High-dose chemotherapy and autologous hematopoietic stem cell transplantation for progressive systemic sclerosis: a retrospective study of outcome and prognostic factors

Purpose Systemic sclerosis (SSc) is a rare autoimmune disease associated with high morbidity and mortality. SSc treatment is still challenging, and evidence is scarce. In the last decades high-dose chemotherapy and autologous stem cell transplantation (HD-ASCT) has proven to be effective. However, treatment related morbidity and mortality (TRM) are high. We conducted a retrospective, single-center analysis of SSc patients following HD-ASCT focusing on TRM and risk factors. Methods 32 patients who underwent HD-ASCT at our hospital between June 2000 and September 2020 were included. Clinical characteristics were evaluated based on chart review before and after HD-ASCT. Analyses focused on overall survival (OS), TRM, and response to HD-ASCT. Results Median OS was 81 months (range 0–243). Within one year, 20 of 32 (76.9%) patients responded to HD-ASCT. Overall, 6 patients (18.8%) died in the context of HD-ASCT. Patients with subjective response to HD-ASCT (p = 0.024) and those with shorter time to platelet engraftment (p = 0.047) had significantly longer OS. Impaired renal function, age at HD-ASCT ≥ 55, disease duration < 12 months, high Hematopoietic cell transplantation-specific comorbidity index (HCT-CI) and Charlton Comorbidity Index (CCI) scores were associated with higher TRM. Patients receiving conditioning chemotherapy with thiotepa needed longer time for neutrophil (p = 0.035) and platelet engraftment (p = 0.021). Conclusion This study confirms the efficacy of HD-ASCT for patients with SSc in a single center real-world setting. High TRM is still a challenge. However, TRM could be reduced by exclusion of high-risk patients and attention to prognostic parameters and scores as suggested in this study.


Introduction
Systemic sclerosis (SSc) is a rare autoimmune disease.It is characterized by pathologic accumulation of extracellular matrix in connective tissue, most probably due to dysregulated and dysfunctional repair mechanisms.This leads to sclerosis of skin, vessels, and organs, especially including heart and lungs.Although an increasing number of immunosuppressive drugs have expanded the therapeutic options, SSc still has the highest mortality rate of all rheumatic diseases with estimated 10-year survival rates between 55 and 73% (Tyndall et al. 2010;Nikpour and Baron 2014;Elhai et al. 2017;Bergamasco et al. 2019).
Among immunosuppressive drugs, intravenous cyclophosphamide is a widely accepted first-line medication due to its positive effects on various organ involvements, especially for SSc-related lung disease (Tashkin et al. 2006(Tashkin et al. , 2016)).In recent years, the concept of high-dose chemotherapy followed by autologous stem cell transplantation (HD-ASCT) has become increasingly important.The simplified rationale behind this treatment is that high-dose chemotherapy depletes autoreactive immune cells, which are thought to be at least partially responsible for the pathogenesis of SSc, and thereby a kind of immunological reset is accomplished (Del Papa et al. 2018).
Three randomized controlled trials demonstrated the superiority of HD-ASCT over cyclophosphamide (ASTIS-Trial, ASSIST-Trial, SCOT-Trial (Burt et al. 2011;van Laar et al. 2014;Sullivan et al. 2018)).Compared to conventional cyclophosphamide protocols, HD-ASCT resulted in longer overall survival (OS), extended progression-free survival (PFS), and improved partial remission rates of organ involvement.Despite these positive effects of HD-ASCT, treatment-related mortality (TRM) and complication rates remain challenging (Henes et al. 2021).Studies have reported TRM rates mostly between 4 and 17% (Binks et al. 2001;Vonk et al. 2008;Burt et al. 2011Burt et al. , 2013;;van Laar et al. 2014;Henes et al. 2014Henes et al. , 2021;;Del Papa et al. 2018;Sullivan et al. 2018;van Bijnen et al. 2020;Henrique-Neto et al. 2021;Blank et al. 2022).In current studies, male sex and older age were associated with higher TRM and morbidity (van Bijnen et al. 2020;Henes et al. 2021;Spierings et al. 2022).Overall, treatment-related deaths are often due to cardiac complications (Burt et al. 2013;van Bijnen et al. 2020;Henes et al. 2021).To select patients who will probably benefit from the advantages of HD-ASCT, there is a need to identify factors predicting risks and long-term outcome after HD-ASCT.However, the predictive power of comorbidity indices such as the Charlson Comorbidity Index (CCI) and the Hematopoietic cell transplantation-specific comorbidity index (HCT-CI) on OS, PFS, or TRM is largely unexplored in the setting of SSc and HD-ASCT (Charlson et al. 1987;Sorror et al. 2005Sorror et al. , 2014)).
Furthermore, the optimal timepoint for HD-ASCT in the course of the disease is still unclear.Considering the risks and potential long-term side effects, HD-ASCT is often performed in rather advanced SSc stages.However, this potentially results in increased complication rates, as SSc-associated organ damage reduces the functional reserve during treatment complications (Vonk et al. 2008;van Laar et al. 2014;Del Papa et al. 2018;van Bijnen et al. 2020).
Regarding outcome measures for HD-ASCT in SSc, research has previously focused on the relevance of patients' general condition and subjective perception.Evaluation with standardized health-related quality of life questionnaires shows an improvement in health-related quality of life in patients treated with HD-ASCT in comparison to conventional therapy (Maltez et al. 2021).
In this retrospective study, we present real-world data of 31 SSc patients treated with HD-ASCT at a tertiary referral center since 2000.Detailed analyses of outcome parameters including TRM, procedure related side effects, OS, and potential risk factors for HD-ASCT are discussed.

