J Nephrol DOI 10.1007/s40620-014-0083-y
ORIGINAL ARTICLE
Serum hepcidin-25 and response to intravenous iron in patients with non-dialysis chronic kidney disease Sourabh Chand • Douglas G. Ward • Zhi-Yan Valerie Ng • James Hodson • Heidi Kirby Patricia Steele • Irina Rooplal • Ferly Bantugon • Tariq Iqbal • Chris Tselepis • Mark T. Drayson • Alison Whitelegg • Marie Chowrimootoo • Richard Borrows
•
Received: 15 November 2013 / Accepted: 18 March 2014 Ó Italian Society of Nephrology 2014
Abstract Background Hepcidin-25 is an iron regulator which reduces iron absorption and promotes sequestration in the reticulo-endothelial system. We investigated hepcidin and traditional iron storage marker utility in predicting haemoglobin increment following bolus intravenous iron. Methods The cohort included 129 consecutive non-dialysis chronic kidney disease patients that attended for intravenous iron over a 6-month period. Serum hepcidin-25
Electronic supplementary material The online version of this article (doi:10.1007/s40620-014-0083-y) contains supplementary material, which is available to authorized users. S. Chand (&) R. Borrows Renal Department, First Floor, Centre for Translational Inflammation Research, Queen Elizabeth Hospital Birmingham, University of Birmingham, Birmingham B15 2WB, UK e-mail: [email protected] S. Chand Z.-Y. V. Ng H. Kirby P. Steele I. Rooplal F. Bantugon M. Chowrimootoo R. Borrows Department of Nephrology and Kidney Transplantation, Queen Elizabeth Hospital Birmingham, Birmingham B15 2WB, UK D. G. Ward C. Tselepis School of Cancer Sciences, University of Birmingham, Birmingham B15 2TT, UK J. Hodson Department of Statistics, Wolfson Laboratory, Old Queen Elizabeth Hospital, Birmingham B15 2TH, UK T. Iqbal Department of Gastroenterology, Queen Elizabeth Hospital Birmingham, Birmingham B15 2WB, UK M. T. Drayson A. Whitelegg Department of Immunology, University of Birmingham, Birmingham B15 2WB, UK
levels (determined by mass spectrometry) pre iron infusion and 6 weeks post were compared with ferritin and transferrin saturation in multivariate models. Results Log10 ferritin [coefficient 0.559 (0.435–0.684) p \ 0.001] and log10 high-sensitive C-reactive protein [coefficient 0.092 (0.000–0.184) p = 0.049] were significantly associated with baseline log10 hepcidin-25 levels. Log10 estimated glomerular filtration rate was the only independent determinant of pre-infusion haemoglobin [coefficient 1.37 (0.16–2.59) p = 0.027]. Log10 hepcidin25 was an independent predictor of haemoglobin increment 6 weeks following iron infusion [coefficient -0.84 (-1.38 to -0.31) p = 0.002]. Ferritin, transferrin saturation and hepcidin had similar predictive utility for a 1 g/dl haemoglobin increase (c-statistics: 0.68, 0.70, 0.69). Conclusions Hepcidin is an iron sensor marker which predicts the magnitude of haemoglobin increment following protocolised intravenous iron infusion. Although displaying similar predictive performance to ferritin and transferrin saturation, hepcidin may also play a mechanistic role. Keywords Hepcidin Iron CKD Haemoglobin Prediction Utility
Introduction There has been a drive to use less erythropoietin in nondialysis renal anaemia due to its cost and potential harmful side-effect profile, leading to more intravenous iron use [1]. However there remain cost and safety concerns, highlighting the importance of using intravenous iron as effectively and efficiently as possible [2, 3]. Hepcidin is an integral component by which iron is absorbed and utilised within the body, predominantly by
123
J Nephrol
reducing the surface expression of the iron transporter ferroportin. Increases in hepcidin result in reduced enteric iron absorption and increased sequestration within the reticuloendothelial system in times of iron repletion. This mechanism protects against iron overload, with hepcidin acting as a ‘‘master regulator’’ of systemic iron homeostasis [4]. In haemodialysis, hepcidin serves no better than ferritin in predicting response to iron supplementation [5]. The performance of this biomarker may be different from that seen in chronic kidney disease (CKD), due to levels of inflammation, degree of renal dysfunction, and the reduction in endogenous erythropoietin production, all of which influence hepcidin levels [6–8]. Therefore the purpose of this study was to investigate the utility of hepcidin in predicting haemoglobin increment after iron infusion in a CKD population.
Methods This study was conducted as part of a service development programme at the Queen Elizabeth Hospital Birmingham, United Kingdom, as part of patients’ routine clinical practice. It was thus exempt from institutional or ethical committee approval requirement. Over a 6-month period from August 2012, 129 consecutive, unselected, non-dialysis CKD patients attended for routine intravenous iron anaemia management. Although 12 patients had received oral iron, no patients had received intravenous iron or blood transfusion in the 6 months prior to iron infusion. Doses of erythropoiesis stimulating agents (ESAs) remained unchanged 1 month prior to iron administration. All received a single dose of ferric carboxymaltose (FerinjectÒ) according to the manufacturer’s guidelines, stratified by body weight to a maximum of 1 g if over 67 kg. Our criteria for iron infusion comprised either ferritin \200 ng/ml or transferrin saturation (TSAT) \20 %. Iron was not administered if ferritin levels were [500 ng/ml, irrespective of TSAT. Eligible patients displayed haemoglobin levels \12 g/dl receiving ESAs, or \11 g/dl in patients not receiving ESAs. This protocol was in line with achieving Renal Association anaemia guideline targets [9]. Patients with coincident sepsis, known active gastrointestinal blood loss or active haematological disorders were excluded. Three patients had non-metastatic prostate cancer on luteinising-hormone-releasing therapy; no other active malignancy was present in this cohort.
[ESAs; antiproliferative medication (azathioprine, mycophenolate and sirolimus); angiotensin converting enzyme inhibitors (ACEI) or angiotensin receptor blockers (ARB)], presence of a functioning kidney transplant, and presence of an arterio-venous fistula (AVF). The following haematological and biochemical measures were collected prior to iron infusion: haemoglobin, hepcidin-25, ferritin, TSAT, serum vitamin B12/folate, estimated glomerular filtration rate (eGFR) and high-sensitive C-reactive protein levels (hsCRP). hsCRP was measured using a cardiac C-reactive protein (latex), high sensitive assay on a CobasÒ c501 turbidimeter (Roche diagnostics, Basel, Switzerland). Serum hepcidin-25 levels were measured by surface enhanced laser desorption/ionisation time-of-flight mass spectrometry using stable isotope labelled synthetic hepcidin-25 as an internal standard as previously described [10]. Haemoglobin, hepcidin-25, ferritin and TSAT were measured again 6 weeks following iron infusion. Bilirubin, alanine transaminase (ALT), alkaline phosphatase, eGFR and serum phosphate levels were measured as safety parameters. Results 6 weeks following iron administration were unavailable in 10 patients due to death (n = 2), non-attendance (n = 5), or commencement of dialysis within the 6-week period (n = 3). In addition, factors influencing serum hepcidin and baseline haemoglobin levels were evaluated with regard to the demographic and biochemical parameters described above. Statistics Data are shown as mean (±standard deviation), or median [interquartile range (IQR)] unless otherwise indicated. The association between predictor variable and outcome measures was first evaluated in a univariate linear regression analysis, and then subsequently in a multivariate model, where no selection process was used. Regression coefficients with 95 % confidence intervals (CI) were used as measures of effect. Skewed data underwent logarithmic transformation prior to analysis. Spearman’s correlation coefficients were used to compare colinearity of the different iron markers. Wilcoxon’s test was used to analyse the changes in liver function measures pre- and post-iron infusion, and paired Student’s t test was used to compare haemoglobin pre- and post-iron infusion. A type 1 error rate of \5 % (p B 0.05) was considered statistically significant. SPSS software version 19 (SPSS Inc., Chicago, IL, USA) was used for analysis.
