Original Paper Acta Haematol 2015;133:310–316 DOI: 10.1159/000365779
Received: May 13, 2014 Accepted after revision: July 4, 2014 Published online: December 2, 2014
Prognostic Value and Clinical Implication of Serum Ferritin Levels following Allogeneic Hematopoietic Cell Transplantation Mika Nakamae Hirohisa Nakamae Shiro Koh Hideo Koh Mitsutaka Nishimoto Yasuhiro Nakashima Takahiko Nakane Asao Hirose Masayuki Hino Hematology, Graduate School of Medicine, Osaka City University, Osaka, Japan
Key Words Hematopoietic cell transplantation · Serum ferritin level · Acute-phase reactants
Abstract Little research has been done on changes in serum ferritin (s-ferritin) levels and clinical implications following allogeneic hematopoietic cell transplantation (HCT). We retrospectively evaluated the correlation of s-ferritin levels after HCT with survival in 203 patients. The s-ferritin level was significantly elevated, with 75% of the patients showing peak levels 90 days after HCT. The level was >10,000 ng/ml in a total of 43% of the patients, a finding that was associated with febrile neutropenia or infection. The s-ferritin level at day 30 and at 1 year after HCT was significantly associated with prognosis. However, this statistically significant relationship was lost after adjusting for acute-phase reactants. We conclude that hyperferritinemia is very common and the degree of influence of a red blood cell transfusion will vary depending on the phase after HCT. A prospective study is needed to determine if iron load in and of itself contributes to a worse prognosis after HCT. © 2014 S. Karger AG, Basel
© 2014 S. Karger AG, Basel 0001–5792/14/1333–0310$39.50/0 E-Mail [email protected] www.karger.com/aha
Introduction
There is a growing body of evidence showing an association between elevated serum ferritin (s-ferritin) levels prior to hematopoietic cell transplantation (HCT) and an adverse outcome in patients with hematological disease [1–13]. In most of the reports, the s-ferritin level before HCT was used as a surrogate marker of the iron burden. However, it is not clear whether iron overload per se contributes to worse survival. Furthermore, s-ferritin is an acute-phase reactant known to increase during acute inflammation, including that caused by infection and/or active hematological disease. It is therefore not an ideal surrogate marker of iron burden. In fact, it might merely represent an indirect marker of a poor prognosis of hematological disease due to its role as an acute-phase reactant that reflects the patient’s inflammatory status in relation to comorbid conditions or advanced disease. A recent report showed that iron overload after HCT had a significant impact on outcome when s-ferritin levels were used as a surrogate maker of iron burden [14]. However, the possibility exists that s-ferritin levels change dynamically as a result of the influence of various conHirohisa Nakamae, MD, PhD Hematology, Graduate School of Medicine Osaka City University 1-4-3 Asahi-machi, Abeno-ku, Osaka 545-8585 (Japan) E-Mail hirohisa @ msic.med.osaka-cu.ac.jp
founding factors including regular red blood cell (RBC) transfusions, infection, hematopoietic recovery, graftversus-host disease (GVHD) and disease progression [15]. We evaluated changes in and the prognostic value of the s-ferritin level after HCT and investigated its clinical implication as a biomarker for outcome.
