Associations with digital ulcers in a large cohort of systemic sclerosis: Results from the Canadian Scleroderma Research Group registry
Digital ulcers are a common complication of systemic sclerosis (SSc; scleroderma). Approximately half of SSc patients have had a digital ulcer, but information is lacking on whether digital ulcers are associated with patient demographics and clinical outcomes. We wanted to determine the associations between digital ulcers and other SSc vascular complications and organ involvement.
Data from the Canadian Scleroderma Research Group are collected annually on SSc patients, including presence, location, and number of digital ulcers; complications from digital ulcers; internal organ involvement; skin score; and laboratory results. Correlation coefficients, chi-square test, and logistic regression modeling were done to determine the associations of digital ulcers with other factors, including internal organ complications.
A total of 938 patients were included; 86% were women, the mean age was 55 years, the mean disease duration was 13.6 years, and 50% had limited cutaneous SSc. Eight percent had a digital ulcer currently and 44% had a digital ulcer ever; 53.1% had digital pitting scars. Digital ulcers were associated with increased modified Rodnan skin score (P = 0.0001), hand and finger skin score (P = 0.0001), Health Assessment Questionnaire score (P = 0.0001) and disease duration (P = 0.001), younger age of SSc onset (P = 0.0001), interstitial lung disease (ILD; P = 0.0001), and topoisomerase I (Scl-70) antibodies (P = 0.0001) in limited and diffuse cutaneous SSc subsets. Digital ulcers were not associated with sex (P = 0.95), smoking (P = 0.9), pulmonary arterial hypertension (PAH; P = 0.35), and scleroderma renal crisis (SRC; P = 0.569). Digital ulcers were further associated with reduced diffusing capacity for carbon monoxide (DLCO; P = 0.0001) and esophageal involvement (P = 0.0001) in diffuse cutaneous SSc only.
Digital ulcers are associated with worse disease, including skin and lung involvement, but are not associated with PAH and SRC. However, the low DLCO that is associated with SRC could represent ILD or microvasculopathy.
Systemic sclerosis (SSc; scleroderma) is a chronic autoimmune systemic disease characterized by a proliferative and obstructive vasculopathy of small blood vessels and tissue fibrosis. While thickening of the skin is prominent, SSc can result in fibrosis of other organs, including the heart, lungs, and gastrointestinal (GI) tract, and vasculopathy of the pulmonary and renal arteries.
Raynaud's phenomenon (RP; reversible changes in the color of the hand) is an almost universal early symptom of SSc (1). Significant vasospasm, structural digital vasculopathy, tissue fibrosis, and local trauma are risk factors for digital ulceration in SSc, with complications such as osteomyelitis and digital amputation (1–3).
Digital ulcers are a significant clinical problem in SSc and occur in 40–50% of patients (4). Digital ulcers occur on the fingers, or less commonly the toes, causing pain and functional impairment. Chronic ulcers have a complication rate of 15%, including infection, gangrene, osteomyelitis, and amputation (4). If there is one amputation, there is a 1–2% likelihood of further amputations in the next 6–12 months (5).
The majority of mortality in SSc relates to lung complications and pulmonary arterial hypertension (PAH) (6). PAH may be a late complication of SSc and occurs in approximately 15% of patients (7). Despite the benefits of current therapy, early diagnosis of PAH is crucial as the prognosis is very poor if diagnosed late (8). Currently, annual echocardiographic screening for PAH is recommended and if suspected, the diagnosis is confirmed by right heart catheterization (RHC) (9). Previous work has identified that limited SSc (1), disease duration >10 years (2), age >50 years (9), male sex (1), limited cutaneous SSc (lcSSc) (9), pulmonary fibrosis (10), and RP onset >3 years prior to SSc (which is more likely to be occurring in lcSSc) (10) may be risk factors for PAH. However, besides echocardiographic screening and possibly using brain natriuretic peptide, it is difficult to find reliable clinical and serologic predictors that identify early PAH and therefore, in whom to suspect PAH before it is symptomatic. Risk factors in limited and diffuse subsets may in fact not be identical. Experts have postulated that digital ulcers are possibly associated with PAH (11). The vasculopathy of digital ulcers is similar to that observed in PAH, i.e., fibrotic intimal hyperplasia, adventitial fibrosis, and compromised arterial lumen (12). Previous studies using the German Network for SSc and the French ItinerAIR-Sclerodermie registry have identified that male sex, topoisomerase I (topo I; Scl-70) antibodies, and early age at onset are associated with digital ulcers (11, 13).