Study design and population
This retrospective, single-center study included 31 SSc patients, who received HD-ASCT at the University Hospital Knappschaftskrankenhaus Bochum, Germany, between June 2000 and September 2020.A total of 31 patients were included, of which one patient had received HD-ASCT twice.For the sake of simplicity, we will therefore refer to 32 performed HD-ASCTs in the further course.Three patients had also been included in the ASTIS trial and some of the patients participated in the DNSS study (van Laar et al. 2014;Blank et al. 2022).The local ethics committee of the Ruhr-University Bochum approved this study .

Data collection
Patient data including several parameters for systemic organ involvement before and after HD-ASCT was collected based on chart reviews.CCI and HCT-CI were calculated for each patient at the timepoint of HD-ASCT.Disease duration was defined as time between the patient's symptom onset or initial diagnosis and HD-ASCT.Skin involvement was present if patients had cutaneous manifestations that could be attributed to SSc or had a pathological value in the modified Rodnan skin score (mRSS).We considered pulmonary involvement of SSc according to imaging if the examining radiologist detected typical SSc related changes on chest X-ray and HR-CT.Functional lung parameters included forced vital capacity (FVC) and diffusing capacity of the lungs for carbon monoxide (DLCO).The presence of pulmonary arterial hypertension was defined as a mean pulmonary arterial pressure greater than 25 mmHg measured via right heart catheterization or estimated either with cardiac MRI or echocardiography.
Joint involvement was diagnosed based on either imaging findings (skeletal scintigram, X-ray, and/or MRI) or if the patient suffered from joint complaints such as pain, stiffness, or swellings due to synovitis, which could be plausibly attributed to SSc in the overall view of clinical symptoms and findings.

Study parameters
Study parameters were overall survival (OS) after HD-ASCT and treatment-related mortality (TRM).
OS was defined as the time between transplantation and patient death as well as last contact or cut-off.Death was classified as treatment related if the patient died within 100 days of transplantation and/or the procedure of HD-ASCT was the major causal influencing factor.
In addition, the patients' subjective responses to HD-ASCT, cutaneous improvement, dynamic development of DLCO and FVC, and overall response (OR) within one year post HD-ASCT were evaluated.
Under the broad parameter "cutaneous response" a response of the mRSS and improvement of the SSc-typical cutaneous manifestations are subsumed.In case of at least 25% mRSS reduction compared to baseline before HD-ASCT, improvement of mRSS was documented.
Lung function was evaluated on basis of DLCO and FVC.Changes in both parameters were categorized as following: "no change" for 0-9%, "improvement" for increases of ≥ 10%, or "deterioration" for decreases of ≥ 10% from baseline before HD-ASCT.A positive subjective response was determined if the individual patient reported a clear improvement of clinical symptoms compared to the status prior to HD-ASCT.OR denotes the presence of any response in at least one category listed above.

