Steroid-refractory acute GVHD — from first-line corticosteroid failure to ruxolitinib and beyond
Acute graft-versus-host disease (GVHD) remains a leading cause of non-relapse morbidity and mortality after allogeneic hematopoietic stem cell transplantation. High-dose systemic corticosteroids have been the standard first-line treatment for over four decades, yet only roughly half of patients achieve a durable response. The interval used to judge that response is the single most consequential checkpoint in the entire GVHD treatment pathway — it determines whether a patient continues on steroids or is escalated to second-line therapy.
Corticosteroids are broad, rapidly acting immunosuppressants that dampen the alloreactive T-cell response driving acute GVHD:
• Mechanism: bind cytoplasmic glucocorticoid receptors → translocate to nucleus → suppress transcription of pro-inflammatory cytokine genes (IL-1, IL-2, IL-6, TNF-α, IFN-γ) and induce lymphocyte apoptosis • Onset: clinically meaningful effect often visible within 3–5 days • Grading at initiation: treatment intensity and monitoring depend on baseline GVHD grade (skin, gut, liver staging combined into an overall grade) • Dosing nuance: some centers use lower "GVHD-directed" doses (0.5–1 mg/kg/d) for lower-grade disease to reduce steroid toxicity while preserving efficacy
Despite being first-line for decades, corticosteroids carry substantial toxicity: hyperglycemia, myopathy, avascular necrosis, mood disturbance, and — critically in the post-transplant setting — heightened susceptibility to invasive fungal and viral infections. This toxicity burden is precisely why timely, well-defined identification of non-response matters: prolonged unnecessary steroid exposure in a patient who will not respond only adds risk without benefit.
Response assessment is organ-based and longitudinal, following the same staging system used at diagnosis:
• Skin: percentage body surface area with rash, bullae presence • Gastrointestinal: stool volume/day (or emesis in children), presence of visible blood • Liver: total bilirubin trend
Each organ is staged 0–4; the overall grade (I–IV) is derived from the worst-affected organs. A response assessment at a given day compares the current grade/stage to baseline:
• Complete response (CR): resolution of all GVHD manifestations • Partial response (PR): improvement in ≥1 organ without progression in others • No response / progression: no improvement, or worsening in any organ, or need for additional immunosuppression
Because this assessment directly gates the decision to escalate therapy, transplant centers increasingly standardize it using consensus frameworks (see Stage 2) rather than informal clinical impression alone.
Steroid-refractory acute GVHD is not a vague clinical impression — it is defined by specific, time-anchored criteria developed to standardize when a patient should move to second-line therapy. These criteria, refined over successive consensus efforts (including the MAGIC and NIH/CIBMTR frameworks), exist because delaying escalation in a truly refractory patient allows ongoing tissue damage, while escalating too early exposes responders to unnecessary additional immunosuppression and infection risk.
Most consensus definitions converge on three overlapping triggers, any one of which is sufficient to classify a patient as steroid-refractory:
1. Progression despite treatment — GVHD grade/stage worsens after ≥3 days of an adequate corticosteroid dose, indicating the current regimen is not controlling alloreactivity.
2. No improvement — no reduction in GVHD grade/stage after ~5–7 days of adequate-dose steroids, indicating a lack of any measurable response.
3. Incomplete response — failure to achieve a complete response by day 28 (or a defined later time point) despite adequate dosing and, often, an initial dose taper attempt, indicating persistent disease that will not resolve with steroids alone.
Some frameworks also incorporate "steroid-dependence" — recurrence of GVHD during steroid taper — as a related-but-distinct category that similarly prompts consideration of additional agents, since it reflects an inability to control disease without ongoing high-dose immunosuppression.
Standardized, time-anchored criteria serve two purposes simultaneously:
• Prevent under-treatment: without a defined trigger, clinicians may continue escalating steroid dose or extending duration in a patient who is unlikely to respond, delaying access to agents with a genuinely different mechanism of action.