The objectives of this study were to determine features associated with digital ulcers and their complications in our Canadian registry and to determine if there were associations of digital ulcers with other evidence of vasculopathy such as PAH and scleroderma renal crisis (SRC). By understanding the features of SSc that are associated with digital ulcers, clinicians may be better able to risk stratify patients and understand the burden of digital ulcers in this disease.
MATERIALS AND METHODS
Data from the Canadian Scleroderma Research Group (CSRG) database are collected annually on an incidence and prevalent SSc cohort. For this analysis, 938 patients were within the database from 12 sites where rheumatologists collected data. Presence, location, and number of digital ulcers are collected annually as well as complications and other internal organ involvement, modified Rodnan skin score (MRSS), and laboratory results (14). This protocol has ethics approval at all participating sites. All of the patients are diagnosed as having SSc by a physician and are given a letter of information and signed consent. All of the values are recorded by a CSRG participating rheumatologist from across Canada.
The relationship between digital ulcers (past or present) or the complications of digital ulcers (pitting scars, digital necrosis/gangrene, and digital amputation) and the presence of PAH, other internal organ involvement, patient outcomes, and laboratory results was assessed at baseline and followup visits using odds ratios, correlations, t-tests, and chi-square tests, as appropriate. All of the comparisons were also done in the subsets of diffuse cutaneous SSc (dcSSc) and lcSSc patients, and these strata were further divided into disease durations of <3 years and >3 years. Multivariate linear regression analysis was performed to ascertain independent associations with digital ulcers. Those variables whose association was found to attain a level of significance of P values less than 0.05 in univariate analysis were retained in the multivariate analysis. Stepwise, forward, and backward multivariate analyses yielded identical results. All statistical work was done using PASW Statistics 17 software (IBM). Patients with missing values were left out of the statistical analysis for the correlation in which their value was missing. However, missing values for the items of interest were rare.
Digital ulcers ever included active and healed ulcers. Active ulcers were defined as denuded areas with defined borders and loss of epithelialization, loss of epidermis, and dermis distal to the proximal interphalangeal joint on the volar aspect of a finger. This definition excluded fissures, paronychia, and extrusions of calcium. Healed ulcers were defined by complete epithelialization of an ischemic ulcer, regardless of residual pain. Digital pitting scars were defined as pinhole-sized depressions with hyperkeratosis. Age at onset of RP and the first non-RP symptom were recorded by the physician. Disease duration was calculated from the onset of the first non-RP symptom. Patients recorded their smoking history, including present and ever smoking. Having smoked ever (past or present) was used in the analysis. Health Assessment Questionnaire (HAQ) score (15), the MRSS, and total bilateral hand and finger skin score (total bilateral hand and finger score are the components of the MRSS for the fingers [range 0–3] and dorsum of the hand [range 0–3] on the right and left hands, so it could range from 0–12). We thought it would be important to determine if worse hand fibrosis was related to more fingertip ulcers and used the skin score on the hands, including the fingers, to explore a possible relationship. The diffusing capacity for carbon monoxide (DLCO) percent predicted and forced vital capacity (FVC) percent predicted were done annually. Baseline antinuclear antibody (ANA) titer and pattern and anticentromere antibody (ACA), anti-RNP, and anti–topo I (Scl-70) were evaluated. ANA was measured by indirect immunofluorescence (IIF) on HEp-2 cells (ImmunoConcepts) at a screening dilution of 1/160, extractable nuclear antigens (ENAs) by an addressable laser bead immunoassay using Inova ENA 9 QuantaPlex kit and a Luminex 100 illuminometer, and ACA was based on the IIF staining pattern on HEp-2 cells. PAH was recorded by the physician at the baseline visit. During subsequent visits, a diagnosis of PAH was assumed using either of 2 criteria: 1) the patient had an echocardiography right ventricular systolic pressure (RVSP) of >40 mm Hg and was sent for RHC and was started on PAH medication, or 2) the patient had an echocardiography RVSP of >45 mm Hg. As defined per agreement of the CSRG, interstitial lung disease (ILD) was considered present when typical Velcro rales were detected or lung fibrosis was seen on a radiograph or computed tomography scan. SRC was asked initially as diagnosed ever and then recorded annually in the physician forms. SRC was not defined because it is a clinical diagnosis that experts in SSc understand. We had asked the physicians to answer if the patient ever had SRC. Dysphagia was assessed by patient forms asking about difficulty swallowing. Gastroesophageal reflux disease (GERD) was assumed present if the patient said yes to experiencing acid-tasting liquid that regurgitated into the mouth. Use of esophageal dilatation was recorded by the physician over the last year at each annual visit.
