Yi-Bin Chen, MD

Looking forward, a lot of room remains for improvement and research in acute graft-versus-host disease [aGvHD]. When we talked about the case that I presented, what I mentioned was the initial assessment of the biology for the patient since the origin of acute graft-versus-host disease is really an adaptive immunological response, which we think takes weeks to develop.

I would think we would be able to develop assays or tests that give us a glimpse into this immune response before actual clinical symptoms. There has been some work done along these lines; the MAGIC [Mount Sinai Acute GVHD International Consortium] biomarkers we mentioned have not shown themselves to be necessarily predictive but can stratify into risk groups, from biomarkers taking a week after transplant to those who will ultimately develop graft-versus-host disease. More sensitive and specific tests would help us intervene in a preemptive fashion before having patients experience the morbidity of clinical graft-versus-host disease.

At diagnosis, we will hopefully be able to develop risk-stratified treatments, so patients who we truly believe are high risk can be treated differently. There are patients who should receive steroids plus another drug up front to help bring about a response. Then, there are patients who are ultimately low risk. These patients may not need steroids at all and can be treated with an alternative, less toxic agent, or perhaps they’re initially treated with steroids and achieve a response. Then, we’re able to decrease the steroid much faster than we already do. This gets into the judgment of response, which is primitive. It’s still based on clinical observation of the body surface area, skin rash, the measurement of the total bilirubin value, and the assessment of the volume of diarrhea in 24 hours.

When patients are in the hospital, we round on them every 24 hours. We’re impatient and want to see improvement on a daily basis. As I mentioned, though, the time it takes for the gastrointestinal epithelium to heal after such severe injury does not mesh with our inpatient time scale. Our estimation of response based on these metrics is wholly inaccurate, so we need to develop better ways to monitor patients, judge if they’re getting better, and gauge whether improvements are working or if they are getting worse and need to add something else.

Hopefully, research can be done to better our clinical judgment, which could lead to a better assessment on clinical trials or if treatments are working. Clinical trials are difficult. There’s no doubt about that for this patient population. It’s a rare disease that has a heterogeneous biology. This is why many clinical trials in the past have ultimately failed or did not accrue and were closed early.

You can see why the judgment of response plays a huge role in these clinical trials. If you start drug X and decide to move to another therapy if the diarrhea volume is not getting better by day 4, then you’ve ultimately declared that patient to be a failure of drug X or drug X to be a failure for that patient. If you had just waited 3 more days, the diarrhea may have gotten better because the kinetics of how the GI [gastrointestinal] epithelium heals is different from what you want it to be. You can see how clinical trials are ultimately dependent on how we behave and judge a response. That’s an obstacle at large in the community that we’re trying to fix.

Lastly, we need different agents. Graft-versus-host disease used to be thought of as just a Th1-driven T-cell response. As it becomes steroid refractory, it involves different pathways of inflammation, immunology, etc, to the point that it is no longer a T-cell–driven process.

The traditional approach of hammering away at the immune system did not lead to success. It only led to incidence of opportunistic infections and death. The newer treatments do target different pathways. Ruxolitinib is an example that targets the JAK1 and JAK2 pathways that signal for a lot of proinflammatory cytokines. That is at the heart of why we think it has some effect, but I’m sure there are other pathways that are involved for patients who have steroid-refractory acute graft-versus-host disease that need to be explored.

Lymphocyte trafficking is another avenue that both myself and others have looked into. Fixing the microbiome is yet another modality that we can address through probiotics or fecal transplantation. Looking at these other pathways of inflammation will be pivotal to our understanding. There are a lot of people with steroid-refractory aGvHD that probably have some evidence of localized TMA [thrombotic microangiopathy], and we see that in biopsies if we look hard enough. Perhaps targeting the endothelium as a target of graft-versus-host disease in this specific organ can help patients, as well.

These are just a few of the avenues that we’re doing research in to hopefully improve outcomes for this unfortunate group of patients.

Case: A 50-Year-Old Woman With Steroid-Refractory Acute Graft Versus Host Disease

A 50-year-old woman presents on day +32 for a routine follow-up visit after myeloablative matched unrelated donor peripheral blood stem cell transplant for AML.She complains of a new erythematous macular rash and some mildly loose stools which she estimates is 450cc in 24 hours.

