Disease progression, hospital readmissions, and clinical outcomes for patients with steroid-refractory acute graft-versus-host disease: A multicenter, retrospective study

Acute graft-versus-host disease (GVHD) is a significant cause of morbidity and mortality following allogeneic hematopoietic cell transplantation (HCT). This analysis of 168 patients (mean age, 54.8 years) from a multicenter, retrospective chart review describes the clinical course, treatment patterns, hospitalizations, and clinical outcomes of patients aged ≥12 years who developed grades II–IV acute GVHD after their first allogeneic HCT (January 1, 2014, to June 30, 2016) and were refractory to or dependent on corticosteroids. Between diagnosis and maximum grade (median, 6.0 days), 53.6% of patients had new organ involvement, particularly lower gastrointestinal tract acute GVHD, or an increase in acute GVHD grade. Eighty-nine patients (53.0%) received additional systemic GVHD therapy (after systemic corticosteroids) within a median of 21.0 days. Hospital readmission(s) was required for 56.5% of patients within 100 days post-HCT (mean inpatient length of readmission stay, 49.5 days); 24.4% had ≥2 readmissions within 100 days post-HCT. From the date of acute GVHD diagnosis, 70.2% of patients died at a median (interquartile range) of 117.5 (49–258) days. In summary, steroid-refractory and steroid-dependent acute GVHD is associated with a rapidly worsening clinical course that leads to high readmission and mortality rates, emphasizing the need for effective and tolerable therapies.

The standard first-line therapy for grades II to IV acute GVHD is systemic corticosteroids, although these provide effective control in only 35-60% of patients [8][9][10]. Patients may experience steroid dependence, typically defined as the inability to taper steroids (ie, reduce prednisone dose to <2 mg/kg/day) or the recurrence of acute GVHD in the same or new organs while tapering [11]. Acute GVHD that progresses within 3 days or fails to improve within 5-7 days of initiation of corticosteroid treatment is generally considered steroid refractory [12].
Early responses to primary acute GVHD treatment are indicative of improved long-term outcomes [13], whereas low rates of survival have been reported for patients with steroid-refractory acute GVHD [13,14]. Previous studies, with study periods ranging from 2004 through 2015, have reported post-HCT 1-year survival rates of ≤30% for patients with steroid-refractory acute GVHD [8,15,16].
The objective of this analysis was to describe the clinical course, treatment patterns, outcomes, and hospital readmissions among patients with acute GVHD who were refractory to or dependent on systemic corticosteroids in a real-world setting based on findings from a multicenter retrospective chart review.

MATERIALS AND METHODS Study design and patients
A multicenter retrospective chart review of patients who received allogeneic HCT between January 1, 2014, and June 30, 2016, (study period) was conducted at 11 US academic and community transplant centers [17]. Study sites were required to have conducted ≥50 adult allogeneic HCTs and to have diagnosed and treated ≥20 patients for acute GVHD during the study period. Institutional review board approval was obtained from participating institutions.
At each site, patients were considered eligible for inclusion if they were aged ≥12 years, had undergone their first allogeneic HCT during the study period, and had subsequently developed grades II, III, or IV acute GVHD, per the International Bone Marrow Transplant Registry Severity Index at any time during the follow-up period (ie, transplant to the end of data availability or death). Exclusion criteria included having undergone >1 allogeneic HCT, participating in a trial for GVHD prophylaxis during the study period (GVHD treatment trial permitted), having used Janus kinase inhibitors for any condition, and being unable to disclose complete GVHDrelated medical history for any reason.

Data collection
Medical records of eligible patients were reviewed by site physicians or clinical research staff. Deidentified patient data were collected through an electronic form from the date of allogeneic HCT to the most recent followup (data collection, March through December 2018) or death. Centers sampled patients with acute GVHD based on transplant date, beginning with the most recent transplant recipients. Data collected included patient demographics, transplant-related characteristics, acute GVHD characteristics (including diagnosis as well as grade and organs involved at diagnosis and at maximum grade), treatments received for acute GVHD, inpatient care utilization, and clinical outcomes (including acute GVHD recurrence, overall rate of infections [viral, bacterial, mycobacterial, fungal, and parasitic] and all-cause mortality).
This analysis evaluated patients who were refractory to or dependent on systemic corticosteroids. The criteria for corticosteroid dependence and refractoriness varied across participating centers. For the purposes of this analysis, corticosteroid refractory was defined as requiring the use of ≥1 additional systemic GVHD therapy. Corticosteroid dependence was defined as not being able to taper high-dose corticosteroids (≥1 mg/kg) by ≥25% or being able to taper corticosteroid dose by ≥25% but not able to taper to <10 mg/day.

Statistical analyses
Frequencies and percentages were reported for categorical variables; mean, SD, median, and interquartile range (IQR) values were calculated for continuous variables. Mortality was calculated using Kaplan-Meier estimates. Descriptive statistics were used for consideration of numeric differences between groups of patients who were corticosteroid-refractory or -dependent; no formal statistical comparisons were performed.

