Predicting Mortality in Systemic Sclerosis

Abstract

Introduction Abbreviations used in this article: ACA, anticentromere autoantibodies, anti-topo I, autoantibodies to DNA-topoisomerase I, CSR, cumulative survival rates, DLCO, carbon monoxide diffusion capacity of the lungs, ECG, electrocardiogram, ESR, erythrocyte sedimentation rate, MCP, metacarpophalangeal joints, NCM, nailfold capillary microscopy, RA, rheumatoid arthritis, SLE, systemic lupus erythematosus, SMR, standardized mortality ratios, SSc, systemic sclerosis Systemic sclerosis (SSc), or scleroderma, is a systemic autoimmune disease characterized by microvascular involvement, fibroblast activation and excessive production of collagen with resulting skin and visceral fibrosis, and the presence of specific autoantibodies. The skin, lungs, gastrointestinal system, heart, and kidneys are major SSc targets. Raynaud phenomenon is present in most patients. The microvascular damage is reflected in nailfold capillary changes readily identified by nailfold capillary microscopy (NCM), including capillary loss and dilatations (19,29). Autoantibodies to centromeres (ACA) and to DNA-topoisomerase I (anti-topo I) are also strongly associated with SSc (47,51). SSc has a worldwide distribution (42). For example, SSc populations have been characterized in patients from the United States (3,43), France (14,33,48), Japan (22,39), Australia (4), and in English Canadians (25). Genetic background modulates the expression of SSc, particularly autoantibodies (21,38). French Canadians are the descendants of approximately 10,000 emigrants from France who settled in Canada between 1608 and 1760 (24). Due to a combination of historic factors, including the 1760 Conquest of New France by England, and the implementation of English rule in 1763, which led to cessation of immigration from France, French Canadians developed relatively in isolation well into the 20th century. Co-assimilation between English and French Canadians did not occur. The resulting relative genetic homogeneity confined in the same geographic environment makes it of interest to study expression of genetically modulated diseases (5,15). Therefore we and others have been studying systemic autoimmune diseases in this population (13,35,41,49). Since, to our knowledge, no detailed study has reported thus far on SSc in French Canadians, the first objective of this report was to describe SSc features at diagnosis in this population. Mortality and factors predictive for decreased survival in patients with SSc have been the focus of interest for over 30 years (32). However, review of the literature revealed limitations in data interpretation due to small sample size, exclusion of certain patients with SSc, lack of uniform definitions for subsets, less than optimal date of entry into study, lack of use of multivariate analysis to identify factors predictive for mortality, and omission of standardized mortality ratios (SMR) to assess the mortality burden due to SSc in comparison with the general population. Relatively few studies have avoided these pitfalls (2,7,18,25). Therefore, the second objective of this report was to use multivariate analysis to identify at diagnosis factors predictive for mortality in our large cohort of well-defined patients encompassing the full spectrum of SSc. We were particularly interested in defining the predictive value of ACA, anti-topo I, and NCM. We also wished to determine SMR in our patient population, and to compare our data with other major series. Patients and Methods Design and patients A prospective cohort design was used. Between 7 April 1984 and 22 September 1999, SSc was diagnosed at first presentation in 309 new, consecutive, and unselected French Canadian patients at the Connective Tissue Diseases and Vascular Medicine Clinics of Notre-Dame Hospital in Montreal, a tertiary care academic center and a major French Canadian hospital founded in 1880. The clinical diagnosis of SSc was made without awareness of NCM, ACA, and anti-topo I results. Patients were classified as French Canadians if their 4 grandparents were French Canadians. SSc subsets Patients were classified as fulfilling American College of Rheumatology (ACR) 1980 preliminary SSc criteria or not, based on the presence or absence of the major criterion, that is, sclerodermatous skin involvement proximal to metacarpophalangeal (MCP) joints (including facial skin thickening), or 2 or more of the minor criteria: 1) sclerodactyly, 2) digital pitting scars or loss of substance of the distal finger pad, and 3) bibasilar pulmonary fibrosis (46). Because ACR SSc criteria exclude some patients who have received a diagnosis of definitive SSc from experienced clinicians (12,36,40,53), patients were also categorized into 1 of 4 SSc subsets using a slightly modified version of the definitions described by Barnett et al (4), Giordano et al (12), and Ferri et al (11) and based on the extent of sclerodermatous skin involvement, as shown in Figure 1: