http://www.kidney-international.org
clinical trial
& 2014 International Society of Nephrology
see commentary on page 676
Association between hemoglobin variability, serum ferritin levels, and adverse events/mortality in maintenance hemodialysis patients Takahiro Kuragano1, Osamu Matsumura2, Akihiko Matsuda3, Taiga Hara4, Hideyasu Kiyomoto5, Toshiaki Murata6, Kenichiro Kitamura7, Shouichi Fujimoto8, Hiroki Hase9, Nobuhiko Joki9, Atushi Fukatsu10, Toru Inoue11, Ikuhiro Itakura12 and Takeshi Nakanishi1 1
Division of Kidney and Dialysis, Department of Internal Medicine, Hyogo College of Medicine, Nishinomiya, Japan; 2Department of Internal Medicine, Musashi Ranzan Hospital, Ranzan, Japan; 3Department of Nephrology and Blood Purification, Saitama Medical Center, Saitama Medical University, Saitama, Japan; 4Division of Nephrology and Dialysis, Department of Cardiorenal and Cerebrovascular Medicine, Faculty of Medicine, Kagawa University, Kagawa, Japan; 5Division of Integrated Nephrology and Telemedicine, Tohoku Medical Megabank Organization, Tohoku University, Sendai, Japan; 6Division of Nephrology and Rheumatology, Department of Internal Medicine, Faculty of Medicine, Fukuoka University, Fukuoka, Japan; 7Department of Nephrology, Kumamoto University, Graduate School of Medical Sciences, Kumamoto, Japan; 8Department of Hemovascular Medicine and Artificial Organs, Faculty of Medicine, University of Miyazaki, Miyazaki, Japan; 9Department of Nephrology, Toho University, Ohashi Medical Center, Tokyo, Japan; 10Department of Internal Medicine, Yachiyo Hospital, Aichi, Japan; 11Department of Internal Medicine, Higashikouri Hospital, Higashikouri, Japan and 12Department of Internal Medicine, Itakura Clinic, Saitama, Japan
In recent times, therapy for renal anemia has changed dramatically in that iron administration has increased and doses of erythropoiesis-stimulating agents (ESAs) have decreased. Here we used a prospective, observational, multicenter design and measured the serum ferritin and hemoglobin levels every 3 months for 2 years in 1086 patients on maintenance hemodialysis therapy. The associations of adverse events with fluctuations in ferritin and hemoglobin levels and ESA and iron doses were measured using a Cox proportional hazards model for time-dependent variables. The risks of cerebrovascular and cardiovascular disease (CCVD), infection, and hospitalization were higher among patients who failed to maintain a target-range hemoglobin level and who exhibited high-amplitude fluctuations in hemoglobin compared with patients who maintained a target-range hemoglobin level. Patients with a higher compared with a lower ferritin level had an elevated risk of CCVD and infectious disease. Moreover, the risk of death was significantly higher among patients with high-amplitude ferritin fluctuations compared with those with a low ferritin level. The risks of CCVD, infection, and hospitalization were significantly higher among patients who were treated with high weekly doses of intravenous iron compared with no intravenous iron. Thus, there is a high risk of death and/or adverse events in patients Correspondence: Takahiro Kuragano, Hyogo College of Medicine, Division of Nephrology and Dialysis, Department of Internal Medicine, 1-1, Mukogawa-cho, Nishinomiya 663-8501, Japan. E-mail: [email protected] Received 7 June 2013; revised 3 March 2014; accepted 6 March 2014; published online 23 April 2014 Kidney International (2014) 86, 845–854
with hemoglobin levels outside the target range, in those with high-amplitude hemoglobin fluctuations, in those with consistently high serum ferritin levels, and in those with high-amplitude ferritin fluctuations. Kidney International (2014) 86, 845–854; doi:10.1038/ki.2014.114; published online 23 April 2014 KEYWORDS: anemia; cardiovascular events; hemodialysis
