Objective To evaluate the possible merit of endothelial markers for the prediction of ischaemic digital ulcers in patients with systemic sclerosis (SSc).
Methods Circulating endothelial progenitor cells (EPC), circulating endothelial cells and serum levels of placental growth factor (PlGF), soluble vascular adhesion molecule and vascular endothelial growth factor were measured in a prospective cohort of 100 SSc patients. The primary outcome was the occurrence of one or more new ischaemic digital ulcers during a planned 3-year follow-up.
Results After the follow-up period, 17 patients developed new digital ulcers. By multivariate analysis focused on biomarkers, high PlGF serum levels and low EPC counts were identified as predictors of the occurrence of at least one new digital ulcer. In a secondary model including biomarkers together with clinical SSc characteristics all predictors of digital ulcers defined by p≤0.1 in univariate analysis, high PlGF serum levels (HR 7.26, 95% CI 1.92 to 27.41) and a history of digital ulcers (HR 9.32, 95% CI 1.51 to 59.83) were identified as independent predictors of a new digital ulcer. In an alternative model excluding patients with a history of digital ulcers at baseline, high PlGF serum levels (HR 13.46, 95% CI 1.58 to 114.73) and low EPC counts (HR 7.95, 95% CI 2.09 to 30.09) remained predictive of new digital ulcer occurrence during follow-up.
Conclusion This study identified high PlGF serum levels and low circulating EPC counts as predictors of new digital ulcers in SSc. It highlights the critical role of angiogenesis in this vascular outcome. These markers may improve digital ulcer risk stratification and therefore allow earlier therapeutic intervention.
Statistics from Altmetric.com
Systemic sclerosis (SSc) is a severe connective tissue disease characterised by vascular, immune and fibrotic changes in the skin and some internal organs.1 Peripheral microvascular and cardiovascular alterations are key features of SSc, with outcome depending on the extent and severity of vascular lesions. Microvascular lesions underlie many manifestations of SSc, including ischaemic digital ulcers, pulmonary hypertension, primary heart involvement and scleroderma renal crisis (SRC).
Digital ulcers are frequent in SSc, averaging 30% prevalence according to the European League Against Rheumatism (EULAR) Scleroderma Trial and Research Group (EUSTAR) registry, and represent a heavy burden due to their major impact on quality of life.2,–,4 In addition, digital ulcers are thought to be a clinical parameter of severe vasculopathy that can be associated with or predict other vascular lesions.5,–,7 Treatment of digital ulcers remains challenging and the identification of reliable predictors of digital ulcers is still an unmet clinical need in SSc.3 4
Endothelial injury and insufficient endothelial repair are strong contributors to the genesis of the vasculopathy underlying digital ulcers. Circulating endothelial cell (CEC) levels in peripheral blood is a direct reflection of vascular damage and more specifically endothelial injury.8 Vascular damage in SSc is insufficiently compensated by endothelial repair processes that relate to angiogenesis and vasculogenesis. Angiogenesis is insufficient in SSc, resulting from defective vascularisation and tissue hypoxia, despite elevated levels of proangiogenic factors in the serum or the skin of SSc patients such as vascular endothelial growth factor (VEGF), placental growth factor (PlGF) or soluble vascular cell adhesion molecule (sVCAM).9,–,12
Previous data have also emphasised quantitative perturbations of vasculogenesis.13,–,15 In particular, we reported high circulating endothelial progenitor cell (EPC) levels in SSc, supporting their general mobilisation. However, low EPC numbers in the peripheral blood of SSc patients were associated with the presence of past and/or current digital ulcers, suggesting increased homing during active vascular disease and overall insufficient vasculogenesis to counterbalance vascular damage.14 16
Subsequent to this first approach, the objective of this study was to determine the possible merit of different endothelial markers reflecting endothelial injury, disturbed angiogenesis and vasculogenesis to predict the occurrence of digital ulcers.10 12,–,15 17,–,21 These biomarkers were also secondarily evaluated for the prediction of other cardiac/vascular events that rely on microvascular complications.
Patients and methods
Extended details of the methodology are shown in the supplementary data (available online only).
We prospectively evaluated 100 consecutive SSc patients who were hospitalised in the Rheumatology A Department, Cochin Hospital, Paris, France, for routine follow-up of the disease from 1 September 2006 to 1 November 2007. SSc was classified as limited cutaneous or diffuse cutaneous subset according to the criteria reported by LeRoy et al.22 Inclusion and exclusion criteria are detailed in the supplementary data (available online only).
