Intra‐arterial methylprednisolone for the management of steroid‐refractory acute gastrointestinal and hepatic graft versus host disease

Abstract

Steroid-refractory or dependent (SR/D) acute graft versus host disease (aGVHD) involving the gastrointestinal (GI) tract or liver is a dire complication of allogeneic hematopoietic stem cell transplantation (HSCT) for which there is no standard therapy [1-3]. Most agents utilized are highly immunosuppressive and thus associated with high rates of opportunistic infections [4]. An alternative approach involves direct infusion of corticosteroids into mesenteric and/or hepatic arteries [5]. While the mechanism of action remains speculative, a few small series suggest reasonable efficacy with this approach [5-10]. Here we describe a retrospective analysis of intra-arterial (IA) methylprednisolone (MePDSL) treatment of 13 patients with SR/D aGVHD involving the GI tract, including four with concurrent liver GVHD. Various combinations of the celiac, superior and inferior mesenteric, and common hepatic arteries were injected with MePDSL, generally 60 mg/vessel, on a median of 3 consecutive days. There were no major procedural complications. Six patients experienced GI complete responses (CRs) and two experienced partial responses. One of four patients with liver GVHD achieved a durable CR. Three patients remain alive and well. These results were obtained in a highly refractory group of patients; we suggest that better results might be obtained with earlier and more standardized IA therapy. Acute GVHD is a common complication of allogeneic HSCT that can result in significant morbidity and mortality [1, 12]. Involvement of the GI tract with aGVHD is particularly problematic; in addition to having a lower response rate than GVHD involving other systems, compromised GI function prevents adequate oral nutrition and provides an obvious portal for infection [13]. When primary therapy with corticosteroids is inadequate, the administration of additional highly immunosuppressive agents often contributes to fatal opportunistic infections. As an alternative, the use of MePDSL infused into regional mesenteric and/or hepatic arteries offers the potential to direct therapy to sites of involvement, thus minimizing systemic effects and secondary complications, thereby improving the long-term outcomes of these patients. The precise mechanism by which IA MePDSL is effective for “steroid-refractory” disease remains unknown. It is possible that a higher regional concentration of glucocorticoid has an enhanced effect on the modulation of local cellular immune response and production of proinflammatory cytokines. It has also been postulated that a high local concentration of steroid overcomes refractoriness caused by down-regulation of steroid receptors in the inflamed GI mucosa of patients with severe GI GVHD [8]. Evidence for such a mechanism comes largely from studies in patients with inflammatory bowel disease, which demonstrate that glucocorticoid receptors are down-regulated in inflamed intestinal mucosa and that this finding correlates with steroid refractoriness [14-16]. During the inclusive period of 26 April 2007 and 15 July 2010, we treated 13 patients with SR/D GI GVHD with IA MePDSL. Four patients had concurrent liver GVHD. Patient demographics, GVHD prophylaxis, and GVHD treatment before, and concurrently with, the initiation of IA MePDSL injections, are summarized in Table I. Details regarding IA MePDSL administration and results are summarized in Table II. The 13 patients were treated on 15 occasions, comprising a total number of IA injections of 106. Two patients received injections of IA MePDSL on two separate occasions, patient 1 five months after the initial treatment and patient 2 two weeks later. All patients except one received all IA injections planned; the IMA could not be cannulated on one occasion in a small child (patient 1). As indicated in Table II, MePDSL was given at a flat dose of 60 mg/vessel in 10 of the 13 patients; two pediatric patients received doses of 1 to 2 mg/kg/vessel and one adult patient received 50 mg/vessel. Most frequently (seven patients), MePDSL was injected into the three main mesenteric vessels (celiac, superior, and inferior mesenteric arteries). Five patients received injections into the common hepatic artery, which also feeds the celiac axis; when the common hepatic artery was injected, a separate injection into the celiac artery was deferred. The majority of patients (eight) received injections on 3 consecutive days. Two pediatric patients received injections on a single day, one due to clinical concerns and one due to parental request. One patient received injections on 4 consecutive days. Consistent with previous reports, we found IA infusion of MePDSL to be safe. Minor complications