Recent large-scale clinical studies have demonstrated that maintenance of a high hemoglobin target level is not necessarily associated with improved survival in patients undergoing maintenance hemodialysis.1 Furthermore, patients who require a high dose of an erythropoiesis-stimulating agent (ESA) to maintain the target hemoglobin level or who are hyporesponsive to an ESA generally have a poor prognosis.2 Factors that have been reported to affect ESA responsiveness include iron deficiency, chronic inflammation, secondary hyperparathyroidism, malnutrition, and inadequate-dose dialysis.3,4 Of the various factors that affect patient responsiveness to an ESA, iron deficiency is thought to be the most prevalent cause of ESA hyporesponsiveness.5 Intravenous iron administration has been shown to improve ESA responsiveness, leading to an increased hemoglobin level and permitting the use of a decreased ESA dose.6,7 In accordance with these observations, the 2012 Kidney Disease Improving Global Outcome (KDIGO) guidelines suggest a trial on intravenous iron in patients with anemia if an increase in hemoglobin concentration is desired, the transferrin saturation (TSAT) is p30%, and serum ferritin is p500 ng/ml (evidence level 2C).8 Another reason for the current decrease in ESA dose is related to reimbursement policies in many countries that 845
clinical trial
bundle dialysis services; at the same time, the frequency of iron use, particularly intravenous iron, is increasing.9,10 Indeed, an increase in iron administration after the introduction of a bundled reimbursement policy was clearly shown in the US Dialysis Outcomes and Practice Patterns Study (DOPPS) report. This study reported that the percentage of patients prescribed intravenous iron increased from 57% in August 2010 to 78% in June 2011, with an increase in the mean serum ferritin level from 556 to 650 ng/ml over the same time period.11 Iron administration for renal anemia treatment is likely to increase in the future. Despite these trends, the frequency and dose of iron supplementation among maintenance hemodialysis patients in Japan is far lower compared with Western countries.12 In the 2007 Japan-DOPPS (J-DOPPS) study, the mean ferritin level for the maintenance hemodialysis patients studied was 244 ng/ml, and the percentage of patients with a ferritin levelo100 ng/ml was 41.3%. Notably, 47.2% of the patients with hemoglobinX11 g/dl had a ferritin levelo100 ng/ml, and only 40.6% received intravenous iron. These observations suggest that a substantial percentage of patients could achieve hemoglobinX11.0 g/dl without iron supplementation.9 Extensive iron administration can result in iron overload, with excessive deposition in the liver, spleen, and heart.13 Furthermore, iron overload is known to promote endothelial dysfunction, cardiovascular disease, and immune dysfunction, which are the leading causes of premature mortality in chronic kidney disease (CKD) and end-stage kidney disease patients receiving maintenance hemodialysis.14 Therefore, exercising caution is imperative when using intravenous iron preparations in these populations. Moreover, no prospective safety trials on intravenous iron have been adequately performed to examine infection rate, cardiovascular events, or survival. In this study, we evaluated the association of renal anemia treatments, specifically doses of ESA and iron, with adverse events and prognosis in maintenance hemodialysis patients in a subanalysis of the Prospective Study of Treatment for Renal Anemia on Prognosis in hemodialysis patients (TRAP) study. The distinguishing feature of this study is that an increase in serum ferritin following iron administration may affect the prognosis of patients receiving maintenance hemodialysis whose ferritin level is in the lower range of the KDIGO guidelines. RESULTS Patients
A total of 1095 patients (mean age, 61.8 years; 60.3% male; 33.7% with diabetes; mean duration of dialysis therapy, 106 months) were enrolled in the study (Table 1). The patients’ mean hemoglobin and ferritin levels were 10.6±1.0 g/dl and 125.4±147.0 ng/ml, respectively. The mean TSAT was 26.7±11.7%, and the mean ESA dose was 3212±2107 IU/ week. Of the patients, 2.6% were treated with oral iron, whereas 20.8% were treated with intravenous iron. There was no significant difference in mean hemoglobin level among the groups (Table 2). 846
T Kuragano et al.: Anemia treatment and adverse events
Incidence of adverse events
A total of 193 patients who received darbepoetin during the study period were excluded depending on the protocol. Seventy-nine CCVD cases, 368 cases of infection, 324 hospitalizations, and 36 deaths occurred during the study period. Impact of hemoglobin changes on the incidence of adverse events