diffuse (must include trunk involvement); intermediate (must include upper extremities proximal to MCP joints without trunk involvement); limited (skin involvement restricted to sclerodactyly, plus Raynaud phenomenon, plus the presence in most patients of at least a third CREST manifestation, that is, calcinosis, SSc-type esophageal involvement, and clinically visible capillary telangiectasias); and normal skin. Inclusion in the normal skin group required the absence of any sclerodermatous skin involvement, the presence of Raynaud phenomenon, plus at least 1 of the following: peripheral vascular manifestation (clinically visible capillary telangiectasias of the hands, mucosal membranes, and/or face, clinically visible nailfold capillaries, digital tip ulcers or pitting scars, loss of distal finger pad) and/or visceral manifestation (as defined below under target organ involvement) (12,36). This classification was selected because its 4 mutually exclusive subsets reflect better the clinical heterogeneity of SSc and may be prognostically superior to the traditional subdivisions into diffuse and limited subsets (26). Patients with overlap syndrome features, mixed connective tissue disease, or anti-RNP autoantibodies are excluded in this classification. Raynaud phenomenon was defined by at least 2 of 3 phases of color changes (white, blue, red), usually induced by cold exposure, and involving at least 1 finger of both upper extremities.Fig. 1: Classification of patients with systemic sclerosis into 4 subsets based on extent of sclerodermatous skin involvement. Involved skin is indicated in black.Nailfold capillaroscopy NCM was performed according to a standardized protocol as previously described (19,28,52). All digits of both hands were examined by a trained research nurse who recorded observations, followed by verification by 1 of the authors (FJ or AR). Interobserver agreement was high (kappa = 0.95). All observers were blinded to the diagnosis, subset classification, and disease duration. They were not blinded to SSc signs of the hands. Observation was performed with an SR stereomicroscope (Carl Zeiss Inc, Montreal) at 8 to 50 magnification using a cool source of illumination and equipped with an eyepiece incorporating a 19.7 μ microscale allowing precise and reproducible measurements of capillary dilatation (28). The following were recorded: degree of capillary dilatation (0 = normal; 1 = borderline, <2 × normal diameter; 2 = definitely dilated, ≥2 × but ≤4 × normal diameter; 3 = extremely dilated, >4 × normal diameter), and avascular areas (A = no capillary loss; B = rare avascular areas; C = moderate capillary loss; D = extensive capillary loss). The capillary origin of clinically visible telangiectasias of the hands was confirmed by NCM, that is, venular telangiectasias were excluded. In our patient population, the of 2 or 3 capillary dilatations for SSc from to as in Raynaud = systemic lupus = and rheumatoid = and were selected because Raynaud the same the of C or D avascular areas for SSc from to at diagnosis A detailed and were performed according to a SSc including a modified skin based on the degree on a of of skin erythrocyte sedimentation and of the hands and pulmonary and carbon monoxide diffusion capacity or or and A was performed in = of and and were performed in = of as clinically were by on Inc, as described previously I were by as using as a anti-topo I from the for The of normal patients with RA, and patients with for target organ involvement were according to et al with SSc changes in the distal of or or of the esophageal bibasilar fibrosis on of normal or syndrome on and to SSc (including or and or SSc with and/or and peripheral and and analysis were on a standardized visceral and NCM protocol encompassing and into an The of SSc was defined as the date of of the first SSc The disease at diagnosis was the between the date of of the first SSc and the first analysis was performed for subsets the and the was used for comparison of group using and were to and cumulative survival were using the The between present at and mortality were using The were identified by using determine if the of in our SSc cohort was to the general population, SMR were using and mortality in the of for the from The SMR was as the of the from in our cohort to the of The for the SMR were by the of as a and its from based on distribution data The or was by 3 patient were to determine the Patients were classified as if were followed on or For patients in the of recorded on the and report was the of patients to were by to in a the of their patient if the date and of was from the to of the and of of the 309 cohort patients were from the the 309 cohort 50 were diagnosed as normal skin, as as and as diffuse SSc The ACR SSc criteria were by patients diffuse and intermediate SSc patients the ACR major criterion, and of the normal skin patients ACR In the limited patients at least 2 ACR minor criteria is, plus at least 1 other minor The exclusion of