Iron Overload Manifesting as Apparent Exacerbation of Hepatic Graft-versus-Host Disease after Allogeneic Hematopoietic Stem Cell Transplantation

Introduction 

Allogeneic hematopoietic stem cell transplantation (allo-SCT) is curative treatment for many patients with malignant and nonmalignant disorders. Transfusion of large quantities of packed red blood cells (PRBCs) during initial diagnosis, chemotherapy, and transplantation can lead to transfusion-associated iron overload. In adults, transfusion-associated iron overload generally occurs with transfusion of more than 100 to 150 units of PRBCs [1]. Iron overload has been reported in the recipients of both autologous SCT and allo-SCT, as well as patients receiving chemotherapy [2, 3, 4, 5, 6, 7, 8, 9]. Lifetime infusion of less than 50 to 60 units of PRBCs led to clinically significant iron overload in most reported patients; increased intestinal absorption due to epithelial injury related to chemotherapy or graft-versus-host disease (GVHD) is implicated [7, 10, 11]. In a patient with an established diagnosis of hepatic chronic GVHD (cGVHD), worsening liver functions are often interpreted as an exacerbation of liver GVHD resulting in empiric intensification of immunosuppressive therapy. Here we provide the first report of 6 patients in whom iron overload–related liver dysfunction was initially considered to be cGVHD exacerbation. Additionally, the role of maintenance phlebotomy after successful iron-depletion therapy is discussed.

Patients and methods 

Clinical and laboratory data of 6 patients who did not respond to continued or intensified hepatic cGVHD treatment in whom iron overload presented as cGVHD exacerbation were prospectively collected (May 2002 to September 2005). The patient characteristics are described in Table 1. Medical charts were reviewed for patient demographics, primary diagnosis, transplant characteristics, and cGVHD status. Clinical organ system involvement including the liver, skin, heart, and pancreas was noted. Serial liver function tests were performed, including alanine aminotransferase (reference range, 7-40 IU/L), alanine aspartate transferase (reference range, 7-40 IU/L), alkaline phosphatase (reference range, 50-136 IU/L), total bilirubin (reference range, 0.3-1.2 mg/dL), albumin (reference range, 3.4-5.0 g/dL), and prothrombin time (reference range, 9.5-12 seconds). Serum ferritin (reference range, 10-240 μg/L), transferrin saturation (reference range, 20%-55%), number of PRBC transfusions, and time to ferritin normalization assessed iron status. Transjugular liver biopsy, performed in 4 patients, was uncomplicated. Pretransplantation stored DNA samples or DNA samples from buccal mucosa were used to exclude hereditary hemochromatosis gene mutations (H63D and C282Y) in all 6 patients.

Table 1. Patient Characteristics

	Patient	Age/Sex	Diagnosis	Conditioning Regimen⁎	Donor/Graft	cGVHD	Life-time PRBC	Pre-Tx Ferritin	Post-Tx Ferritin	Transferrin Saturation %	Time to IOL (M)	IOL Organs	Liver Biopsy
	1	48/F	AML-CR2	CY/TBI	MUD/BM	Limited	18	NA	7761	51	20	Liver, skin, pancreas	+
	2	44/F	AML-CR2	CY/TBI	MUD/BM	Extensive	46	NA	6702	51	9	Liver, skin, pancreas	+
	3	29/M	AML-CR1	CY/TBI	MUD/BM	Limited	16	NA	8678	89	18	Liver, skin	NA
	4	62/M	AML-relapse-2	BU/CY	MRD/BM	Limited	NA	999	6321	96	12	Liver, skin, pancreas	+
	5	42/M	MDS-RAEB-CR-1	BU/CY	MRD/BM	Limited	20	1552	2398	65	6	Liver, skin, pancreas	NA
	6	41/M	CML-CP	BU/CY	MRD/BM	Limited	32	NA	11,159	96	35	Liver, skin, pancreas	+

IOL = iron overload, Tx = transplant, CP = chronic phase, CRI = first complete remission, CR2 = second complete remission, relapse-2 = second relapse, BM = bone marrow, PB = peripheral blood.

⁎ Conditioning regimen [25].