The risk for CCVD was significantly higher in the lowhemoglobin group than in the target-hemoglobin group (Hazards ratio (HR) 3.78, P ¼ 0.003). The risks for CCVD (HR 4.23, P ¼ 0.024), infection (HR 1.96, P ¼ 0.005), and hospitalization (HR 2.29, P ¼ 0.001) were significantly higher in the LAL-hemoglobin (low-amplitude fluctuation with low hemoglobin level) group than in the target-hemoglobin group. Furthermore, the risks for CCVD (HR 3.94, P ¼ 0.033), infection (HR 1.73, P ¼ 0.027), and hospitalization (HR 2.01, P ¼ 0.012) were significantly higher in the HA-hemoglobin (high-amplitude fluctuation across the target hemoglobin range) group than in the target-hemoglobin group (Figure 1). Impact of ESA dose on hemoglobin level and the incidence of adverse events
The hemoglobin levels at 0, 3, 6, 9, 12, 21, and 24 months were significantly (Po0.005) higher in the low-dose ESA group than in the high-dose ESA group (Figure 2). The risks for infection (HR 1.48, P ¼ 0.015) and hospitalization (HR 1.80, P ¼ 0.001) were significantly higher in the high-dose ESA group than in the low-dose ESA group (Figure 3). Impact of ferritin changes on the risk for adverse events
The risks of hospitalization (HR 1.59, P ¼ 0.013) and death (HR 6.18, P ¼ 0.002) were significantly higher in the L-H (change from a low to a high ferritin level) ferritin group than in the L-ferritin group. The risks for CCVD (HR 2.22, P ¼ 0.033) and infection (HR 1.76, Po0.001) were significantly higher in the high-ferritin group than in the lowferritin group. The risk of death was significantly (HR 3.75, P ¼ 0.029) higher in the HA-ferritin group than in the lowferritin group (Figure 4). Impact of iron dose on the incidence of adverse events
The risks for CCVD (HR 6.02, P ¼ 0.038), infection (HR 5.22, P ¼ 0.001), and hospitalization (HR 2.77, P ¼ 0.015) were significantly higher in the patients treated with highdose intravenous iron than in those not treated with iron. The risk for infection (HR 1.78, P ¼ 0.037) was significantly higher in the patients treated with low-dose intravenous iron than in those not treated with iron (Figure 5). Propensity-score matching
Because the baseline hemoglobin (less than or greater than 10 g/dl) and ferritin (less than or greater than 100 ng/ml) levels are dependent variables, a propensity score was created using a logistic regression model, adjusting for sex, age, Kidney International (2014) 86, 845–854
clinical trial
T Kuragano et al.: Anemia treatment and adverse events
Table 1 | Characteristics of patients
All patients A area B area C area D area E area F area G area H area I area
Number of patients
Age, years
BMI, kg/m2
Sex (% of male)
Etiology (% of DM)
Time on hemodialysis (month)
albumin (g/dl)
hCRP (mg/dl)
1095 166 122 106 50 126 104 44 273 103
61.8±9.9 60.9±10.3 62.4±10.9 63.7±8.8 63.7±8.1 59.6±11.0 62.4±9.6 59.9±9.9 61.6±8.9 63.2±11.1
21.7±12.3 21.0±3.7 21.5±3.4 22.1±5.9 20.5±2.9 20.9±3.0 24.8±3.8 21.6±3.6 21.6±3.3 21.6±3.0
60.3 52.4 66.4 65.1 46 65.7 65.4 65.9 63.7 47.6
33.7 33.7 36.1 38.7 26 30.2 25 38.6 37.7 29.1
106±101 86±67 97±107 96±90 1544±95 125±108 116±165 97±86 108±117 113±89
3.7±0.31 3.9±0.32 3.8±0.48 3.7±0.24 3.6±0.25 3.8±0.28 3.6±0.31 3.7±0.26 3.6±0.31 3.8±0.35
0.29±1.23 0.13±0.16 0.36±1.03 0.28±0.79 0.29±0.65 0.27±1.28 0.48±3.21 0.11±0.14 0.42±2.91 0.28±0.92
Abbreviations: BMI, body mass index; DM, diabetes mellitus; hCRP, high-sensitive C-reactive protein. Data are the mean±s.d.
Table 2 | Anemia-related factors % of iron administration
All patients A area B area C area D area E area F area G area H area I area
Number of patients
Hb (g/dl)
Ferritin (ng/ml)
TSAT (%)
Dose of ESA (IU/week)
ESA/Hb
Oral
IV
1095 166 122 106 50 126 104 44 274 103
10.6±1.0 10.2±0.9 10.6±1.1 10.6±0.9 10.3±1.0 10.6±1.1 10.4±1.1 11.0±0.9 10.8±0.9 10.7±0.7
125.4±147.0 145.7±119.5 139.5±151.2 169.7±94.9 139.2±263.7 121.9±182.0 155.7±164.6 55.64±76.0 99.4±132.1 97.5±111.3
26.7±11.7 28.0±11.6 27.9±14.2 28.3±9.4 26.1±13.5 25.7±12.7 26.8±13.0 26.9±9.2 26.3±11.5 23.9±8.3
3212±2107 3569±2477 3304±2366 3170±1932 2190±1468 3590±2394 2676±1996 3460±1651 3055±1737 3478±2165
308±219 378±256 292±255 312±216 215±150 336±244 258±213 313±143 289±174 327±216
2.6 2.4 1.6 0 0 2.4 0 31.2 1.5 1
20.8 30.7 31.9 13.2 8 30.2 14.4 4.5 21.3 4.9
Abbreviations: ESA, erythropoiesis-stimulating agent; Hb, hemoglobin; IV, intravenous; TSAT, transferrin saturation. Data are the mean±s.d.
