Daclizumab, an efficient treatment for steroid-refractory acute graft-versus-host disease
P. Bordigoni, Blood and Marrow Transplantation Unit, Department of Paediatric Onco-Hematology, Hôpital d'enfants, rue du Morvan, 54511 Vandoeuvre les Nancy, France. E-mail: email@example.com
In a phase II study, daclizumab was given as single second-line agent to 62 patients with steroid-refractory acute graft-versus-host disease (aGVHD). Complete resolution of aGVHD was achieved in 68·8% of patients. This response rate was significantly associated with a lower number of involved organs and smaller extent of skin involvement. The 4-year event-free survival (EFS) was 54·6%. Grade ≥III aGVHD, ≥2 involved organs at baseline and patient age >18 years were independently associated with lower EFS. Daclizumab could be a suitable alternative treatment for aGVHD, particularly when limited to the skin or gastrointestinal tract.
No standard effective treatments are currently available for patients with steroid-refractory acute graft-versus-host disease (aGVHD) (Martin et al, 1991; Antin et al, 2004). A variety of immunosuppressive agents including interleukin-2 receptor alpha (IL-2Rα) antagonists have been used in this setting but with a disappointing long-term outcome (Przepiorka et al, 2000; Willenbacher et al, 2001; Schmidt-Hieber et al, 2005).
The results of our experience with daclizumab, as a single second-line agent, in the management of corticosteroid-refractory aGVHD in 62 patients are reported. This prospective, single-centre phase II trial aimed to evaluate the safety and efficacy of daclizumab in this setting.
Patients and methods
Patient characteristics are depicted in Table I. Primary treatment of aGVHD consisted of methylprednisolone (MP) at 2·5 mg/kg/d. Patients were excluded if they were given systemic immunosuppressive drugs for aGVHD treatment other than systemic corticosteroids. Twenty-six patients were children. Recipients with steroid-resistant (SR) aGVHD were given MP (2·5 mg/kg/d) for a maximum of 14 d. The patients with steroid-dependant (SD) aGVHD were given a reduced dose of MP (1 mg/kg/d) for at least 14 d. The patients also received daclizumab (1 mg/kg i.v. on days 1, 4, 8, 15 and 22). The response criteria were defined according to Przepiorka et al (2000).
Table I. Patients’ characteristics
| Median (range)||25·4 (1·5–53)|
| >18 years||36|
| Acute lymphoblastic leukemia||13|
| Acute myeloid leukemia||21|
| Chronic myeloid leukemia||3|
| Other malignancies||5|
| Inherited diseases||5|
| Mismatched-related||1 (cord-blood)|
| Mismatched-unrelated||18 (cord-blood = 8)|
| CSA + MTX||35|
| CSA + MMF||9|
| CSA + MP + ATG||9|
|CMV serostatus of donor and recipient|
|Status of the disease before daclizumab|
|Source of hamopoietic stem cells|| |
| Cord blood||9|
|Types of cortico-refractory aGVHD|
| Steroid-resistant (SR)||27 (53·2)|
| Steroid-dependent (SD)||35|
|aGVHD grade|| |
| At baseline ≥II||59 (95)|
| At baseline ≥III||19 (30·6)|
| At time of daclizumab ≥II||62 (100)|
| At time of daclizumab ≥III||21 (33·9)|
|Types of organ involvement at baseline|
| Skin-only||25 (40·3)|
| Gut-only||15 (24·2)|
|Types of organ involvement at initiation of daclizumab|
| Skin-only||19 (30·6)|
| Gut-only||21 (33·9)|
Treatment with daclizumab was administered at a median of 23·7 d (range, 5–69) after the diagnosis of aGVHD. Characteristics of aGVHD at baseline and at initiation of daclizumab are depicted in Table II. Eight patients received fewer than five doses of daclizumab (3, n = 6; 4, n = 2).
Table II. Multivariate analyses of risk factors for complete response by day 30 and with mortality
|Initial organ involvement|
| Gut only|| || || ||1||0·074–0·918||0·036|
| Multiorgan|| || || ||3·8|| || |
|Organ involvement at initiation of daclizumab|
| Skin or gut only||12·6||0·008–0·738||0·026|| || || |
| Multiorgan||1|| || || || || |
|Initial grade of aGVHD|
| II|| || || ||1||1·52–8·55||0·003|
| ≥III|| || || ||3·6|| || |
|Extent of skin involvement at initiation of daclizumab|
| Stages 1–2||15·8||1·6–154·8||0·018|| || || |
| Stages 3–4||1|| || || || || |
| ≤18|| || || ||1||0·164–0·973||0·04|
| >18|| || || ||2·5|| || |
On study day 30, 68·8% patients had achieved a complete remission (CR) of aGVHD and 21·3% achieved a partial remission. Five complete responders presented with aGVHD recurrence between days 45 and 105 after daclizumab initiation. Among the 59 patients who survived >90 d after the first dose of daclizumab, 40 (57·6 ± 5·6%) [95% confidence interval (CI)] developed chronic GVHD, which was extensive in 87% of these cases. In Cox regression analysis, patients with isolated skin or gut aGVHD, responded more frequently (respectively 73·7% and 90·5%) than those with multiorgan involvement (42·9%) (P = 0·026). Furthermore, CR was obtained in 72·7% of patients with stage 1–2 skin aGVHD, versus 33·3% of patients with stage 3–4 (P = 0·018).
