Article Text

Download PDFPDF

Predictive factors for treatment-related mortality and major adverse events after autologous haematopoietic stem cell transplantation for systemic sclerosis: results of a long-term follow-up multicentre study
  1. Sandra van Bijnen1,
  2. Jeska de Vries-Bouwstra2,
  3. Cornelia H van den Ende1,
  4. Maaike Boonstra2,
  5. Lucie Kroft3,
  6. Bram Geurts4,
  7. Miranda Snoeren4,
  8. Anne Schouffoer5,
  9. Julia Spierings6,
  10. Jacob M van Laar6,
  11. Tom WJ Huizinga2,
  12. Alexandre Voskuyl7,
  13. Erik Marijt8,
  14. Walter van der Velden9,
  15. Frank HJ van den Hoogen10,
  16. Madelon C Vonk1
  1. 1 Department of Rheumatic Diseases, Radboud University Medical Center, Nijmegen, Netherlands
  2. 2 Department of Rheumatology, Leiden University Medical Center, Leiden, Netherlands
  3. 3 Radiology, Leiden University Medical Center, Leiden, Netherlands
  4. 4 Radiology, Radboud University Medical Center, Nijmegen, Netherlands
  5. 5 Rheumatology, Haga Hospital, the Hague, Netherlands
  6. 6 Rheumatology and Clinical Immunology, Universitair Medisch Centrum Utrecht, Utrecht, Netherlands
  7. 7 Department of Rheumatology, Amsterdam UMC, Amsterdam, Noord-Holland, Netherlands
  8. 8 Haematology, Leiden University Medical Center, Leiden, Netherlands
  9. 9 Haematology, Radboud University Medical Center, Nijmegen, Netherlands
  10. 10 Rheumatology, Radboud University Medical Center, Nijmegen, Gelderland, Netherlands
  1. Correspondence to Dr Madelon C Vonk, Department of Rheumatic diseases, Radboud UMC, Nijmegen, Netherlands; Madelon.Vonk{at}


Background Autologous haematopoietic stem cell transplantation (HSCT) improves survival in systemic sclerosis (SSc) with poor prognosis, but is hampered by treatment-related mortality (TRM).

Objective To evaluate event-free survival (EFS), TRM, response to treatment, disease progression and patient characteristics associated with events.

Methods All patients treated with HSCT for SSc in The Netherlands until 2017 (n=92) were included. Data on skin involvement (modified Rodnan skin score (mRSS), pulmonary function (forced vital capacity (FVC) and diffusion capacity of the lungs for carbon monoxide (DLCO)), extent of interstitial lung disease on high-resolution CT using Goh scores and left ventricular ejection fraction (LVEF) were collected at baseline, 1, 2 and 5 years. Occurrence of events, defined as death or major organ failure, were collected until 2019. As control, a comparison between patients treated with cyclophosphamide (CYC) and patients with HSCT who participated in the Autologous Stem Cell Transplantation International Scleroderma (ASTIS) trial was performed.

Results Median follow-up was 4.6 years. EFS estimates at 5, 10 and 15 years were 78%, 76% and 66%, respectively. Twenty deaths occurred. Mean FVC, DLCO, mRSS and Goh scores all improved significantly. Disease progression occurred in 22 patients. Frequency of TRM decreased over time and occurred more often in males. Events were independently associated with male sex, LVEF <50% and older age. In ASTIS, patients treated with HSCT (n=23) 7 events occurred versus 13 in the CYC group (n=22).

Conclusion Our data confirm long-term efficacy of HSCT in improving survival, skin and lung involvement in SSc. Male sex, lower LVEF and older age at baseline were identified as risk factors for events.

  • systemic sclerosis
  • treatment
  • autoimmune diseases

Statistics from

Key messages

What is already known about this subject?

  • Autologous haematopoietic stem cell transplantation (HSCT) improves survival in systemic sclerosis (SSc) with poor prognosis, but is hampered by treatment-related mortality.