Once included in the study cohort, patients were seen by his/her usual physician on a regular basis of 3–6 month intervals, as indicated by their disease severity, and were hospitalised at least annually for clinical evaluation, including echocardiography and pulmonary function tests.
The final study observation was planned as the last observation of the last trimester of the year 2010 (or the last available observation) or the last observation at the time of death from any cause, with analyses and completion of the database performed during the first trimester of the year 2011.
The primary outcome was the occurrence of at least one or more ischaemic digital ulcers, defined by a painful area of 2 mm or greater in diameter with visible depth and loss of dermis, amenable to healing and localised by fingertip. Digital ulcers caused by conditions other than SSc and non-ischaemic digital ulcers were not taken into consideration.
The secondary outcome was the occurrence of at least one new cardiac or vascular event, as a consequence of the generalised vasculopathy. Cardiac/vascular events were assessed by an exploratory index, defined as the occurrence of at least one of the following events: (1) one or more new ischaemic digital ulcers; (2) pre-capillary pulmonary hypertension confirmed by resting mean pulmonary artery pressure of 25 mm Hg or greater together with a pulmonary capillary wedge pressure of 15 mm Hg or less; (3) left ventricular dysfunction, defined by a left ventricular ejection fraction less than 50%, as determined using the Simpson method; (4) SRC, defined by a sudden and marked increase in systemic blood pressure and acute renal failure, with or without significant microangiopathic haemolytic anaemia or thrombocytopenia.6 23 We also considered as a secondary outcome the occurrence of death from any cause during the follow-up period.
EPC and CEC sorting and flow cytometry quantification
We used a method previously described and validated to enrich mononuclear cells and quantify EPC and CEC by flow cytometry.14 24 This method is described in detail in the supplementary data (available online only).
VEGF, PlGF and sVCAM ELISA
Serum levels of VEGF, PlGF and sVCAM were measured using the quantitative sandwich ELISA technique (Quantikine kits; R&D Systems, Minneapolis, MN).
To estimate the cut-off of each endothelial marker, we compared levels of EPC, CEC, VEGF, PlGF and sVCAM between SSc patients and a group of 20 healthy controls (15 women, 55±16 years old), assessed during the same study period. These results revealed that the median value of all the markers for SSc patients were above the 95th percentile of the levels observed in the controls; therefore, to evaluate their predictive value on clinical outcomes, levels of EPC, CEC, VEGF, PlGF and sVCAM were considered as categorical variables, and associations were assessed according to the median value of each endothelial marker. Predictors of digital ulcers and cardiac/vascular outcomes were evaluated by univariate and multivariate Cox proportional hazards models and summarised as HR and 95% CI. Survival was evaluated using the Kaplan–Meier method.
The mean±SD age of the 100 patients (89 women) was 56±13 years and the mean±SD disease duration was 9±8 years. Forty patients had the diffuse cutaneous subset, and 60 the limited. Their baseline characteristics are shown in table 1.
Baseline values and correlation between vascular markers
The median (IQR) number of EPC and CEC in peripheral blood of SSc patients was 99 (64–164) and 215 (12–708)/106 Lin-mononuclear cells, respectively. The median (IQR) serum levels of VEGF, PlGF and sVCAM were 352 (183–474) pg/ml, 61 (33–87) pg/ml and 614 (4869–819) ng/ml, respectively. Among these vascular markers, only baseline values of EPC and CEC were correlated (r=0.52, p<0.0001).
Primary outcome, occurrence of ischaemic digital ulcers
During a follow up period of 29±13 months (median 36 months), new ischaemic digital ulcers occurred in 17 SSc patients (10 patients with a history of previous digital ulcers and seven patients with no previous digital ulcers). The mean time to the development of new digital ulcers after study inclusion was 18 months (median 17 months).
By univariate analysis, low EPC numbers (HR 3.94, 95% CI 1.41 to 11.03, p=0.009), high PlGF (HR 8.25, 95% CI 1.78 to 38.26, p=0.007) and sVCAM (HR 3.14, 95% CI 1.01 to 9.83, p=0.04) serum levels were identified as predictive biomarkers of the occurrence of at least one new digital ulcer (table 2). Kaplan–Meyer analyses of freedom from the occurrence of new digital ulcers for these three markers are shown in figures 1A–C. Multivariate Cox analysis confirmed PlGF levels (HR 7.95, 95% CI 2.09 to 30.10, p=0.002) and EPC levels (HR 2.33, 95% CI 1.44 to 12.22, p=0.03) as independent predictors of the development of digital ulcers (table 2).
