Intended for healthcare professionals


Are the risks of treatment to cure a child with severe sickle cell disease too high?

BMJ 2017; 359 doi: (Published 23 November 2017) Cite this as: BMJ 2017;359:j5250
  1. Mariane de Montalembert, professor1,
  2. Valentine Brousse, consultant paediatrician1,
  3. Subarna Chakravorty, consultant paediatric haematologist2,
  4. Antonio Pagliuca, professor3,
  5. John Porter, professor4,
  6. Paul Telfer, consultant haematologist5,
  7. Ajay Vora, professor6,
  8. David C Rees, professor2
  1. 1Department of Pediatrics, Reference Centre for Sickle Cell Disease, Hôpital Universitaire Necker-Enfants Malades, APHP, Paris, France
  2. 2Department of Paediatric Haematology, King’s College Hospital, London SE5 9RS, UK
  3. 3Department of Haematological Medicine, King’s College Hospital, London
  4. 4Department of Haematology, University College London Hospitals
  5. 5Department of Paediatric Haematology and Oncology, Barts Health NHS Trust, Royal London Hospital
  6. 6Department of Haematology, Great Ormond Street Hospital for Children, London
  1. Correspondence to: D Rees david.rees2{at}

Haematopoietic stem cell transplantation from alternative donors is curative for children with sickle cell disease, but risk of death is high. David Rees and colleagues call for greater caution in its use

Key messages

  • Curative treatment for sickle cell disease is potentially available to all patients using haematopoietic stem cell transplantation from alternative donors, but the risks of these procedures are very high

  • We don’t know when these high risk transplants should be offered to children or how parents should be counselled to give appropriate consent

  • In high income countries, non-transplant treatments are being developed for sickle cell disease that are likely to improve prognosis

  • We need to develop safer transplants and to explore the ethics of offering high risk procedures to children with sickle cell disease and other chronic conditions

Sickle cell disease is one of the most common severe inherited conditions in the world. Around 300 000 babies are born with sickle cell disease each year, mostly in Africa, although there are roughly 100 000 affected people in the US and 50 000 in Europe.1 The prognosis for children born with the condition today varies enormously, particularly with geography: only about 20% of babies born in Africa survive to adulthood,2 whereas more than 93% of children survive to adulthood in Europe3 and the US,4 thanks to basic medical care, screening programmes, vaccinations, prophylactic antibiotics, blood transfusions, stroke prevention, and hydroxyurea.1 Median survival is 60 years in some high income countries, such as the UK .5

Children severely affected with sickle cell disease (around 5%) fail to respond to treatment with hydroxyurea or regular blood transfusion. These children may have acute complications, including frequent attacks of severe pain and acute chest syndrome, or have evidence of progressive organ damage such as cerebrovascular disease. They are admitted to hospital frequently, which can negatively affect siblings and parents. Even in children with few overt complications, quality of life may be substantially impaired by fatigue, nocturnal enuresis, jaundice, and delayed puberty. Multiple comorbidities develop over time, with further reductions in quality of life.1 These observations have justified the use of high risk treatments such as haematopoietic stem cell transplantation (HSCT).

HSCT is the only curative option for sickle cell disease and has been routinely used for this indication in well resourced countries for more than 40 years.6 Transplantation from HLA matched siblings has the best results, with mortality of 5% and event free survival of more than 90%.78 These excellent outcomes have justified systematically offering HSCT to severely affected patients, especially those requiring chronic transfusion, who have a sibling donor with identical HLA.89 It is also increasingly being offered to patients with less severe disease.10 But we lack long term analyses of the relative benefits of HSCT over medical therapy.

HSCT is limited by the availability of suitable donors, with HLA identical siblings available in only 10-20% of cases.8 To expand the number of children who might benefit from HSCT, investigators have developed new protocols using alternative donors, where the donor is either an unrelated HLA identical sibling or haploidentical (parent or sibling with one matched HLA haplotype). These procedures are available only in high resource settings and are offered to children with the most severe disease. The few existing publications11121314 show that these alternative donor transplants are associated with significantly higher risks of death and rejection. We consider the acceptable risk of curing sickle cell disease in children.