Statistical analysis
Data was analyzed with central tendency measures, boxplots, and frequencies.Due to small numbers, all continuous variables were tested for normal distribution using the Shapiro-Wilk test.The probabilities of OS were calculated using the Kaplan-Meier method.Two formed groups of patients on different independent variables were compared for the probability of OS after HD-ASCT using log-rank tests.An association between continuous/ordinal variables and OS was evaluated using Pearson's or Spearmann's correlation, depending on the scale level and the distribution of variable's values.Association between TRM and categorical variables was examined using the chi-square test, unpaired T test for continuously normally distributed variables, and Mann-Whitney U test for continuously non-normally distributed variables respectively.
Parameters before and after HD-ASCT were compared using paired-samples T tests for normally distributed variables and Mann-Whitney U tests for non-normally distributed variables.For multivariate analysis we applied Cox regression analysis.Statistical significance was defined as p < 0.05.All analyses were performed using IBM SPSS Statistics 28.Figures were generated using R.
Patients reporting a subjective response after HD-ASCT had a highly significant longer OS as compared to those not reporting a subjective response (n = 18, 75.0%; log-rank p < 0.001; Fig. 3a).Likewise, patients with a present overall response (OR) had a significantly longer OS as compared to patients with no OR (n = 20, 76.9%; p < 0.001, Fig. 3b).However, response in one of the other categories did not correlate with OS.
Interestingly, there was a significant negative association between time to platelet engraftment and OS (r = −0.431;p = 0.025).

Multivariate analysis
Multivariate Cox regression analysis was performed with the variables subjective response, gastrointestinal involvement, time to platelet engraftment and HCT-CI for OS.In multivariate analysis patients with subjective response (p = 0.024) or shorter time to platelet engraftment (p = 0.047) had a significantly higher OS.The presence of a subjective response within one year reduced the risk of death by more than 98% (HR: 0.016, 95%-CI 0.0-0.59).For every extra day to platelet engraftment, the risk of death increased approximately 2.3-fold (HR: 2.26, 95%-CI: 1.01-5.04).Gastrointestinal involvement (p = 0.973) and HCT-CI (p = 0.985) did not reach significance in multivariate analyses.