In a second multivariate Cox model including these two biomarkers and all other predictors of digital ulcers defined by a p value of 0.1 or less in univariate analysis, high PlGF serum levels (HR 7.26, 95% CI 1.92 to 27.41) and a history of digital ulcers (HR 9.32, 95% CI 1.51 to 59.83) were identified in multivariate analysis as independent predictors of new digital ulcers (table 2).
In a third model excluding patients with a history digital ulcers at baseline, high PlGF serum levels (HR 13.46, 95% CI 1.58 to 114.73) and low EPC counts (HR 7.95, 95% CI 2.09 to 30.09) were found to be predictors of new digital ulcers (table 2).
During the follow up period, 27 of the 100 patients developed at least one cardiac or vascular event: new ischaemic digital ulcers in 17, precapillary pulmonary hypertension in five, left ventricular dysfunction in four and SRC in a single patient. The mean time to the development of a cardiac/vascular event was 19 months (median 20 months).
Seven patients died during the follow-up period. Causes of death were precapillary pulmonary hypertension in four patients (three patients with pulmonary arterial hypertension and one with pulmonary hypertension secondary to interstitial lung disease), right heart failure secondary to myocardial fibrosis in one patient, serious infection in one patient and severe inflammatory myositis associated with SSc in one patient.
According to our exploratory index, low circulating EPC numbers, high PlGF and high sVCAM serum levels were identified in univariate and multivariate Cox analyses as predictors of cardiac and vascular events in SSc (table 3). In a second model including these three biomarkers and other predictors of cardiac/vascular events defined by a p value of 0.1 or less in univariate analysis, low EPC counts (HR 5.14, 95% CI 1.21 to 21.80) and high PlGF serum levels (HR 5.32, 95% CI 1.24 to 22.91) were identified in multivariate analysis as independent predictors of the occurrence of cardiac and vascular events in SSc patients (table 3).
In addition, low EPC counts, high PlGF and sVCAM serum levels were predictive in univariate Cox analysis of the occurrence of death during the study period, but this was not confirmed by multivariate analysis.
This is the first prospective study to demonstrate that increased serum PlGF levels and circulating EPC levels predict the subsequent development of digital ulcers in patients with SSc. The identification of relevant predictive biomarkers may allow the detection of a subset of patients who might be candidates for preventive therapeutic strategies.3 4
Digital ulcers are a frequent complication in patients with SSc. The origin of ulcers is multifactorial. Ischaemia due to vascular disease, sclerodactily, calcinosis and local trauma may all contribute to ulcer initiation.25 Then, the development of digital ulcers results from poor blood flow and tissue hypoxia. Whatever its origin, digital ulcers are often extremely painful, cause significant impairment of hand function and activities of daily living, infection, digital resorption, and as a consequence markedly alter quality of life.3
Scarce data are available regarding the incidence of new ischaemic digital ulcers in SSc. The number of patients who developed new digital ulcers was less important in our cohort than in the placebo group of the RAPIDS-1 study (17 of 100 patients with at least one new digital ulcers vs 26 of 43 patients in the RAPIDS-1 study), a study that assessed the preventive effect of bosentan for ischaemic digital ulcers.3 Based on the high number of patients with active digital ulcers (24/43 patients from the placebo group of the RAPIDS-1 study had active digital ulcers at baseline), we assume that the exclusion of patients with active digital ulcers in our study accounts for this difference.
Our results have identified high serum PlGF levels as a predictor of the development of digital ulcers in SSc patients with or without a history of digital ulcers. PlGF, a secreted dimeric glycoprotein very similar to VEGF, has been shown to be chemotactic and mitogenic for endothelial cells in vitro and proangiogenic in vivo.18 19 PlGF has been reported to be dysregulated in preeclampsia, a severe vascular complication of pregnancy. Two recent cross-sectional studies have reported elevated serum PlGF levels in SSc patients compared with healthy controls.12 20 Together with our results, these findings suggest a possible contribution of this proangiogenic protein in SSc vascular disturbances. Increased levels of PlGF, reflecting increased angiogenesis, were predictive of the development of digital ulcers, in which there is a loss of vessels. Digital ulcers are associated with a destructive vasculopathy, which is a source of tissue hypoxia, one of the major stimuli of angiogenesis.17 25 Therefore, elevated angiogenic markers, including PlGF, may appear as an attempt at a compensatory response to maintain vascular function with the formation of new vessels, which might lead to enhanced capillary formation. Increased serum levels of angiostatic markers, such as thrombospondin 1, endoglin or endostatin, have been reported in SSc and might also contribute to the lack of angiogenesis despite increased levels of angiogenic biomarkers.