Baseline data for the entire group (n = 938) and dcSSc and lcSSc subsets are shown in Table 1. Eighty-six percent were women, the mean age was 55 years, and the mean disease duration was 14 years. Fifty percent had lcSSc, 43% had dcSSc, and the rest were not defined. Eight percent had a digital ulcer currently, 44% had a digital ulcer ever, and 53% had digital pitting scars. As expected, digital ulcers occurred more frequently in dcSSc patients (P = 0.0001). More patients with dcSSc were topo I (Scl-70) positive. Fewer patients with dcSSc had an FVC percent predicted (P = 0.018) and a DLCO percent predicted (P = 0.018) greater than 75% predicted compared to lcSSc patients. Patients with dcSSc were more likely to have ILD (P = 0.0001). SRC (P = 0.0001) and GERD (P = 0.038) were more prevalent in dcSSc. The HAQ score was greater in dcSSc patients (P = 0.0001). If a fingertip ulcer was present, the mean number of digital ulcers per patient with a prevalent digital ulcer was 1.9 (Table 1).
Table 1. Baseline characteristics, including digital ulcers, in the SSc patients from the Canadian Scleroderma Research Group database divided also by lcSSc and dcSSc subsets*
|Total||938 (100)||412 (43.4)||469 (50.0)||0.064|
|Age, mean ± SD years||55.4 ± 0.34||53.1 ± 0.60||57.6 ± 0.55||0.0001|
|Disease duration, mean ± SD years||13.6 ± 0.31||12.3 ± 0.44||14.9 ± 0.45||0.0001|
|Sex|| || || || |
| Male||131 (14.0)||71 (17.2)||56 (11.9)||0.026|
| Female||807 (86.0)||341 (82.8)||413 (88.1)||0.026|
|Smoking history|| || || || |
| Smoked ever||550 (62.1)||247 (62.1)||279 (62.8)||0.816|
| Smoked currently||122 (14.5)||58 (14.6)||64 (14.4)||0.82|
|Age at first RP symptom, mean ± SD years||40.9 ± 0.52||41.3 ± 0.75||40.5 ± 0.73||0.411|
|Age at first non-RP symptom, mean ± SD years||44.5 ± 0.47||43.7 ± 0.67||45.2 ± 0.65||0.118|
|ACA positive||269 (36.2)||57 (16.7)||190 (52.0)||0.0001|
|Topo I (Scl-70) positive||126 (16.0)||84 (23.3)||41 (10.6)||0.0001|
|RNP positive||40 (5.1)||16 (4.4)||21 (5.4)||0.536|
|PAH||84 (9.4)||35 (8.7)||42 (9.3)||0.350|
|Echocardiography PAP, mean ± SD mm Hg||57.2 ± 1.5||55.6 ± 2.12||57.1 ± 2.1||0.637|
|ILD||342 (38.3)||207 (50.2)||127 (27.1)||0.0001|
| Radiographic evidence of ILD||233 (38.0)||150 (39.2)||77 (17.7)||0.0001|
| Lung fibrosis on CT||246 (31.9)||155 (44.4)||84 (21.4)||0.0001|
|Scleroderma renal crisis||49 (5.3)||38 (9.3)||10 (2.1)||0.0001|
|Dysphagia||534 (60.5)||242 (61.3)||264 (60.0)||0.679|
|GERD||628 (71.0)||296 (74.7)||301 (68.3)||0.038|
|Esophageal dilatation||125 (13.7)||53 (13.0)||67 (14.4)||0.543|
|MRSS, mean ± SD||11.2 ± 0.24||15.0 ± 0.38||8.5 ± 0.27||0.0001|
|Hand and finger skin score, mean ± SD||5.7 ± 0.11||7.8 ± 0.16||4.3 ± 0.12||0.0001|
|Digital pitting scars||489 (53.1)||260 (63.2)||218 (46.6)||0.0001|
