 411 eligible subjects, 116 did not satisfy any SSc criteria. Of these patients, 59 did not have any SSc-related history at all. A history of urinary tract infection (ICD-10 code N34) or back pain (ICD-10 code M54) was common among them (ie, coding errors). Another 31 patients had a medical history of probable SSc but did not meet any SSc criteria, and 26 patients had a history of localised scleroderma (eg, morphea).

Figure 1

Flow chart for identification of patients with systemic sclerosis (SSc).

We identified 231 subjects without any M34 diagnosis but with a history of RP who had been assessed at the rheumatology clinic in Lund. Of these, no one fulfilled the 1980 ARA criteria while seven fulfilled the proposed ACR-EULAR criteria.

None of the private practicing rheumatologists could recall any case of potential SSc which they had not referred to the rheumatology clinic in Lund.

Hence, based on our case ascertainment, the prevalence of SSc among adults in the Skåne region by December 2010 was 305 per 1 million, with a male to female ratio of 1 : 6.0 (table 1). The corresponding figure according to the 1980 ARA criteria only was 235 per 1 million adults, with a male to female ratio of 1 : 5.7.

Table 1

Prevalence of systemic sclerosis (SSc) in southern Sweden 2010 based on the proposed American College of Rheumatology–European League Against Rheumatism (ACR-EULAR) criteria* according to sex

Incidence

Between 1 January 2006 and 31 December 2010, 66 patients were diagnosed with SSc according to the 1980 ARA criteria and 94 patients were diagnosed according to the proposed ACR-EULAR criteria (table 2). If we applied the 1980 ARA-criteria, the annual mean cumulative incidence was 14 per 1 million inhabitants, with a male to female ratio of 1 : 3.6. The corresponding figure for the ACR-EULAR criteria was 19 per 1 million inhabitants, with a male to female ratio of 1 : 4.2.

Table 2

Annual incidence of systemic sclerosis (SSc) based on the proposed American College of Rheumatology–European League Against Rheumatism (ACR-EULAR) criteria in southern Sweden according to sex and disease subtype

Patient characteristics

A majority of the prevalent SSc patients, 249 cases (82%), had the lSSc subtype. The most common serological abnormality was positivity for ANA, which was prevalent in 262 (87%) of the 302 patients alive in 2010 (table 3). Mean (SD) age at disease onset was 48 (15) years with a mean (SD) of 4.1 (5.4) years until diagnosis. The mean (SD) age on 31 December 2010 was 61 (14) years (figure 2). Of the total 302 ascertained SSc cases, 290 (96%) had been examined by at least one of four senior doctors at the SSc unit at the rheumatology clinic in Lund.

Table 3

Immunological characteristics of systemic sclerosis (SSc) patients according to disease subtype among the subjects alive on the 31 December 2010.

Figure 2

Age and sex distribution of systemic sclerosis at point prevalence date (31 December 2010) in the Skåner region, Sweden.

Among the 233 prevalent subjects fulfilling the 1980 ARA criteria, the vast majority (n=207, 89%) satisfied the so-called major criterion, namely skin fibrosis proximal to the metacarpophalangeal joints as assessed by skin palpation (table 4). Of the additional 69 subjects who only fulfilled the proposed ACR-EULAR criteria, all had the lSSc subtype, a majority (35/69) were positive for anti-centromere antibodies (ACA) and 65/69 (94%) had reported RP at the time of diagnosis.

Table 4

Distribution of systemic sclerosis (SSc) patients in relation to what sets of criteria they satisfied

Discussion

We have studied the prevalence and incidence of SSc in Skåne, a well-defined region and population in southern Sweden by careful case ascertainment using two criteria sets. We found the disease to be more common than previously reported from northern Europe, with a prevalence and incidence of 235 and 14 per million according to the 1980 ARA criteria.2–7 These estimates are similar or higher than reports from southern Europe based on the same criteria.12–15 Our results do not support the concept that SSc is less frequent in northern Europe compared to southern Europe.17

By applying the proposed ACR-EULAR criteria, a substantially higher number of patients were classified as having SSc resulting in a prevalence and incidence estimate of 305 and 19 per million. Our study supports the contention that the proposed ACR-EULAR criteria succeeds in identifying SSc subjects which the 1980 ARA criteria fail to recognise, of which a majority were ACA+ patients with lSSc.24

From a methodological perspective, our study has several strengths compared to other European reports.2–7 12–15 The SHR covers all inhabitants in Skåne. Medical records from every eligible patient have been retrieved for classification in reference to the 1980 ARA and the proposed ACR-EULAR criteria. The latter criteria set has, to our knowledge, not been applied in a prevalence study before.

Our findings are different compared to a recent study from south-east Norway that, in addition to close geographical proximity, shares several methodological features with our study.4 However, some differences between these studies deserve to be mentioned. Most notably, the Norwegian study included subjects under 18 years of age but excluded patients with other co-existing connective tissue diseases. We agree with the committee from 1980 that a possible SSc subtype comprising SSc-overlap patients may exist but is problematic to define. Furthermore, we also agree that such a subtype would show substantial clinical heterogeneity, possibly questioning the clinical rationale of defining such a subtype.20 Still, overlap syndrome has been estimated to occur in 11% of SSc patients.31 Our inclusion of overlap patients has contributed to a higher total SSc prevalence estimate.