Outcome measures Results The primary outcome measure was the increment in haemoglobin 6 weeks following iron administration. Demographics of the patients included age, sex, comedication
123
Baseline demographics are shown in Table 1. About onethird (34 %) of patients had type 2 diabetes mellitus, with
J Nephrol Table 1 Baseline characteristics of the non-dialysis chronic kidney disease cohort Characteristic Number of patients
129
Age, years
62 (44–77)
Age range, years
19–92
Male, %
54
Iron dose/weight, mg/kg
13.2 (11.6–14.0)
eGFR, ml/min/1.73 m2
23 (15–33)
Number of renal transplant patients (%) Number of patients on ESAs (%)
27 (21) 43 (33)
Dose of ESA, equivalent to epoetin, IU/kg/week
53 (31–75)
Mean Hb, g/dl (±SD)
10.3 (±1.4)
Ferritin, ng/ml
82 (36–164)
TSAT, %
18.2 (12.1–24.2)
Hepcidin-25, ng/ml
47 (26–90)
hsCRP, mg/l
3.25 (1.28–9.53)
Albumin, g/l
42 (38–45)
Weight, kg
75.8 (64.6–86.6)
BMI, kg/m2
27.2 (24.8–30.5)
Comorbidity number, n patients (%) 0
14 (11)
1
44 (34)
2
32 (25)
3 [4
26 (20) 13 (10)
Renal diagnoses, n patients (%) Glomerular disease
23 (18)
Systemic disease
47 (36)
Tubulointerstitial disease
16 (12)
Familial nephropathies
10 (8)
Miscellaneous
33 (26)
Values are median (and interquartile range) unless otherwise stated ESA erythropoiesis stimulating agent, Hb haemoglobin, TSAT transferrin saturation, hsCRP high-sensitive C-reactive protein, BMI body mass index, SD standard deviation
the majority of diabetic patients (82 %) having a known concurrent vascular disease (hypertension, ischaemic heart disease, cerebral or peripheral vascular disease). Over half of the patients had two or more comorbidities on top of CKD diagnosis. No patients had indwelling intravenous dialysis catheters or peritoneal dialysis catheters, or had a failed kidney transplant in situ. Determinants of serum hepcidin and haemoglobin levels prior to iron infusion Baseline predictors of serum hepcidin-25 and haemoglobin level were studied in all 129 patients in the cohort. As shown in Table 2, increased log-transformed ferritin and
elevated hsCRP were significantly associated with pre-iron infusion log-transformed hepcidin-25 level after adjusting age, gender, comedication (including ESA use), transplant status, the presence of AVFs, B12 and folate levels (low vs. normal), and eGFR. eGFR was significantly associated with serum hepcidin on univariate analysis, but not in the adjusted model as shown in previous studies [7, 11]. Further examination suggested this was likely due to the correlation between eGFR and ferritin (r = -0.31), and between eGFR and hsCRP (r = -0.22). TSAT and ferritin were significantly correlated (r = 0.55; p \ 0.001) and were therefore not analysed simultaneously, although a univariate association between TSAT and serum hepcidin level was seen (effect per 1 % TSAT: 0.02; 95 % CI 0.01–0.02; p \ 0.001). Because of the correlation between hepcidin-25, ferritin and TSAT (p \ 0.001 for all comparisons) only hepcidin25 was evaluated as a predictor variable in the analysis of the factors associated with baseline haemoglobin. In the adjusted model (see Supplementary material 1) lower eGFR was associated with lower haemoglobin level (p = 0.03), with some effect seen between male gender and higher haemoglobin (p = 0.06). Predictors of haemoglobin increment following intravenous iron Hepcidin-25, ferritin, TSAT, and haemoglobin levels all rose significantly 6 weeks following iron infusion (p \ 0.001 respectively; Fig. 1), in keeping with previous studies [12]. The median haemoglobin increment was 0.7 g/dl (IQR 0.1–1.55). In the adjusted analysis, log-transformed hepcidin-25 and the pre-iron infusion haemoglobin level were the only variables significantly associated with the magnitude of the haemoglobin increment, although there was some evidence of an effect for hsCRP (Table 3). In addition, there was a significant interaction between log10 hsCRP and log10 hepcidin-25 [coefficient 0.998 (0.090–1.906) p = 0.031]. Figure 2 demonstrates that in patients displaying even minor degrees of inflammation (hsCRP C3 mg/l; 53 % of patients) the relationship between hepcidin and haemoglobin increment was less evident. Hepcidin-25, ferritin and TSAT had comparative predictive utility with similar c-statistic values [receiver operating characteristic (ROC) curve analysis] when assessed against an outcome increment in haemoglobin of either 0.5 g/dl (c-statistics: 0.67, 0.65, 0.67 respectively) or 1 g/dl (c-statistics: 0.69, 0.68, 0.70 respectively) (Fig. 3). Although the haemoglobin increment was less at higher values of baseline hepcidin-25, ferritin and TSAT, a clinically relevant increase in haemoglobin was observed in
123
J Nephrol Table 2 Univariate and multivariate analysis of predictors of log-transformed serum hepcidin-25 Variable
Univariate
Multivariate
Coefficient (95 % CI)
p
Coefficient (95 % CI)
p
Ferritina
0.568 (0.461, 0.675)
hsCRPa
0.168 (0.057, 0.280)
\0.001
0.559 (0.435, 0.684)
\0.001
0.003
0.092 (0.000, 0.184)
eGFRa
0.05
-0.454 (-0.724, -0.183)
0.001
-0.131 (-0.391, 0.128)
0.32
Ageb
-0.001 (-0.004, 0.003)
0.72
-0.003 (-0.007, 0.000)
0.06
Gender (male)
-0.001 (-0.136, 0.134)
0.99
-0.002 (-0.106, 0.103)
0.98
ESA use (yes)
0.065 (-0.077, 0.207)
0.37
-0.071 (-0.184, 0.041)
0.21
Anti-proliferative (yes) ACEI/ARB (yes)
0.014 (-0.139, 0.167) -0.003 (-0.139, 0.133)
0.86 0.97
-0.088 (-0.249, 0.073) -0.004 (-0.107, 0.098)
0.28 0.94
Functioning renal Tx (yes)
0.100 (-0.064, 0.264)
0.23
0.130 (-0.032, 0.292)
0.12
Access (AVF)
0.088 (-0.110, 0.287)
0.38
-0.031 (-0.187, 0.124)
0.69
B12 (low)
-0.083 (-0.334, 0.168)
0.51
-0.074 (-0.259, 0.110)
0.43
Folate (low)
-0.013 (-0.188, 0.163)
0.89
0.046 (-0.087, 0.178)
0.49