|Digital ulcers ever||413 (44.0)||225 (54.6)||180 (38.4)||0.0001|
| Mean ± SD||1.9 ± 0.14||2.0 ± 0.20||1.7 ± 0.19||0.003|
| Digital ulcers currently||76 (8.2)||49 (11.9)||24 (5.1)||0.0001|
| Elsewhere||180 (19.5)||131 (31.9)||45 (9.6)||0.0001|
|Digital necrosis||17 (1.8)||9 (2.2)||8 (1.7)||0.609|
|Digital amputation||67 (7.3)||28 (6.8)||38 (8.1)||0.458|
|HAQ score, mean ± SD||0.74 ± 0.02||0.91 ± 0.04||0.62 ± 0.03||0.0001|
|Calcinosis||313 (35.6)||168 (41.1)||145 (30.9)||0.002|
|Telangectasia||760 (87.4)||349 (85.7)||411 (88.8)||0.182|
Associations with digital ulcers.
There was no significant association between the history of digital ulcers and any definition of PAH (P = 0.350) (Table 2). By univariate analysis, a history of digital ulcers was associated with the following patient and demographic factors: younger age at visit (P = 0.0001), longer disease duration (P = 0.001), younger age at onset of RP (P = 0.0001) and first non-RP symptom (P = 0.0001), and positive topo I (P = 0.0001). The following features of SSc were associated with digital ulcers: reduced DLCO (P = 0.0001), ILD (P = 0.0001), esophageal dilatation ever (P = 0.019), increased MRSS (P = 0.0001), increased hand and finger skin score (P = 0.0001), and a higher HAQ score (P = 0.0001).
Table 2. Patient and demographic factors associated with digital ulcers in all Canadian Scleroderma Research Group patients and in dcSSc and lcSSc*
|Age, years||−4.32|| ||0.0001||−5.4|| ||0.0001||−2.7|| ||0.019|
|Disease duration, years||+2.01|| ||0.001||+3.4|| ||0.0001||+2.7|| ||0.032|
|Age at first RP symptom, years||−4.92|| ||0.0001||−8.7|| ||0.0001||−2.6|| ||0.800|
|Age at first non-RP symptom, years||−6.14|| ||0.0001||−8.2|| ||0.0001||−4.5|| ||0.0001|
|Male sex|| ||0.989||0.951|| ||1.484||0.13|| ||0.745||0.304|
|Smoking|| ||0.984||0.909|| ||1.06||0.783|| ||0.883||0.535|
|MRSS||+3.8|| ||0.0001||+2.8|| ||0.0001||+1.9|| ||0.0001|
|Hand and finger skin score||+2.5|| ||0.0001||+1.5|| ||0.0001||+1.6|| ||0.01|
|HAQ DI score||+0.15|| ||0.0001||+0.16|| ||0.023||+0.18|| ||0.002|
|PAH|| ||0.81||0.350|| ||0.949||0.882|| ||0.776||0.459|
|DLCO, % predicted||−6.40|| ||0.0001||−7.00|| ||0.0001||−4.20|| ||0.068|
|ILD on chest radiograph|| ||2.11||0.0001|| ||2||0.001|| ||1.61||0.059|
|ILD on CT|| ||1.83||0.0001|| ||1.72||0.013|| ||1.29||0.307|
|Scleroderma renal crisis|| ||1.18||0.569|| ||0.801||0.514|| ||1.06||0.924|
|GERD|| ||1.18||0.270|| ||2.05||0.002|| ||0.725||0.121|
|Dysphagia|| ||1.1||0.478|| ||1.37||0.131|| ||0.889||0.555|
|Esophageal dilatation|| ||1.57||0.019|| ||3.69||0.0001|| ||0.947||0.842|
|ACA|| ||0.752||0.064|| ||1.54||0.155|| ||0.963||0.863|
|RNP|| ||0.73||0.345|| ||1.31||0.605|| ||0.513||0.194|
|Topo 1 (Scl-70)|| ||3.06||0.0001|| ||2.34||0.001|| ||2.67||0.003|
Associations with digital ulcers in lcSSc versus dcSSc subsets.