Patients and Methods We retrospectively reviewed evaluable patients who underwent HCT at our institute between February 2004 and December 2011. The impact of s-ferritin levels after HCT was evaluated in patients for whom both pre- and post-HCT s-ferritin data were available. Due to an ongoing prospective study on haploidentical stem cell transplantation in our institute, patients who had undergone this particular procedure were excluded from this study. Patients with low-risk diseases included those with chronic myelogenous leukemia (chronic phase), refractory anemia and aplastic anemia. Intermediate-risk diseases included chronic myelogenous leukemia (i.e. the accelerated or chronic phase after a blastic phase), acute leukemia or lymphoma (remission), refractory anemia with excess blasts and chronic lymphocytic leukemia. High-risk diseases included chronic myelogenous leukemia (blastic phase), acute leukemia or lymphoma (in relapse), refractory anemia with excess blasts in transformation and myeloma [16]. The levels of s-ferritin measured at least 1 month prior to the start of HCT conditioning after the completion of conventional chemotherapy were used as the ‘pre-HCT s-ferritin value’. Values taken at day 30 ± 5 days, day 60 ± 7 days, day 90 ± 10 days, 6 months ± 1 month and 12 months ± 2 months were used as the s-ferritin values at day 30, day 60, day 90, day 180 and 1 year after HCT, respectively. The context of this study was disseminated to the public by posting a notice at our hospital and on our website in accordance with the ethical guidelines for epidemiological research compiled by the Ministry of Education, Culture, Sports, Science and Technology and the Ministry of Health, Labour and Welfare in Japan. This retrospective study was approved by our Institutional Review Board. Statistical Analysis A two-tailed Pearson correlation test was applied to evaluate the relationship between s-ferritin levels and cumulative RBC transfusion volume. The effect on overall survival (OS) of s-ferritin levels taken preHCT, on days 30, 60, 90, 180 and at 1 year was analyzed using univariate analysis with a Cox proportional hazards model. We performed a test of linearity assumption for s-ferritin values in Cox regression to show that the relationship between s-ferritin and OS was linear. Nonlinear effects of s-ferritin values were evaluated using quadratic, square-root and log transformations. In the study population, patients who did not achieve remission regarding the effects of s-ferritin prior to HCT on transplantation-related mortality (TRM) and relapse were excluded from the analysis. For TRM, relapse was treated as the competing event, and for relapse, any deaths without relapse were treated as the
Serum Ferritin Level after HCT
Table 1. Patient characteristics Characteristics
n
Total number of patients Median age at HCT, years (range) Male/female Disease Acute myeloid leukemia Acute lymphoblastic leukemia Myelodysplastic syndrome Chronic myeloid leukemia Chronic neutrophilic leukemia Non-Hodgkin’s lymphoma Adult T cell leukemia/lymphoma Diffuse large B cell lymphoma NK/T cell lymphoma Peripheral T cell lymphoma, unspecified Anaplastic large-cell lymphoma Follicular lymphoma Mantle-cell lymphoma Other Aplastic anemia Primary myelofibrosis Chronic active EBV infection Disease status Low Intermediate High Stem cell source Bone marrow Peripheral blood Cord blood Donor type Related Unrelated HLA disparity (at A, B, DR antigen) Matched Mismatched CMV seropositivity of recipient Positive Negative Missing GVHD prophylaxis Calcineurin inhibitor alone Calcineurin inhibitor + MTX Calcineurin inhibitor + MMF None Others Conditioning regimen High-dose total-body irradiation-based Busulfan + cyclophosphamide Fludarabine-containing Others Year of transplant 2004 – 2008 2008 – 2011
203 45 (16 – 69) 113/90 81 34 23 8 1 47 20 7 1 2 3 6 1 7 7 1 1 18 116 69 111 39 53 52 151 140 63 173 23 7 48 139 7 2 7 61 32 96 14 116 87
EBV = Epstein-Barr virus; MMF = mycophenolate mofetil; MTX = methotrexate.
Acta Haematol 2015;133:310–316 DOI: 10.1159/000365779
311