Clinically significant iron overload was defined as evidence of hepatic dysfunction attributed to excess iron deposition with serum ferritin levels of >1000 μg/L [12, 13, 14]. Patients were evaluated for viral hepatitis. Chronic GVHD was classified as extensive or limited [15]. Treatment of cGVHD included a combination of prednisone, cyclosporine, or tacrolimus and mycophenolate sodium (CellCept; Roche Pharmaceuticals, Nutley, NJ) in variable dosage. Phlebotomy was performed by using a venesection kit (Baxter Health Corporation, Deerfield, IL). With each phlebotomy, approximately 350 to 500 mL of whole blood was removed. Phlebotomies were repeated every 1 to 2 weeks until serum ferritin normalized (phlebotomy program). Once normal serum ferritin levels were achieved, maintenance phlebotomy was performed every 3 to 6 months to maintain serum ferritin in the low normal range (10-40 μg/L) [16]. The complete blood count was checked before each phlebotomy, and the procedure was not performed if the hemoglobin was <11 g/dL. Recombinant erythropoietin in the form of either darbepoetin (Aranesp, Amgen Inc., Thousand Oak, CA) 200 μg subcutaneously every 2 weeks or epoetin alfa (Procrit, Ortho Biotech Products, LP, Raritan, NJ) 40 000 IU subcutaneously weekly was used intermittently to facilitate planned phlebotomy. Erythropoietin therapy was initiated for hemoglobin levels ≤11 g/dL and was discontinued at 12 g/dL. The institutional review board approved the study.

Results 

Table 1, Table 2 describe patient characteristics and responses to phlebotomy. The median time to clinical manifestations of iron overload from transplantation was 15 months (range, 6-35 months). The last transfusion preceding the diagnosis of iron overload occurred at a median of 8 months (range, 4-33 months). Lifetime PRBC transfusion data were incomplete in 1 patient; 5 patients received a median of 20 units of PRBC transfusion (range, 16-46 units). Analysis of C282Y and H63D hemochromatosis genes from pretransplantation DNA samples (n = 4) and buccal mucosa (n = 2) revealed no mutations in 4 patients. One patient was heterozygous for both the C282Y and H63D genotypes; another patient was heterozygous for H63D but not for C282Y.

Table 2. Response to Phlebotomy

	Patients	Liver Functions at the Diagnosis of IOL				Liver Functions Following Phlebotomy				Phlebotomy Units†	Time to Liver Enzyme Normalization (M)	Time to Ferritin Normalization (M)	Total Post-Tx Follow-up (M)
		ALT (IU)	AST (IU)	Albumin (gm/L)	Bilirubin (mg/dL)	ALT (IU)	AST (IU)	Albumin (gm/dL)	Bilirubin (mg/dL)
	1	625	406	1.5	4.1	22	31	3.9	0.4	29	12	14	71
	2	457	508	1.6	2.9	26	37	3.6	0.4	27	7	11	59⁎
	3	182	109	3.4	0.6	26	21	4.1	0.5	22	2	8	74
	4	630	253	2.9	1.7	36	46	3.3	0.4	24	7	11	24
	5	323	485	2.6	1.1	36	43	2.9	0.5	22	12	active	17
	6	220	217	2.6	0.6	36	37	3.4	0.3	15	6	active	39

⁎ Died of extensive cGVHD and chronic respiratory failure. † One unit phlebotomy = 300 ml, whole blood.

Apart from abnormal liver function tests, the clinical manifestations of iron overload included bronze skin (n = 6), glucose intolerance (n = 4), and hepatic encephalopathy and ascites (n = 2; patients 1 and 2). Cardiac status, evaluated in 4 patients with 3-dimensional echocardiography, was normal. Liver biopsy in 4 patients revealed hepatic GVHD, along with a marked iron deposition in the Kupffer cells, macrophages, and periportal hepatocytes, with a varied degree of fibrosis. One patient (No. 1) had advanced cirrhosis. In 2 patients (Nos. 3 and 5) in whom liver biopsy was not performed, improvement in organ function with iron-reduction therapy (in the absence of other specific therapy) strongly suggested iron overload. The most severe liver dysfunction beyond day 100 after allo-SCT was observed at a median of 760 days (range, 180-1060 days). The hepatitis profile was negative at the diagnosis of iron overload; 1 patient (No. 3) developed hepatitis B after successful iron depletion. All patients had cGVHD (limited, n = 5; extensive, n = 1). The liver function abnormality persisted despite continued or intensified immunosuppressive therapy in all patients. The median time to successful ferritin normalization with weekly or biweekly phlebotomy was 11 months (range, 8-14 months) in the 4 patients who completed the phlebotomy program (Figure 1). The median time to liver function normalization was 7 months (range, 2-12 months). Baseline median hemoglobin levels before phlebotomy were 14.5 g/dL (12.3-16.5 g/dL), and they decreased to 12 g/dL (10.9-13.8 g/dL) during phlebotomy. The median hemoglobin concentration at ferritin normalization (n = 4) was 14.1 g/dL (13.1-14.9 g/dL).

View Large Image Download to PowerPoint

Figure 1. Time to serum ferritin normalization in 4 patients who completed the phlebotomy program.