• Prevent over-treatment: without criteria, patients showing early signs of response might be prematurely switched to second-line agents that carry their own toxicity and cost, when watchful continuation of steroids would have sufficed.
• Enable trial comparability: because "steroid-refractory" was historically defined inconsistently across studies, pivotal trials for second-line agents (including REACH2) explicitly built their eligibility criteria around a harmonized definition — anchoring day-3, day-5/7, and day-28 assessment points. This is part of why interpreting the phrase "second-line therapy indicated" requires knowing exactly which criterion was met and when.
A patient is not "steroid-refractory" simply because GVHD is still present — refractory status is earned by meeting a specific time-and-response threshold. This distinction is exactly what the days-on-steroid and response sliders in this simulator are built to illustrate.
Ruxolitinib, an oral JAK1/2 inhibitor already established in myeloproliferative disease, became the first agent with randomized, controlled evidence of superiority over investigator-chosen best available therapy (BAT) specifically for steroid-refractory acute GVHD. The phase III REACH2 trial reshaped clinical guidelines, positioning ruxolitinib as the preferred second-line agent for most patients with steroid-refractory acute disease.
Acute GVHD is propagated by a cascade of pro-inflammatory cytokines (IFN-γ, IL-6, IL-2, TNF-α, IL-15) signaling through type I/II cytokine receptors that rely on Janus kinases (JAK1, JAK2, JAK3, TYK2) to transmit signal from the receptor to STAT transcription factors:
• Cytokine binds receptor → receptor-associated JAKs trans-phosphorylate each other → JAKs phosphorylate STAT proteins → phosphorylated STATs dimerize → translocate to nucleus → drive transcription of further inflammatory genes and amplify donor T-cell activation and expansion
• Ruxolitinib is a small-molecule ATP-competitive inhibitor selective for JAK1 and JAK2, occupying the kinase active site and preventing phosphorylation of the receptor complex
• Because JAK1/2 sit downstream of multiple distinct cytokine receptors, a single small molecule simultaneously dampens signaling from several of the cytokines driving alloreactive T-cell expansion and tissue-directed cytotoxicity — rather than neutralizing one cytokine at a time as biologic agents do
This broad, receptor-proximal blockade is thought to explain much of ruxolitinib's efficacy advantage over agents that target a single inflammatory mediator.
REACH2 randomized patients with steroid-refractory acute GVHD (grades II–IV) to ruxolitinib or investigator-chosen best available therapy:
• Primary endpoint (day 28 overall response rate): 62% ruxolitinib vs. 39% BAT • Durable response at day 56: also favored ruxolitinib • Response occurred across skin, gut, and liver involvement, though absolute rates vary by organ and baseline severity • Common toxicities: cytopenias (thrombocytopenia, anemia), increased infection risk consistent with added immunosuppression, and need for blood-count monitoring during dose titration
Because of this evidence, ruxolitinib is now recommended by major consensus guidelines as the preferred second-line agent for most patients with steroid-refractory acute GVHD, reserving other agents for patients with contraindications (e.g., severe cytopenias), inadequate ruxolitinib response, or specific organ/logistic considerations.
REACH2 remains the only large randomized trial in steroid-refractory acute GVHD to show a clear response-rate advantage for a single agent over investigator-chosen alternatives — the basis for ruxolitinib's current "preferred second-line" status.
No second-line agent works for every patient, and ruxolitinib is not always appropriate — severe cytopenias, prior JAK inhibitor exposure, drug interactions, or an inadequate response may all prompt consideration of alternative approaches. A range of immunosuppressive and biologic agents, cellular therapies, and physical modalities have accumulated supportive (though generally less robust, non-randomized) clinical experience and remain part of the second-line armamentarium.
Selection among alternative agents is guided by which organs are involved, prior treatment exposure, and comorbidities:
• Extracorporeal photopheresis (ECP): leukapheresed cells are exposed to a photosensitizing agent and UV-A light, then reinfused; induces immunomodulatory (not simply cytotoxic) effects. Favored for skin- and liver-predominant disease and in patients where added pharmacologic immunosuppression is a concern, given its comparatively favorable infection profile.