PE: rash noted on her neck, shoulders, and upper trunk (~60% BSA)



Labs: total bilirubin 2.7 mg/dl, AST 60 U/L, ALT 75 U/L

Stool testing negative for bacterial/viral infection

Skin biopsy of the rash showed sparse inflammation and abundant dyskeratotic keratinocytes

Flexible sigmoidoscopy showed patchy erosion and biopsy showed inflammatory cells with cryptitis

Modified Glucksberg Criteria: grade II; MAGIC Criteria: grade II

Admitted as an inpatient for evaluation and initiated methylprednisolone 1.0 mg/kg and topical steroids

No treatment response after 3 days, after dose increase of up to 2.0 mg/kg/day IV methylprednisolone

After 7 days of systemic steroids diarrhea was around 1500 cc/day, rash was better

She was started on ruxolitinib 5 mg PO BID which was tolerated well; increased to 10 mg PO BID 3 days later

Videos

Yi-Bin Chen, MD, comments on the need for improvements in patient assessment tools, measurements of response, and therapeutic agents for the management of acute GvHD in the future.

Unmet Needs in the Management of aGvHD

Once this patient was deemed to be steroid refractory, which was obvious in this case, there were a few treatment options. None of them are truly FDA approved, except for ruxolitinib, which was recently approved in 2019. In the past, we used other modalities. These include drugs like sirolimus; mycophenolate; ATG [antithymocyte globulins]; and etanercept, which is the TNF [tumor necrosis factor] blocker. Newer treatments have involved ECP [extracorporeal photopheresis]; a monoclonal antibody called vedolizumab, which targets lymphocyte trafficking and is used for inflammatory bowel disease; and ruxolitinib. Ruxolitinib is the only one that’s been tested in a large phase 3 study, as we’ll discuss. It has the most data to support its use in this setting.

The REACH2 study was pivotal for all in the transplant field, specifically for those of us who do research in graft-versus-host disease [GvHD]. It was a large international phase 3 study that enrolled 294 patients with steroid-refractory acute GvHD and randomized patients in a 1:1 fashion to ruxolitinib 10 mg twice a day vs the treating clinician’s choice of best appropriate therapy. There is a list of options that clinicians could choose from.

The primary end point of this study was overall response rate at day 28, which correlates with long-term outcomes, such as nonrelapse mortality and, more importantly, overall survival.

The key secondary end points were overall survival, overall response rate at other times (eg, day 14, day 56), duration of response, and overall survival.

The authors or investigators need to be applauded for carrying out this trial. Large trials with steroid-refractory acute graft-versus-host disease, much less randomized ones, have been very difficult to do for many reasons. Our patients have many competing risks and toxicities that are hard to judge, adjudicate, and attribute, and they all factor in. This is a huge step going forward in validating the findings of each.

The primary end point of the study, as I said before, was overall response rate at day 28. In the ruxolitinib arm, this was experienced by 62% of patients, with a complete response rate of 34%. In the control arm, this was experienced by 39% of patients, with a complete response rate of about 19%. This was highly statistically significant and validated the experience we were all having using ruxolitinib, as well as the findings of the initial REACH-1 trial and several observational ports.

Other secondary end points showed the durable response at day 56 was higher in the ruxolitinib arm—about 40% compared with 22% in the control arm, thus underlying the durable response that ruxolitinib has in this population—more than helping patients respond the first month and have disease flare.

The findings of this study were important. It’s the first big study in steroid-refractory acute graft-versus-host disease that showed a benefit of one therapy over others. The investments from all investigators merit applause and will help make progress in the field. It was ultimately FDA approved for this indication.

We knew ruxolitinib was a standard of care for steroid-refractory acute graft-versus-host disease, if not the standard of care. There are other agents not yet tested in large trials that many of us have used that brought about clinical responses. It’s not to say that they haven’t had negative results. Many agents have not been tested for a lot of the reasons I talked about in terms of the difficulty in conducting these trials. For the average patient with steroid-refractory acute graft-versus-host disease, ruxolitinib appears to be the standard of care.

There are some clinical factors to take into account. The main toxicity with using ruxolitinib has been cytopenias, specifically anemia and thrombocytopenia, mainly because of its blockade of the signaling cascades of EPO [erythropoietin] and TPO [thyroid peroxidase]. We do see this, but most of these cytopenias can be resolved without dose reductions or transfusions. There are other considerations, too, such as if the GI [gastrointestinal] disease is severe, as well as patients unable to tolerate oral agents. Then, you move to a different therapy, but that is rare in this population.

In our experience using it, we have not had a lot of other adverse events, let alone ones that have caused us to stop ruxolitinib and use another agent. This is not to say that it’s a home run. The overall response rate, while impressive, leaves room for improvement. The durable response rate is less than 50%, so we have more room to figure out what other treatments we can use—for patients who either fail ruxolitinib, or possibly develop assays in predicting which patients will respond to ruxolitinib and which will respond to other therapies.