Significant associations with digital ulcers were then calculated in the dcSSc and lcSSc subsets (Table 2). The following patient and demographic factors were associated with digital ulcers in lcSSc and dcSSc: age at onset of SSc, disease duration, age at first non-RP symptom and positive topo I, ILD, increased MRSS and hand and finger skin score, and HAQ score. Additionally, age at first RP symptom, reduced DLCO, GERD, and esophageal dilatation were associated with digital ulcers, but only in dcSSc patients.
Although the number of patients was small (69 patients), none of the associations with digital ulcers in dcSSc patients were significant in a subset of patients with a disease duration <3 years. All of the associations in dcSSc and lcSSc held in the subset of patients with a disease duration >3 years (data not shown).
Multivariate associations with digital ulcers.
A logistic regression model including the significant associations in univariate analysis was performed. In the multivariate analysis, the most important variables in the model to predict digital ulcers were younger age of onset, ILD, higher hand and finger skin score, and higher HAQ score, with an R2 for the model of 0.29 (Table 3). In multivariate analysis in a subset of dcSSc patients, the most important variables in the model to predict digital ulcers were younger age of onset, ILD, esophageal dilatation, and hand and finger skin score, with an R2 of 0.28 (Table 3). In lcSSc, the most important variables to predict digital ulcers were higher hand and finger skin score, younger age of onset, and higher HAQ score, with an R2 of 0.20 (Table 3). Therefore, many associations were similar in the subsets.
Table 3. Multivariate analysis of associations of patient and demographic factors and clinical features of SSc with digital ulcers*
|Age at first non-RP symptom, years||0.045||0.008||0.029, 0.061||0.0001||0.0001||0.001|
|ILD||−0.761||0.219||−1.190, −0.332||0.001||0.003|| |
|Hand and finger skin score||−0.118||0.032||−0.181, −0.055||0.0001||0.0001||0.0001|
|HAQ score||−0.549||0.157||−0.857, −0.241||0.0001|| ||0.001|
|Esophageal dilatation|| || || || ||0.001|| |
Complications of digital ulcers and associations.
Fingertip pitting scars occurred in 53%; they occurred more frequently in dcSSc than lcSSc patients (P = 0.0001) (Table 1) and were associated with younger age (P = 0.0001), longer disease duration (P = 0.0001), younger age at first RP (P = 0.0001) and non-RP symptom (P = 0.0001), positive topo I (P = 0.0001), DLCO (P = 0.0001), ILD (P = 0.0001), increased MRSS (P = 0.0001), increased hand and finger skin score (P = 0.0001), and higher HAQ score (P = 0.0001). Associations were then measured in a subset of diffuse and limited patients. Most features were also associated with both subsets except in diffuse alone, where there was an association of pitting scars and topo I (P = 0.0001) and GERD (P = 0.002). All of the associations remained in the subsets of patients with >3 years of disease duration.
Digital necrosis was only present in 1.8% of patients in the database and was not associated with a subtype of SSc (P = 0.609) (Table 1), but was statistically related to positive topo I (P = 0.017), higher hand and finger skin score (P = 0.031), and HAQ score (P = 0.002). Seven percent had a digital amputation, which was not different in dcSSc versus lcSSc (P = 0.458). Digital amputation was associated with longer disease duration (P = 0.001), younger age at first non-RP symptom (P = 0.001), positive topo I (P = 0.010), esophageal dilatation (P = 0.0001), higher hand and figure skin score (P = 0.0001), and higher HAQ score (P = 0.0001). Therefore, associations with digital ulcers were similar to digital ulcer complications.
Digital ulcers were associated with ILD, increased MRSS, the subset of the MRSS, i.e., hand and finger skin score, and HAQ score overall. In both subtypes of SSc, digital ulcers were associated with longer disease duration, younger age of first non-RP symptom, ILD, topo I antibodies, and younger age at onset of SSc. We found no association between digital ulcers and PAH using either definition, nor was complicated digital ischemia related to PAH.
