Confirmed efficacy of etoposide and dexamethasone in HLH treatment: long-term results of the cooperative HLH-2004 study

Abstract

Hemophagocytic lymphohistiocytosis (HLH) is a life-threatening hyperinflammatory syndrome comprising familial/genetic HLH (FHL) and secondary HLH. In the HLH-94 study, with an estimated 5-year probability of survival (pSu) of 54% (95% confidence interval, 48%-60%), systemic therapy included etoposide, dexamethasone, and, from week 9, cyclosporine A (CSA). Hematopoietic stem cell transplantation (HSCT) was indicated in patients with familial/genetic, relapsing, or severe/persistent disease. In HLH-2004, CSA was instead administered upfront, aiming to reduce pre-HSCT mortality and morbidity. From 2004 to 2011, 369 children aged <18 years fulfilled HLH-2004 inclusion criteria (5 of 8 diagnostic criteria, affected siblings, and/or molecular diagnosis in FHL-causative genes). At median follow-up of 5.2 years, 230 of 369 patients (62%) were alive (5-year pSu, 61%; 56%-67%). Five-year pSu in children with (n = 168) and without (n = 201) family history/genetically verified FHL was 59% (52%-67%) and 64% (57%-71%), respectively (familial occurrence [n = 47], 58% [45%-75%]). Comparing with historical data (HLH-94), using HLH-94 inclusion criteria, pre-HSCT mortality was nonsignificantly reduced from 27% to 19% (P = .064 adjusted for age and sex). Time from start of therapy to HSCT was shorter compared with HLH-94 (P =020 adjusted for age and sex) and reported neurological alterations at HSCT were 22% in HLH-94 and 17% in HLH-2004 (using HLH-94 inclusion criteria). Five-year pSu post-HSCT overall was 66% (verified FHL, 70% [63%-78%]). Additional analyses provided specific suggestions on potential pre-HSCT treatment improvements. HLH-2004 confirms that a majority of patients may be rescued by the etoposide/dexamethasone combination but intensification with CSA upfront, adding corticosteroids to intrathecal therapy, and reduced time to HSCT did not improve outcome significantly.

Conflict of interest statement

Conflict-of-interest disclosure: M.A. and J.-I.H. have been unpaid consultants to NovImmune. The remaining authors declare no competing financial interests.

© 2017 by The American Society of Hematology.

Figures

Figure 1.

Overview of the treatment protocols. (A) HLH-94 and (B) HLH-2004. (A) Both HLH-94 and HLH-2004 consist of an initial therapy of 8 weeks, with immunosuppressive and cytotoxic agents, and continuation therapy thereafter, for patients with familial, relapsing, or severe and persistent, aiming at a HSCT as soon as an acceptable donor is available. In both HLH-94 and HLH-2004, daily dexamethasone (Dexa) (10 mg/m2 per day weeks 1-2; 5 mg/m2 per day weeks 3-4; 2.5 mg/m2 per day weeks 5-6; 1.25 mg/m2 per day week 7, and tapering during week 8), and etoposide (VP-16) (150 mg/m2, twice weekly weeks 1-2, then once weekly) is administrated during the initial therapy. The continuation therapy for both HLH-2004 and HLH-94 consists of Dexa every second week (10 mg/m2 per day for 3 days), VP-16 (150 mg/m2) every second week, and CSA (aiming at 200 µg/L trough value). For patients with progressive neurological symptoms during the first 2 weeks, or if an abnormal cerebrospinal fluid value at onset has not improved after 2 weeks, intrathecal (I.T.) treatment is recommended (up to 4 doses, weeks 3, 4, 5, 6). In the HLH-94 protocol, I.T. MTX (doses by age: <1 year, 6 mg; 1-2 years, 8 mg; 2-3 years, 10 mg; >3 years, 12 mg each dose) is recommended. (B) In HLH-2004, CSA (aiming at 200 µg/L trough value) is administered already upfront during the initial therapy, a modification from HLH-94 where CSA is not administered until the continuation therapy. It is recommended to start CSA with 6 mg/kg daily orally (divide in 2 daily doses), if normal kidney function. Moreover, in the HLH-2004 protocol, in addition to I.T. MTX, I.T. prednisolone (doses by age: <1 year, 4 mg; 1-2 years, 6 mg; 2-3 years, 8 mg; >3 years, 10 mg each dose) is recommended. In HLH-2004, the total treatment period is reduced to 40 weeks as compared with 52 weeks in HLH-94. Reprinted from Henter et al (A) and Henter et al (B) with permission. BMT, bone marrow transplantation.

Figure 2.

Overall survival in HLH-2004. pSu for all patients and for different subgroups in the HLH-2004 study. The 5-year pSu is displayed with a 95% CI. (A) Five-year pSu for the entire HLH-2004 cohort (n = 369). (B) Five-year pSu for patients with an affected sibling or genetically verified FHL in HLH-2004 (n = 168) and for patients without verified FHL (n = 201) (dashed line); (P = .42, log rank). (C) Five-year-pSu for patients in HLH-2004 that fulfilled the HLH-94 inclusion criteria (n = 240) compared with patients in HLH-94 (n = 232) (dashed line); (P = .15, log rank). (D) Five-year pSu for familial patients (affected sibling) in HLH-2004 (n = 47) compared with familial patients (affected sibling) in HLH-94 (n = 52) (dashed line); (P = .86, log rank).

Figure 3.

Analyses of patients who died without HSCT with regard to time of death after onset of therapy and type of patient. All patients who died without having a HSCT (n = 75) were divided into 3 groups; (a) verified FHL (verified biallelic mutations = 23, affected sibling = 5), (b) without verified FHL and ≥1-year old at onset (n = 32), and (c) without verified FHL and <1 year old at onset (n = 15). Notably, 10 of 17 (59%) who died days 11 to 28 were aged ≥1 year and without verified FHL as opposed to 3 of 17 (18%) who died the first 10 days (P = .032), and of these 10 patients, 7 had findings possibly associated with toxicity related to overtreatment. Analyzing the treatment period days 11 to 42, this pattern was further strengthened (P = .018). In contrast, among the 8 (4 with verified FHL) who died days 43 to 60, 6 (75%) were reported to suffer active HLH, possibly suggesting a need of less reduction of initial treatment during this period. Finally, 8 of 16 (50%) who died after day 120 had verified FHL, many of whom died of disease reactivation, stressing the importance of an early HSCT for this cohort.

Figure 4.

Survival after HSCT. Survival data after HSCT (n = 185). The 5-year pSu is displayed with a 95% CI. (A) Five-year pSu after HSCT for the entire HLH-2004 cohort (n = 185). (B) Five-year pSu for patients with an affected sibling or genetically verified FHL in HLH-2004 (n = 133) and for patients without verified FHL (n = 52) (dashed line); (P = .06, log rank). (C) Five-year pSu after HSCT for patients in HLH-2004 who fulfilled the HLH-94 inclusion criteria (n = 133), compared with patients in HLH-94 (n = 118) (dashed line); (P = .77, log rank). (D) Five-year pSu after HSCT for familial patients (affected sibling) in HLH-2004 (n = 33) compared with familial patients (affected sibling) in HLH-94 (n = 40) (dashed line); (P = .36, log rank).