competing event. The relationship between s-ferritin prior to HCT and TRM or relapse was evaluated using the proportional hazards model of Fine and Gray for subdistribution of competing risk. In all the models, the hazard ratio (HR) per 1-SD-magnitude of each valuable was calculated for continuous variables. The Mann-Whitney U test was used to compare s-ferritin levels against disease status for acute leukemia. To assess the compounding effects of s-ferritin and various positive and negative acute-phase reactants, the effect of s-ferritin levels pre- and post-HCT was evaluated using a multivariate Cox proportional hazards model after adjusting for statistically significant variables at each point pre- or post-HCT in a univariate Cox proportional hazards model. The acute-phase reactant variables included C-reactive protein (CRP), fibrinogen, serum albumin and haptoglobin levels. A Kruskal-Wallis test was employed to compare s-ferritin levels prior to HCT, on days 30, 60, 90, 180 and at 1 year. Dunn’s multiple comparison test was used as a post hoc test. The alteration in s-ferritin values was analyzed between pre-HCT and 1 year postHCT for the patients (n = 44) for whom values were available (at 1 year post-HCT). To compare the incidence of coexisting active infection or febrile neutropenia, active GVHD and active hematological disease, and the proportion of patients who received systemic steroid treatment in the 3 groups classified by peak s-ferritin level, χ2 tests were used for categorical variables. All p values were two-tailed and considered statistically significant at a value of
Received: May 13, 2014 Accepted after revision: July 4, 2014 Published online: December 2, 2014
Prognostic Value and Clinical Implication of Serum Ferritin Levels following Allogeneic Hematopoietic Cell Transplantation Mika Nakamae Hirohisa Nakamae Shiro Koh Hideo Koh Mitsutaka Nishimoto Yasuhiro Nakashima Takahiko Nakane Asao Hirose Masayuki Hino Hematology, Graduate School of Medicine, Osaka City University, Osaka, Japan
Key Words Hematopoietic cell transplantation · Serum ferritin level · Acute-phase reactants
Abstract Little research has been done on changes in serum ferritin (s-ferritin) levels and clinical implications following allogeneic hematopoietic cell transplantation (HCT). We retrospectively evaluated the correlation of s-ferritin levels after HCT with survival in 203 patients. The s-ferritin level was significantly elevated, with 75% of the patients showing peak levels 90 days after HCT. The level was >10,000 ng/ml in a total of 43% of the patients, a finding that was associated with febrile neutropenia or infection. The s-ferritin level at day 30 and at 1 year after HCT was significantly associated with prognosis. However, this statistically significant relationship was lost after adjusting for acute-phase reactants. We conclude that hyperferritinemia is very common and the degree of influence of a red blood cell transfusion will vary depending on the phase after HCT. A prospective study is needed to determine if iron load in and of itself contributes to a worse prognosis after HCT. © 2014 S. Karger AG, Basel
© 2014 S. Karger AG, Basel 0001–5792/14/1333–0310$39.50/0 E-Mail [email protected] www.karger.com/aha
Introduction
There is a growing body of evidence showing an association between elevated serum ferritin (s-ferritin) levels prior to hematopoietic cell transplantation (HCT) and an adverse outcome in patients with hematological disease [1–13]. In most of the reports, the s-ferritin level before HCT was used as a surrogate marker of the iron burden. However, it is not clear whether iron overload per se contributes to worse survival. Furthermore, s-ferritin is an acute-phase reactant known to increase during acute inflammation, including that caused by infection and/or active hematological disease. It is therefore not an ideal surrogate marker of iron burden. In fact, it might merely represent an indirect marker of a poor prognosis of hematological disease due to its role as an acute-phase reactant that reflects the patient’s inflammatory status in relation to comorbid conditions or advanced disease. A recent report showed that iron overload after HCT had a significant impact on outcome when s-ferritin levels were used as a surrogate maker of iron burden [14]. However, the possibility exists that s-ferritin levels change dynamically as a result of the influence of various conHirohisa Nakamae, MD, PhD Hematology, Graduate School of Medicine Osaka City University 1-4-3 Asahi-machi, Abeno-ku, Osaka 545-8585 (Japan) E-Mail hirohisa @ msic.med.osaka-cu.ac.jp
founding factors including regular red blood cell (RBC) transfusions, infection, hematopoietic recovery, graftversus-host disease (GVHD) and disease progression [15]. We evaluated changes in and the prognostic value of the s-ferritin level after HCT and investigated its clinical implication as a biomarker for outcome.