Associations were measured in subsets of diffuse and limited cutaneous subtypes because the former may be a confounding variable, as digital ulcers and other complications of SSc occur more frequently in the diffuse subtype (Table 2). In the diffuse subset, there was also a relationship between low DLCO and esophageal involvement. A low DLCO is nonspecific and thus could indicate ILD, early PAH, or both, or something else such as the effects of smoking. However, there was not an association between PAH and digital ulcers, nor with digital ulcer complications and PAH. The CSRG database records a diagnosis of PAH at the baseline visit, and the results of any RHC in the annual review thereafter are documented. Nine percent of patients in the database had PAH under these criteria, which is similar to the 13% prevalence in a study using RHC testing (16).
One study of 52 patients with dcSSc and 39 patients with lcSSc observed that digital ulcers were more common in dcSSc (17). In a previous study of 87 patients with SSc, digital pitted scars were found in 39% and were associated with RP, skin thickening, and joint disease. The number of pitted scars was not associated with disease duration (3). Antiendothelial cell antibodies have been correlated with severe digital ischemia and PAH in some SSc studies (18–21). They were not measured in our study.
As some of the patients had a long disease duration and the maximum degree of skin involvement may not have been recorded, the investigators may have chosen not to classify the patient (e.g., now limited but likely was diffuse in past) or the patient may have had disease distal to the elbows but was of new onset and thought they would eventually be diffuse (e.g., skin to mid forearm is rare in the limited subset). This could explain the nearly 7% of unclassified patients. Also, only 69 patients had early dcSSc, so the lack of statistically significant associations may be due to a lack of power because the sample size is small.
The prevalence of digital ulcers ever and currently was similar to other studies (4, 5). Our large database found several associations that were not due to dcSSc as a confounder (more skin involvement, younger age of onset, more ILD, and more digital ulcers all can cluster together), as many of the observed associations were also seen in the lcSSc subset. Table 4 shows the comparisons between the CSRG and other SSc databases with respect to associations with digital ulcers. We could not determine the usual earliest time after diagnosis that digital ulcers occurred, as we had mostly a longstanding prevalent population that enrolled in the CSRG. For current fingertip ulcers, we did not include calcinosis. However, for past ulcers we could only go by history and the judgment of the investigator as to whether they had an ischemic fingertip ulcer. The fact that the frequency of digital ulcers ever and the presence of digital pitting scars was very similar is reassuring that we are discussing not calcinosis but ischemic digital ulcers. Also, the rate of digital ulcers is similar to other series of SSc (4, 5). Calcinosis ever occurred less often than digital ulcers, at 36%.
Table 4. Comparison of CSRG digital ulcer associations to other registries*
|Age, years (lower)||+ (lower age at onset of RP)||+ (lower age at onset of SSc)||+ (lower age at onset of SSc, lower age of first non-RP)|
|Male sex, increased digital ulcers||+||+||−|
|More skin involvement||More in dcSSc||Increased MRSS||Increased in dcSSc and increased with higher MRSS in both lcSSc and dcSSc|
|Low DLCO||Not measured||+||+|
|ILD on chest radiograph or HRCT||+||Not studied||+|
|Esophageal involvement||+||Not studied||+ (dcSSc only)|
|Longer disease duration||−||+||+|
|Topo I (Scl-70)||+||+||+|
|HAQ score||Not studied||Not studied||+|
Digital ulcers have a negative impact on the physical (4) and mental (22) components of quality of life. We observed that the HAQ score was higher in both lcSSc and dcSSc patients with digital ulcers versus without digital ulcers. Although a hand function scale is not used in the CSRG database, previous analysis has shown that hand function is reduced in patients with digital ulcers (22). Digital ulcers are more likely to occur in patients with ILD (in both lcSSc and dcSSc) and result in increased disease burden, as there are also associations with GI complications. Similar to other reports (4), 44% of patients in the CSRG database had a history of digital ulcers; however, unlike previous publications (11, 13), data from the CSRG showed no association between smoking and digital ulcers or sex and digital ulcers. An early onset of disease as assessed by age at first non-RP symptom was associated with digital ulcers. This occurred in both lcSSc and dcSSc, so the results are not confounded by dcSSc, which has an earlier age of onset and has more digital ulcers overall (Table 1).