included three hematomas that did not require transfusion, one episode of gastrointestinal bleeding that subsided with supportive care, and one episode of abdominal pain and bacteremia that resolved with appropriate therapy. One patient suffered a mild myocardial infarction a few days after the procedure; this event was felt to be unrelated to the procedure. While there were no serious immediate complications, the relative contribution, if any, of this therapy to more delayed and/or systemic complications (e.g., infection) is more difficult to ascertain. Of particular note in this regard, all but one of our patients received monoclonal antibody therapy (generally infliximab) either before, or concurrently with, IA MePDSL. This additional immunosuppressive therapy, as well as the inherently high mortality rate associated with severe SR/D GI GVHD, makes it virtually impossible to determine whether or not treatment with IA MePDSL contributed to any of the adverse outcomes we observed. For assessing therapeutic response, we chose only two parameters: achievement of a CR of GI (and hepatic) signs and symptoms by day 28 following IA treatment and overall survival. Even so, responses may have been over- or underestimated; to some degree this is inherent in grading aGVHD of the GI tract, as the signs and symptoms can be affected by other inciting factors, as well as ancillary therapies. In addition to limiting response parameters, we also applied a relatively stringent definition of response: the absence of GI symptoms while tolerating enteral nutrition and medications. Of note in this regard, many of our patients were acutely ill at the time of IA MePDSL treatment, making the assessment of response problematic. Indeed, two patients died prior to day +28 and four additional patients who died between day +29 and +43 were so unwell that accurate retrospective assessments of response were not completely feasible. Six of our 13 patients had GI CRs. Two have had durable CRs: one is 8 months postprocedure, tapering immunosuppressive therapy; the other is more than 3 years postprocedure and off all immunosuppression. One patient (1) had a CR following a single IA infusion, but then suffered a recurrence of GVHD; he received another infusion 5 months after the first, resulting in a PR. He survived for 15 months before succumbing to an infection. Two patients had resolution of GI symptoms at day +28, but died of infection shortly thereafter. One patient had a GI CR, but subsequently developed progressive skin GVHD and liver dysfunction and died from a complication of therapy. Of the remaining seven patients, three had no response, two had PRs, and two died before day +28 and were therefore not evaluable. Of the eight patients achieving less than a CR, one who achieved a good PR is doing well 8 months post-IA MePDSL, tapering off immunosuppression. Four patients received hepatic IA MePDSL for liver GVHD. One of these patients achieved a durable CR, one had no response, and two expired before day +28. While the responses and outcomes we observed may be considered modest, our patients were generally treated late in the disease process, diminishing the likelihood of a good response or of long-term survival. The two patients who were treated relatively early in the disease process both had good responses and remain alive and well. Patient 4 received IA MePDSL as second line therapy and is now more than 3 years post treatment, off all immunosuppression. Patient 13 was treated with IA MePDSL soon after institution of other second line agents; he achieved both GI and hepatic CRs and remains well 8 months post therapy. All of our other patients had received two to four lines of prior therapy and were receiving multiple other agents concurrently with IA MePDSL. Considering our results, as well as those of others, we believe this therapeutic strategy warrants follow-up studies. Optimizing patient selection and administration of IA therapy to permit the aggressive tapering of systemic immunosuppression would likely be a crucial to the success of such studies. One approach would be to incorporate IA MePDSL into the primary treatment of aGVHD therapy in selected patients; even if this resulted in treatment of some patients destined to respond to standard therapy, it is possible that production of a more rapid CR would allow earlier discontinuation of systemic steroids. More importantly, primary use of IA steroids might be beneficial in those patients destined to have inadequate responses to systemic steroids. In addition to the use of IA MePDSL as previously described, it may also be useful to consider whether alternative administration strategies and/or the use of other IA agents might be more efficacious. For example, dose escalation or response-based treatment schedules could be