Ruxolitinib is an effective steroid sparing agent in children with steroid refractory/dependent bronchiolitis obliterans syndrome after allogenic hematopoietic cell transplantation

Bronchiolitis obliterans syndrome (BOS), chronic graft versus host disease (cGVHD) of the lungs, confers a dismal prognosis [1, 2]. Steroids are the backbone therapy for BOS, but prolonged use is associated with increased mortality and is not recommended [2, 3]. However, there is no consensus therapy for patients who are steroid dependent or refractory (SR).

Ruxolitinib, a JAK1/2 inhibitor, exerts potent anti-inflammatory effects and is FDA approved to treat acute GVHD in adults. Ruxolitinib therapy for cGVHD in adults is also described [4], though reports of ruxolitinib use in children for GVHD are rare [5, 6]. We describe a single center experience of pediatric patients with SR-BOS treated with ruxolitinib.

In this IRB approved retrospective study, patients treated with ruxolitinib for SR-BOS after allogenic hematopoietic cell transplantation (HCT) were identified from January 2010 to July 2018 using departmental records. 2014 NIH consensus conference criteria were used to define BOS, severity of disease, response to therapy, and steroid refractory/dependent disease [7–9]. BOS pulmonary function testing (PFT) criteria included: FEV1/VC <0.7, FEV1 <75% predicted with a ≥10% decline, absence of infection, and consistent chest computed tomography (CT) findings. Severe BOS was defined as an FEV1% of predicted of ≤39% and moderate 40–50% [7]. An improvement in FEV1% predicted of ≥10% was a partial response (PR), and a complete response (CR) was an FEV1 of ≥80% predicted [9].

BOS was considered steroid refractory if patients had declining PFTs on steroids for 2 weeks and steroid dependent ≥0.5 mg/kg/day if steroids were required for >4 weeks, and/or if the patient was unable to taper steroids without additional immunosuppressive medications [8]. Common terminology criteria for adverse events (CTCAE) version 5 were used to define an adverse event.

Treatment was at the discretion of the provider. Ruxolitinib was dosed 5 mg twice daily (BID) and escalated to 10 mg BID as tolerated with close monitoring of blood counts and infectious complications. As all patients weighed ≥40 kg, the final dose of ruxolitinib was 10 mg BID in four patients. In one patient with severe renal insufficiency, ruxolitinib was dosed at 2.5 mg Monday, Wednesday/Friday. No pharmacokinetic or biomarker data were collected.

Five patients received ruxolitinib for SR-BOS. The median age at HCT was 10 years (range 7–21 years) (Supplementary Table 1). All received an unrelated bone marrow transplant (BMT) and identical acute GVHD prophylaxis: cyclosporine, methotrexate on days + 1, 3, 6, and 11, and steroids of 1 mg/kg until day +21. In the absence of aGVHD, steroids were tapered off over 10 weeks. Three patients were matched at 10/10 HLA-loci and 2 at 9/10. Transplant indications included BALL (n = 1), AML (n = 2), immunodeficiency (n = 1), and MDS (n = 1). Prior to BOS diagnosis, four patients had a single HCT. One patient (#5) received a second HCT for relapsed MDS. Two patients had aGVHD that was steroid responsive and resolved prior to BOS diagnosis.

BOS was diagnosed a median of 315 days after stem cell infusion (range 131–875 days). All patients met PFT and chest CT criteria, and one patient (#5) also had a lung biopsy consistent with BO. Four patients presented with dyspnea, and one patient (#4) was developmentally delayed and unable to report symptoms but developed a new oxygen requirement. Median FEV1% of predicted at BOS diagnosis was 38% (range 23–64%). Four patients had severe BOS, and one had moderate BOS. Three patients had identifiable triggers for BOS including toxic epidermal necrolysis (TEN) (#2), pulmonary fungal infection with biopsy proven concurrent BO (#5), and donor lymphocyte infusion given for relapsed disease (#3). BOS was the only manifestation of cGVHD in 2 patients (#2, 5), the remainder had multiorgan cGVHD. Two patients were on steroids prior to BOS diagnosis, patient #2 for treatment for TEN and patient 4 for extensive multiorgan GVHD that delayed his BOS diagnosis.

All patients were treated with steroids and fluticasone, azithromycin, and montelukast, (FAM) after diagnosis with BOS, except patient #2. Azithromycin was a potential trigger for his TEN, so he received fluticasone and montelukast only. Initial steroid dose varied from 1–2 mg/kg/ day. One patient (#5) was steroid refractory with progressive disease on steroids. The remainder were steroid dependent. The median duration of steroid therapy from BOS diagnosis prior to initiation of ruxolitinib was 4 months (range 2.5 months to 2 years). Ruxolitinib ranged from the third to sixth line therapy.

Two patients had a PR to ruxolitinib with an increase in predicted FEV1% of 10 and 24% (#2 and 3 respectively) (Fig. 1). One patient (#4) had an increase of FEV1 of 9% and was able to stop steroids. Patient #1 was able to wean steroids to 0.3 mg/kg/day with stable PFTs but continued to report severe dyspnea when steroid weans were made. Low dose interleukin 2 was added and steroids were ultimately discontinued. Patient 2 continues to wean off steroids. Since initiating ruxolitinib his dose was reduced by 70% (current dose 0.3 mg/kg/day). The median time to response in the two patients with a PR was 4 weeks, and the median duration of therapy with ruxolitinib was 7 months (range 2 months to 1 year). Four patients continue on ruxolitinib at date of last follow up.

Fig. 1.

Changes in FEV1% predicted over time in children with SR-BOS. Legend: graphically depicted in figure

Patient 5 was treated for 4 months with ruxolitinib and was unable to have any PFTs during this period because of a pneumothorax following lung biopsy and treatment for fungal infection. When the pneumothorax resolved, he had new nodules on chest CT, worsened PFTs and was found to have culture confirmed additional pulmonary infections with Nocardia and Hypodontia palmae. Ruxolitinib was discontinued in the setting of new infection. Without interim PFTs, his response was considered non-evaluable. He continues to have progressive BOS on steroids and had a third relapse of MDS.

Four patients tolerated ruxolitinib with no side effects. There were no grade 3 or dose limiting cytopenias. One patient (#5) had a grade 3 adverse event while on ruxolitinib, a fungal infection. However, this occurred after months of treatment with steroids and was not directly attributed to ruxolitinib. Otherwise, there were no infectious complications or viral reactivations. Of the 3/5 patients with extensive cGVHD, all other organ systems had a PR or CR to ruxolitinib. All patients are alive at time of last follow up, a median time of 2 years (range 10–32 months).

To our knowledge, there is only one case report of ruxolitinib use in a child for cGVHD published in abstract form [6]. A case series of ruxolitinib use in children for aGVHD reported a response rate of 45%. However, 13/13 children experienced toxicities, highlighting the need for determining the maximum tolerated dose in pediatrics [5]. Our cohort was comprised of children who weighed >40 kg, allowing adult FDA approved dosing. Perhaps this explains the discrepancies in toxicities reported in the two cohorts.

In conclusion, we describe 5 pediatric and young adult patients with SR-BOS treated with ruxolitinib. Ruxolitinib was steroid sparing in all 4 patients with an evaluable response; 3 were able to stop steroids, and 1 weaned significantly. Responses occurred early, and the medication was well tolerated with minimal side effects. Further prospective studies are needed to evaluate the role of ruxolitinib in children with cGVHD, particularly BOS.