CI confidence interval, hsCRP high-sensitive C-reactive protein, eGFR estimated glomerular filtration rate, ESA erythropoiesis stimulating agent, ACEI angiotensin converting enzyme inhibitor, ARB angiotensin receptor blocker, Tx transplant, AVF arterio-venous fistula a b
Continuous log-transformed variable continuous linear variable. Otherwise all variables are binary
Fig. 1 Pre- and post-iron infusion mean ± standard error mean levels of haemoglobin, hepcidin-25, ferritin and transferrin saturation (TSAT). p values derived from two-tailed, paired t tests
123
J Nephrol Table 3 Univariate and multivariate analysis of predictors of haemoglobin increment following intravenous iron infusion CI confidence interval, hsCRP high sensitive C-reactive protein, eGFR estimated glomerular filtration rate, ESA erythropoiesis stimulating agent, ACEI angiotensin converting enzyme inhibitor, ARB angiotensin receptor blocker, Tx transplant, AVF arterio-venous fistula
Variable
Univariate
Multivariate
Coefficient (95 % CI)
p
Coefficient (95 % CI)
Hepcidina
-1.05 (-1.57, -0.54)
\0.001
hsCRPa
-0.58 (-0.93, -0.23)
0.001
eGFRa
0.39 (-0.52, 1.31) -0.01 (-0.02, 0.00)
-0.84 (-1.38, -0.31)
p 0.002
-0.36 (-0.75, 0.02)
0.07
0.40
0.28 (-0.73, 1.29)
0.59
0.21
0.00 (-0.02, 0.01)
0.66
-0.32 (-0.46, -0.17)
\0.001
-0.33 (-0.48, -0.19)
\0.001
0.03 (-0.08, 0.14)
0.62
0.01 (-0.09, 0.12)
0.78
Gender (male)
-0.14 (-0.57, 0.28)
0.50
0.09 (-0.32, 0.5)
0.66
ESA use (yes)
-0.31 (-0.76, 0.13)
0.16
0.05 (-0.38, 0.49)
0.81
0.37 (-0.10, 0.84)
0.13
-0.01 (-0.6, 0.59)
0.98
-0.08 (-0.51, 0.35)
0.70
0.09 (-0.3, 0.47)
0.67
0.31 (-0.20, 0.81) 0.18 (-0.47, 0.84)
0.24 0.58
0.28 (-0.32, 0.89) 0.44 (-0.17, 1.05)
0.36 0.15
Ageb b
Pre-infusion haemoglobin Iron dose/weightb
Anti-proliferative (yes) ACEI/ARB (yes)
a
Continuous log-transformed variable
Functioning renal Tx (yes) Access (AVF)
b
B12 (low)
0.90 (0.12, 1.68)
0.02
0.46 (-0.26, 1.18)
0.21
Folate (low)
0.17 (-0.38, 0.72)
0.54
0.22 (-0.29, 0.72)
0.40
continuous linear variable. Otherwise, all variables are binary
Fig. 2 Scatter plot describing the relationship between haemoglobin increment and hepcidin-25 levels and demonstrating impact of inflammation status (hsCRP \3 mg/l and hsCRP C3 mg/l). hsCRP high-sensitive C-reactive protein
some patients even when the iron markers suggested iron repletion (Table 4). Short term safety outcomes No immediate or delayed complications of intravenous iron administration were observed by the clinical team. Bilirubin, ALT and eGFR showed no significant difference from baseline at 6 weeks post-infusion (data not shown). A marginal increase in alkaline phosphatase from 79.5 l/l
(IQR 65.0–109.8) to 88.0 l/l (IQR 68.0–121.0) was seen (p \ 0.001). A decrease in serum phosphate level was observed 1.22 mmol/l (IQR 1.08–1.46) to 1.20 mmol/l (IQR 0.93–1.39) following iron infusion (p = 0.002).
Discussion This is the first study to describe the utility of serum hepcidin-25 in predicting the haemoglobin increment
123
J Nephrol
Fig. 3 Receiving operation characteristics (ROC) curve analysis of pre-iron infusion, ferritin, transferrin saturation (TSAT) and hepcidin-25 levels to 0.5 and 1 g/dl rise in haemoglobin (Hb)
Table 4 Haemoglobin levels pre- and post-intravenous iron are shown for each baseline range for: (a) ferritin, (b) transferrin saturation and (c) hepcidin-25
(a) Ferritin (ng/ml) Ferritin levels
Number of patients
\50
40
50–99
SD
Mean post iron Hb (g/dl)
SD p*
% with 0.5 g/ % with 1 g/ dl Hb rise dl Hb rise
9.9
1.5
11.3
1.7 \0.001
75
63 38
Mean pre iron Hb (g/dl)
24
10.9
1.3
11.3
1.3
0.003
58
100–149 21
10.4
1.1
11.2
1.2
0.004
57
38
150–199 18
10.7
1
11.1
0.9
0.16
33
28
200–500 16
9.6
1.4
10.4
1.1
0.02
56
31
% with 0.5 g/ dl Hb rise
% with 1 g/ dl Hb rise
(b) Transferrin saturation (TSAT %) TSAT levels
Number of patients
\10
21
10–19
46
20–24
23
25–29
19
C30
10
Mean pre iron Hb (g/dl)
p*
SD Mean post iron Hb (g/dl)
SD
9.6
1.2
11.3
1.6 \0.001
86
76
10.2
1.3
11.1
1.4 \0.001
61
43
10.9
1.4
11.6
1.7 \0.001
52
43
10.6
1.1
11
0.8
0.13
42
16
10
1.4
10.5
1.1
0.06
50
30
(c) Hepcidin-25 (ng/ml)
Also shown are the percentages of patients demonstrating a haemoglobin increment of 0.5 g/dl or 1.0 g/dl for each baseline range * Two-tailed paired t test comparing haemoglobin level pre- and post-intravenous iron
Number of patients
Mean pre iron Hb (g/ dl)
SD
Mean post iron Hb (g/dl)
SD
\20
16
10.2
1.3
11.6
1.6 \0.001
81
69
20–29
21
10.2
1.6
11.8
1.6 \0.001
81
71
30–49 50–99
23 31
10.1 10.5
1.2 1.5
10.8 11.1
1.3 \0.001 1.3 \0.001
57 48
39 32
C100
28
10.3
1.2
10.7
1.2
39
25
following bolus iron infusion in patients with non-dialysis CKD. The aim of the study was to identify an iron marker that can identify individuals most likely to benefit from
123
p*
Hepcidin levels
0.10
% with 0.5 g/dl Hb rise
% with 1 g/ dl Hb rise
intravenous iron, thereby improving the efficacy whilst minimising risk of iron toxicity. In this regard, hepcidin-25 is a potential candidate as it serves as a sensor of iron
J Nephrol