What does this study add?

  • All patients treated with HSCT for SSc in The Netherlands were included with a median follow-up of 4.6 years.

  • Frequencies of mortality, organ failure and disease progression were determined and univariate and multivariate Cox regression analysis was used to evaluate factors associated to events.

  • Response to treatment was analysed for event-free survival, skin and pulmonary involvement.

How might this impact on clinical practice or future developments?

  • Baseline characteristics were identified as risk factors for events, attributing to further improve benefit–risk ratio of HSCT.


Systemic sclerosis (SSc) is a heterogeneous autoimmune disease of unknown aetiology, in which inflammation, fibrosis and microvascular damage affect the skin and vital organs, including the lungs, heart, kidney and the gastrointestinal tract. Particularly, the rapidly progressive diffuse form, in which skin sclerosis is localised proximally of knees and elbows, is a devastating disease with poor prognosis. Interstitial lung disease (ILD) and cardiac involvement are the major causes of death among patients with SSc.1

Treatment of rapidly progressive diffuse SSc consists of immunosuppression, including mycophenolate mofetil, cyclophosphamide (CYC) or autologous haematopoietic stem cell transplantation (HSCT).2 Multiple studies, including three randomised trials, have shown the superior efficacy of treatment with HSCT above intravenous CYC pulse therapy in selected patients with diffuse cutaneous SSc (dcSSc) and disease duration <4 years.3–5 Two of these studies describe increased long-term event-free survival (EFS), that is, survival without respiratory, cardiac or renal failure, increased overall survival and greater improvement in modified Rodnan skin score (mRSS) in patients who underwent HSCT compared with control treatment. However, treatment-related mortality (TRM) with HSCT is variable up to 10% and factors predicting TRM are largely unknown. Identification of patients at higher risk for TRM or major organ failure despite HSCT is highly important to enable adequate patient selection and improve the outcome of HSCT in SSc. The haematopoietic cell transplant comorbidity index (HCT-CI), a weighted score of 15 pretransplant comorbidities, has shown value to predict survival after HSCT in patients with haematological malignancies.6 7 Currently, its value in SSc is unknown.

Here, we evaluate the long-term outcome of 92 patients with SSc treated with HSCT in the Netherlands, in particular EFS, TRM, response to treatment and disease progression. Furthermore, we have investigated whether the HCT-CI, in addition to baseline characteristics, contributes to identify patients at high risk for complications related to HSCT.



This cohort includes all patients treated with autologous HSCT for SSc between 1998 and 1 July 2017 in the Netherlands (n=92). All patients fulfilled the American College of Rheumatology/European League Against Rheumatism classification criteria and, except those who died within 1 year, were required to have at least 1 year of follow-up.8 HSCT procedures were performed in four hospitals: Radboud University Medical Centre in Nijmegen (n=40), Leiden University Medical Centre (n=39), University Medical Centre Utrecht (n=7) and Amsterdam University Medical Centre (n=6). Twenty-six patients had participated in the Autologous Stem Cell Transplantation International Scleroderma (ASTIS) trial.4 Due to the retrospective nature of this study, informed consent was not deemed required, but was obtained from all alive patients not lost to follow-up.

Autologous HSCT procedure

In all centres, HSCT was performed as previously described in the ASTIS trial.4 Briefly, the procedure consisted of mobilising haematopoietic progenitor cells using CYC (2 g/m2/day for 2 days) followed by filgrastim (10 μg/kg/day for 5 days). CD34+ cells were harvested by leukapheresis and subsequent immunomagnetic separation. Conditioning consisted of CYC (50 mg/kg/day for 4 days) followed by rabbit antithymocyte globulin (2,5 mg/kg/day for 3 days) and methylprednisolone (1 mg/kg/day). Thereafter, autologous CD34+ cells were reinfused.