considered. In terms of other therapeutic agents, methotrexate has been administered intra-arterially in conjunction with MePDSL for treatment of liver GVHD [17] and other chemotherapeutic agents are sometimes administered by this route for the treatment of malignancies. Thus, one could consider concurrent administration of intra-arterial MePDSL and antibody modulators of inflammation (e.g., infliximab) for the treatment of GI GVHD, particularly in those patients with refractory disease. Certainly, a better understanding of the mechanism of action for IA MePDSL would be helpful in optimizing its use. Based on the presumption that regional treatment with IA MePDSL applied early in the course of GI aGVHD would: (1) be of therapeutic benefit, (2) be associated with minimal acute toxicities, and (3) reduce the need for prolonged systemic immunosuppressive therapy, we are pursuing a Phase II study. Hopefully, this will lead to a more standardized method of administration of IA MePDSL, generate better guidelines for its use, and provide insights for future studies and improvements in this therapy. In this retrospective analysis, charts of all patients treated with IA MePDSL for the management of GI GVHD were reviewed according to institutional guidelines following Institutional Review Board approval. Patient demographics are summarized in Table I. Patients receiving IA MePDSL had clinical evidence of acute GVHD of the GI tract (stool volume >1500 ml/day, >5 stools/day, abdominal pain, and/or GI bleeding.) In addition, all but one had biopsy-documented GI GVHD (one patient had GVHD documented by skin biopsy, but was too ill to undergo GI biopsy.) All patients received systemic corticosteroids and were gauged either steroid-refractory or dependent, defined as follows: (1) use of MePDSL or prednisone at 2 mg/kg/day; (2) in those so treated, resistance was defined as progression after three, no change after seven and/or persistence after 14 days; dependence was defined as the inability to taper to a dose of <0.5 mg/kg/day without progression of GVHD. As indicated in Table I, all patients except one received routine GVHD prophylaxis with a calcineurin inhibitor and either mycophenolate or methotrexate. One patient undergoing a second transplant for relapsed acute leukemia received tacrolimus alone. When acute GVHD developed, in addition to restarting or continuing prior therapy with a calcineurin inhibitor and MMF, systemic MePDSL, and oral nonabsorbable steroids (i.e., budesonide) [18] were started. (In a single case—a small child—treatment with oral budesonide was not feasible.) Routine supportive care included administration of systemic antibacterial, antiviral, and antifungal prophylaxis. A proton pump inhibitor was given as ulcer prophylaxis. Total parenteral nutrition, enteral supplementation, antiemetics, and antidiarrheals were used when indicated. We followed the method of Shapira et al. for administration of IA MePDSL [6]. In 10 cases, a vascular sheath was left in place for intra-arterial therapy on subsequent days. Doses and schedules of MePDSL administration are summarized in Table II. We assessed three outcome parameters: (1) GI ± hepatic CR at d +28 [18]; (2) local complications, including inability to cannulate an artery and post-IA events, such as a clinically-significant hematoma or new onset GI bleeding; and (3) current status and/or cause of death. GI CR was defined as the absence of GI symptoms and signs, and ability to tolerate full oral nutrition and medications. Partial response (PR) was defined as >50% reduction in diarrhea or improvement in other GI symptoms while tolerating partial enteral nutrition. Hepatic CR was defined as complete normalization, and PR as >50% reduction, in liver transaminase and bilirubin levels. Laurie A. Milner [email protected]* , Michael W. Becker , Steven H. Bernstein , Lauren Bruckner , Jonathan W. Friedberg , George A. Holland§, J.J. Ifthikharuddin , Jane L. Liesveld , Edward J. Mathes§, Heather L. Menchel¶, Craig A. Mullen , Talia Sasson§, Gordon L. Phillips II , * Pathology and Laboratory Medicine, University of Rochester Medical Center, James P. Wilmot Cancer Center, and Strong Memorial Hospital, Rochester, New York, Hematology Oncology, University of Rochester Medical Center, James P. Wilmot Cancer Center, and Strong Memorial Hospital, Rochester, New York, Pediatric Hematology Oncology, University of Rochester Medical Center, James P. Wilmot Cancer Center, and Strong Memorial Hospital, Rochester, New York, § Interventional Radiology, University of Rochester Medical Center, James P. Wilmot Cancer Center, and Strong Memorial Hospital, Rochester, New York, ¶ Hematology Oncology, Nursing, University of Rochester Medical Center, James P. Wilmot Cancer Center, and Strong Memorial Hospital, Rochester, New York.