Discussion
SSc is a systemic autoimmune disease, for which HD-ASCT has proven to be superior to standard cyclophosphamide in several prospective clinical trials.Up to now only few studies exist, analyzing the risks and benefits of this infrequently used therapy in a real-world setting.Therefore, with this retrospective, single-center study we aimed to analyze the effectivity and safety of HD-ASCT in a real-world setting with a special focus on TRM and factors associated with TRM.
Notably, in our cohort 1-and 5-year OS rates were lower than those of other studies with 71.9% and 59.5%, respectively (Fig. 1).In the literature, for example 1-year survival rates are found in the range of 80-95% (van Laar et al. 2014;Henes et al. 2021).This was driven by an increased TRM of 18.8% in our cohort, as compared to other studies with approximately 0-17%, but mostly in the range of 5-10% (Binks et al. 2001;Vonk et al. 2008;Burt et al. 2011Burt et al. , 2013;;van Laar et al. 2014;Henes et al. 2014Henes et al. , 2021;;Del Papa et al. 2018;Sullivan et al. 2018;van Bijnen et al. 2020;Henrique-Neto et al. 2021;Blank et al. 2022).However, due to the lack of treatment alternatives the potentially life-saving HD-ASCT was offered to some patients in our center, which would have been excluded in clinical trials.To strengthen this assumption about our cohort, we followed the ASTIS study and applied its exclusion criteria to our patient population.According to the ASTIS study, patients with severe organ involvement or with severe comorbidities were excluded (van Laar et al. 2014).Translating these criteria to our patient cohort, we defined a group of patients who had a mPAP > 50 mmHg, CCI ≥ 4, DLCO < 25%, GFR < 45 ml/min, FVC < 40%, oxygen mandatory pulmonary insufficiency, or severe comorbidities such as COPD or myocardial infarction.Those 16 patients would have been excluded following ASTIS criteria.For those severely ill patients, median OS was 29 months whereas median was not reached for the fitter patients.Among these, only one treatment-related death occurred.Accordingly, TRM was reduced from 18.8% (n = 6) to 6.3% (n = 1), which is in the range of the prospective studies mentioned above.
Our retrospective observations are supported by another study by Binks et al. including a comparable patient population which had similar survival rates (1-year OS 73%) and TRM (NRM 17%) (Binks et al. 2001).Not all of our patients who were excluded according to ASTIS criteria or with comorbidity scores died related to HD-ASCT.For example, one patient with a CCI ≥ 4 was still alive 60 months after HD-ASCT.Another patient had exhibited an HCT-CI > 9, however, he survived HD-ASCT and had a favorable clinical response.Accordingly, these cases OR for TRM for all factors significant in chi-square-test, which were associated with a higher prevalence of TRM.Shown in each case is the group of HD-ASCT procedures that included patients at higher risk for TRM.HD-ASCT procedures including patients with an high age-adjusted hematopoetic cell transplantation (HCT)-Comorbidity Index ≥10 (n=4, 12.5%) had a 25-fold increased risk for TRM (Chi 2 (1)=9.5, p=0.002).Similarly 24-fold increased risk for TRM was noticed for patients with a high conventional HCT-Comorbidity Index ≥9 (n=6, 18.8%) (Chi 2 (1)=11.13,p<0.001) illustrate that even high-risk patients potentially benefit from HD-ASCT in advanced stages, which is line with the findings of a recently published study (Spierings et al. 2022).
A major purpose of our work was to identify factors potentially associated with TRM, in order to provide a rationale for selecting patients likely benefitting from HD-ASCT (Table 2).That is why different scores, such as the CCI and its further development the HCT-CI have been established in the last years.The HCT-CI was originally developed in the setting of allogeneic stem cell transplantation.Its validity in relation to autologous stem cell transplantation is still debated and its predictive power for TRM especially in systemic sclerosis has not yet been clarified (Saad et al. 2014;Sorror et al. 2015;Berro et al. 2017;van Bijnen et al. 2020;Barth et al. 2022).Here, we demonstrated that patients with a high CCI and especially high HCT-CI were at significantly increased risk for TRM (OR 11, OR 24, respectively (Charlson et al. 1987;Sorror et al. 2005;Sorror et al. 2014)).Therefore, calculating those scores before HD-ASCT in SSc patients seems to be a reasonable tool to predict survival and could therefore be used to facilitate decisions for intensive therapies.
Furthermore, low GFR turned out to be significantly associated with TRM in our study cohort.It is debatable whether it is nevertheless useful to perform HD-ASCT in patients with severe renal involvement, as they have a very poor prognosis with a one-year survival rate of only 70% (Guillevin et al. 2012).In the present study, patients with impaired renal function (GFR < 90 ml/min) showed a oneyear survival rate of 58%.No other organ involvement is associated with survival rates as low as scleroderma renal crisis.Therefore, HD-ASCT does harbor a high risk for TRM in these patients; however, the partial remission of the clinical picture due to transplantation may also prevent a future scleroderma renal crisis and terminal renal insufficiency in these patients as well as prolong survival in the long term.
Another important and still unanswered question is when to perform HD-ASCT during the course of disease.To address the question of HD-ASCT, we analyzed the time between symptom onset as well as initial diagnosis and transplantation in terms of OS and TRM.It could be shown that a particularly short period between SSc diagnosis and HD-ASCT (< 12 months) was associated with a higher risk of TRM.This can most likely be explained by the fact that a very short period between diagnosis and HD-ASCT is indicative for a very aggressive disease course and therefore leading to a worse outcome.Nevertheless, a better physical condition with less organ involvement improves patients´ outcome.Therefore, it could be important to have a close monitoring of the patients´ organ functions to ensure that the right timepoint for HD-ASCT is not missed.This is in line with the analyses of the recently published, large DNSS study that showed that patients with a disease duration of more than 3 years do not necessarily have a worse outcome after HD-ASCT (Blanks et al. 2022).
In this study, 75% of patients showed a therapeutic response to HD-ASCT within one year.Furthermore, there was no significant deterioration in the various organ parameters in the study cohort after HD-ASCT.In line with other studies, one can conclude that HD-ASCT has great therapeutic benefit for patients with SSc, which also holds true for patients in the real-world setting, even if the definition of response differs between studies (Vonk et al. 2008;Burt et al. 2011;Henes et al. 2021).Nevertheless 25% of patients do not respond and the risk of TRM is still eminent.Therefore, it is important to know, which groups of patients are more likely to respond after HD-ASCT than others.The question becomes even more relevant considering that in our study, patients who responded to HD-ASCT and especially patients with subjective response within one year after HD-ASCT had significantly longer OS (Fig. 3a).Overall and subjective response also proved to be the strongest independent factors influencing OS in multivariate analysis.In our setting it could be seen that clinical response was associated with less SSc organ involvement.This emphasizes the fact, that close monitoring for organ involvement should be performed and HD-ASCT should not be delayed when organ functions are deteriorating.Furthermore, it highlights the importance of patients´ subjective response, which is not easily quantified.A fact that should be addressed in future prospective trials.
In the last years two different conditioning regimens have been used: The standard CyATG protocol was extended by thiotepa (Henes et al. 2014) or fludarabine (Burt et al. 2021) in order to reduce cyclophosphamide dosage and thereby cardiac and other toxicities.Evidence regarding the TrCy-ATG protocol is based on a single prospective study comparing survival data from a small number of patients using this protocol with the results of other studies using the CyATG protocol (Giorgetti et al. 2004;Komatsuda et al. 2006;Henes et al. 2014).Henes et al. showed that patients treated with thiotepa needed longer time for hematologic recovery (Henes et al. 2014).This can be confirmed by our study since patients with a thiotepa-based conditioning protocol also had longer time for neutrophil and platelet engraftment.It can therefore be assumed that if TRM of cardiac etiology decreases, an increase in infectious complications must be accepted.
The retrospective design of our study may be considered for limited interpretation of our results.Nevertheless, with this real-world data, important information about patients excluded in other studies due to severe comorbidities or advanced disease stages has been gathered.
In summary, with this study we demonstrated the success and potent therapeutic potential of HD-ASCT in SSc patients.TRM remains a challenge that can be reduced to some degree by eliminating high-risk factors and could be even less, if patients´ organ functions were monitored more closely, in order to avoid missing the right timepoint for HD-ASCT.