Our results also show that decreased EPC levels independently predict the occurrence of digital ulcers in SSc patients without any sign of previous digital ulcers. This is in accordance with our previous results, which showed a higher likelihood of digital ulcers in patients with low EPC levels. In contrast to the measurement of a single serum marker for the prediction of risk, the use of a cellular marker of risk, such as the level of EPC, unifies the complex interactions of multiple negative factors and may yield a better picture of in-vivo mechanisms.
While we hypothesise that an accurate identification of high-risk patients using PlGF and EPC levels will hasten the detection of digital ulcers and the introduction of adequate treatment at a time period when they may be of greater efficacy, this remains to be confirmed in dedicated studies. Further explorations are therefore now needed to assess PlGF and EPC as biomarkers of digital ulcers in terms of screening, diagnostic evaluation and response to therapy.
SSc is the most life threatening of the connective tissue disorders, although survival rates have improved in recent years.26 27 Heart involvement and pulmonary fibrosis are the main causes of morbidity and mortality.7 28,–,30 Consistent with these observations, five of seven deaths that occurred in our cohort were related to primary heart involvement and precapillary pulmonary hypertension.
The development of new cardiac/vascular events was frequent in our cohort after a mean follow-up of 29 months, with 27 new events. EPC and PlGF levels were also predictive of cardiac/vascular involvement defined as an exploratory composite index. This latter included digital ulcers, pulmonary hypertension, left ventricular dysfunction and SRC, as these complications all rely on the generalised vasculopathy. This observation is sustained by the well-established association between digital ulcers and pulmonary hypertension, as well as the most recent data showing an association between left ventricular dysfunction and digital ulcers.5,–,7 31 However, such a composite index, including both cardiac and vascular manifestations, will require further validation.
CEC levels have been correlated with endothelial dysfunction and were found to be elevated in patients with atherosclerosis, cancer or SSc.13 15 24 32 Elevated serum VEGF and sVCAM levels, such as elevated skin VEGF levels,10 17 have also been reported in SSc. However, none of these three biomarkers was predictive of the development of digital ulcers in SSc patients.
This study should be interpreted within its limitations. We cannot comment on patients with known renal or cardiac failure as they were excluded from the study. Despite the recent elaboration of dedicated recommendations, the detection of EPC by flow cytometry in peripheral blood remains challenging. No consensual technique of cell enrichment before fluorescence-activated cell sorter analysis or a combination of surface markers has yet been validated.33 Therefore, further studies are necessary regarding EPC quantification before being used routinely. Conversely, serum PlGF measurement by quantitative sandwich ELISA is easier to handle and more standardised, which suggests an easier application in routine care.
The number of patients who developed precapillary pulmonary hypertension, left ventricular dysfunction and SRC was too low to allow the firm identification of predictors. Further studies with a greater sample size and larger population are thus warranted to assess the predictive value of endothelial biomarkers for the development of these cardiac and vascular complications. This would also allow subgroup analyses according to the different causes of pulmonary hypertension and left ventricular dysfunction.
We did not perform a systematic capillaroscopic evaluation at inclusion, which prevented us from determining the value of this examination for the development of ischaemic digital ulcers. However, the merit of capillaroscopy for digital ulcer prediction is now under investigation in a European observational study (CAP study, protocol number AC-052-521).
In conclusion, our results suggest that circulating EPC counts and serum PlGF levels can be used to identify SSc patients who are at risk of the development of digital ulcers during a median 36-month follow-up period. This may have important clinical implications, in that non-invasive tests may be used to identify high-risk patients who should undergo more careful evaluation. Further studies assessing the use of these markers in therapeutic strategies aimed at the selection of patients for early treatment or preventive therapy are indicated.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Funding This study received financial support from the Société Française de Rhumatologie, Association des Sclérodermiques de France, Groupe Français de Recherche sur la Sclérodermie (bourse AMPLI), INSERM, Agence Nationale pour la Recherche (grant no R07094KS) and Fond d'Etude et de Recherche du Corps Médical des Hôpitaux de Paris. It was also supported by an investigator-initiated research grant from Pfizer Ltd.
Competing interests None.
Ethics approval Ethics approval was received from CPP Ile de France 3.
Provenance and peer review Not commissioned; externally peer reviewed.
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.