limited SSc patients by ACR SSc criteria has been 1: at diagnosis in 309 patients with systemic cohort to was shown in the ratios were in the normal skin in the in the and in the diffuse subsets, The at disease was years for years for = The at disease was in The at diagnosis was years for years for The disease at diagnosis was in than in years = The disease at diagnosis were and 1 in the 4 SSc subsets, from normal to This was in the diffuse intermediate or limited features 2 the subsets in to the extent of skin involvement and the of target organ involvement at Raynaud phenomenon was the most manifestation of SSc, in of patients in the normal skin, and intermediate However, the of Raynaud decreased to in patients with diffuse SSc, hands were more The skin was from in the normal skin to and in the other subsets, involvement was in of patients. involvement and/or on at diagnosis was with from the normal skin to diffuse subsets, that is, in and of the SSc subsets, involvement shown by in and of intermediate and diffuse SSc patient with but this was in 7 patients 1 of involvement at diagnosis was in of and more in the intermediate and diffuse subsets = and were most in patients with intermediate and diffuse and target organ involvement at diagnosis in 309 patients with systemic and anti-topo I autoantibodies 3 that the of decreased from in patients with normal skin and limited SSc to and in the intermediate and diffuse subsets, is that the of was in limited intermediate SSc = and in intermediate diffuse SSc = The of anti-topo I without was in limited intermediate and diffuse SSc In our anti-topo I the for was defined as 8 the in the at the did not the or distribution of anti-topo I in patients without skin limited intermediate diffuse SSc the to 4 the was not associated with a of anti-topo I in diffuse SSc and decreased the of the skin limited intermediate and diffuse SSc The of and anti-topo I was in of as reported (ACA) and I (anti-topo I) autoantibodies in 309 patients with systemic of anti-topo I autoantibodies with pulmonary fibrosis patients with pulmonary fibrosis, = anti-topo In of patients without pulmonary fibrosis, = anti-topo I = of nailfold capillaroscopy for SSc 4 that the of moderate or extensive capillary loss C or at diagnosis was in the normal skin in the limited and in the other subsets The of NCM at diagnosis for the subsets from to with an of of nailfold capillary microscopy at diagnosis in 309 patients with systemic of capillary loss with more extensive skin involvement The skin was in patients with more capillary loss for C or D for A or B capillary was between the skin and extent of capillary nailfold capillary in more SSc subsets determine if more NCM were a of disease the disease at diagnosis was in the cohort according to moderate or extensive capillary loss C or normal A or was in patients with C or D = A or B = avascular areas = However, in patients with C or D capillary the disease at diagnosis was in diffuse than in normal skin limited and intermediate SSc analysis of = disease was in patients with capillary telangiectasias = than in without = in patients with more capillary dilatations the disease was in diffuse than in normal skin limited SSc and intermediate SSc = data that the of of nailfold capillary according to SSc as shown in Figure revealed that the of of moderate or extensive capillary loss the 4 SSc subsets The of patients no or rare avascular areas A or decreased more in diffuse with intermediate subsets = and in intermediate with limited subsets = in the of 2 or 3) capillary the of were the 4 SSc subsets = data not with more in diffuse with intermediate subsets = and in intermediate with limited subsets = data not of for avascular areas by nailfold capillaroscopy in for the 4 subsets, normal skin = limited = intermediate diffuse SSc, = the 309 SSc The of subset from in the normal skin group to and in the other subsets, = The major of was SSc, for of the was in the normal skin and of the were in the and diffuse subsets, = and of in 309 patients with systemic second of was for = of in our The were of the = = = = = and = The third most of was = of these patients of in the limited and 3 in the intermediate and in the 3 other patients the of was for and of in the normal skin, and intermediate subsets, other for 8 to pulmonary disease = and The of to SSc were of pulmonary involvement with and small involvement and pulmonary fibrosis pulmonary involvement was the second of for In of the did not the Mortality that SMR are in the and diffuse SSc subsets in comparison with the general population. In the SMR are in in these 3 The SMR of is in with diffuse SSc. The SMR is mortality ratios (SMR) in the French Canadian systemic sclerosis cohort to the of a of survival was using as the date of diagnosis of SSc. The cumulative survival for of in the 4 subsets, from normal to were and at and and at the specific of SSc on were using and the are shown in Figure The for the 4 subsets from normal to diffuse were and