Fig. 1
Fig. 1 Overall survival.Kaplan-Meier curve showing survival after high-dose chemotherapy and autologous stem cell transplantation (day 0) for all patients (n = 32)

Fig. 3 a
Fig. 3 a Overall survival by subjective response.Kaplan-Meier curve showing survival of patients with (n = 18, 75%) or without (n = 6, 25%) subjective response to HD-ASCT.Patients with subjective response had significantly longer OS than patients, who showed no subjective response (p < 0.001).b Overall survival by overall response (OR).Kaplan-Meier curve showing survival of patients with (n = 20, 76.9%) or without (n = 6, 23.1%) clinical response to HD-ASCT in at least one category (overall response; OR).Patients with OR had significantly longer OS than patients without (p < 0.001)

Table 1
Baseline characteristics before HD-ASCT Incomplete data, % from patients with available data # Percentages based on the number of HD-ASCT procedures of the heading category as reference population HD-ASCT High dose chemotherapy with autologous stem cell transplantation, HCT-CI hematopoietic cell transplantation specific comorbidity index, DLCO diffusion capacity of lung for carbon monoxide, FVC forced vital capacity, HR-CT high-resolution computer tomography, TTE transthoracic echocardiogram, LVEF left ventricular ejection fraction, GFR glomerular filtration rate, CDK-Epi equation chronic kidney disease epidemiology collaboration equation, SSc systemic sclerosis *

Table 2
Odds Ratios (OR) for treatment-related mortality (TRM)