Event
Cerebro-cardiovascular disease
Infection disease
Hospitalization
Death
All events
Hb fluctuation Target Low LAL High LAH HA Low LAL High LAH HA Low LAL High LAH HA Low LAL High LAH HA Low LAL High LAH HA
Hazard ratio (95% CI) 1 3.78 (1.12–12.8) 4.23 (1.21–14.76) 0.74 (0.08–7.05) 1.6 (0.45–5.71) 3.94 (1.12–13.87) 1.68 (0.95–2.96) 1.96 (1.22–3.16) 1.13 (0.63–2.02) 1.09 (0.68–1.75) 1.73 (1.06–2.82) 1.84 (1.00–3.37) 2.29 (1.38–3.81) 0.6 (0.25–1.45) 0.81 (0.47–1.41) 2.01 (1.16–3.46) 1.19 (0.19–7.29) 1.6 (0.44–5.80) NA NA 0.63 (0.15–2.65) 1.84 (0.55–6.10) 1.44 (0.93–2.22) 1.55 (1.10–2.18) 0.95 (0.60–1.50) 0.88 (0.62–1.25) 1.53 (1.07–2.18) 0.01 0.1
P-value 0.033 0.024 0.797 0.473 0.033 0.072 0.005 0.676 0.731 0.027 0.05 0.001 0.257 0.452 0.012 0.855 0.477 NA 0.525 0.319 0.1 0.012 0.832 0.469 0.018 1
10
100
Figure 1 | Impact of hemoglobin fluctuation pattern on the risk for adverse events and mortality. HA, high-amplitude fluctuation; High, high hemoglobin level; LAH, low-amplitude fluctuation with high hemoglobin level; LAL, low-amplitude fluctuation with low hemoglobin level; Low, low hemoglobin level; NA, not available; Target, target hemoglobin level. Kidney International (2014) 86, 845–854
847
clinical trial
T Kuragano et al.: Anemia treatment and adverse events
etiology, and albumin, C-reactive protein (CRP), and intact parathyroid hormone levels. According to the matching caliper (s.d. ¼ 0.3 and 1:1), matched pairs of 355 (710 patients) and 201 (402 patients) estimate the variation pattern for ferritin and hemoglobin. We analyzed the data for each matched subject pair using a time-dependent hazard model. Compared with the low-hemoglobin group, the risks for hospitalization (P ¼ 0.039, P ¼ 0.010) and all events (P ¼ 0.013, P ¼ 0.001) were significantly lower for the target-hemoglobin or LAH-hemoglobin (low-amplitude fluctuation with high hemoglobin level) groups. Compared with the low-ferritin group, the L-H ferritin group showed a significantly (P ¼ 0.035) higher risk for hospitalization (P ¼ 0.035) and death (P ¼ 0.006). In addition, the HA ferritin group showed significantly higher risks for hospitalization (P ¼ 0.043) and all events (P ¼ 0.008). Although matching resulted in a decreased number of patients analyzed, these results also indicated that patients with low hemoglobin and high ferritin levels could have a higher risk for adverse events. DISCUSSION
In this study, we demonstrated that maintenance hemodialysis patients with a stable target hemoglobin level had 11.5
Hb (g/dl)
11
*
*
*
*
0M
3M
6M
9M
*
*
*
10.5 10 9.5 9
12M 15M 18M 21M 24M
Figure 2 | Comparison of hemoglobin level in patients treated with a high erythropoiesis-stimulating agent (ESA) dose with those treated with a low ESA dose. The patients in this study were treated with epoetin-a or epoetin-b. K: Low-dose ESA (o4500 IU/ week) group; ’: high-dose ESA (X4,500 IU/week) group. The error bars indicate s.d.s. *Po0.05.
Event
Dose of ESA
Low dose
Hazard ratio
less risk for adverse events compared with maintenance hemodialysis patients without a stable hemoglobin level. We also found that the risks for infection and hospitalization were significantly higher in patients who were treated with a higher ESA dose. Furthermore, the risks for CCVD, infection, hospitalization, and death were significantly higher in patients with an elevated serum ferritin and in those with high-amplitude fluctuations in ferritin level. Moreover, the risks for CCVD, infection, and hospitalization were significantly higher in patients who were treated with high weekly doses of intravenous iron. Comparison of the mean hemoglobin, ferritin, and TSAT levels and iron doses in the study with those from the DOPPS report
According to the 2011 DOPPS report,9 the global mean hemoglobin level is 11.5 g/dl, compared with the mean level of 10.3 g/dl in Japan. The mean hemoglobin level at the start of the present study was 10.6±1.0 g/dl, which is lower than the global mean value but higher than the mean value in the Japanese population reported in J-DOPPS. The Japanese Society of Dialysis Therapy (JSDT) suggested that the difference in hemoglobin levels between Japan and other countries might be due to the timing of blood collection (the beginning of the week in Japan was the middle of the week in other countries), the method used for blood collection, and the patient’s position (supine position in Japan, sitting position in the other countries) during blood collection.15 The sampling time could affect the observed rate of weight gain. Furthermore, the use of the supine position for blood collection may decrease the apparent hemoglobin level. On the basis of these reports, we presumed that there was no significant difference in the hemoglobin levels between Japan and other countries.16 The 2008 JSDT guidelines for treating anemia in CKD recommend that iron be administered to patients with TSATp20% and serum ferritinp100 ng/ml. Furthermore, the recommended frequency of administration is up to once a week for 3 months or a total of 13 times, depending on the patient’s hemoglobin level.17 The 2008 JSDT guidelines for treating anemia in CKD also recommended that a minimal amount of iron be given to CKD patients with