Yi-Bin Chen, MD, considers potential treatment options for steroid-refractory aGvHD and examines the results of the phase 3 REACH2 study of ruxolitinib.

Treatment Options for Steroid Refractory aGvHD

There have been several definitions of steroid-refractory acute graft-versus-host disease that have been used throughout the years and for clinical trials. They can be separated into 3 categories.

Everyone would agree that after 3 days of high-dose systemic steroids, if your patient is getting worse, that meets the definition for steroid refractory. Another definition is just not getting better. Different times have been used for this. We generally use 7 days; some trials have used 14 days of not achieving a response to qualify as steroid refractory. The third definition is a bit more controversial. That involves patients who initially achieve a response and then, upon tapering of steroids, have a flare-up of their symptoms. That has been called steroid refractory in some trials. I tend to call that steroid dependent, and I think those patients have a less aggressive biology than the first 2 classes of patients who are regarded as truly steroid refractory.

The actual grade of the graft-versus-host disease at the time we declare people to be steroid refractory doesn’t factor into how I think about treatment very much. At that point, once a patient is steroid refractory, that’s their risk. I think about what the grade is at that point, though. I wouldn’t be surprised if, at the time of being steroid refractory, those with clinically more severe disease do worse than those with less severe disease. The important thing is that the graft-versus-host disease has shown itself to be steroid refractory, and you have to move on to another therapy.

In our routine practice, we go by the definitions I just said, so after adding high-dose steroids, we wait 3 days. If a patient gets worse, either with new organ involvement or worsening of their involved organs, then we move to therapy. In general, we wait at least 1 week to give treatments a chance to work.

Just as I talked about at the initial diagnosis, we do not have a good measurement in terms of biology of the disease. We’re very crude at judging if patients are responding. If I see a patient who’s being treated for acute graft-versus-host disease, I’m still looking at their skin rash, which is a good judgment. Is it fading? Is it getting darker? Less body surface involved? If the liver is involved, I’m looking at the level of bilirubin. If their GI [gastrointestinal] tract is involved, which are the majority of cases with steroid-refractory disease?

Unfortunately, we were looking at the volume of diarrhea in 24 hours. For many reasons, that is a crude and primitive way to judge response. Volume of diarrhea is influenced by many things: oral intake, use of narcotics, use of antibiotics. Even little swings in the volume of diarrhea that may or may not be accurately measured can dictate your judgment of response.

It emphasizes the point that you need to look at the whole patient to try to figure out if your treatment is working. Oftentimes, even though they might meet the definition of steroid refractory, you might wait because you think other signals are pointing to patients who are recovering.

For this patient, it was clear that at day 7, with things getting worse, she clearly met the definition of steroid refractory. In retrospect, we might have wanted to start with 2 mg/kg per day on day 1. Nobody knows if that would have made a difference, but there was already evidence of multiorgan involvement with all 3 organs, and especially the lower GI [gastrointestinal] tract, so that probably merited starting at 2 mg/kg per day instead of 1 mg/kg. We don’t know if that would have made a difference. It does emphasize our need to have better biomarkers for prognosis and response to help us ultimately be able to tailor our therapies appropriately.

Yi-Bin Chen, MD, discusses diagnostic criteria and measurements of response as they relate to the case of a 50-year-old woman with steroid-refractory acute GvHD.

Defining Steroid Refractory aGvHD

Acute graft-versus-host disease [aGvHD] is a complication of allogeneic transplantation in the mechanism of action. For most patients, we transplant. The potential curative modality here is an immunological attack of the donor graft upon the patient’s underlying malignancy, this graft vs malignancy effect.

The corollary to that is graft-versus-host disease. Graft-versus-host disease is a donor’s immune system attacking the host’s healthy body. You can view that as organ rejection from within because we’re transplanting the host with a new immune system. It is a potential complication of all allogeneic transplants, and every allogeneic transplant recipient will receive some form of graft-versus-host disease prevention.

There are several well-defined clinical risk factors for acute graft-versus-host disease. We know things like older age of the donor and recipient predisposal are associated with increased risk of aGvHD. We know the degree of HLA [human leukocyte antigen] mismatch, so if the recipient and donor have a higher degree of incompatibility, that intuitively leads to a higher incidence of acute graft-versus-host disease. The intensity of the conditioning regimen used before transplant is associated with a higher incidence of aGvHD, specifically the use of higher doses of total body range.