Patients and Methods We retrospectively reviewed evaluable patients who underwent HCT at our institute between February 2004 and December 2011. The impact of s-ferritin levels after HCT was evaluated in patients for whom both pre- and post-HCT s-ferritin data were available. Due to an ongoing prospective study on haploidentical stem cell transplantation in our institute, patients who had undergone this particular procedure were excluded from this study. Patients with low-risk diseases included those with chronic myelogenous leukemia (chronic phase), refractory anemia and aplastic anemia. Intermediate-risk diseases included chronic myelogenous leukemia (i.e. the accelerated or chronic phase after a blastic phase), acute leukemia or lymphoma (remission), refractory anemia with excess blasts and chronic lymphocytic leukemia. High-risk diseases included chronic myelogenous leukemia (blastic phase), acute leukemia or lymphoma (in relapse), refractory anemia with excess blasts in transformation and myeloma [16]. The levels of s-ferritin measured at least 1 month prior to the start of HCT conditioning after the completion of conventional chemotherapy were used as the ‘pre-HCT s-ferritin value’. Values taken at day 30 ± 5 days, day 60 ± 7 days, day 90 ± 10 days, 6 months ± 1 month and 12 months ± 2 months were used as the s-ferritin values at day 30, day 60, day 90, day 180 and 1 year after HCT, respectively. The context of this study was disseminated to the public by posting a notice at our hospital and on our website in accordance with the ethical guidelines for epidemiological research compiled by the Ministry of Education, Culture, Sports, Science and Technology and the Ministry of Health, Labour and Welfare in Japan. This retrospective study was approved by our Institutional Review Board. Statistical Analysis A two-tailed Pearson correlation test was applied to evaluate the relationship between s-ferritin levels and cumulative RBC transfusion volume. The effect on overall survival (OS) of s-ferritin levels taken preHCT, on days 30, 60, 90, 180 and at 1 year was analyzed using univariate analysis with a Cox proportional hazards model. We performed a test of linearity assumption for s-ferritin values in Cox regression to show that the relationship between s-ferritin and OS was linear. Nonlinear effects of s-ferritin values were evaluated using quadratic, square-root and log transformations. In the study population, patients who did not achieve remission regarding the effects of s-ferritin prior to HCT on transplantation-related mortality (TRM) and relapse were excluded from the analysis. For TRM, relapse was treated as the competing event, and for relapse, any deaths without relapse were treated as the
Serum Ferritin Level after HCT
Table 1. Patient characteristics Characteristics
n
Total number of patients Median age at HCT, years (range) Male/female Disease Acute myeloid leukemia Acute lymphoblastic leukemia Myelodysplastic syndrome Chronic myeloid leukemia Chronic neutrophilic leukemia Non-Hodgkin’s lymphoma Adult T cell leukemia/lymphoma Diffuse large B cell lymphoma NK/T cell lymphoma Peripheral T cell lymphoma, unspecified Anaplastic large-cell lymphoma Follicular lymphoma Mantle-cell lymphoma Other Aplastic anemia Primary myelofibrosis Chronic active EBV infection Disease status Low Intermediate High Stem cell source Bone marrow Peripheral blood Cord blood Donor type Related Unrelated HLA disparity (at A, B, DR antigen) Matched Mismatched CMV seropositivity of recipient Positive Negative Missing GVHD prophylaxis Calcineurin inhibitor alone Calcineurin inhibitor + MTX Calcineurin inhibitor + MMF None Others Conditioning regimen High-dose total-body irradiation-based Busulfan + cyclophosphamide Fludarabine-containing Others Year of transplant 2004 – 2008 2008 – 2011
203 45 (16 – 69) 113/90 81 34 23 8 1 47 20 7 1 2 3 6 1 7 7 1 1 18 116 69 111 39 53 52 151 140 63 173 23 7 48 139 7 2 7 61 32 96 14 116 87
EBV = Epstein-Barr virus; MMF = mycophenolate mofetil; MTX = methotrexate.
Acta Haematol 2015;133:310–316 DOI: 10.1159/000365779
311
competing event. The relationship between s-ferritin prior to HCT and TRM or relapse was evaluated using the proportional hazards model of Fine and Gray for subdistribution of competing risk. In all the models, the hazard ratio (HR) per 1-SD-magnitude of each valuable was calculated for continuous variables. The Mann-Whitney U test was used to compare s-ferritin levels against disease status for acute leukemia. To assess the compounding effects of s-ferritin and various positive and negative acute-phase reactants, the effect of s-ferritin levels pre- and post-HCT was evaluated using a multivariate Cox proportional hazards model after adjusting for statistically significant variables at each point pre- or post-HCT in a univariate Cox proportional hazards model. The acute-phase reactant variables included C-reactive protein (CRP), fibrinogen, serum albumin and haptoglobin levels. A Kruskal-Wallis test was employed to compare s-ferritin levels prior to HCT, on days 30, 60, 90, 180 and at 1 year. Dunn’s multiple comparison test was used as a post hoc test. The alteration in s-ferritin values was analyzed between pre-HCT and 1 year postHCT for the patients (n = 44) for whom values were available (at 1 year post-HCT). To compare the incidence of coexisting active infection or febrile neutropenia, active GVHD and active hematological disease, and the proportion of patients who received systemic steroid treatment in the 3 groups classified by peak s-ferritin level, χ2 tests were used for categorical variables. All p values were two-tailed and considered statistically significant at a value of