at and and at of the survival by revealed = Figure is that the between limited and intermediate SSc was not = In the between intermediate and diffuse SSc survival revealed a = as well as between limited and diffuse survival = Figure survival for systemic sclerosis subsets, in of 50 normal skin, of of and of diffuse SSc patients. for the 4 subsets, = normal skin = normal skin = normal skin limited = limited = intermediate = for survival We used first a to factors present at diagnosis for their to mortality shown in associated with mortality were sclerodermatous skin involvement proximal to MCP joints, skin involvement of the ECG, of the normal and presence of anti-topo For in of was a in The in mortality were associated with trunk involvement and ratios and survival was also associated with an or to years at diagnosis data not A and C or D capillary loss were not associated with We that the presence of was associated with a in mortality, but this was not analysis of the between factors at diagnosis and mortality by in 309 patients with systemic identify multivariate analysis was performed using of NCM data were in patients. shown in the following at diagnosis skin involvement of the of normal and this analysis was in the patients in NCM data were C or D capillary loss did not as predictive of the mortality according to the of predictive factors identified in The mortality from in patients with of the factors to in patients with 4 of for by analysis in 309 patients with systemic mortality from according to predictive present at diagnosis in 309 patients with systemic the subset of intermediate SSc based on of skin involvement below and the Because of the in SSc and mortality between the intermediate subset and other subsets, we if the features of the intermediate subset be that is, due to the of more limited patients with more diffuse patients Therefore, the intermediate subset was into patients with below the skin involvement = and with the skin involvement or without involvement, = of target organ involvement as in 2 no between the 2 not is that the of was in both = = = the of anti-topo I was in both = = = The of from was in both = = = as was the of SSc = = = our knowledge, this is the first report on SSc in a large cohort of French Canadians. between French Canadian SSc patients and other include a of the disease, in the or of in the subsets, and between pulmonary fibrosis and anti-topo I autoantibodies the of anti-topo I in French Canadians with diffuse SSc is the In other the of anti-topo I in diffuse SSc from in patients and in and to over in patients In the of anti-topo I in French Canadians with limited SSc is to the in other This of anti-topo I in French Canadians with diffuse SSc is The of SSc autoantibodies on and et al that patients are according to SSc subsets, in the of certain autoantibodies are was in background were for these this of anti-topo I, and the anti-topo I is genetically modulated and to and it be of interest to study in French Canadian SSc and to determine other are present in these patients with diffuse SSc. anti-topo I is associated with pulmonary fibrosis in our patients it be of interest to determine if the of anti-topo I is associated with a of pulmonary fibrosis on than in other mortality data the present study and other SSc studies selected for their large of subset and use of multivariate to identify factors predictive for mortality from to 1999, allowing comparison of SMR and over of in a in survival SSc subsets over In the patients by Barnett et al between and the survival in the and diffuse subsets were and In patients by et al intermediate and diffuse SSc patients between and survival of and In the present study, patients were between 1984 and 1999, the in the and diffuse subsets were and the of diffuse SSc to be with high mortality as shown by SMR in English and French Canadian patients In SMR mortality in the intermediate and limited subsets as mortality data for systemic sclerosis in studies selected for multivariate we have identified at diagnosis factors strongly and predictive for sclerodermatous skin involvement of the decreased DLCO, ESR, and the combination of these factors was associated with a on shown in most of these predictive factors have been identified in other as that is, decreased DLCO, ESR, and of skin involvement of the trunk in this cohort and not in other is due to is that predictive factors identified in the present study were also identified by et al in the analysis of their that is, diffuse skin disease, and However, these did not as predictive for mortality in the multivariate analysis of et due to of we did not in their multivariate predictive factors are to determine at presence patients at high for mortality and may and a for these as well as the SMR data and geographic and American SSc patient populations factors and of survival associated with the presence of ACA, and decreased survival in the presence of anti-topo I, we were interested in these as in this in