Data collection

Data were collected pretreatment (≤6 months before HSCT), at baseline and at years 1, 2 and 5 (±6 months) thereafter. Survival status and the occurrence of major organ failure were assessed up to 2019. We collected data on mRSS, respiratory involvement (forced vital capacity % predicted (FVC % pred); diffusion capacity % predicted (DLCO % pred)), cardiac involvement (left ventricular ejection fraction (LVEF)), presence of pulmonary hypertension (cardiac ultrasound, and if available, right heart catheterisation), renal involvement (creatinine, dialysis) and post-transplantation immunosuppressive treatment. At baseline, the HCT-CI was determined.6 Of note, in this index, a rheumatological disease provides two points and any pulmonary disease resulting in DLCO between 66% and 80% pred provides two points, or three points when DLCO is below 66% pred.

Evaluation of outcomes

Duration of EFS was defined as the absence of all-cause mortality or cardiac, respiratory or renal failure (table 1). All deaths were discussed in a consensus meeting (SB, MB, JV, MV) and classified as TRM, due to progression of SSc or other reasons. Response to treatment was evaluated in terms of changes in mRSS, pulmonary function and extent of ILD on high-resolution CT (HRCT) scan using Goh scores without taking the FVC % pred into account.

Table 1

Definition of mortality, major organ failure and disease progression

Disease progression was defined as progression of skin or pulmonary involvement or the necessity of immunosuppressive drugs for any manifestation of SSc (table 1). An additional analysis was performed to address the possible benefits of HSCT compared with standard of care. For this, we used the subgroup of ASTIS patients from the two main Dutch centres in ASTIS and compared EFS of patients randomised to HSCT (n=23) with EFS of ASTIS patients randomised to CYC (n=22).

Severity of ILD on HRCT, according to Goh scores

HRCT extent of ILD was determined by mean Goh scores without taking the FVC values into account.9 All patients with available HRCT scans pretreatment and at 1, 2 and/or 5 years of follow-up were evaluated by two pairs of experienced observers. To ensure consensus between the observers, all observers completed a training set of 10 HRCTs. Differences in scores were discussed and final Goh scores were determined by averaging the Goh scores of two observers, or when differing >10%, by consensus.

Statistical analysis

Frequencies of all-cause mortality, major organ failure, TRM and disease progression were determined. A Kaplan-Meier event-free survival (EFS) curve was constructed. Univariate Cox regression analysis was used to evaluate the association between baseline characteristics and EFS and TRM (complete cases). Baseline characteristics included age, sex, disease duration, smoking (ever), FVC % pred, DLCO % pred, LVEF, HCT-CI, mRSS and change in mRSS during 6 months prior to HSCT treatment. For EFS, baseline factors with an association with p<0.1 were entered in a multivariate Cox regression analysis. To reduce the impact of missing data, data at baseline were imputed using multiple imputations to create 20 datasets. For TRM, only univariate cox regression could be performed, given the low number of events. Linear mixed model analysis was used to estimate the effect of time on response to treatment, that is, the course of Goh scores, mRSS, FVC % pred and DLCO % pred during the 5 years after HSCT excluding deceased patients.


Patients and treatment

Ninety-two patients who initiated HSCT treatment were included. Baseline characteristics are shown in table 2. The median follow-up time was 4.6 years (IQR 2.1–12.4 years). The majority of the patients (96%) had dcSSc. The mean mRSS (SD) was 26(10) and median disease duration was 1.5 years (IQR 0.9–3.0 years). At baseline, the mean (SD) HCT-CI was 4.9 (1.3). Three patients did not complete the HSCT procedure because of progressive disease after stem cell mobilisation and reaching exclusion criteria, consisting of cardiac failure (n=2) and respiratory failure (n=1). Baseline characteristics of ASTIS patients (n=26) did not differ from patients not participating in ASTIS (n=66; online supplementary tables 1 and 2 and figure 1).