Immunosuppressive therapy was discontinued without worsening of liver functions in all 4 patients who completed the phlebotomy program. Two patients continued phlebotomy for 6 and 12 months. At median follow-up of 50 months (range, 17-74 months) from transplantation and 25 months (range, 5-36 months) after successful ferritin normalization, all 4 patients require maintenance phlebotomy every 3 to 6 months to maintain ferritin <40 μg/dL. Figure 2 illustrates ferritin increases in 3 patients after ferritin normalization (1 patient [No. 4] completed the phlebotomy program only 3 months before this writing). One patient died of extensive cGVHD and chronic respiratory failure 19 months after ferritin normalization.

View Large Image Download to PowerPoint

Figure 2. Ferritin elevation after serum ferritin normalization in 4 patients who completed the phlebotomy program.

Discussion 

Apart from hepatocellular, cardiac, and other organ dysfunction [16], iron overload may worsen the natural course of liver GVHD, similar to that of chronic hepatitis, and its response to therapy [17, 18]. Iron overload also increases the risk of opportunistic infections in immunocompromised patients, including Yersinia enterocolitica, Vibrio vulnificus, and Listeria monocytogenes infections, aspergillosis, and mucormycoses [19]. Although acute inflammation [20], vitamin C deficiency [21], and liver disease [12] influence serum ferritin, it provides a simple noninvasive means to assess body iron stores [5, 8, 13, 16]. Hyperferritinemia is a common late effect of hematopoietic transplantation, with a reported incidence of 88% [2]. Most patients, however, do not develop clinical evidence of iron overload. Moreover, serum ferritin declines with time [5, 8, 17, 22], and therapeutic phlebotomy for chemical hyperferritinemia [8, 9] cannot be justified. Also in these reports, the distinction between chemical hyperferritinemia and clinically significant iron overload (excess iron leading to organ dysfunction) is lacking [8, 9]. Confounding factors in the form of viral hepatitis (25%-50% incidence in some series) also interfere with interpretation of both liver function and serum ferritin results [2, 4]. Iron overload has been reported in the recipients of both autologous SCT and allo-SCT, as well as in patients receiving chemotherapy [2, 3, 4, 5, 6, 7, 8, 9]. It is interesting to note that lifetime infusion of <50 to 60 units of PRBCs has led to clinically significant iron overload in most reported patients. Its is speculated that intestinal absorption is increased as a result of epithelial injury related to chemotherapy or GVHD [7, 10, 11].

In our patients, erythropoietin support allowed uninterrupted phlebotomy by preventing iatrogenic anemia and may have mobilized stored iron. The most profound manifestations of iron overload were seen in the first 2 patients and included ascites and grade 2 hepatic encephalopathy. Early recognition in the subsequent 4 patients may have prevented such an advanced presentation. Although steroids and immunosuppression therapy with cyclosporine and tacrolimus can account for hyperglycemia and glucose intolerance in these patients, hyperglycemia persisted despite their tapering or discontinuation. After successful iron depletion, however, all 4 patients had normal blood glucose levels. Hemochromatosis gene analysis was negative for the homozygous state of both the C282Y and H63D genotypes. Because further genetic studies for other mutations were not performed, genetic predisposition in these patients cannot be completely ruled out.

Gradual ferritin elevation after successful iron depletion (Figure 2) is fascinating. In the absence of oral iron supplementation, transfusion, inflammation, or abnormal liver functions, ferritin elevation is likely due to increased gastrointestinal absorption of bioavailable iron in meals, somewhat similar to that in primary hemochromatosis. The gastrointestinal mucosal dysfunction related to cGVHD [10], mucosal injury related to previous chemotherapy treatments [7, 11], and/or alteration in hepcidin regulation [23] may explain the increased iron absorption. Such reaccumulation, along with iron overload with merely 20 units (median) of PRBC transfusion in our patient, may also the indicate potential role of GVHD in iron dysregulation. Chronic liver disease (cGVHD)–related increased gastrointestinal iron absorption in these patients might explain the need for maintenance phlebotomy after successful iron depletion [24].

Prospective well-designed studies are necessary to estimate the prevalence of clinically significant iron overload and to establish guidelines for screening in allo-SCT recipients. Such studies will define the role of routine screening with serum ferritin and the evaluation of hepatic iron load with noninvasive magnetic resonance imaging [25] and will identify a subgroup of patients who may benefit from currently available oral chelating agents [26, 27]. We recommend serum ferritin estimation in all patients with unexplained abnormal liver functions and routine screening 6 to 12 months after transplantation. We conclude that iron overload can manifest as GVHD exacerbation resulting in unnecessary continuation or intensification of immunosuppressive therapy and that maintenance phlebotomy is necessary after initial iron-reduction therapy to maintain normal ferritin values.