Autologous hematopoietic stem cell transplantation (AHSCT) for Diffuse Systemic Sclerosis. Health Technology Assessment
Health technology assessment
|Published
Autologous hematopoietic stem cell transplantation (AHSCT) is suggested for rapid and severe development of diffuse systemic sclerosis. We have summarized the effect and safety, conducted a cost analysis and discussed ethical issues related to the treatment.
Key message
Autologous hematopoietic stem cell transplantation (AHSCT) is suggested for rapid and severe development of diffuse systemic sclerosis. We have summarized the effect and safety, conducted a cost analysis and discussed ethical issues related to the treatment. We included five studies comparing AHSCT with standard treatment, three of which were randomized. Compared with standard treatment we found that:
- Transplant-related mortality was between seven and ten percent after AHSCT. We have low confidence in the effect estimate.
- Total mortality and organ failure rates at two years and beyond were lower in the AHSCT group than in the standard treatment group. We have moderate confidence in the result.
- AHSCT resulted in more side effects and adverse events including viral infections than standard treatment measured after two and six years. We have moderate confidence in the effect estimates.
- Skin involvement improved after AHSCT measured two and five years after treatment. We have moderate confidence in the result.
- Lung function measured by forced vital capacity and total lung capacity also improved to a certain extent after AHSCT, while the diffusion capacity for carbon monoxide (DLCO) and residual volume were not different from standard treatment measured two and five years after treatment. We have moderate confidence in the effect estimate for DLCO, and low for the others.
- AHSCT improved self-perceived physical health measured two and five years after treatment. We have moderate confidence in the effect estimates.
- Costs associated with AHSCT in the treatment-year are approximately 600,000 kroner per year per patient.
- AHSCT is ethically challenging as it may lead to premature death, but it may also lead to improved function and may be life-prolonging.
Total mortality and organ failure were lower after AHSCT than after standard treatment at two-year follow-up and beyond. AHSCT had a positive effect on multiple outcomes, but transplant-related mortality was high, and the selection of appropriate patients for treatment is challenging.
Summary
Background
Systemic sclerosis, also called scleroderma, belongs to the group of rheumatoid diseases, and is a rare autoimmune systemic connective tissue disease. In systemic sclerosis, connective tissues are attacked in the skin, blood vessels and internal organs with a progressive fibrosis formation. Autologous hematopoietic stem cell transplant (AHSCT) has been suggested for a small group of patients with diffuse systemic sclerosis who have insufficient effect of standard pharmacological treatment to prevent serious and irreversible organ damage. Ordering Forum, Bestillerforum RHF, has commissioned the Norwegian Institute of Public Health to prepare a health technology assessment comparing AHSCT with standard treatment for patients with systemic sclerosis.
Objective
To perform a systematic review of efficacy and safety, an economic evaluation and ethical evaluation related to AHSCT for systemic sclerosis.
Method
We conducted a systematic search for randomized and non-randomized controlled studies as well as registry studies on 18.02.2016. In addition, we regularly checked PubMed for publications on an expected randomized controlled study. The last search was performed on 06.10.2017. The inclusion criteria were: patients above 18 years with systemic sclerosis. Intervention: autologous AHSCT. Comparison: standard pharmacological treatment. Outcomes: mortality and organ failure, adverse events, disease progression measured by skin, lung, heart and kidney involvement, and health-related quality of life. Risk of bias and methodological quality were assessed with Cochrane's Risk of Bias form and checklist. Effect estimates were calculated as risk ratio for dichotomous outcomes and mean difference or standardized mean difference (SMD) for continuous outcomes. Hazard ratios were used in survival analysis. The quality of the evidence was assessed using the GRADE tool for each outcome. The quality of the evidence, i.e., whether we have confidence in that an effect estimate is close to a theoretically true underlying effect, can be considered as high, moderate, low or very low.
Based on the low number of patients potentially eligible for AHSCT according to the suggested criteria in Norway, and lack of a complete cost estimate for standard treatment, we chose to perform a cost analysis from a healthcare perspective. In the description of costs related to AHSCT treatment, we included pre-treatment evaluation of the patient, costs of the procedure itself, and costs of follow-up and infection prophylaxis up to one year after AHSCT.
We also searched for literature to discuss selected questions from the "Checklists for ethical evaluations". The issues were discussed from different perspectives: that of the patient, the health care provider and -system, and that of the society.