Table 2

Baseline characteristics

Figure 1

Event-free survival of all patients initiated with haematopoietic stem cell transplantation procedure.


EFS estimates (95% CI) at 5, 10 and 15 years were 78% (67% to 86%), 76% (64% to 84%) and 66% (48% to 79%), respectively (figure 1). During follow-up, 28 events occurred in 22 patients, including 20 deaths and major organ failure in 8 patients (figure 2). Ten deaths (10/92, 11%) were classified as TRM, including 7/51 (13.7%) before 2012 and 3/41 (7.3%) after 2012. Four deaths (4/92, 4%) were classified as a consequence of SSc and six (6/92, 7%) were attributed to other causes. Eight cases of TRM occurred within 3 weeks after HSCT, the two other cases occurred at 2 months after HSCT. The majority of patients with TRM (n=9, 90%) died from cardiorespiratory failure.

Figure 2

Events during follow-up in 92 patients. SSc, systemic sclerosis; TRM, treatment-related mortality.

Eight patients developed major organ failure during follow-up. Median duration to an event was 0.3 years (IQR 0.1–4.4 years). These included renal failure (n=2), respiratory failure (n=5, of whom one patient received lung transplantation), and cardiac failure (n=1). Six of the eight patients with major organ failure died; one patient with cardiac failure and one patient with respiratory failure survived. We found no difference in survival between patients with baseline Goh scores >20% compared with ≤20%. (online supplementary figure 2). In our cohort, 33 patients had early ILD, defined as FVC ≥80% pred and DLCO ≤70% pred. Seven of the patients with early ILD had an event: six died and one developed cardiac failure (online supplementary figure 3 and online supplementary tables 3 and 4).

Table 3

Univariate and multivariate associations (Cox regression analysis) between patient and disease-related characteristics and events, presented as HRs (95% CI)

In total, 7 events (30%) occurred in ASTIS patients randomised to HSCT (n=23) versus 13 events (59%) in ASTIS patients randomised to CYC (n=22; online supplementary figure 4 and table 5).

Factors associated with events and TRM

Male sex, older age, higher HCT-CI and LVEF <50% at baseline were associated with lower EFS (table 3). Increase of mRSS >5 points in the last 6 months before HSCT was associated with higher EFS (table 3). For TRM, only male sex was predictive (HR 8.6 (95% CI 1.1 to 68.0)) Previous or current smoking was not related to events or TRM. Multivariate analysis (Cox hazard) revealed that male sex, LVEF <50% and older age were independently associated with events (table 3).

Response to treatment

During follow-up, skin involvement, pulmonary function tests and Goh scores all improved significantly in patients who survived the first 5 years (figure 3). The median (IQR) mRSS decreased from 26 (18–32, n=69) at baseline to 6 (2–7, n=35) at 5 years of follow-up. The median (IQR) FVC and DLCO % pred increased from 84% (68%–102%, n=66) and 55% (42%–67%, n=67) at baseline to 94% (81%–107%, n=40) and 61% (53%–73%, n=38) at 5-year follow-up. Median Goh scores improved from 14% (7%–34%, n=39) at baseline to 8% (3%–23%, n=16) at 5 years of follow-up (figure 3). Estimates for the mean (95% CI) improvement per year for mRSS, FVC % pred, DLCO % pred and Goh scores were −3.9 (−4.4 to −3.3), 2.5 (1.9 to 3.0), 1.6 (1.0 to 2.2) and −1.0 (−1.9 to 0.0), respectively (figure 3).

Figure 3

Course over time excluding deceased patients in the first 5 years (n=75). x-axis: time in years. y-axis: (A) skin involvement as measured by modified Rodnan skin score (mRSS). (B) Forced vital capacity (FVC) pred: forced vital capacity % predicted. (C) Diffusion capacity of the lungs for carbon monoxide (DLCO) pred: diffusion capacity % predicted. (D) Total Goh: extent of pulmonary involvement on HRCT scan. Lines represent medians (25–75 percentiles).

