Long-term outcome following cyclosporine-related neurotoxicity in paediatric allogeneic haematopoietic stem cell transplantation

Long-term outcome following cyclosporine-related neurotoxicity in paediatric allogeneic haematopoietic stem cell transplantation

With increasing numbers of children and adults undergoing allogeneic haematopoietic stem cell transplantation (HSCT), neurotoxicity is emerging as an important cause of transplant-related morbidity. [1][2][3] The calcineurin inhibitor cyclosporine A (CI CSA) is the most frequently used agent for prevention of GvHD in both adult and paediatric patients. CSA is an established cause of posttransplantation central nervous system (CNS) toxicity, typically characterised by posterior reversible encephalopathy syndrome (PRES). 4 Even though neurological complications of CSA seldom result in mortality, the necessitated withdrawal of this potent anti-GvHD agent can have major implications on clinical outcome particularly in the face of on-going GvHD. The spectrum of CSA-related neurological complications has been well characterised in several case series of paediatric HSCT recipients, [5][6][7] however, longterm follow-up data of those children who experience neurotoxicity are lacking. Here, we present the outcome of 26 children who developed CSA-neurotoxicity following allogeneic HSCT.
From a cohort of 569 consecutive paediatric allogeneic HSCT recipients at our institution between 1 January 2003 and 31 December 2013, we retrospectively identified those children who developed neurological complications at any time after the start of conditioning whilst receiving CSA using a computerised search of our patient database. Children with clinical symptoms of PRES or neurological manifestations of thrombotic microangiopathy (TMA) 7 were included. Those with obvious alternate causes for neurotoxicity were excluded. All children who suffered CNS complications were clinically evaluated by a neurologist. Supplementary investigations such as computed tomography (CT) or magnetic resonance imaging (MRI) scan of head, cerebrospinal fluid analysis and electroencephalogram were used as appropriate to identify the cause. PRES was diagnosed in accordance with its well described MRI appearance. 8 Hypodense lesions in the parietooccipital areas were considered to be CT findings consistent with PRES. TMA was diagnosed based on clinical and laboratory features of microangiopathic haemolytic anaemia and thrombocytopenia. GvHD was diagnosed clinically and supported with histological confirmation where possible. 9 IV CSA administration at a dose of 5 mg/kg/day in two divided doses was commenced either on day − 1 (2003-2008) or day − 3 (2009 onwards). Levels were monitored twice weekly. Target trough serum levels were 100-150 ng/mL for HLA-identical sibling transplants and 150-200 ng/mL for unrelated donor transplants.
During the 11-year study period, 569 paediatric allogeneic HSCTs were performed at this centre. Twenty-six children developed neurotoxicity associated with CSA (4.6%). Their demographics, disease characteristics and transplant-related parameters are summarised in Table 1. Median age at HSCT was 8.3 years (range 0.5-12.8 years). CSA+MMF was the most commonly employed GvHD prophylaxis (n = 19), followed by CSA+MTX (n = 4) and CSA alone (n = 3). In one patient (subject #19), CSA was changed to Tacrolimus because of renal toxicity before the development of neurotoxicity that is, in this patient neurotoxicity occurred while being treated with Tacrolimus. Twenty-one children experienced CIrelated neurotoxicity during their primary transplant procedure and 5 children following their second HSCT, having tolerated CSA without adverse events during their first HSCT.
Median time from HSCT to neurotoxicity was 47.5 days (range day -1 to +545) ( Table 1). Clinical symptoms of neurotoxicity and severity are listed in Table 1. Twenty-four out of 26 patients had radiological evaluation of the brain with MRI, CT or both; 19 had classical findings of PRES, 2 showed non-specific changes consistent with encephalopathy and 3 had normal appearances of the brain. CSA level in peripheral blood at the time of presentation with neurotoxicity ranged from 64 to 256 ng/mL ( Table 1).
Management of CI-induced neurotoxicity in addition to supportive care consisted of discontinuation of CSA/Tacrolimus in 25/26 patients. Neurological symptoms fully resolved in all patients except in patient #9 who developed TMA and died shortly thereafter and patient #3 who continues to require anticonvulsants despite normalisation MRI appearance. One patient (#12) who presented with seizures on day +24 and had non-specific features on MRI with a normal EEG was continued on IV CSA along with anticonvulsants and had no further neurological events.