In 2 other studies, digital ulcers were associated with topo I antibodies (Scl-70) (13, 23). It is known that topo I antibodies are associated with higher skin scores, lung fibrosis, and contractures, so some of the associations we found with digital ulcers could be confounded with the known clustering of topo I (23). We did find many associations with digital ulcers that were consistent in the lcSSc group, who were topo I negative, so it appears that patients with digital ulcers are more likely to have ILD and high skin scores, even if topo I is not present. Unlike these reports, we did not find an association between ACAs and digital ulcers.
A strength of the data is that this is a large database with excellent annual characterization of digital ulcers documented annually. Furthermore, the questionnaire used for the CSRG database provides a strict definition of active and healed ulcers. This is an important point in this study, given the interobserver variability among clinicians in defining digital ulcers (24). However, our patient numbers in early SSc are too small to have meaningful results. Weaknesses may include the lack of standardization (no central laboratory or central reading) for pulmonary function tests (PFTs) and echocardiograms and the definitions of PAH. The data were the same whether we used truly diagnosed PAH (confirmed by RHC) or suspected PAH by echocardiogram estimates of elevated PA pressures. Each site orders PFTs in certified pulmonary laboratories. This is a multiple-site observational study and therefore, we did not have central standardization. DLCO percent predicted was used from all PFTs and if there are different methods at various laboratories, then this could be a systematic bias (measurement bias), but the bias should not be different in the groups with and without digital ulcers. All echocardiograms are read by trained specialists. In a cohort where many have long disease duration at the time of entry to the study, we could miss other associations because there could be recall bias where the number with digital ulcers may be under- or overestimated. We also cannot determine a cause and effect relationship, as both antibodies and skin involvement may have occurred before, after, or simultaneously with respect to the timing of the first digital ulcer.
Digital ulcers are associated with a larger disease burden as measured by higher skin scores overall and on the hands and higher HAQ scores. The worse HAQ score could be from the ulcer, but we studied ever digital ulcers and it was still associated with a higher current HAQ score, so the HAQ score may be confounded by those with worse hand function as already manifested by a higher skin score of the hand in those with digital ulcers compared to those without.
Some have thought that digital ulcers are a marker of damage (vascular and fibrotic changes) (5), but interestingly, they were not associated with SRC or PAH, which are other major vascular complications.
Our numbers were large enough to separately study dcSSc and lcSSc subsets to determine if the observations were similar in each group such as age, ILD, and skin score when comparing those with and without digital ulcers.
In conclusion, there are some associations with digital ulcers but a surprising lack of association between digital ulcers and other manifestations of structural vasculopathy, including PAH and SRC. However, one could argue that a low DLCO could be associated with either ILD or pulmonary vascularity changes. We did not study macrovascular disease in the extremities. Digital ulcers are not more associated with PAH but accompany SSc with a higher disease burden as measured by higher skin scores in both lcSSc and dcSSc, worse function as reported by the HAQ, younger age at onset, longer disease duration, and increased anti–topo I.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Pope had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Khimdas, Baron, Pope.
Acquisition of data. Khimdas, Harding, Zummer, Baron, Pope.
Analysis and interpretation of data. Khimdas, Harding, Bonner, Pope.
INVESTIGATORS OF THE CANADIAN SCLERODERMA RESEARCH GROUP
The investigators of the Canadian Scleroderma Research Group are as follows: M. Baron, Marie Hudson: Jewish General Hospital, McGill University, Montreal, Quebec; J. Pope: St. Joseph's Health Care, University of Western Ontario, London, Ontario; J. Markland: Saskatoon, Saskatchewan; P. Docherty: Moncton, New Brunswick; M. Fritzler: Advanced Diagnostics Laboratory, Calgary, Alberta; N. Jones: Edmonton, Alberta; E. Kaminska, N. Khalidi: Hamilton, Ontario; S. Ligier, J.-P. Mathieu, T. Grodzicky: Montreal, Quebec; A. Masetto: Sherbrooke, Quebec; D. Robinson, S. Mittoo: Winnipeg, Manitoba; D. Smith: Ottawa, Ontario; E. Sutton: Halifax, Nova Scotia, Canada. Former members: M. Abu-Hakim, S. LeClercq: Calgary, Alberta, Canada.