Disease progression

Disease progression occurred in 22 (24%) patients during the first 5 years. Progression of the skin occurred in 5 (5%) patients, of the lungs in 11 (12%) and in 12 (13%) patients immunosuppressive therapy was started or increased; however, 7/12 did not fulfil prespecified criteria for disease progression of skin or lung. The majority of cases of disease progression (19/22) occurred within the first year after HSCT.


With this large, multicentre, long-term follow-up study including 92 patients, we confirm the clinical efficacy of HSCT as treatment for selected cases of rapidly progressive diffuse SSc which justifies addition of HSCT to the therapeutic arsenal. Despite characteristics indicating severe disease at baseline, EFS was 78% at 5, 76% at 10 years and 66% at 15 years after HSCT. In line with previous trials, we observe relatively high TRM of 11% which decreased over time. We identified male sex, impaired LVEF and older age as important predictors of worse outcome.

Our results are comparable to the HSCT arms in the ASTIS (EFS 81% (n=79) at 5.8 years) and Scleroderma: Cyclophosphamide or Transplantation (SCOT) trial (EFS 79% (n=36) at 4.5 years), and better than in previous longitudinal cohort studies.4 5 10–13 TRM in patients with SSc undergoing HSCT postulated to be influenced by factors including transplant regimen and thus cardiotoxicity, patient selection and centre experience.14 15 The majority of cases of TRM in our study were due to cardiorespiratory failure during the early post-transplant phase, confirming the need for extensive cardiac screening before HSCT.11 14 The presence of cardiac disease might reflect more extensive organ involvement and poor performance state both associated to poor outcome.16 In our cohort, cardiac screening consisted of echocardiography and ECG. In line with published guidelines, since 2012, this was extended with left and right heart catheterisation including fluid challenge, 24 hours Holter registration and cardiac MRI when indicated.15 We found a decrease in TRM from 14% (n=7/51) before 2012 to 7% after 2012 (n=3/41). This indicates that, besides improvement of supportive care, better selection of patients by more intensive screening contributes to reduction of TRM.

One of the aims of our study was to identify factors that can discriminate patients at high risk for TRM. Because of relatively low number of TRM, we lack power to identify all relevant factors. With univariate analysis, male sex was identified as a risk factor. We could not confirm the association between smoking and TRM or other events, as described in both the ASTIS and the SCOT study.4 5 Whether male sex increased the risk for treatment-related complications or whether this observation reflects that males have more severe disease compared with females remains to be elucidated.

Another aim of this study was to identify factors that help predict the outcome of HSCT in order to improve selection of patients that may have the best benefit of HSCT. In our multivariate analysis, male sex, older age and LVEF <50% were associated with events. Strikingly, patients with a significant increase in skin involvement in the 6 months prior to HSCT seemed to have higher EFS. This might indicate that this specific subset of patients with clinically active disease could benefit most from HSCT, which is in line with observations on skin signatures of the SCOT trial. However, regarding the retrospective nature of our study, this finding should be interpreted with care.

We specifically evaluated whether the HCT-CI can contribute to patient selection. In a large prospective observational study of HSCT recipients for haematological malignancy, an HCT-CI score of ≥3 was associated with a higher risk of mortality.7 Mean HCT-CI score in our cohort was 4.9 (SD 1.3). Although there was an association between events and HCT-CI, in the multivariate analyses HCT-CI did not contribute. Probably, the contribution of other comorbidities is relatively low compared with cardiac and pulmonary involvement related to SSc itself, resulting in little discriminative value of the HCT-CI compared with LVEF alone.