patients we by analysis that were associated with a decreased but not in mortality, anti-topo I were associated with a in However, was to survival by multivariate the to use this in the analysis of C or D capillary loss was associated with mortality by this was not predictive for no predictive value for mortality be to NCM we have identified of of capillary loss SSc years of the first SSc of diffuse SSc patients developed capillary loss in with and in the intermediate and limited subsets, data also that capillary loss on NCM is not specific for diffuse SSc most SSc patients more capillary loss with more extensive skin involvement. We have shown previously that of capillary loss by NCM may as well with visceral involvement organ involvement in diffuse SSc most the first 3 years of disease these data that the degree of capillary loss as by NCM may the of of skin and visceral involvement in SSc and its We the full spectrum of SSc in our patient population to 2 due to classification The first is the exclusion of limited SSc patients by the ACR preliminary criteria for the classification of SSc. criteria were not for but with the to a for disease to comparison of of patients from (46). in clinical these criteria have been and used as criteria is not that these criteria are specific but not for SSc as by that of the patients with limited SSc not ACR This has been for years in the that of involvement not the diagnosis of We and others have criteria for the diagnosis of SSc to this In the present study, the of SSc patients with normal skin or with is of the in SSc clinical is that the of and NCM in our patients with normal skin or were in with other on SSc patients fulfilling ACR criteria for SSc, that our patients were that both the and disease were in our normal skin limited SSc it be of interest to determine if patients in the normal skin group some skin involvement over The second is the of an subset of SSc, based on extent of skin involvement. This was by intermediate SSc the limited and diffuse SSc subsets (26). data for the of an intermediate SSc In comparison with intermediate SSc our diffuse SSc patients more skin and involvement at diagnosis, and less Raynaud phenomenon, calcinosis, and All in diffuse SSc were in intermediate SSc. In survival was in patients with diffuse SSc. In patients with limited SSc less skin, and involvement, and more telangiectasias and than with intermediate SSc. less in patients with limited SSc, and survival was better than in with intermediate SSc. our intermediate patients into subsets of and skin involvement revealed no in the of target organ involvement, ACA, anti-topo I, from or from SSc. This that the intermediate subset be due to the of more limited patients with more diffuse patients the of an intermediate SSc subset is also strongly by our NCM data that the of of of capillary loss SSc the data and in of a intermediate SSc first objective was to report the features of systemic sclerosis in a large cohort of French Canadian patients encompassing the full spectrum of the second objective was to identify these features factors predictive for mortality and to compare these data with of other series. Between 1984 and 1999, SSc was diagnosed at first presentation in 309 and unselected French Canadian patients at a SSc the patients were in a prospective study and followed September Patients were categorized into 4 subsets based on sclerodermatous skin diffuse = intermediate to metacarpophalangeal joints without = limited = or normal skin = was according to a protocol encompassing and including nailfold capillary mortality ratios were using and mortality in the of general population. was from the The cohort to was features in in were most in the normal skin and limited SSc subsets of and less in intermediate and diffuse subsets and The of to DNA-topoisomerase I (anti-topo I) in diffuse SSc was nailfold capillary microscopy, capillary loss was not restricted to any However, the of of capillary loss the subsets in patients. SSc was the major of = followed by and In the and diffuse subsets, and of the were = For due to SSc, cumulative survival for these subsets were and at and and at were = for mortality were skin involvement of the carbon monoxide diffusion capacity of the of normal erythrocyte sedimentation and The mortality from in patients with of these to in patients with mortality ratios were and in the and diffuse subsets, of the intermediate group into skin involvement below the revealed no in the of clinical anticentromere anti-topo I, and our knowledge, the present study is the first report on a large French Canadian SSc The of anti-topo I in diffuse SSc is the reported thus The data in of a intermediate SSc factors predictive for mortality were identified at factors are to determine and may a for and in patients at high for with other that and geographic American and patients with SSc factors and SSc to be associated with high We the who in patient and for this We also for for data and for on and for