Second-line therapy for acute GvHD (aGvHD) remains an unmet need. Clinical trials for second-line therapy typically enroll patients with steroid refractory (SR) aGvHD, defined as worsening after 3 days or no improvement after 7-10 days of corticosteroid based therapy. Approximately 60% patients respond by day 28 of initial corticosteroid based therapy. A significant proportion of these patients lose their response during corticosteroid taper, and are thus steroid dependent (SD). We hypothesized that patients who have recurrent or progressive aGvHD prior to reaching 50% of the initial corticosteroid dose (early failure, SD-EF) will have an outcome similar to aGvHD that is upfront SR (USR) at initial presentation, and thus should be enrolled on clinical trials for SR aGvHD without re-challenging with corticosteroids to prove refractoriness.


This was a retrospective, single-center study of allo-HCT recipients between 2005 and 2013. Records of all patients with a diagnosis of aGvHD were reviewed. Patients were included if they were on a calcineurin inhibitor and received systemic corticosteroids for clinical or histologically proven aGvHD and were USR or SD. Patients were excluded if they had progression or relapse of their underlying disease before aGvHD or if they received >1 allo-HCT. USR aGvHD was defined as worsening of aGvHD within the first 3 days or no improvement after 10 days of starting ≥1 mg/kg of prednisone-equivalent. SD aGvHD was defined as initial response to corticosteroids with recurrence or worsening on tapering. The SD group was subdivided into early failure (SD-EF) and late failure (SD-LF) if they recurred at ≥50% of the starting corticosteroid dose and <50% of the starting corticosteroid dose, respectively. The primary outcomes were overall survival (OS) and non-relapse mortality (NRM) at 1 y and 2 y after initiation of corticosteroids for aGvHD. Secondary outcomes were cause of death and use of second line therapy. Cause of death due to GvHD included patients who died from infection while on therapy for active GvHD.


Of 596adult patients who received allo-HCT, 465 developed aGvHD (78%). Patients were excluded for: recurrent or progressive disease prior to aGvHD onset (6), receiving >1 allo-HCT (6), and not being on a calcineurin inhibitor (4). Systemic corticosteroids were used in 345 (74%). Incidence of USR and SD aGvHD among those treated with corticosteroids was 13% (43/345) and 29% (100/345). SD-EF and SD-LF accounted for 45% and 55% of all SD aGvHD. Table 1 outlines baseline characteristics. Median follow up of the cohort was 1.7 y (range, 0.03 to 12.2) and OS was 1.8 y (95% CI, 1.2 to 3.1). Second line therapy was required in 65%, 42%, and 27% (p=0.001) of USR, SD-EF and SD-LF patients, respectively.

Primary endpoints were OS and NRM at 1 y and 2 y. OS at 1 y and 2 y for USR vs SD were 47% vs 68% (p=0.002) and 44% vs 51% (p=0.077), respectively. NRM at 1 y and 2 y for USR vs SD groups were 40% vs 20% (p=0.007) and 42% vs 28% (p=0.04). As hypothesized, outcomes for USR and SD-EF were similar: OS at 1 y and 2 y in USR vs SD-EF were 47% vs 60% (p=0.08) and 44% vs 42% (p=0.49), respectively; NRM at 1 y and 2 y for USR vs SD-EF were 40% vs 29% (p=0.19) and 42% vs 40% (p=0.53). GvHD leading to NRM was similar in USR and SD-EF (63% vs 44%, p=0.24).

The OS was similar for SD-EF vs SD-LF at 1 y (60% vs 75%, p=0.1) and 2 y (42% vs 58%, p=0.09). NRM was higher for SD-EF compared to SD-LF at both 1 y (29% vs 13%, p=0.04) and 2 y (40% vs 18%, p=0.02), primarily due to a higher rate of GvHD related deaths (SD-EF vs SD-LF 44% vs 16%, p=0.03). This was in contrast to the SD-LF group, where relapse accounted for a higher proportion of death compared with 53% of deaths due to relapse for SD-LF vs 28% for SD-EF (p=0.03)


Forty-five percent of patients with SD aGvHD have SD-EF. Our data suggests that patients with USR and SD-EF remain at a high risk of death from complications of GvHD, and both groups should be considered for second line therapy of aGvHD. The current approach of requiring SD aGvHD patients to fulfill SR aGvHD criteria prior to enrolling on a study of second-line therapy increases corticosteroid exposure and related morbidity, without clear benefit of long-term outcome. As interventions with safer adverse effect profiles are developed, it is imperative that the eligibility criteria of second-line therapy be modernized to include SD-EF aGvHD, to allow this high-risk group to derive the benefit of experimental therapeutics.


Savani: Jazz Pharmaceuticals: Speakers Bureau. Byrne: Pfizer: Membership on an entity's Board of Directors or advisory committees; Karyopharm: Research Funding; Concert Pharmaceuticals: Membership on an entity's Board of Directors or advisory committees. Jagasia: Therakos: Consultancy, Research Funding; Janssen: Consultancy, Research Funding; Mallinckrodt: Consultancy.

Author notes


Asterisk with author names denotes non-ASH members.