Results
We included a total of five studies, three randomized controlled studies; the ASTIS- (n = 156), SCOT- (n = 75) and ASSIST-studies (n = 19), one small non-randomized controlled study (n = 16), and one large European multicenter registry study, EBMT (n = 175). Based on the results from the ASTIS and the registry study, we estimated transplant-related mortality to be between 7 and 10%. In the ASTIS study, with a median follow-up of 5.8 years, mortality was highest during the first three to six months in the AHSCT group, for then to gradually level off. From two years of follow-up and beyond, mortality (hazard ratio = 0.29 [0.13 to 0.65]) and the sum of mortality and organ failure (hazard ratio = 0.35 [0.13 to 0.74]), was lower in the AHSCT group than in the control group. A meta-analysis of total mortality in the ASTIS- and SCOT studies gave a risk ratio of 0.61 (0.40 to 0.93) in favor of the AHSCT group after approximately five years. The three randomized controlled studies reported that side effects and adverse events, as well as viral infections, occurred more frequently in the AHSCT group than in the control group, RR = 1.4 [1.04 to 1.87] and RR = 11.74 (3.76 to 36.70), respectively, measured after two and six years. We have low confidence in the results of transplant-related mortality, and moderate confidence for total mortality and organ failure, adverse events and virus infections. Skin involvement measured by the modified Rodnan skin score (mRSS) and forced vital capacity (FVC) improved after AHSCT, estimated as a large (SMD = -0.99, [-1, 33 to -0.66]) and a medium effect (SMD = 0.51 [0.19 to 0.83]) by standardized mean difference (SMD). Total lung capacity (TLC) (SMD = 0.40 [0.09 to 0.72]) and physical health-related quality of life (SF-36: SMD = 0.44 [0.12 to 0.76]) was also significantly better. No differences were found between the AHSCT and the control group in left ventricular echocardiography (MD = -0.30 [-4.77 to 4.17]) and in the lung function measurement diffusion capacity for carbon monoxide, DLCO, (SMD = -0, 04 [-0.35 to 0.28]). We have moderate confidence in the quality of evidence for skin involvement (mRSS), DLCO, and physical health (SF-36) and low confidence in the quality of evidence for the other outcomes (FVC, TLC and mental health SF-36).
The total cost of treatment with AHSCT in a one year perspective is approximately 600,000 kroner per patient.
Discussion
We conclude that transplant-related mortality is between seven and ten percent based on the results from the ASTIS- and the registry study. It has been pointed out that a thorough examination of the heart, especially for pulmonary hypertension, is required prior to treatment with AHSCT in order to better select patients suitable for the treatment. In the ASTIS study, the 2004 guidelines were followed, which were applicable at the time the patients were included in the study, but points out that further cardiac investigations may be necessary to reduce transplant-related death. Further, the ASTIS study found that total mortality and organ failure were lower in the intervention group than in the control group from two years of follow-up and beyond. Although transplant-related mortality was high, AHSCT could thus have a positive effect on mortality and severe organ failure in the long term. In Norway, the selection of patients is closely linked to the inclusion criteria in the three randomized controlled studies ASTIS, SCOT and ASSIST. We therefore assume that the results can be transferable to the Norwegian conditions.
A number of transient side effects are expected in stem cell treatment that do not occur to the same extent in standard treatment. Some serious and life-threatening adverse events may possibly be attributed to these. In the long term, with a median follow-up of approximately six years, more virus infections occurred in the AHSCT group than in the control group, indicating that the immune system has been impaired for a long time. The results for clinical effect indicated improvement after AHSCT in several areas, especially for the skin, but whether this improvement lasts to the same extent beyond two years is more uncertain.
We chose to limit the economic analysis to describe costs. The patient group is very heterogeneous and has varying needs for health care services. Patients who might be relevant for AHSCT treatment are severely ill, often with involvement of one or more internal organs, and requiring examinations and treatment across multiple medical specialties. In order to estimate pharmacological costs alone for a single patient with conventional treatment, we used estimates from a mini-HTA conducted by Oslo University Hospital in 2015. These amount to approximately 730,000 Norwegian kroner per year in addition to other direct costs such as repeated hospitalizations, wound care, oxygen devices and pacemaker insertions. It is unclear to what extent the use of health care is altered after the patients have been treated with AHSCT. It was therefore impossible to carry out a reliable comparative analysis of the two treatment options. We have reported the budget impact per patient for AHSCT for standard pharmacological treatment during the first treatment year as a result.
AHSCT is a comprehensive treatment that can both be harmful and cause premature death, but may also improve the patient's functioning, quality of life and be life-prolonging provided appropriate patient selection. The treatment is not without ethical issues that the physician and the patient should discuss carefully before the treatment begins.
The quality of the evidence in this HTA is mainly low and moderate because neither the patients nor the staff were blinded, and the number of participants included in the comparisons/analyses is a little too low to be considered as sufficient in the GRADE assessments. However, the evidence for clinical effect pointed in the same direction in all studies, which strengthens the confidence in the results.
Diffuse systemic sclerosis is a rare disease, and patients relevant to stem cell treatment are only a small sample of this group. Therefore, it is not likely to include large patient groups in new randomized controlled studies. The need for long-term follow-up in the future may possibly be met by the EBMT registry in terms of mortality, side effects and adverse events.
Conclusion
Our conclusion is that transplant-related mortality after AHSCT is high, but risk of organ failure and death, for whatever reason, is nevertheless lower after two years follow-up and beyond compared to standard treatment. AHSCT provides a significant improvement of skin involvement and lung function measured by forced vital capacity compared to standard treatment with cyclophosphamide injections two to five years after initiation of treatment, but side effects and adverse events occur more frequently than in standard treatment. Our confidence in the quality of the evidence is low and moderate, mainly due to the limited number of patients in the studies, and that blinding is not possible in this type of study.