(95% CI)
P-value
1
Cerebro-cardiovascular disease
High dose
1.59 (0.88–2.85)
0.122
Infectious disease
High dose
1.48 (1.08–2.02)
0.015
Hospitalization
High dose
1.8 (1.33–2.43)
clinical trial
& 2014 International Society of Nephrology
see commentary on page 676
Association between hemoglobin variability, serum ferritin levels, and adverse events/mortality in maintenance hemodialysis patients Takahiro Kuragano1, Osamu Matsumura2, Akihiko Matsuda3, Taiga Hara4, Hideyasu Kiyomoto5, Toshiaki Murata6, Kenichiro Kitamura7, Shouichi Fujimoto8, Hiroki Hase9, Nobuhiko Joki9, Atushi Fukatsu10, Toru Inoue11, Ikuhiro Itakura12 and Takeshi Nakanishi1 1
Division of Kidney and Dialysis, Department of Internal Medicine, Hyogo College of Medicine, Nishinomiya, Japan; 2Department of Internal Medicine, Musashi Ranzan Hospital, Ranzan, Japan; 3Department of Nephrology and Blood Purification, Saitama Medical Center, Saitama Medical University, Saitama, Japan; 4Division of Nephrology and Dialysis, Department of Cardiorenal and Cerebrovascular Medicine, Faculty of Medicine, Kagawa University, Kagawa, Japan; 5Division of Integrated Nephrology and Telemedicine, Tohoku Medical Megabank Organization, Tohoku University, Sendai, Japan; 6Division of Nephrology and Rheumatology, Department of Internal Medicine, Faculty of Medicine, Fukuoka University, Fukuoka, Japan; 7Department of Nephrology, Kumamoto University, Graduate School of Medical Sciences, Kumamoto, Japan; 8Department of Hemovascular Medicine and Artificial Organs, Faculty of Medicine, University of Miyazaki, Miyazaki, Japan; 9Department of Nephrology, Toho University, Ohashi Medical Center, Tokyo, Japan; 10Department of Internal Medicine, Yachiyo Hospital, Aichi, Japan; 11Department of Internal Medicine, Higashikouri Hospital, Higashikouri, Japan and 12Department of Internal Medicine, Itakura Clinic, Saitama, Japan
In recent times, therapy for renal anemia has changed dramatically in that iron administration has increased and doses of erythropoiesis-stimulating agents (ESAs) have decreased. Here we used a prospective, observational, multicenter design and measured the serum ferritin and hemoglobin levels every 3 months for 2 years in 1086 patients on maintenance hemodialysis therapy. The associations of adverse events with fluctuations in ferritin and hemoglobin levels and ESA and iron doses were measured using a Cox proportional hazards model for time-dependent variables. The risks of cerebrovascular and cardiovascular disease (CCVD), infection, and hospitalization were higher among patients who failed to maintain a target-range hemoglobin level and who exhibited high-amplitude fluctuations in hemoglobin compared with patients who maintained a target-range hemoglobin level. Patients with a higher compared with a lower ferritin level had an elevated risk of CCVD and infectious disease. Moreover, the risk of death was significantly higher among patients with high-amplitude ferritin fluctuations compared with those with a low ferritin level. The risks of CCVD, infection, and hospitalization were significantly higher among patients who were treated with high weekly doses of intravenous iron compared with no intravenous iron. Thus, there is a high risk of death and/or adverse events in patients Correspondence: Takahiro Kuragano, Hyogo College of Medicine, Division of Nephrology and Dialysis, Department of Internal Medicine, 1-1, Mukogawa-cho, Nishinomiya 663-8501, Japan. E-mail: [email protected] Received 7 June 2013; revised 3 March 2014; accepted 6 March 2014; published online 23 April 2014 Kidney International (2014) 86, 845–854
with hemoglobin levels outside the target range, in those with high-amplitude hemoglobin fluctuations, in those with consistently high serum ferritin levels, and in those with high-amplitude ferritin fluctuations. Kidney International (2014) 86, 845–854; doi:10.1038/ki.2014.114; published online 23 April 2014 KEYWORDS: anemia; cardiovascular events; hemodialysis