• Anti-thymocyte globulin (ATG): polyclonal antibody preparation causing broad T-lymphocyte depletion. More aggressively immunosuppressive, with meaningfully higher infection risk; generally reserved for more refractory or higher-grade presentations.
• Gut-selective and TNF-directed biologics: infliximab (anti-TNF-α) and vedolizumab (anti-α4β7 integrin, blocking gut-homing lymphocyte trafficking) are considered preferentially for GI-predominant steroid-refractory disease, where localized mechanisms may offer benefit with somewhat more targeted toxicity than systemic agents.
• Mesenchymal stromal cells (MSCs): infused allogeneic stromal cells with broad immunomodulatory paracrine effects; clinical trial results have been mixed across studies and endpoints, and regulatory approval status varies by region.
• Mycophenolate mofetil, sirolimus, and other calcineurin/mTOR-pathway agents: sometimes added or substituted based on prior GVHD prophylaxis regimen and renal/hepatic tolerance.
Even with strong randomized evidence for ruxolitinib, alternative agents remain clinically relevant for several reasons:
• Contraindications: significant baseline cytopenias, active uncontrolled infection, or prior intolerance to JAK inhibition • Partial or absent ruxolitinib response: patients who do not respond adequately to ruxolitinib still need a next option, often selected from this same alternative list • Access and logistics: ECP requires specialized apheresis infrastructure not available at every center; some biologics require specific infusion scheduling • Combination strategies: in practice, agents are sometimes layered or sequenced (e.g., ECP added to ruxolitinib) rather than used strictly in isolation, particularly for organ-refractory disease
The evidence base for most alternatives is considerably less robust than REACH2 — predominantly retrospective series, single-arm studies, or smaller randomized trials — so agent choice at this stage relies heavily on individualized clinical judgment.
Starting a second-line agent is not the end of the decision process — it opens a new monitoring window with the same fundamental question as Stage 1: is the disease responding adequately, and on what timeline? When second-line therapy also fails to achieve adequate response, patients face a materially worse prognosis, and management shifts toward additional treatment lines, combination approaches, and structured clinical trial enrollment.
Once second-line therapy is initiated, the same organ-based staging used at diagnosis and during first-line assessment continues on a structured cadence:
• Skin: body-surface-area involvement trend • GI: stool output/day, presence of blood, ability to tolerate oral intake • Liver: bilirubin trajectory
Clinicians look not just at a single time point but at the trajectory — is the disease burden falling, plateaued, or rising — since kinetics inform whether to continue the current agent, add a complementary modality, or move to a different line entirely. Infection surveillance runs in parallel, since escalating immunosuppression compounds an already elevated infection risk in the post-transplant setting.
Patients who do not achieve adequate response to second-line therapy (including ruxolitinib) represent a particularly high-risk population, sometimes referred to as having multi-line-refractory or steroid- and ruxolitinib-refractory GVHD:
• Prognosis: non-relapse mortality rises substantially in this group compared with patients who respond to first- or second-line therapy • No single established "third-line standard": unlike the ruxolitinib-versus-BAT comparison settled by REACH2, evidence for specific third-line agents is far more limited, drawn mostly from smaller studies and case series • Clinical trial enrollment: because of the scarcity of high-quality evidence beyond second-line therapy, enrollment in a clinical trial is frequently prioritized at this point, both to access investigational agents and to contribute to the evidence base for this difficult-to-treat population • Combination and sequencing strategies: agents from Stage 4 (ECP, ATG, gut-selective biologics, MSCs) are often layered or trialed in sequence, guided by which organs remain most affected and what has already been used
The core message for this stage is procedural, not just pharmacologic: continued, structured reassessment — and a low threshold to pursue additional lines of therapy or trial enrollment when second-line treatment is inadequate — is itself part of the standard of care for steroid-refractory GVHD.