An interesting risk factor that has emerged over the last decade is the health of the microbiome of the recipient, meaning the diversity of the bacteria in the stool or the colon of the patient. Patients who have a restrictive diversity, which we call dysbiosis, seem to be at a much higher risk of developing acute graft-versus-host disease than those who have a diverse microbiome.

It’s tough to know if the microbiome is just a marker of certain things or if there’s a cause-and-effect relationship. Exciting research is ongoing to see if it can be manipulated.

There are different regimens used for graft-versus-host disease prophylaxis, and those can shape the risk and incidence after transplant. Traditionally, the more immunosuppression you have around a transplant to prevent graft-versus-host disease, the more it would result in a higher incidence of infection or even relapse of malignancy out back. Outcomes are fairly similar, but more modern regimens may be changing this. We hope to see results of clinical trials going forward that help improve outcomes for patients, ultimately sparing them graft-versus-host disease.

Acute graft-versus-host disease is common; it probably affects about 30% to 35% of patients after conventional transplant. When I talk about aGvHD, I’m talking about grades 2 to 4, which usually require the initiation of systemic treatment and have been shown to correlate with outcomes, as well as have an effect on patient outcomes in the long run.

Every recipient does receive some form of prevention. There are 2 real categories of graft-versus-host disease prevention. One is ex vivo manipulation of the graft. That means our cytotherapy lab gets the cells from the donor and manipulates them. This has traditionally been called T-cell depletion, so you can either take a graft and select for the stem cells or select the T cells of that donor graft to try to prevent graft-versus-host disease.

The much more common way we prevent graft-versus-host disease is through medications. Traditionally, we use a combination of a calcineurin inhibitor, such as tacrolimus or cyclosporine, added to short doses of methotrexate out back. There have been variations on this platform of a calcineurin inhibitor, methotrexate. We’ve added other drugs, such as sirolimus or mycophenolate, and have even used polyclonal T-cell globulins called antithymocyte globulins routinely.

The newest regimen that has emerged over the last 10-plus years is based on giving high doses of post-transplant cyclophosphamide on days 3 and 4 after transplant to not only deplete the allo-reactive cells of the graft in vivo but also encourage the reconstitution of regulatory T cells. These post-transplant cyclophosphamide platforms were initially developed and have become the standard for use in haploidentical transplants, meaning half mismatched-related donor transplants. They have been hugely successful, to the point that we’re not testing such regimens in the conventional match donor setting to see if they can alter the standard of care that we’ve used for a couple of decades.

Yi-Bin Chen, MD, defines acute graft-versus-host disease (aGvHD) as a complication of allogeneic transplant, discusses the clinical risk factors, and reviews prophylactic treatment options.

Clinical Risk Factors and Prophylaxis for aGvHD

Even though she was only overall grade 2, we feared that she had much more aggressive disease, given we were already seeing visceral involvement of her liver and lower GI [gastrointestinal] tract. Specifically, the lower GI disease is the most difficult organ to treat. If she hadn’t come that day and instead saw you 2 days later when the stool was maybe 1100 cc a day, her grade would have automatically been much worse, and you might think of her differently. That’s why you have to consider the perspective of when you’re seeing the patient and what the actual biology is.

Given the snapshot, we don’t have a measure of the biology of her disease to accurately predict the prognosis. Although the clinical grade does loosely correlate with prognosis, it is not a very good predictor. What we lack is something like noninvasive biomarkers that can accurately predict the prognosis of our patient. More importantly, we can have risk-stratified treatments for the different groups.

There are biomarkers that are commercially available and developed by MAGIC [Mount Sinai Acute GVHD International Consortium]. These measure 2 proteins called REG3-alpha and ST2, which are thought to be measures of mucosal injury in the bowel. They have been shown to predict for prognosis, but we don’t have proof that risk stratifying our treatments based on these biomarkers helps our patients. We have no proof that we should treat any of these patients differently based on their biomarker group. Future clinical trials and efforts can help us better accurately prognosticate patients, so we can develop these risk-stratified treatments.

Yi-Bin Chen, MD, gives insight into the current accuracy of predictive prognostics and the potential use of biomarkers to risk stratify patients with acute GvHD in the future.

Predictive Prognosis and Risk-Stratification in aGvHD

Hello. I’d like to welcome everybody to our 

™ case discussion. We have a very interesting case illustrating current standards of treatment and evaluation for steroid-refractory acute graft-versus-host disease.