Recent large-scale clinical studies have demonstrated that maintenance of a high hemoglobin target level is not necessarily associated with improved survival in patients undergoing maintenance hemodialysis.1 Furthermore, patients who require a high dose of an erythropoiesis-stimulating agent (ESA) to maintain the target hemoglobin level or who are hyporesponsive to an ESA generally have a poor prognosis.2 Factors that have been reported to affect ESA responsiveness include iron deficiency, chronic inflammation, secondary hyperparathyroidism, malnutrition, and inadequate-dose dialysis.3,4 Of the various factors that affect patient responsiveness to an ESA, iron deficiency is thought to be the most prevalent cause of ESA hyporesponsiveness.5 Intravenous iron administration has been shown to improve ESA responsiveness, leading to an increased hemoglobin level and permitting the use of a decreased ESA dose.6,7 In accordance with these observations, the 2012 Kidney Disease Improving Global Outcome (KDIGO) guidelines suggest a trial on intravenous iron in patients with anemia if an increase in hemoglobin concentration is desired, the transferrin saturation (TSAT) is p30%, and serum ferritin is p500 ng/ml (evidence level 2C).8 Another reason for the current decrease in ESA dose is related to reimbursement policies in many countries that 845
clinical trial
bundle dialysis services; at the same time, the frequency of iron use, particularly intravenous iron, is increasing.9,10 Indeed, an increase in iron administration after the introduction of a bundled reimbursement policy was clearly shown in the US Dialysis Outcomes and Practice Patterns Study (DOPPS) report. This study reported that the percentage of patients prescribed intravenous iron increased from 57% in August 2010 to 78% in June 2011, with an increase in the mean serum ferritin level from 556 to 650 ng/ml over the same time period.11 Iron administration for renal anemia treatment is likely to increase in the future. Despite these trends, the frequency and dose of iron supplementation among maintenance hemodialysis patients in Japan is far lower compared with Western countries.12 In the 2007 Japan-DOPPS (J-DOPPS) study, the mean ferritin level for the maintenance hemodialysis patients studied was 244 ng/ml, and the percentage of patients with a ferritin levelo100 ng/ml was 41.3%. Notably, 47.2% of the patients with hemoglobinX11 g/dl had a ferritin levelo100 ng/ml, and only 40.6% received intravenous iron. These observations suggest that a substantial percentage of patients could achieve hemoglobinX11.0 g/dl without iron supplementation.9 Extensive iron administration can result in iron overload, with excessive deposition in the liver, spleen, and heart.13 Furthermore, iron overload is known to promote endothelial dysfunction, cardiovascular disease, and immune dysfunction, which are the leading causes of premature mortality in chronic kidney disease (CKD) and end-stage kidney disease patients receiving maintenance hemodialysis.14 Therefore, exercising caution is imperative when using intravenous iron preparations in these populations. Moreover, no prospective safety trials on intravenous iron have been adequately performed to examine infection rate, cardiovascular events, or survival. In this study, we evaluated the association of renal anemia treatments, specifically doses of ESA and iron, with adverse events and prognosis in maintenance hemodialysis patients in a subanalysis of the Prospective Study of Treatment for Renal Anemia on Prognosis in hemodialysis patients (TRAP) study. The distinguishing feature of this study is that an increase in serum ferritin following iron administration may affect the prognosis of patients receiving maintenance hemodialysis whose ferritin level is in the lower range of the KDIGO guidelines. RESULTS Patients
A total of 1095 patients (mean age, 61.8 years; 60.3% male; 33.7% with diabetes; mean duration of dialysis therapy, 106 months) were enrolled in the study (Table 1). The patients’ mean hemoglobin and ferritin levels were 10.6±1.0 g/dl and 125.4±147.0 ng/ml, respectively. The mean TSAT was 26.7±11.7%, and the mean ESA dose was 3212±2107 IU/ week. Of the patients, 2.6% were treated with oral iron, whereas 20.8% were treated with intravenous iron. There was no significant difference in mean hemoglobin level among the groups (Table 2). 846
T Kuragano et al.: Anemia treatment and adverse events
Incidence of adverse events