status, but also has a unique position as a biological regulator of iron absorption and utilisation. A significant relationship was found between hepcidin-25 and haemoglobin increment following protocolised iron infusion, with higher hepcidin-25 levels associated with a smaller haemoglobin increment across the studied population. However, before any suggestion that this iron marker be adopted in routine clinical practice, three caveats should be borne in mind. First, prediction was imperfect but, reassuringly in an era of ESA minimisation, even at the highest levels of serum hepcidin-25 encountered in the studied population iron infusion was often associated with a clinically relevant increase in haemoglobin. This finding is in line with the ‘DRIVE’ study which demonstrated the utility of intravenous iron in a proportion of haemodialysis patients deemed ‘‘iron-replete’’ who had received an initial increase in dose of erythropoietin at randomisation [13]. It seems that to some extent the effect of intravenous iron infusion may overwhelm or bypass the effects of iron sequestration as a result of elevated hepcidin levels, and therefore it influences marrow production of red blood cells. However, it may also be the case that iron deposition occurs elsewhere, and future studies are afoot to address the long-term safety profile of intravenous iron use in CKD [14]. The second caveat is that the prognostic performance of serum hepcidin-25 was almost identical to that of the established markers serum ferritin and TSAT (ROC-curve analysis), which for most clinicians will be more easily available and interpretable. Thus it is unlikely that hepcidin-25 has a role to play as an isolated marker of iron status in this setting. However hepcidin-25 has the advantage of being potentially mechanistic in regard to iron utilisation [15], although this will need formal assessment in prospective studies. The importance of this study is that it suggests the possibility that greater increments in haemoglobin may be achieved in this setting by antagonising the effect of high hepcidin-25 pre-iron infusion. Whilst an advantage of hepcidin antagonism may be improved efficacy of oral iron due to increased intestinal absorption, it should be remembered that hepcidin antagonism may also potentiate the effect of intravenous iron administration (i.e. where a block to intestinal absorption is not an issue) by reducing iron sequestration in the reticulo-endothelial system [16]. A study conducted in 56 haemodialysis patients failed to show an association between hepcidin-25 level and haemoglobin response to intravenous iron [5]. However, the current study describes a larger cohort of CKD patients not on dialysis, and who were therefore not subject to the blood loss and intra-individual variation of inflammation and hepcidin-25 seen in maintenance haemodialysis patients [17]. However, the relationship between hepcidin and haemoglobin increment in our cohort did depend on
prevailing levels of inflammation, with evidence that the relationship is less evident in patients displaying hsCRP C3 mg/l. Despite 21 % of the study population being transplant recipients, the use of anti-proliferative medications and ESAs did not appear to significantly affect baseline hepcidin-25 levels and post-iron infusion haemoglobin increment in multivariate analysis. Though the current study benefits from its ‘‘real world’’ setting and potential generalisability, the results would need to be reproduced in view of their heterogeneity. In summary, this study demonstrates the relationship between hepcidin-25 and traditional markers of iron storage and inflammation in patients with non-dialysis dependent CKD. It also shows an association between hepcidin25 levels and the response to intravenous iron, which may be of utility in predicting the expected response to iron. Although the prognostic performance of hepcidin-25 was non-superior to traditional iron storage markers, its mechanistic role points to a potential rationale for hepcidin antagonism in patients with CKD. Conflict of interest On behalf of all authors, the corresponding author states that there is no conflict of interest.
References 1. Pfeffer MA, Burdmann EA, Chen CY et al (2009) A trial of darbepoetin alfa in type 2 diabetes and chronic kidney disease. N Engl J Med 361:2019–2032 2. Fievet P, Coppin M, Brazier F, Lefevre M, Stephan R, Demontis R (2012) Treatment of iron deficiency in predialysis state by low molecular weight iron dextran high doses intravenously. Nephrol Ther 8:41–46 3. Zager RA, Johnson AC, Hanson SY, Wasse H (2002) Parenteral iron formulations: a comparative toxicologic analysis and mechanisms of cell injury. Am J Kidney Dis 40:90–103 4. Sarafidis PA, Rumjon A, MacLaughlin HL, Macdougall IC (2012) Obesity and iron deficiency in chronic kidney disease: the putative role of hepcidin. Nephrol Dial Transpl 27:50–57 5. Tessitore N, Girelli D, Campostrini N et al (2010) Hepcidin is not useful as a biomarker for iron needs in haemodialysis patients on maintenance erythropoiesis-stimulating agents. Nephrol Dial Transpl 25:3996–4002 6. Gama-Axelsson T, Heimburger O, Stenvinkel P, Barany P, Lindholm B, Qureshi AR (2012) Serum albumin as predictor of nutritional status in patients with ESRD. Clin J Am Soc Nephrol 7:1446–1453 7. Uehata T, Tomosugi N, Shoji T et al (2012) Serum hepcidin-25 levels and anemia in non-dialysis chronic kidney disease patients: a cross-sectional study. Nephrol Dial Transpl 27:1076–1083 8. Zaritsky J, Young B, Wang HJ et al (2009) Hepcidin––a potential novel biomarker for iron status in chronic kidney disease. Clin J Am Soc Nephrol 4:1051–1056 9. Mikhail A, Shrivastava R, Richardson D (2010) Anaemia in CKD. http://www.renal.org/clinical/guidelinessection/AnaemiaInCKD. aspx. Edited by Renal Association 2010. Accessed 12 Sept 2013
123
J Nephrol 10. Ward DG, Roberts K, Stonelake P et al (2008) SELDI–TOF-MS determination of hepcidin in clinical samples using stable isotope labelled hepcidin as an internal standard. Proteome Sci 6:28 11. Peters HP, Laarakkers CM, Swinkels DW, Wetzels JF (2010) Serum hepcidin-25 levels in patients with chronic kidney disease are independent of glomerular filtration rate. Nephrol Dial Transpl 25:848–853 12. Ashby DR, Gale DP, Busbridge M et al (2009) Plasma hepcidin levels are elevated but responsive to erythropoietin therapy in renal disease. Kidney Int 75:976–981 13. Coyne DW, Kapoian T, Suki W et al (2007) Ferric gluconate is highly efficacious in anemic hemodialysis patients with high serum ferritin and low transferrin saturation: results of the
123
14.
15. 16. 17.