After a median of 11 days post onset of CSA neurotoxicity (range 0-597 days) a CI was reintroduced in 18 patients: in 10 patients CSA was used and in 12 patients Tacrolimus, 3 of whom had failed re-challenge with CSA (Table 2). One patient received Tacrolimus after having been successfully re-challenged with CSA (patient #25). Eight of 18 patients re-challenged with a CI had recurrent symptoms of neurotoxicity: 3 following re-challenge with CSA, 2 following re-challenge with Tacrolimus and 3 following re-challenge with CSA as well as Tacrolimus sequentially. In addition to attempted re-challenge with a CI, all but one patient with grade I skin GvHD only before the onset of neurotoxicity, received corticosteroids as GvHD prophylaxis/treatment. Other immunosuppressive agents used are listed in Table 2.
Despite these measures, 7/9 patients (78%) who had no prior GvHD, developed acute/late acute GvHD after onset of CSA neurotoxicity whereas 13/17 had recurrence, persistence or progression of acute/late acute GvHD. Twenty-three patients were evaluable for chronic GvHD as 2 died before day +100 and one patient had second HSCT before day +100. Eighteen out of 23 (78%) had chronic GvHD and this was extensive in 14 patients: 13 patients developed chronic GvHD following CSA-neurotoxicity whereas pre-existing chronic GvHD persisted or progressed in 5 patients (Table 2).
At a median of 176 days (range 7-1889 days) following development of CSA-induced neurotoxicity 15/26 (58%) patients died, most due to progressive GvHD or its complications ( Table 2). There were no acute deaths from neurotoxicity except in one patient (subject #9) who had CSA-related TMA on day +61 and died of encephalopathy and respiratory failure on day +68. Currently, 11/26 (42%) of children are alive a median of 8.2 years (range 5.5-13.1 years) after HSCT and a median of 7.8 years (range 5.5-12.1 years) after development of CSA-neurotoxicity. In contrast, during the same study period in our institution, the overall survival at 5 years following allogeneic HSCT was 67.7% for haematological   Letter to the Editor malignancy, 82.1% for immunodeficiency and 87.2% for metabolic disorders. Of the 11 children who survived despite developing CSA-related neurotoxicity, 4 (36%) have a significantly impaired quality-of-life due to sequelae of extensive chronic GvHD. This case series of paediatric HSCT recipients who developed CSA-related neurotoxicity illustrates that outcome following this complication is notably poor: high non-relapse mortality of 58% and significant morbidity with 36% of survivors living with late effects of extensive chronic GvHD. These findings are similar to those in large case series of adult allogeneic HSCT recipients who developed CSA-related neurotoxicity reporting 43-52% mortality due to progressive GvHD/infection. 10,11 In our study, outcome was particularly poor in the 10 patients who had severe GvHD (Grade III/IV) before the development of CSA-neurotoxicity: 8/10 died and 2 are alive with extensive vitiligo. In the same time period at our institution, other patients with severe GvHD, but without the added complication of CSA neurotoxicity had a far superior survival rate of 70%. 12 Hence it would seem that the development of CSA neurotoxicity and consequent inability to tolerate CI, adversely affected their prognosis. In future studies with larger patient numbers, it would be useful to substantiate our findings with multivariate analyses.
Development of CSA-related neurotoxicity poses a complex clinical management situation as one of the most effective drugs in the treatment and prevention of GvHD needs to be discontinued promptly; sometimes in patients with on-going GvHD. As symptoms of CSA-neurotoxicity usually resolve over several days, re-challenge could be a viable option. Tacrolimus has been used as alternative agent in the event of CSA-related neurotoxicity but as it is also a CI, its use is similarly associated with significant neurotoxicity 7,13,14 as seen in 5 patients in our case series. Nevertheless, re-challenge with a CI was tolerated in 56% of patients in this series, a similar proportion as that reported by others. 10 In those patients where re-challenge resulted in recurrence of symptoms and use of CI was precluded permanently, outcome appears particularly dismal with overall survival of 13% (1/8) in this case series.
Evidence for the optimal approach of prophylaxis/management of GvHD when use of CI is contra-indicated is not available. In this case series, all but one patient received corticosteroids. In addition, as the combination of CSA and MMF is the most common GvHD prophylaxis regimen in our centre, the majority of patients were already receiving MMF at the time of development of CSA neurotoxicity. The efficacy of MMF as sole anti-GvHD agent is limited, 13 and on its own it does not provide satisfactory GvHD prophylaxis or treatment. Sirolimus, which has a completely different mechanism of action to the CI provides another option. Although Sirolimus can also cause neurotoxicity, this has mostly been reported when it has been used in combination with CSA. 15 In this series, Sirolimus was only used in five patients but this agent may potentially be increasingly used in this clinical situation in future.
In conclusion, this case series illustrates the dismal prognosis in patients following the development of CI-related neurotoxicity and the complexities of managing GvHD in this situation. There is a need for further studies to determine the optimal treatment approach to improve outcome following this rare but serious complication.