With a median follow-up of 44 months (range, 4–74), the Kaplan–Meier probability of 4-year event-free survival (EFS) was 54·6 ± 6·6% (95% CI). In multivariate analysis (MVA), initial grades III–IV aGVHD, ≥2 involved organs at baseline and patient age >18 years were independently associated with poorer EFS. The survival for children and adults was 61·8 ± 10·4% and 46·4 ± 7·5% (95% CI) (P = 0·04), respectively. Patients with an initial isolated skin or gut aGVHD had a higher probability of EFS (53·6 ± 8·4% and 66·7 ± 11·5%), respectively, than recipients with multiorgan (≥2) involvement (36·4 ± 4·5%) (95% CI) (P = 0·036). In addition, the probability of survival was 56·1 ± 8·6% and 39·7 ± 6·7% for patients with grades I–II and grades III–IV GVHD (95% CI) (P = 0·003), respectively. The 4-year cumulative risk of relapse was 23 ± 8·6% (95% CI) [median: 4·5 months (range, 2–9 months)]. By study day 360, 17 patients had died without relapsing (median time to death, 7 months (range, 0·75–36) after initiation of daclizumab. The primary cause of death was aGVHD or cGVHD in two and four patients, respectively, and infections in seven patients (11·3%) [adenovirus infection (4), fungal infection (2), respiratory syncytial virus (1)]. The cumulative incidence of non-leukaemic transplant-related mortality was 11·7 ± 3·1% and 27·4 ± 8·4% (95%CI), respectively, on days 100 and 360 after initiation of daclizumab. Cytomegalovirus (CMV) reactivation occurred in 38·7% OF patients and 6·4% of them developed non-lethal CMV disease.
In the present analysis, the use of daclizumab produced a primary overall complete response rate of 68·8% by day 30, at the expense of a 57·6% rate of early, progressive chronic GVHD. Complete resolution of aGVHD was documented in 72·2, 73 and 42·8% of patients with skin, gut and liver GVHD, respectively. Of note, few patients suffered from severe grades III–IV aGVHD at the beginning of daclizumab therapy. Surprisingly, and in contrast with other series (Patriarca et al, 2004; Greinix et al, 2006), achievement of CR by day 30 was not predictive of improved long-term EFS, highlighting the need for aggressive monitoring and prophylaxis of infections in this population.
The 4-years EFS of 54·6% compared favourably with published series examining the role of second-line therapy of patients with steroid-refractory aGVHD (Przepiorka et al, 2000; Willenbacher et al, 2001; Srinivasan et al, 2004; Bay et al, 2005; Schmidt-Hieber et al, 2005; Greinix et al, 2006). Of note, as less than half of the patients received daclizumab as single agent in these studies, the estimation of the independent role of daclizumab is difficult to elucidate. Numerous prior series have shown that infection-related mortality is a major cause of failure in this setting. The striking difference in TRM, particularly the risk of fatal infections (Willenbacher et al, 2001), could be attributed to our stringent policy of infection control.
In contrast with other trials (Przepiorka et al, 2000; Willenbacher et al, 2001; Bay et al, 2005; Schmidt-Hieber et al, 2005), we observed no significant difference in response rates and EFS according to GVHD localisation (skin or gut). Another characteristic of our trial was the proportion of SD aGVHD. This cohort of patients was included because, in the retrospective trial of Martin et al (1991), after a secondary rechallenge of cortico-steroid, complete responses were seen in only 24% of patients. A sparing cortico-steroid effect was demonstrated in this setting in our study. Paediatric studies using daclizumab for steroid-refractory aGVHD are limited (Teachey et al, 2006). The present study included the largest prospective series of children with promising results, particularly in terms of long-term EFS (61·8%).
The relapse of the original malignancy in our patients was 23 ± 8·6%, similar to what has been previously reported (Stem cell trialists’ collaborative group, 2005). Furthermore, no published study has reported a higher relapse-rate in patients who received daclizumab treatment for steroid-refractory aGVHD (Lee et al, 2004).
In conclusion, our results suggest that, patients with low-grade steroid-refractory aGVHD, especially when limited to the skin or gastrointestinal tract, were the best responders (Willenbacher et al (2001). Conversely, patients, particularly adults, with overall grades III–IV aGVHD as well as with extensive skin involvement benefited less from daclizumab therapy.