Dialysis Patients’ Response to IV Iron with Elevated Ferritin (DRIVE) Study. J Am Soc Nephrol 18:975–984 Macdougall IC (2013) Ferric carboxymaltose assessment in subjects with iron deficiency anaemia and non-dialysis-dependent CKD (FIND-CKD). http://clinicaltrials.gov/ct2/show/study/NCT009943 18. Edited by ClinicalTrials.gov 2013. Accessed 21 June 2013 Coyne DW (2011) Hepcidin: clinical utility as a diagnostic tool and therapeutic target. Kidney Int 80:240–244 Macdougall IC (2012) New anemia therapies: translating novel strategies from bench to bedside. Am J Kidney Dis 59:444–451 Ford BA, Eby CS, Scott MG, Coyne DW (2010) Intra-individual variability in serum hepcidin precludes its use as a marker of iron status in hemodialysis patients. Kidney Int 78:769–773
ORIGINAL ARTICLE
Serum hepcidin-25 and response to intravenous iron in patients with non-dialysis chronic kidney disease Sourabh Chand • Douglas G. Ward • Zhi-Yan Valerie Ng • James Hodson • Heidi Kirby Patricia Steele • Irina Rooplal • Ferly Bantugon • Tariq Iqbal • Chris Tselepis • Mark T. Drayson • Alison Whitelegg • Marie Chowrimootoo • Richard Borrows
•
Received: 15 November 2013 / Accepted: 18 March 2014 Ó Italian Society of Nephrology 2014
Abstract Background Hepcidin-25 is an iron regulator which reduces iron absorption and promotes sequestration in the reticulo-endothelial system. We investigated hepcidin and traditional iron storage marker utility in predicting haemoglobin increment following bolus intravenous iron. Methods The cohort included 129 consecutive non-dialysis chronic kidney disease patients that attended for intravenous iron over a 6-month period. Serum hepcidin-25
Electronic supplementary material The online version of this article (doi:10.1007/s40620-014-0083-y) contains supplementary material, which is available to authorized users. S. Chand (&) R. Borrows Renal Department, First Floor, Centre for Translational Inflammation Research, Queen Elizabeth Hospital Birmingham, University of Birmingham, Birmingham B15 2WB, UK e-mail: [email protected] S. Chand Z.-Y. V. Ng H. Kirby P. Steele I. Rooplal F. Bantugon M. Chowrimootoo R. Borrows Department of Nephrology and Kidney Transplantation, Queen Elizabeth Hospital Birmingham, Birmingham B15 2WB, UK D. G. Ward C. Tselepis School of Cancer Sciences, University of Birmingham, Birmingham B15 2TT, UK J. Hodson Department of Statistics, Wolfson Laboratory, Old Queen Elizabeth Hospital, Birmingham B15 2TH, UK T. Iqbal Department of Gastroenterology, Queen Elizabeth Hospital Birmingham, Birmingham B15 2WB, UK M. T. Drayson A. Whitelegg Department of Immunology, University of Birmingham, Birmingham B15 2WB, UK
levels (determined by mass spectrometry) pre iron infusion and 6 weeks post were compared with ferritin and transferrin saturation in multivariate models. Results Log10 ferritin [coefficient 0.559 (0.435–0.684) p \ 0.001] and log10 high-sensitive C-reactive protein [coefficient 0.092 (0.000–0.184) p = 0.049] were significantly associated with baseline log10 hepcidin-25 levels. Log10 estimated glomerular filtration rate was the only independent determinant of pre-infusion haemoglobin [coefficient 1.37 (0.16–2.59) p = 0.027]. Log10 hepcidin25 was an independent predictor of haemoglobin increment 6 weeks following iron infusion [coefficient -0.84 (-1.38 to -0.31) p = 0.002]. Ferritin, transferrin saturation and hepcidin had similar predictive utility for a 1 g/dl haemoglobin increase (c-statistics: 0.68, 0.70, 0.69). Conclusions Hepcidin is an iron sensor marker which predicts the magnitude of haemoglobin increment following protocolised intravenous iron infusion. Although displaying similar predictive performance to ferritin and transferrin saturation, hepcidin may also play a mechanistic role. Keywords Hepcidin Iron CKD Haemoglobin Prediction Utility
Introduction There has been a drive to use less erythropoietin in nondialysis renal anaemia due to its cost and potential harmful side-effect profile, leading to more intravenous iron use [1]. However there remain cost and safety concerns, highlighting the importance of using intravenous iron as effectively and efficiently as possible [2, 3]. Hepcidin is an integral component by which iron is absorbed and utilised within the body, predominantly by
123
J Nephrol
reducing the surface expression of the iron transporter ferroportin. Increases in hepcidin result in reduced enteric iron absorption and increased sequestration within the reticuloendothelial system in times of iron repletion. This mechanism protects against iron overload, with hepcidin acting as a ‘‘master regulator’’ of systemic iron homeostasis [4]. In haemodialysis, hepcidin serves no better than ferritin in predicting response to iron supplementation [5]. The performance of this biomarker may be different from that seen in chronic kidney disease (CKD), due to levels of inflammation, degree of renal dysfunction, and the reduction in endogenous erythropoietin production, all of which influence hepcidin levels [6–8]. Therefore the purpose of this study was to investigate the utility of hepcidin in predicting haemoglobin increment after iron infusion in a CKD population.
Methods This study was conducted as part of a service development programme at the Queen Elizabeth Hospital Birmingham, United Kingdom, as part of patients’ routine clinical practice. It was thus exempt from institutional or ethical committee approval requirement. Over a 6-month period from August 2012, 129 consecutive, unselected, non-dialysis CKD patients attended for routine intravenous iron anaemia management. Although 12 patients had received oral iron, no patients had received intravenous iron or blood transfusion in the 6 months prior to iron infusion. Doses of erythropoiesis stimulating agents (ESAs) remained unchanged 1 month prior to iron administration. All received a single dose of ferric carboxymaltose (FerinjectÒ) according to the manufacturer’s guidelines, stratified by body weight to a maximum of 1 g if over 67 kg. Our criteria for iron infusion comprised either ferritin \200 ng/ml or transferrin saturation (TSAT) \20 %. Iron was not administered if ferritin levels were [500 ng/ml, irrespective of TSAT. Eligible patients displayed haemoglobin levels \12 g/dl receiving ESAs, or \11 g/dl in patients not receiving ESAs. This protocol was in line with achieving Renal Association anaemia guideline targets [9]. Patients with coincident sepsis, known active gastrointestinal blood loss or active haematological disorders were excluded. Three patients had non-metastatic prostate cancer on luteinising-hormone-releasing therapy; no other active malignancy was present in this cohort.
[ESAs; antiproliferative medication (azathioprine, mycophenolate and sirolimus); angiotensin converting enzyme inhibitors (ACEI) or angiotensin receptor blockers (ARB)], presence of a functioning kidney transplant, and presence of an arterio-venous fistula (AVF). The following haematological and biochemical measures were collected prior to iron infusion: haemoglobin, hepcidin-25, ferritin, TSAT, serum vitamin B12/folate, estimated glomerular filtration rate (eGFR) and high-sensitive C-reactive protein levels (hsCRP). hsCRP was measured using a cardiac C-reactive protein (latex), high sensitive assay on a CobasÒ c501 turbidimeter (Roche diagnostics, Basel, Switzerland). Serum hepcidin-25 levels were measured by surface enhanced laser desorption/ionisation time-of-flight mass spectrometry using stable isotope labelled synthetic hepcidin-25 as an internal standard as previously described [10]. Haemoglobin, hepcidin-25, ferritin and TSAT were measured again 6 weeks following iron infusion. Bilirubin, alanine transaminase (ALT), alkaline phosphatase, eGFR and serum phosphate levels were measured as safety parameters. Results 6 weeks following iron administration were unavailable in 10 patients due to death (n = 2), non-attendance (n = 5), or commencement of dialysis within the 6-week period (n = 3). In addition, factors influencing serum hepcidin and baseline haemoglobin levels were evaluated with regard to the demographic and biochemical parameters described above. Statistics Data are shown as mean (±standard deviation), or median [interquartile range (IQR)] unless otherwise indicated. The association between predictor variable and outcome measures was first evaluated in a univariate linear regression analysis, and then subsequently in a multivariate model, where no selection process was used. Regression coefficients with 95 % confidence intervals (CI) were used as measures of effect. Skewed data underwent logarithmic transformation prior to analysis. Spearman’s correlation coefficients were used to compare colinearity of the different iron markers. Wilcoxon’s test was used to analyse the changes in liver function measures pre- and post-iron infusion, and paired Student’s t test was used to compare haemoglobin pre- and post-iron infusion. A type 1 error rate of \5 % (p B 0.05) was considered statistically significant. SPSS software version 19 (SPSS Inc., Chicago, IL, USA) was used for analysis.