I’ll start with a case. This is somewhat based on a true story. This is a typical case that we see for those who take care of transplant patients. We have a 50-year-old woman who presents on day 32 for a routine follow-up visit after a myeloablative matched unrelated donor peripheral blood stem cell transplant for underlying diagnosis of acute myeloid leukemia in first remission.

When she comes into the office, she complains of a new, erythematous macular rash that had developed over the last 3 days and some mildly loose stools, which she estimates are approximately a little less than 0.5 L in the previous 24 hours, though obviously it’s a bit difficult to estimate that.

Her past medical history is not significant. She’s healthy otherwise and was a very active person before her diagnosis of leukemia. Her therapy for leukemia was rather uncomplicated until she proceeded to transplant.

On physical exam, I noted the aforementioned rash, which had a distribution on her neck, shoulders, and upper trunk. By estimation, it covered about 60% of her body surface area. Further clinical work-up that came back showed laboratory values significant for a total bilirubin of 2.7 mg/dL, with mildly elevated transaminases around twice the upper limit of normal.

Given this presentation, we admitted her to the hospital for inpatient expedited work-up with the suspicion that these findings could represent the initial onset of acute graft-versus-host disease. With our first battery of tests that we sent in the first 24 hours, we found that her stool testing was negative for any bacteria or viral infections, including 

]. Blood tests were negative for any evidence of infection or reactivation of viruses, including hepatitis B and C, CMV [cytomegalovirus], EBV [Epstein-Barr virus], as well as HHV6 [human herpesvirus 6].

We did perform a biopsy of the affected area, and it came back showing sparse inflammation and abundant dyskeratotic keratinocytes. Given her diarrhea present in the hospital, which went along with her estimated volume of about 400 to 500 cc in 24 hours and was quite watery, we did ask our colleagues in GI [gastrointestinal] to perform a flexible sigmoidoscopy with random mucosal biopsies. Macroscopically, we did see some areas of patchy erosive ulceration in her sigmoid colon. The biopsy itself showed an inflammatory infiltrate with evidence of cryptitis and apoptosis.

Based on these findings, she was diagnosed with the initial onset of acute graft-versus-host disease. In terms of consensus staging criteria, given the body surface area of her skin, she will be a skin stage II. Given the volume of her diarrhea was less than 500 cc a day, she’d be a GI stage I and with lower GI involvement. Based on her total bilirubin level at 2.7 mg/dL, she’d be a liver stage I. Considering the individual organ stages, translation into an overall grade would be consistent with a grade 2 overall acute graft-versus-host disease.

Given this diagnosis, we initiated therapy with IV [intravenous] steroids at a dose of 1 mg/kg of intravenous methylprednisolone per day. We did give topical steroids for the skin rash as well, with the hope that this initial therapy would provide a response. At the same time, we feared—given the evidence of visceral involvement already in terms of the liver and the diarrhea—that we were seeing the beginning of it, where the snapshot we were getting was not helping us get a glimpse into the actual biology. We feared that our initial steroid dose would not be enough.

This was proven correct; after 3 days, there was no response in terms of her diarrhea or her liver tests. In fact, the diarrhea did progress to reach almost about 1 L a day, and the bilirubin stayed right around 3 mg/dL. Her skin rash did fade a bit and didn’t spread. It turned a bit darker, but it didn’t disappear right away. Because there was no response and things were worsening after 3 days, her steroid dosage was increased to more conventional dosing of 2 mg/kg per day in divided doses of IV methylprednisolone.

Even with the increased steroid dosing, after a total of 7 days of systemic steroids, she still had not experienced a response with her graft-versus-host disease. The diarrhea was now between 1000 and 1500 cc a day, even on bowel rest. Her bilirubin stood around 4 mg/dL. Her skin rash faded and seemed to be resolving. That may have been the only organ that was experiencing any type of response.

At this point, a week into therapy with this course, she certainly met the definition of steroid-refractory acute graft-versus-host disease. She was then started on ruxolitinib 5 mg orally twice a day, which was well tolerated from a subjective, as well as laboratory, point of view. Three days later, we were able to increase to up to 10 mg twice a day.

Yi-Bin Chen, MD, provides an overview of the case of a 50-year-old woman with steroid-refractory acute GvHD, including initial presentation, clinical work-up, and treatment.

Yi-Bin Chen, MD | Authors

Articles

Unmet Needs in the Management of aGvHD

Treatment Options for Steroid Refractory aGvHD

Defining Steroid Refractory aGvHD

Clinical Risk Factors and Prophylaxis for aGvHD

Predictive Prognosis and Risk-Stratification in aGvHD

Case Overview: Steroid Refractory aGvHD