A total of 193 patients who received darbepoetin during the study period were excluded depending on the protocol. Seventy-nine CCVD cases, 368 cases of infection, 324 hospitalizations, and 36 deaths occurred during the study period. Impact of hemoglobin changes on the incidence of adverse events
The risk for CCVD was significantly higher in the lowhemoglobin group than in the target-hemoglobin group (Hazards ratio (HR) 3.78, P ¼ 0.003). The risks for CCVD (HR 4.23, P ¼ 0.024), infection (HR 1.96, P ¼ 0.005), and hospitalization (HR 2.29, P ¼ 0.001) were significantly higher in the LAL-hemoglobin (low-amplitude fluctuation with low hemoglobin level) group than in the target-hemoglobin group. Furthermore, the risks for CCVD (HR 3.94, P ¼ 0.033), infection (HR 1.73, P ¼ 0.027), and hospitalization (HR 2.01, P ¼ 0.012) were significantly higher in the HA-hemoglobin (high-amplitude fluctuation across the target hemoglobin range) group than in the target-hemoglobin group (Figure 1). Impact of ESA dose on hemoglobin level and the incidence of adverse events
The hemoglobin levels at 0, 3, 6, 9, 12, 21, and 24 months were significantly (Po0.005) higher in the low-dose ESA group than in the high-dose ESA group (Figure 2). The risks for infection (HR 1.48, P ¼ 0.015) and hospitalization (HR 1.80, P ¼ 0.001) were significantly higher in the high-dose ESA group than in the low-dose ESA group (Figure 3). Impact of ferritin changes on the risk for adverse events
The risks of hospitalization (HR 1.59, P ¼ 0.013) and death (HR 6.18, P ¼ 0.002) were significantly higher in the L-H (change from a low to a high ferritin level) ferritin group than in the L-ferritin group. The risks for CCVD (HR 2.22, P ¼ 0.033) and infection (HR 1.76, Po0.001) were significantly higher in the high-ferritin group than in the lowferritin group. The risk of death was significantly (HR 3.75, P ¼ 0.029) higher in the HA-ferritin group than in the lowferritin group (Figure 4). Impact of iron dose on the incidence of adverse events
The risks for CCVD (HR 6.02, P ¼ 0.038), infection (HR 5.22, P ¼ 0.001), and hospitalization (HR 2.77, P ¼ 0.015) were significantly higher in the patients treated with highdose intravenous iron than in those not treated with iron. The risk for infection (HR 1.78, P ¼ 0.037) was significantly higher in the patients treated with low-dose intravenous iron than in those not treated with iron (Figure 5). Propensity-score matching
Because the baseline hemoglobin (less than or greater than 10 g/dl) and ferritin (less than or greater than 100 ng/ml) levels are dependent variables, a propensity score was created using a logistic regression model, adjusting for sex, age, Kidney International (2014) 86, 845–854
clinical trial
T Kuragano et al.: Anemia treatment and adverse events
Table 1 | Characteristics of patients
All patients A area B area C area D area E area F area G area H area I area
Number of patients
Age, years
BMI, kg/m2
Sex (% of male)
Etiology (% of DM)
Time on hemodialysis (month)
albumin (g/dl)
hCRP (mg/dl)
1095 166 122 106 50 126 104 44 273 103
61.8±9.9 60.9±10.3 62.4±10.9 63.7±8.8 63.7±8.1 59.6±11.0 62.4±9.6 59.9±9.9 61.6±8.9 63.2±11.1
21.7±12.3 21.0±3.7 21.5±3.4 22.1±5.9 20.5±2.9 20.9±3.0 24.8±3.8 21.6±3.6 21.6±3.3 21.6±3.0
60.3 52.4 66.4 65.1 46 65.7 65.4 65.9 63.7 47.6
33.7 33.7 36.1 38.7 26 30.2 25 38.6 37.7 29.1
106±101 86±67 97±107 96±90 1544±95 125±108 116±165 97±86 108±117 113±89
3.7±0.31 3.9±0.32 3.8±0.48 3.7±0.24 3.6±0.25 3.8±0.28 3.6±0.31 3.7±0.26 3.6±0.31 3.8±0.35
0.29±1.23 0.13±0.16 0.36±1.03 0.28±0.79 0.29±0.65 0.27±1.28 0.48±3.21 0.11±0.14 0.42±2.91 0.28±0.92
Abbreviations: BMI, body mass index; DM, diabetes mellitus; hCRP, high-sensitive C-reactive protein. Data are the mean±s.d.
Table 2 | Anemia-related factors % of iron administration
All patients A area B area C area D area E area F area G area H area I area
Number of patients
Hb (g/dl)
Ferritin (ng/ml)
TSAT (%)
Dose of ESA (IU/week)
ESA/Hb
Oral
IV
1095 166 122 106 50 126 104 44 274 103
10.6±1.0 10.2±0.9 10.6±1.1 10.6±0.9 10.3±1.0 10.6±1.1 10.4±1.1 11.0±0.9 10.8±0.9 10.7±0.7
125.4±147.0 145.7±119.5 139.5±151.2 169.7±94.9 139.2±263.7 121.9±182.0 155.7±164.6 55.64±76.0 99.4±132.1 97.5±111.3
26.7±11.7 28.0±11.6 27.9±14.2 28.3±9.4 26.1±13.5 25.7±12.7 26.8±13.0 26.9±9.2 26.3±11.5 23.9±8.3
3212±2107 3569±2477 3304±2366 3170±1932 2190±1468 3590±2394 2676±1996 3460±1651 3055±1737 3478±2165
308±219 378±256 292±255 312±216 215±150 336±244 258±213 313±143 289±174 327±216
2.6 2.4 1.6 0 0 2.4 0 31.2 1.5 1
20.8 30.7 31.9 13.2 8 30.2 14.4 4.5 21.3 4.9
Abbreviations: ESA, erythropoiesis-stimulating agent; Hb, hemoglobin; IV, intravenous; TSAT, transferrin saturation. Data are the mean±s.d.