Patient demographics and clinical characteristics
The analysis included 168 patients with corticosteroid-refractory (n = 113) or -dependent (n = 55) acute GVHD (Table 1) from a cohort of 475 patients with grades II to IV acute GVHD, with data collected from 11 transplant centers (see Acknowledgements). Mean (SD) age among the 168 patients at HCT was 54.8 (12.5) years, including 1.8%, 13.7%, and 84.5% in age groups <18 years, 18-40 years, and >40 years, respectively. Most patients (63.7%) were male. The most common underlying malignancies were acute myeloid leukemia (35.1%), myelodysplastic syndrome (19.0%), and acute lymphoid leukemia (16.1%). The main stem cell source was peripheral blood (73.8%). The most commonly used transplant conditioning regimens were high-dose myeloablative regimens among patients who were steroid-refractory (58/113 [51.3%]) and reduced-intensity regimens among patients who were steroid-dependent (26/55 [47.3%]). Median (IQR) time from transplant to acute GVHD diagnosis was 30.5 (21-49) days and from acute GVHD diagnosis to death/last visit was 194 (58-720) days. ) developed acute GVHD in other organs. During progression from acute GVHD diagnosis to maximum grade among the 168 patients, there was an increase in the proportion who had lower gastrointestinal involvement (from 37.5% to 54.2%) and ≥2 organs involved (from 36.9% to 54.2%; Fig. 2). Between the time of diagnosis and maximum acute GVHD grade, 53.6% (90/168) of patients had new organ involvement or an increase in acute GVHD grade. Median (IQR) time from acute GVHD diagnosis to maximum grade was 6.0 (0-29.5) days; patients who were steroidrefractory progressed to maximum grade over a longer period of time than those who were steroid-dependent (median, 10.0 vs 2.0 days).

Treatment Patterns
All 168 patients received corticosteroids only as first-line therapy; 22 patients were diagnosed with grade I acute GVHD that later progressed to grades II-IV. Among 146 patients with grades II to IV acute GVHD at diagnosis, 81.5% (n = 119) were given systemic corticosteroids as first-line therapy; the remaining 18

DISCUSSION
Despite the use of systemic treatment(s) and recent advances in supportive care and availability of novel investigational agents, the majority of patients with corticosteroid-refractory or -dependent acute GVHD assessed in this large, multicenter, retrospective chart review developed severe (grades III/IV) disease. Additionally, many patients experienced rapid disease progression, with particularly high rates of progression observed in the lower gastrointestinal tract (from 38% with involvement at diagnosis to 54% at maximum grade), highlighting the challenges in treating patients with lower gastrointestinal acute  GVHD involvement. A high rate of mortality was observed in patients with steroid-refractory and steroid-dependent disease. The mortality rate of approximately 70% over a median of 117.5 days since acute GVHD diagnosis in the present study is comparable to other real-world data [15,18,19], suggesting that more successful strategies are needed for prevention of steroidrefractory and -dependent acute GVHD and that additional second-line and later treatment options are needed for patients who have progressed on steroids.
Higher and cumulative steroid doses are associated with an increased incidence of infection [20]. Despite this, 36% of patients included in these analyses had an increase in steroid dose during follow-up, and 88% were unable to taper below 10 mg/day. It follows that half of patients experienced infection between acute GVHD diagnosis and 100 days post-HCT, and 13% of deaths were attributed to infections. Although there is no standard second-line treatment for acute GVHD [21,22], steroid-refractory acute GVHD is typically diagnosed within 7 days of first-line therapy [12,22]. In these analyses, median time from systemic corticosteroid initiation to additional therapy was 21 days, and 42% of 89 patients who received ≥1 line of any additional systemic GVHD therapy first had an increase in corticosteroid dose. These data suggest that second-line therapies have delayed initiation. Furthermore, only half of the patients (89/168) included in this study received additional therapy. Treatment patterns were at the discretion of physicians, indicating that clinicians may not have perceived any benefit of additional therapy, again highlighting the need for more effective second-line treatment strategies. In May 2019, ruxolitinib, a Janus kinase (JAK) 1/JAK2 inhibitor, became the first US Food and Drug Administration-approved treatment for steroid-refractory acute GVHD in patients ≥12 years old [23]. It would be of interest to re-evaluate clinical management practices of steroid-refractory and -dependent acute GVHD from June 2019 to evaluate management and outcomes after ruxolitinib approval.
In the current analyses, more than half of the patients (56.5%) required hospital readmission with an extended length of stay. In a retrospective review of a hospital discharge database, high hospital readmission rates (77.2%) were reported for patients with high-risk or steroid-refractory acute GVHD within 100 days post-HCT [14]. Although not comparable, owing in part to differences in study design, these findings suggest that hospital readmission rates remain high for patients with steroidrefractory acute GVHD, further supporting that disease management has not yet been optimized. Effective treatments that block the major pathogenic pathways and provide rapid control of disease may help with reducing frequency and length of hospitalizations.
Limitations to this study include the retrospective nature of the analysis. Data collection was limited by available information in medical charts; treatment decisions were at the discretion of physicians from 11 different centers. Furthermore, results may not be generalizable to patients beyond those with acute GVHD who were refractory to or dependent on corticosteroids per the definitions used in this analysis. Finally, the number of patients who were steroid-dependent was small, limiting comparison with patients who were steroid-refractory.
In conclusion, a rapidly worsening clinical course and high mortality rate was observed in real-world patients with steroidrefractory and steroid-dependent acute GVHD from 11 US centers. These findings further emphasize the need for therapies that effectively prevent or reverse disease progression.   Fig. 3 Additional systemic anti-GVHD therapy (n = 89). *GVHD Graft-versus-host disease. *25.8% of patients used ≥2 additional therapies. † Polyclonal antibodies included antithymocyte globulin and antilymphocyte globulin.