Outcome measures Results The primary outcome measure was the increment in haemoglobin 6 weeks following iron administration. Demographics of the patients included age, sex, comedication
123
Baseline demographics are shown in Table 1. About onethird (34 %) of patients had type 2 diabetes mellitus, with
J Nephrol Table 1 Baseline characteristics of the non-dialysis chronic kidney disease cohort Characteristic Number of patients
129
Age, years
62 (44–77)
Age range, years
19–92
Male, %
54
Iron dose/weight, mg/kg
13.2 (11.6–14.0)
eGFR, ml/min/1.73 m2
23 (15–33)
Number of renal transplant patients (%) Number of patients on ESAs (%)
27 (21) 43 (33)
Dose of ESA, equivalent to epoetin, IU/kg/week
53 (31–75)
Mean Hb, g/dl (±SD)
10.3 (±1.4)
Ferritin, ng/ml
82 (36–164)
TSAT, %
18.2 (12.1–24.2)
Hepcidin-25, ng/ml
47 (26–90)
hsCRP, mg/l
3.25 (1.28–9.53)
Albumin, g/l
42 (38–45)
Weight, kg
75.8 (64.6–86.6)
BMI, kg/m2
27.2 (24.8–30.5)
Comorbidity number, n patients (%) 0
14 (11)
1
44 (34)
2
32 (25)
3 [4
26 (20) 13 (10)
Renal diagnoses, n patients (%) Glomerular disease
23 (18)
Systemic disease
47 (36)
Tubulointerstitial disease
16 (12)
Familial nephropathies
10 (8)
Miscellaneous
33 (26)
Values are median (and interquartile range) unless otherwise stated ESA erythropoiesis stimulating agent, Hb haemoglobin, TSAT transferrin saturation, hsCRP high-sensitive C-reactive protein, BMI body mass index, SD standard deviation
the majority of diabetic patients (82 %) having a known concurrent vascular disease (hypertension, ischaemic heart disease, cerebral or peripheral vascular disease). Over half of the patients had two or more comorbidities on top of CKD diagnosis. No patients had indwelling intravenous dialysis catheters or peritoneal dialysis catheters, or had a failed kidney transplant in situ. Determinants of serum hepcidin and haemoglobin levels prior to iron infusion Baseline predictors of serum hepcidin-25 and haemoglobin level were studied in all 129 patients in the cohort. As shown in Table 2, increased log-transformed ferritin and
elevated hsCRP were significantly associated with pre-iron infusion log-transformed hepcidin-25 level after adjusting age, gender, comedication (including ESA use), transplant status, the presence of AVFs, B12 and folate levels (low vs. normal), and eGFR. eGFR was significantly associated with serum hepcidin on univariate analysis, but not in the adjusted model as shown in previous studies [7, 11]. Further examination suggested this was likely due to the correlation between eGFR and ferritin (r = -0.31), and between eGFR and hsCRP (r = -0.22). TSAT and ferritin were significantly correlated (r = 0.55; p \ 0.001) and were therefore not analysed simultaneously, although a univariate association between TSAT and serum hepcidin level was seen (effect per 1 % TSAT: 0.02; 95 % CI 0.01–0.02; p \ 0.001). Because of the correlation between hepcidin-25, ferritin and TSAT (p \ 0.001 for all comparisons) only hepcidin25 was evaluated as a predictor variable in the analysis of the factors associated with baseline haemoglobin. In the adjusted model (see Supplementary material 1) lower eGFR was associated with lower haemoglobin level (p = 0.03), with some effect seen between male gender and higher haemoglobin (p = 0.06). Predictors of haemoglobin increment following intravenous iron Hepcidin-25, ferritin, TSAT, and haemoglobin levels all rose significantly 6 weeks following iron infusion (p \ 0.001 respectively; Fig. 1), in keeping with previous studies [12]. The median haemoglobin increment was 0.7 g/dl (IQR 0.1–1.55). In the adjusted analysis, log-transformed hepcidin-25 and the pre-iron infusion haemoglobin level were the only variables significantly associated with the magnitude of the haemoglobin increment, although there was some evidence of an effect for hsCRP (Table 3). In addition, there was a significant interaction between log10 hsCRP and log10 hepcidin-25 [coefficient 0.998 (0.090–1.906) p = 0.031]. Figure 2 demonstrates that in patients displaying even minor degrees of inflammation (hsCRP C3 mg/l; 53 % of patients) the relationship between hepcidin and haemoglobin increment was less evident. Hepcidin-25, ferritin and TSAT had comparative predictive utility with similar c-statistic values [receiver operating characteristic (ROC) curve analysis] when assessed against an outcome increment in haemoglobin of either 0.5 g/dl (c-statistics: 0.67, 0.65, 0.67 respectively) or 1 g/dl (c-statistics: 0.69, 0.68, 0.70 respectively) (Fig. 3). Although the haemoglobin increment was less at higher values of baseline hepcidin-25, ferritin and TSAT, a clinically relevant increase in haemoglobin was observed in
123
J Nephrol Table 2 Univariate and multivariate analysis of predictors of log-transformed serum hepcidin-25 Variable
Univariate
Multivariate
Coefficient (95 % CI)
p
Coefficient (95 % CI)
p
Ferritina
0.568 (0.461, 0.675)
hsCRPa
0.168 (0.057, 0.280)
\0.001
0.559 (0.435, 0.684)
\0.001
0.003
0.092 (0.000, 0.184)
eGFRa
0.05
-0.454 (-0.724, -0.183)
0.001
-0.131 (-0.391, 0.128)
0.32
Ageb
-0.001 (-0.004, 0.003)
0.72
-0.003 (-0.007, 0.000)
0.06
Gender (male)
-0.001 (-0.136, 0.134)
0.99
-0.002 (-0.106, 0.103)
0.98
ESA use (yes)
0.065 (-0.077, 0.207)
0.37
-0.071 (-0.184, 0.041)
0.21
Anti-proliferative (yes) ACEI/ARB (yes)
0.014 (-0.139, 0.167) -0.003 (-0.139, 0.133)
0.86 0.97
-0.088 (-0.249, 0.073) -0.004 (-0.107, 0.098)
0.28 0.94
Functioning renal Tx (yes)
0.100 (-0.064, 0.264)
0.23
0.130 (-0.032, 0.292)
0.12
Access (AVF)
0.088 (-0.110, 0.287)
0.38
-0.031 (-0.187, 0.124)
0.69
B12 (low)
-0.083 (-0.334, 0.168)
0.51
-0.074 (-0.259, 0.110)
0.43
Folate (low)
-0.013 (-0.188, 0.163)
0.89
0.046 (-0.087, 0.178)
0.49
CI confidence interval, hsCRP high-sensitive C-reactive protein, eGFR estimated glomerular filtration rate, ESA erythropoiesis stimulating agent, ACEI angiotensin converting enzyme inhibitor, ARB angiotensin receptor blocker, Tx transplant, AVF arterio-venous fistula a b
Continuous log-transformed variable continuous linear variable. Otherwise all variables are binary