Event
Cerebro-cardiovascular disease
Infection disease
Hospitalization
Death
All events
Hb fluctuation Target Low LAL High LAH HA Low LAL High LAH HA Low LAL High LAH HA Low LAL High LAH HA Low LAL High LAH HA
Hazard ratio (95% CI) 1 3.78 (1.12–12.8) 4.23 (1.21–14.76) 0.74 (0.08–7.05) 1.6 (0.45–5.71) 3.94 (1.12–13.87) 1.68 (0.95–2.96) 1.96 (1.22–3.16) 1.13 (0.63–2.02) 1.09 (0.68–1.75) 1.73 (1.06–2.82) 1.84 (1.00–3.37) 2.29 (1.38–3.81) 0.6 (0.25–1.45) 0.81 (0.47–1.41) 2.01 (1.16–3.46) 1.19 (0.19–7.29) 1.6 (0.44–5.80) NA NA 0.63 (0.15–2.65) 1.84 (0.55–6.10) 1.44 (0.93–2.22) 1.55 (1.10–2.18) 0.95 (0.60–1.50) 0.88 (0.62–1.25) 1.53 (1.07–2.18) 0.01 0.1
P-value 0.033 0.024 0.797 0.473 0.033 0.072 0.005 0.676 0.731 0.027 0.05 0.001 0.257 0.452 0.012 0.855 0.477 NA 0.525 0.319 0.1 0.012 0.832 0.469 0.018 1
10
100
Figure 1 | Impact of hemoglobin fluctuation pattern on the risk for adverse events and mortality. HA, high-amplitude fluctuation; High, high hemoglobin level; LAH, low-amplitude fluctuation with high hemoglobin level; LAL, low-amplitude fluctuation with low hemoglobin level; Low, low hemoglobin level; NA, not available; Target, target hemoglobin level. Kidney International (2014) 86, 845–854
847
clinical trial
T Kuragano et al.: Anemia treatment and adverse events
etiology, and albumin, C-reactive protein (CRP), and intact parathyroid hormone levels. According to the matching caliper (s.d. ¼ 0.3 and 1:1), matched pairs of 355 (710 patients) and 201 (402 patients) estimate the variation pattern for ferritin and hemoglobin. We analyzed the data for each matched subject pair using a time-dependent hazard model. Compared with the low-hemoglobin group, the risks for hospitalization (P ¼ 0.039, P ¼ 0.010) and all events (P ¼ 0.013, P ¼ 0.001) were significantly lower for the target-hemoglobin or LAH-hemoglobin (low-amplitude fluctuation with high hemoglobin level) groups. Compared with the low-ferritin group, the L-H ferritin group showed a significantly (P ¼ 0.035) higher risk for hospitalization (P ¼ 0.035) and death (P ¼ 0.006). In addition, the HA ferritin group showed significantly higher risks for hospitalization (P ¼ 0.043) and all events (P ¼ 0.008). Although matching resulted in a decreased number of patients analyzed, these results also indicated that patients with low hemoglobin and high ferritin levels could have a higher risk for adverse events. DISCUSSION
In this study, we demonstrated that maintenance hemodialysis patients with a stable target hemoglobin level had 11.5
Hb (g/dl)
11
*
*
*
*
0M
3M
6M
9M
*
*
*
10.5 10 9.5 9
12M 15M 18M 21M 24M
Figure 2 | Comparison of hemoglobin level in patients treated with a high erythropoiesis-stimulating agent (ESA) dose with those treated with a low ESA dose. The patients in this study were treated with epoetin-a or epoetin-b. K: Low-dose ESA (o4500 IU/ week) group; ’: high-dose ESA (X4,500 IU/week) group. The error bars indicate s.d.s. *Po0.05.
Event
Dose of ESA
Low dose
Hazard ratio
less risk for adverse events compared with maintenance hemodialysis patients without a stable hemoglobin level. We also found that the risks for infection and hospitalization were significantly higher in patients who were treated with a higher ESA dose. Furthermore, the risks for CCVD, infection, hospitalization, and death were significantly higher in patients with an elevated serum ferritin and in those with high-amplitude fluctuations in ferritin level. Moreover, the risks for CCVD, infection, and hospitalization were significantly higher in patients who were treated with high weekly doses of intravenous iron. Comparison of the mean hemoglobin, ferritin, and TSAT levels and iron doses in the study with those from the DOPPS report
According to the 2011 DOPPS report,9 the global mean hemoglobin level is 11.5 g/dl, compared with the mean level of 10.3 g/dl in Japan. The mean hemoglobin level at the start of the present study was 10.6±1.0 g/dl, which is lower than the global mean value but higher than the mean value in the Japanese population reported in J-DOPPS. The Japanese Society of Dialysis Therapy (JSDT) suggested that the difference in hemoglobin levels between Japan and other countries might be due to the timing of blood collection (the beginning of the week in Japan was the middle of the week in other countries), the method used for blood collection, and the patient’s position (supine position in Japan, sitting position in the other countries) during blood collection.15 The sampling time could affect the observed rate of weight gain. Furthermore, the use of the supine position for blood collection may decrease the apparent hemoglobin level. On the basis of these reports, we presumed that there was no significant difference in the hemoglobin levels between Japan and other countries.16 The 2008 JSDT guidelines for treating anemia in CKD recommend that iron be administered to patients with TSATp20% and serum ferritinp100 ng/ml. Furthermore, the recommended frequency of administration is up to once a week for 3 months or a total of 13 times, depending on the patient’s hemoglobin level.17 The 2008 JSDT guidelines for treating anemia in CKD also recommended that a minimal amount of iron be given to CKD patients with
(95% CI)
P-value
1
Cerebro-cardiovascular disease
High dose
1.59 (0.88–2.85)
0.122
Infectious disease
High dose
1.48 (1.08–2.02)
0.015
Hospitalization
High dose
1.8 (1.33–2.43)