Correct application of the 1980 ARA criteria demands that the medical examiner has accurate knowledge of the assessment of skin thickness.32 Our study revealed that the vast majority of the SSc subjects have been examined at the referral centre in Lund by a limited number of doctors using a standard assessment routine. This fact lends support to the validity of the medical records studied regarding data on skin thickness and nailfold capillary microscopy.

Of the total population in Skåne, 15% were born outside the Nordic countries, which might be a higher proportion than in other studies from this region. Although ethnical discrepancies hardly can explain our main findings, a minor influence cannot be ruled out.

Although we report higher prevalence estimates compared to previous studies in the region, the age distribution in this study is relatively similar to other reports from northern Europe, including Iceland, Norway and England.4 ,5 ,7 However, age comparisons should be interpreted carefully. In particular, disease onset is hard to determine and has not been defined in the same way in these studies. Nevertheless, the average age at disease onset seems to be higher in the European studies compared to reports from North America.9 Furthermore, our study shows a female predominance of 1 : 6, which is less marked than in other reports but is in agreement with the reports from Norway and England.4 ,7 ,17 Of note, these studies also showed a high proportion of the lSSc disease subtype.

Our data on immunological characteristics should be interpreted with caution. These prevalence data are biased towards SSc survivors, who possibly have less severe manifestations of the disease, and against antitopoisomerase positive patients with more severe SSc. The total number of ANA-positive patients is lower than in some other studies. It is possible that this reflects local decisions regarding cut-off levels for immunology testing, as well as the fact that the majority of patients only were subject to one HEp2-based ANA assay system. Indeed, our data are fairly similar to a recent large study in Germany, where every case was subject to at least four assay systems.33

A limitation of this study is that we only made case ascertainments using the two preselected patient categories that we considered to be the most likely to hold the vast majority of true SSc cases. The availability of both general and rheumatic healthcare in the Skåne region is adequate and includes a tertiary referral centre for SSc located within the region. Thus we judge the risk that SSc subjects might have escaped appropriate diagnosis to be relatively low.27 Still, subjects whose SSc diagnosis was not recognised by any doctor have not been identified through our screening using the ICD code M34. This is especially true for patients with a primary diagnosis of, for example, systemic lupus erythematosus, dermatomyositis or Sjögren's syndrome but with a clinical picture overlapping SSc. RP is a cardinal feature of SSc that is relatively easy to recognise and is also a common secondary diagnosis in patients with other connective tissue diseases overlapping SSc. Indeed, among 231 subjects with a history of both RP and at least one clinical visit to the rheumatology clinic in Lund, nine subjects (seven of whom were alive in 2010) were identified as having SSc according to the proposed ACR-EULAR criteria.

The proposed ACR-EULAR classification criteria have been presented in 2012. They are, however, not yet finalised and published, and we cannot rule out that there may be additional steps that should be undertaken before they are applied to independent cohorts. While our data show that the new criteria do identify patients not recognised by the 1980 ARA criteria, it falls outside the scope of this report to compare the different criteria sets with respect to sensitivity and specificity. The purpose of this report is not to evaluate the performance of the proposed 2012 ACR-EULAR criteria but to investigate the epidemiology of SSc in the Skåne region using the latest available SSc classification criteria.

In conclusion, we report robust population-based estimates of prevalence and incidence of SSc in a well-defined region in Scandinavia. We suggest that SSc is not less common in northern Europe compared to southern Europe. Application of the proposed ACR-EULAR criteria resulted in substantially higher estimates of the prevalence and incidence of SSc compared to the 1980 ARA criteria.

Acknowledgments

The authors acknowledge Professor Emeritus Frank Wollheim and associate Professor Agneta Scheja for providing helpful input on the manuscript and for their dedicated and enthusiastic work establishing and managing the regional SSc unit in Lund, Sweden. We also express our sincere gratitude to Professor Janet Pope for her kind support in providing us the necessary details and for allowing us to apply the new ACR-EULAR classification criteria for SSc in our study.

References

Footnotes

  • Handling editor Tore K Kvien

  • Contributors Conception and design: KA, TS, RH and ME. Acquisition of data: CB, KA and RH. Analysis and interpretation of the data: all authors. Drafting of article: KA. Reviewing for important intellectual content: TS, CB, RH and ME. Final approval of submitted version: all authors.

  • Funding The study was supported by the Swedish Research Council, the Medical Faculty at Lund University, Skåne Region, the Swedish Rheumatism Association, King Gustaf V 80-year Fund, the Österlund Foundation and the Kock Foundation.

  • Competing interests None.

  • Ethics approval The study was approved by the Lund University Ethics Committee, Lund, Sweden. Our ethical permit does not allow us to collect, register or analyse data on ethnicity. We reviewed all medical records with the written permission from the head of the department.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.