As for response to treatment, we found that the decrease of mRSS was comparable to the decrease reported in the systematic review of the published trials.17 However, we found a clinically significant increase of pulmonary function test results which has not been reported before.17 In our cohort, the FVC increased during follow-up (median, % pred) from 84% at baseline to 94% at 5 years of follow-up. The DLCO (median, % pred) increased from 55% at baseline to 61% at 5 years, which is significant although probably not clinically meaningful. Assessment of HRCT scan images for ILD over longer time in patients with SSc treated with HSCT has not been performed before. One small study in nine patients evaluating HRCT images at baseline, 6 and 12 months using the Wells score, showed an increase of pulmonary involvement at 1 year after a transient decrease after 6 months.18 Burt et al evaluated pulmonary involvement in 10 HSCT patients using volumetric software and showed a decrease in pulmonary involvement at 1 year compared with CYC.3 Our study is the first to evaluate Goh scores before and after HSCT and our findings suggest an improvement over time.

Disease progression occurred at similar or higher frequencies in our cohort during the first 5 years of follow-up (24%) compared with other longitudinal studies.4 5 12 This is partly explained by the stringent definition of disease progression. Our definition of progressive disease is in accordance with the proposed Outcome Measures in Rheumatology (OMERACT) definition of clinically meaningful progression in ILD.19 This definition was met in 12% of patients in the first 5 years after HSCT. Skin relapse occurred in 5%. The majority of cases of disease progression consisted of patients initiating or increasing the dosage of immunosuppressive drugs, which accounted for 13% relapses after 5 years, regardless of fulfilling the relapse criteria for skin or lung.

This study has some limitations that should be taken into account. First, our comparison of EFS between HSCT and CYC was based on a small sample, impeding firm conclusions about difference in long-term survival. Second, it is a retrospective study with consequent methodological issues including, for example, survival bias in the responder analyses. However, we think that the findings are valuable as they reflect results of HSCT in daily clinical practice as the majority of patients did not participate in a trial. Third, all patients were treated according to the ASTIS scheme, so results cannot be generalised to other transplant regimens.4 On the other hand, the fact that all patients were treated using the same regimen enables and strengthens current responder and event analyses. As compared with the ASTIS study, described TRM was lower in the SCOT trial which applied a conditioning regimen using a smaller dosage of CYC combined with ATG and total body irradiation.4 5 Strikingly, in the SCOT trial, TRM was 4% between 54 and 72 months compared with 3% during the first 54 months.5 With the ASTIS regimen, the majority of TRM occurs in the first months after HSCT.4 Whether and how these differences are accounted for by differences in conditioning regimens, learning effects or patient selection, is currently unclear.

In conclusion, the results of this large historical cohort study of patients with severe SSc confirm efficacy of HSCT in clinical practice in terms of EFS, response to treatment and disease progression rate. This study also shows that EFS at longer term after HSCT continues to be relatively high (ie, 66% at 15 years). We show that male sex, older age and worse cardiac function are associated with higher risk for events despite HSCT. We were able to show a decrease of TRM over time, but still identification of additional factors that can identify patients at risk of TRM are of key importance to further improve risk–benefit ratio of HSCT in clinical practice.



  • Handling editor Josef S Smolen

  • Contributors SvB and JKdV-B contributed equally and share first authorship. All authors contributed to drafting and reviewing the manuscript and approved the final version for publication. CE and MV had access to all data and analyses. Study conception and design: MV, SvB, CE, JKdV-B and MB. Acquisition of data, analysis and interpretation of data: all authors.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

  • Patient consent for publication Not required.

  • Ethics approval The current study has been approved by the medical ethics committee CMO region Arnhem-Nijmegen, Nijmegen, the Netherlands NL-2016-2915. Due to the retrospective nature of this study, informed consent was not deemed required. Informed consent was obtained from all alive patients not lost to follow-up.

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

  • Data availability statement Data are available upon reasonable request. All data are available on request to Investigators whose proposed use of the data has been approved by an independent review committee identified for this purpose and for meta-analysis, as well as the study protocol and analysis plan.

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.