Fig. 1 Pre- and post-iron infusion mean ± standard error mean levels of haemoglobin, hepcidin-25, ferritin and transferrin saturation (TSAT). p values derived from two-tailed, paired t tests
123
J Nephrol Table 3 Univariate and multivariate analysis of predictors of haemoglobin increment following intravenous iron infusion CI confidence interval, hsCRP high sensitive C-reactive protein, eGFR estimated glomerular filtration rate, ESA erythropoiesis stimulating agent, ACEI angiotensin converting enzyme inhibitor, ARB angiotensin receptor blocker, Tx transplant, AVF arterio-venous fistula
Variable
Univariate
Multivariate
Coefficient (95 % CI)
p
Coefficient (95 % CI)
Hepcidina
-1.05 (-1.57, -0.54)
\0.001
hsCRPa
-0.58 (-0.93, -0.23)
0.001
eGFRa
0.39 (-0.52, 1.31) -0.01 (-0.02, 0.00)
-0.84 (-1.38, -0.31)
p 0.002
-0.36 (-0.75, 0.02)
0.07
0.40
0.28 (-0.73, 1.29)
0.59
0.21
0.00 (-0.02, 0.01)
0.66
-0.32 (-0.46, -0.17)
\0.001
-0.33 (-0.48, -0.19)
\0.001
0.03 (-0.08, 0.14)
0.62
0.01 (-0.09, 0.12)
0.78
Gender (male)
-0.14 (-0.57, 0.28)
0.50
0.09 (-0.32, 0.5)
0.66
ESA use (yes)
-0.31 (-0.76, 0.13)
0.16
0.05 (-0.38, 0.49)
0.81
0.37 (-0.10, 0.84)
0.13
-0.01 (-0.6, 0.59)
0.98
-0.08 (-0.51, 0.35)
0.70
0.09 (-0.3, 0.47)
0.67
0.31 (-0.20, 0.81) 0.18 (-0.47, 0.84)
0.24 0.58
0.28 (-0.32, 0.89) 0.44 (-0.17, 1.05)
0.36 0.15
Ageb b
Pre-infusion haemoglobin Iron dose/weightb
Anti-proliferative (yes) ACEI/ARB (yes)
a
Continuous log-transformed variable
Functioning renal Tx (yes) Access (AVF)
b
B12 (low)
0.90 (0.12, 1.68)
0.02
0.46 (-0.26, 1.18)
0.21
Folate (low)
0.17 (-0.38, 0.72)
0.54
0.22 (-0.29, 0.72)
0.40
continuous linear variable. Otherwise, all variables are binary
Fig. 2 Scatter plot describing the relationship between haemoglobin increment and hepcidin-25 levels and demonstrating impact of inflammation status (hsCRP \3 mg/l and hsCRP C3 mg/l). hsCRP high-sensitive C-reactive protein
some patients even when the iron markers suggested iron repletion (Table 4). Short term safety outcomes No immediate or delayed complications of intravenous iron administration were observed by the clinical team. Bilirubin, ALT and eGFR showed no significant difference from baseline at 6 weeks post-infusion (data not shown). A marginal increase in alkaline phosphatase from 79.5 l/l
(IQR 65.0–109.8) to 88.0 l/l (IQR 68.0–121.0) was seen (p \ 0.001). A decrease in serum phosphate level was observed 1.22 mmol/l (IQR 1.08–1.46) to 1.20 mmol/l (IQR 0.93–1.39) following iron infusion (p = 0.002).
Discussion This is the first study to describe the utility of serum hepcidin-25 in predicting the haemoglobin increment
123
J Nephrol
Fig. 3 Receiving operation characteristics (ROC) curve analysis of pre-iron infusion, ferritin, transferrin saturation (TSAT) and hepcidin-25 levels to 0.5 and 1 g/dl rise in haemoglobin (Hb)
Table 4 Haemoglobin levels pre- and post-intravenous iron are shown for each baseline range for: (a) ferritin, (b) transferrin saturation and (c) hepcidin-25
(a) Ferritin (ng/ml) Ferritin levels
Number of patients
\50
40
50–99
SD
Mean post iron Hb (g/dl)
SD p*
% with 0.5 g/ % with 1 g/ dl Hb rise dl Hb rise
9.9
1.5
11.3
1.7 \0.001
75
63 38
Mean pre iron Hb (g/dl)
24
10.9
1.3
11.3
1.3
0.003
58
100–149 21
10.4
1.1
11.2
1.2
0.004
57
38
150–199 18
10.7
1
11.1
0.9
0.16
33
28
200–500 16
9.6
1.4
10.4
1.1
0.02
56
31
% with 0.5 g/ dl Hb rise
% with 1 g/ dl Hb rise
(b) Transferrin saturation (TSAT %) TSAT levels
Number of patients
\10
21
10–19
46
20–24
23
25–29
19
C30
10
Mean pre iron Hb (g/dl)
p*
SD Mean post iron Hb (g/dl)
SD
9.6
1.2
11.3
1.6 \0.001
86
76
10.2
1.3
11.1
1.4 \0.001
61
43
10.9
1.4
11.6
1.7 \0.001
52
43
10.6
1.1
11
0.8
0.13
42
16
10
1.4
10.5
1.1
0.06
50
30
(c) Hepcidin-25 (ng/ml)
Also shown are the percentages of patients demonstrating a haemoglobin increment of 0.5 g/dl or 1.0 g/dl for each baseline range * Two-tailed paired t test comparing haemoglobin level pre- and post-intravenous iron
Number of patients
Mean pre iron Hb (g/ dl)
SD
Mean post iron Hb (g/dl)
SD
\20
16
10.2
1.3
11.6
1.6 \0.001
81
69
20–29
21
10.2
1.6
11.8
1.6 \0.001
81
71
30–49 50–99
23 31
10.1 10.5
1.2 1.5
10.8 11.1
1.3 \0.001 1.3 \0.001
57 48
39 32
C100
28
10.3
1.2
10.7
1.2
39
25
following bolus iron infusion in patients with non-dialysis CKD. The aim of the study was to identify an iron marker that can identify individuals most likely to benefit from
123
p*
Hepcidin levels
0.10
% with 0.5 g/dl Hb rise
% with 1 g/ dl Hb rise
intravenous iron, thereby improving the efficacy whilst minimising risk of iron toxicity. In this regard, hepcidin-25 is a potential candidate as it serves as a sensor of iron
J Nephrol
status, but also has a unique position as a biological regulator of iron absorption and utilisation. A significant relationship was found between hepcidin-25 and haemoglobin increment following protocolised iron infusion, with higher hepcidin-25 levels associated with a smaller haemoglobin increment across the studied population. However, before any suggestion that this iron marker be adopted in routine clinical practice, three caveats should be borne in mind. First, prediction was imperfect but, reassuringly in an era of ESA minimisation, even at the highest levels of serum hepcidin-25 encountered in the studied population iron infusion was often associated with a clinically relevant increase in haemoglobin. This finding is in line with the ‘DRIVE’ study which demonstrated the utility of intravenous iron in a proportion of haemodialysis patients deemed ‘‘iron-replete’’ who had received an initial increase in dose of erythropoietin at randomisation [13]. It seems that to some extent the effect of intravenous iron infusion may overwhelm or bypass the effects of iron sequestration as a result of elevated hepcidin levels, and therefore it influences marrow production of red blood cells. However, it may also be the case that iron deposition occurs elsewhere, and future studies are afoot to address the long-term safety profile of intravenous iron use in CKD [14]. The second caveat is that the prognostic performance of serum hepcidin-25 was almost identical to that of the established markers serum ferritin and TSAT (ROC-curve analysis), which for most clinicians will be more easily available and interpretable. Thus it is unlikely that hepcidin-25 has a role to play as an isolated marker of iron status in this setting. However hepcidin-25 has the advantage of being potentially mechanistic in regard to iron utilisation [15], although this will need formal assessment in prospective studies. The importance of this study is that it suggests the possibility that greater increments in haemoglobin may be achieved in this setting by antagonising the effect of high hepcidin-25 pre-iron infusion. Whilst an advantage of hepcidin antagonism may be improved efficacy of oral iron due to increased intestinal absorption, it should be remembered that hepcidin antagonism may also potentiate the effect of intravenous iron administration (i.e. where a block to intestinal absorption is not an issue) by reducing iron sequestration in the reticulo-endothelial system [16]. A study conducted in 56 haemodialysis patients failed to show an association between hepcidin-25 level and haemoglobin response to intravenous iron [5]. However, the current study describes a larger cohort of CKD patients not on dialysis, and who were therefore not subject to the blood loss and intra-individual variation of inflammation and hepcidin-25 seen in maintenance haemodialysis patients [17]. However, the relationship between hepcidin and haemoglobin increment in our cohort did depend on
prevailing levels of inflammation, with evidence that the relationship is less evident in patients displaying hsCRP C3 mg/l. Despite 21 % of the study population being transplant recipients, the use of anti-proliferative medications and ESAs did not appear to significantly affect baseline hepcidin-25 levels and post-iron infusion haemoglobin increment in multivariate analysis. Though the current study benefits from its ‘‘real world’’ setting and potential generalisability, the results would need to be reproduced in view of their heterogeneity. In summary, this study demonstrates the relationship between hepcidin-25 and traditional markers of iron storage and inflammation in patients with non-dialysis dependent CKD. It also shows an association between hepcidin25 levels and the response to intravenous iron, which may be of utility in predicting the expected response to iron. Although the prognostic performance of hepcidin-25 was non-superior to traditional iron storage markers, its mechanistic role points to a potential rationale for hepcidin antagonism in patients with CKD. Conflict of interest On behalf of all authors, the corresponding author states that there is no conflict of interest.
References 1. Pfeffer MA, Burdmann EA, Chen CY et al (2009) A trial of darbepoetin alfa in type 2 diabetes and chronic kidney disease. N Engl J Med 361:2019–2032 2. Fievet P, Coppin M, Brazier F, Lefevre M, Stephan R, Demontis R (2012) Treatment of iron deficiency in predialysis state by low molecular weight iron dextran high doses intravenously. Nephrol Ther 8:41–46 3. Zager RA, Johnson AC, Hanson SY, Wasse H (2002) Parenteral iron formulations: a comparative toxicologic analysis and mechanisms of cell injury. Am J Kidney Dis 40:90–103 4. Sarafidis PA, Rumjon A, MacLaughlin HL, Macdougall IC (2012) Obesity and iron deficiency in chronic kidney disease: the putative role of hepcidin. Nephrol Dial Transpl 27:50–57 5. Tessitore N, Girelli D, Campostrini N et al (2010) Hepcidin is not useful as a biomarker for iron needs in haemodialysis patients on maintenance erythropoiesis-stimulating agents. Nephrol Dial Transpl 25:3996–4002 6. Gama-Axelsson T, Heimburger O, Stenvinkel P, Barany P, Lindholm B, Qureshi AR (2012) Serum albumin as predictor of nutritional status in patients with ESRD. Clin J Am Soc Nephrol 7:1446–1453 7. Uehata T, Tomosugi N, Shoji T et al (2012) Serum hepcidin-25 levels and anemia in non-dialysis chronic kidney disease patients: a cross-sectional study. Nephrol Dial Transpl 27:1076–1083 8. Zaritsky J, Young B, Wang HJ et al (2009) Hepcidin––a potential novel biomarker for iron status in chronic kidney disease. Clin J Am Soc Nephrol 4:1051–1056 9. Mikhail A, Shrivastava R, Richardson D (2010) Anaemia in CKD. http://www.renal.org/clinical/guidelinessection/AnaemiaInCKD. aspx. Edited by Renal Association 2010. Accessed 12 Sept 2013
123
J Nephrol 10. Ward DG, Roberts K, Stonelake P et al (2008) SELDI–TOF-MS determination of hepcidin in clinical samples using stable isotope labelled hepcidin as an internal standard. Proteome Sci 6:28 11. Peters HP, Laarakkers CM, Swinkels DW, Wetzels JF (2010) Serum hepcidin-25 levels in patients with chronic kidney disease are independent of glomerular filtration rate. Nephrol Dial Transpl 25:848–853 12. Ashby DR, Gale DP, Busbridge M et al (2009) Plasma hepcidin levels are elevated but responsive to erythropoietin therapy in renal disease. Kidney Int 75:976–981 13. Coyne DW, Kapoian T, Suki W et al (2007) Ferric gluconate is highly efficacious in anemic hemodialysis patients with high serum ferritin and low transferrin saturation: results of the
123
14.
15. 16. 17.
Dialysis Patients’ Response to IV Iron with Elevated Ferritin (DRIVE) Study. J Am Soc Nephrol 18:975–984 Macdougall IC (2013) Ferric carboxymaltose assessment in subjects with iron deficiency anaemia and non-dialysis-dependent CKD (FIND-CKD). http://clinicaltrials.gov/ct2/show/study/NCT009943 18. Edited by ClinicalTrials.gov 2013. Accessed 21 June 2013 Coyne DW (2011) Hepcidin: clinical utility as a diagnostic tool and therapeutic target. Kidney Int 80:240–244 Macdougall IC (2012) New anemia therapies: translating novel strategies from bench to bedside. Am J Kidney Dis 59:444–451 Ford BA, Eby CS, Scott MG, Coyne DW (2010) Intra-individual variability in serum hepcidin precludes its use as a marker of iron status in hemodialysis patients. Kidney Int 78:769–773