Risks of alternative donors

A trial of unrelated bone marrow transplantation for children with sickle cell disease by Shenoy and colleagues found that seven of 29 (24%) children died and 11 (38%) had extensive graft versus host disease at one year.11 The largest cohort study of HSCT for children with sickle cell disease using haploidentical bone marrow donors, by Dhedin et al, was not a formal clinical trial and used different approaches at different times.14 Of 22 children, aged 3-18 years, three died (14%) and two had graft failure (9%); event-free survival was 82%. Overall in these two trials, 51 children received transplants from alternative donors, and 10 died (20%).

These studies show higher rates of transplant related death and graft rejection than with HLA identical transplants (7% in a study of HLA identical HSCT that included several different protocols from 1986 to 2013).7 The excellent benefit-risk ratio of HLA identical siblings is much reduced when alternative donors are the source of haematopoietic stem cells.

Randomised controlled trial data show that the some of the indications for transplantation used in Shenoy’s study (>2 episodes of acute pain a year (n=12), abnormal transcranial Doppler velocities, (n=2), and >1 episode of acute chest syndrome (n=4)) can be effectively managed in most cases with hydroxyurea1516 or regular blood transfusions,17 although neither of these options is curative. Chronic transfusion carries substantial risks, particularly iron overload and red cell alloimmunisation, although these can be effectively managed using iron chelation and erythrocytapheresis, and extensive blood group typing, respectively.

Balancing risk and benefit

Parents value curative treatments highly and are prepared to take high risks on behalf of their children to achieve a cure. In interviews, 12% parents of children with sickle cell disease were willing to accept a short term transplant related mortality of more than 50%.18 But research on the opinions of families and children on the risks of potentially curative treatments in sickle cell disease is sparse, particularly in light of improving medical outcomes and the possibility of transplantation as an adult. Similarly, few ethical analyses have tried to balance the benefits of cure against the risks of death, and studies on how to present these difficult choices to children and parents are lacking. These aspects should be developed in parallel with medical and transplantation advances.

We think that the considerable risks of using alternative donors are not justified in most cases. Alternative donor HSCT may be appropriate in about 5% of children, including those with progressive cerebrovascular disease or other organ failure despite optimal non-transplant treatment, including transfusion, hydroxyurea, or both.

In high income countries, the life expectancy of children with sickle cell disease is roughly comparable with that for other chronic illnesses such as insulin dependent diabetes mellitus and cystic fibrosis. The timing of lung transplantation in patients with cystic fibrosis is a similar dilemma—a lifesaving but high risk procedure that has around 30% mortality at five years. Guidelines recommend that lung transplantation is considered in cystic fibrosis with significant impairment of lung function (FEV1 <30%) corresponding to a median survival of five to six years.19 No such guidelines exist for sickle cell disease, partly because survival is hard to predict and partly because very few children with sickle cell disease in high income countries die in childhood (<7%).3 We need equivalent validated prognostic markers in sickle cell disease to facilitate the development of appropriate guidelines. People have different views on what comprises an acceptable mortality rate for HSCT in children with sickle cell disease, but few are likely to accept a mortality rate >10% at one year.

Timing of alternative donor transplantation

The optimal time to offer HSCT using alternative donors is unclear. Studies of HSCT in patients with haemoglobinopathies consistently show that transplants in adulthood are associated with higher toxicity and poorer outcomes, largely because older patients have more complications, as severity of sickle cell disease increases with age.7 Reduced adherence to treatment during adolescence and the transition to adult care may also contribute to worsening disease.20 Older patients are less tolerant of transplantation, but early transplants may be responsible for the death of children who would not have experienced severe complications and could have lived more than 60 years. Moreover, adult transplantation enables patients to give consent themselves.

The decision to perform a high risk procedure in children is complicated by the improving prognosis with medical management and emergence of new treatments—in 10 years there may be many more effective non-transplant options. A better understanding of the pathophysiology of the disease has led to the development of drugs acting on inflammatory, coagulation, and endothelial damage pathways.21 An unprecedented number of clinical trials are under way, with already promising results, such as the recent evidence for the effectiveness of P-selectin inhibition,22 which may substantially reduce the number of patients with refractory pain. Gene therapy and editing also offer real promise, with active clinical trials.23 We need more time to assess the effectiveness and middle and long term benefits of these new treatments.

Clinical trials

Research on the long term effects of HSCT for sickle cell disease is lacking, particularly using alternative donors.24 We need multicentre randomised controlled trials comparing alternative HSCT with optimal medical care. Such studies are very difficult to perform, owing to the scarcity of patients who are very severely affected, the absence of a clear definition of “severely affected”, and the need for long term follow-up. An adequately powered, randomised controlled trial may take more than five years to complete, by which time both transplant and non-transplant arms may be using redundant treatments.

Clinicians who perform transplants, paediatricians, haematologists, patient groups, and ethicists should debate the indications for entry to these high risk trials. We must identify prognostic factors that will reliably identify children likely to follow a severe clinical course. Currently such children are only identified after substantial organ damage has occurred, such as cerebrovascular disease causing overt stroke. New genomic and proteomic approaches to identify reliable prognostic markers for these severe complications will enable high risk HSCTs to be offered to those who most need them, and this should improve the risk to benefit ratio.

Notably, the Shenoy study reported mortality of 24%,11 whereas even one procedure related death in a clinical trial of non-transplant treatments is likely to result in the suspension or permanent stopping of the study. Although it could be argued that the same standards should apply to all novel treatments, including transplantation, sequential trials of transplantation may lead to the evolution of more effective and safer treatments, as happened in the treatment of thalassaemia major.25 We must develop appropriate entry and stopping criteria to avoid inappropriate management of children with sickle cell disease. Clinical trials will need close collaboration with health economists, who have validated tools for comparing different prospects.

Low income countries

HSCT is not widely available in low income countries, where most patients with sickle cell disease live, owing to high costs and the lack of necessary infrastructure. The high risks of alternative donor HSCT are arguably more justifiable in African countries because of the poor prognosis of sickle cell disease. However, far greater benefit is likely to result from investing in increasing access to cheaper effective treatments, including prophylaxis against infection, safe blood transfusion, and hydroxyurea.2


In theory every patient with sickle cell disease could now be offered curative treatment for their disease with HSCT. But for the majority of people, who don’t have an HLA identical sibling, the risk of death is 10-24%. We cannot currently identify which children will develop severe disease without transplantation, making it difficult to weigh the risks and benefits. We encourage the development of randomised controlled trials and more research into offering these difficult decisions to parents and children. Recent research in decision theory for implementing high risk treatment shows that physicians should defer to the considered preferences of the participants rather than rely on their own judgment of the benefit-risk ratio, although this is more complicated when the participants are children.26

The medical community must work with patients and parents on how to increase the availability of transplants, while tackling ethical questions with the same priority as the technical ones.


  • Contributors and sources: This article arose out of concern among the authors after the publication high mortality rates using alternative donor HSCT in children with sickle cell disease. DCR wrote an initial draft, which was fully revised and improved by MdM. MdM summarised the results of previous trials of alternative donor transplantation. DCR circulated the different authors for their views, including authors from the fields of paediatrics, haematology, and transplantation. All authors made significant changes and contributions, and approved the final manuscript. DCR is the guarantor.

  • Competing interests: We have read BMJ policy on competing interests and declare the following: MdM has received funding for taking part in clinical trials and advisory boards from Novartis, Pfizer, and Addmedica. She has received funding to attend academic meetings from Novartis and Addmedica. VB has received funding from Addmedica for taking part in an advisory board. SC has received funding from Novartis and Pfizer for advisory boards and other meetings. AP has received funding from Bluebirdbio for attending an advisory board. JP has received funding for from Novartis, Celgene and Bluebirdbio for attending advisory boards and clinical studies. PT has received funding from Bluebirdbio, Global Blood Therapeutics and Pfizer for attending advisory boards; Novartis and Apopharma for attending meetings; and Napp and Kyora Kirin for research studies. AV has received funding from Amgen, Pfizer, Jazz Pharmaceuticals and Medac to attend academic meetings and take part in advisory boards. DR has received funding for taking part in clinical trials and advisory boards from Novartis, Eli Lilly and Bluebirdbio. He received funding to attend academic meetings from Eli Lilly, Novartis, Biogen, and Addmedica.

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


View Abstract