ABSTRACT
Background
The optimal range of serum iron markers and usefulness of iron supplementation are uncertain in patients with pre-dialysis chronic kidney disease (CKD). We investigated the association between serum iron indices and risk of cardiovascular disease (CVD) events and the effectiveness of iron supplementation using Chronic Kidney Disease Japan Cohort data.
Methods
We included 1416 patients ages 20–75 years with pre-dialysis CKD. The tested exposures were serum transferrin saturation and serum ferritin levels and the outcome measures were any cardiovascular event. Fine–Gray subdistribution hazard models were used to examine the association between serum iron indices and time to events. The multivariable fractional polynomial interaction approach was used to evaluate whether serum iron indices were effect modifiers of the association between iron supplementation and cardiovascular events.
Results
The overall incidence rate of CVD events for a median of 4.12 years was 26.7 events/1000 person-years. Patients with serum transferrin saturation <20% demonstrated an increased risk of CVD [subdistribution hazard ratio (HR) 2.13] and congestive heart failure (subdistribution HR 2.42). The magnitude of reduction in CVD risk with iron supplementation was greater in patients with lower transferrin saturations (P = .042).
Conclusions
Maintaining transferrin saturation >20% and adequate iron supplementation may effectively reduce the risk of CVD events in patients with pre-dialysis CKD.
Keywords: cardiovascular disease, CKD, iron deficiency, iron supplementation, pre-dialysis
Graphical Abstract
Graphical Abstract.
KEY LEARNING POINTS.
What was known:
Although there is a high association between anaemia and cardiovascular disease (CVD) events in pre-dialysis CKD patients, interventional trial data supporting a reduction in CVD events with agents that raise haemoglobin levels are lacking.
In pre-dialysis CKD patients with anaemia, optimal levels of serum iron markers have not been established, and there is also a lack of evidence for the clinical benefit of iron supplementation.
This study adds:
Maintaining serum transferrin saturation >20% may reduce the incidence of CVD in pre-dialysis CKD patients.
The clinical benefit of iron supplementation may be stronger in pre-dialysis CKD patients with lower serum transferrin saturation.
Potential impact:
The findings from our study may be useful in stratifying the risk of CVD in pre-dialysis CKD patients and may suggest new optimal levels of serum iron markers and the clinical benefit of iron supplementation.
INTRODUCTION
Cardiovascular disease (CVD) is the leading cause of morbidity and mortality in patients with pre-dialysis chronic kidney disease (CKD). Even patients with mild CKD have a greater risk of CVD than the age-matched general population [1, 2]. Anaemia is the most common comorbidity in patients with pre-dialysis CKD [3] and is associated with increased CVD-related mortality [4]; therefore, cardiorenal anaemia syndrome has been proposed.
A previous study reported that approximately half of CKD stage 3–5 patients with anaemia (haemoglobin <11 g/dl) who were not receiving epoetin and iron supplementation were deficient in bone marrow iron [5]. Further, a low estimated glomerular filtration rate (eGFR) is associated with low haemoglobin levels and iron deficiency [6]. Iron deficiency, defined as the depletion of iron stores in the bone marrow, is a prognostic factor for CVD independent of anaemia or CKD. There are two types of iron deficiency: absolute iron deficiency that involves low iron stores and functional iron deficiency due to iron sequestration [7]. In vivo iron stores are estimated based on transferrin saturation (TSAT) and serum ferritin levels. The Kidney Disease: Improving Global Outcomes guidelines recommend monitoring TSAT and serum ferritin levels in anaemic CKD patients [8]. The recent guidelines of the European Society of Cardiology also recommend measuring TSAT and serum ferritin levels in patients with congestive heart failure (CHF) [9]. Iron deficiency can be corrected by iron supplementation; however, patients with pre-dialysis CKD are often inadequately treated for iron deficiency [10].
Moreover, evidence regarding the optimal range of serum iron markers in patients with pre-dialysis CKD is lacking. There is also limited knowledge on the clinical benefits of iron supplementation in patients with pre-dialysis CKD and iron deficiency. Japanese patients with pre-dialysis CKD are less likely to receive iron supplementation than those in other countries [11, 12], thus this study population will allow appropriate investigation of the clinical impact of baseline iron deficiency and the effectiveness of iron supplementation. We investigated the association between serum indices of iron metabolism and the incidence of CVD events and the clinical effectiveness of iron supplementation in patients with pre-dialysis CKD using data from the Chronic Kidney Disease Japan Cohort (CKD-JAC) study.
MATERIALS AND METHODS
Study design, setting and participants
We obtained data from the CKD-JAC study, a multicentre prospective cohort study conducted in 17 clinical centres in Japan. The CKD-JAC study enrolled 2996 patients, ages 20–75 years, with pre-dialysis CKD and an estimated glomerular filtration rate (eGFR) of 10–59 ml/min/1.73 m2. We calculated eGFR based on corrected facility-measured serum creatinine levels and age using the following formula: eGFR (ml/min/1.73 m2) = 194 × age−0.287 × serum creatinine−1.094 × (0.739 for women) [13]. The exclusion criteria were the presence of polycystic kidney disease, human immunodeficiency virus infection, cirrhosis or active cancer or receipt of cancer treatment within the past 2 years; receipt of a transplant and previous chronic dialysis; pregnancy; and refusal to provide informed consent. Detailed information on the CKD-JAC study protocol is available elsewhere [14].
Exposures to be tested
We used baseline TSAT and serum ferritin levels obtained after study enrolment as indices of serum iron metabolism. We used the following formula to calculate TSAT: TSAT = [serum iron concentration (Fe)/total iron-binding capacity (TIBC)] × 100%. We set TSAT and serum ferritin as exposures to be tested and evaluated their association with CVD outcomes. Considering the clinically used thresholds [15] and data distribution, we categorised TSAT levels as <20, 20–30, 30–40 and >40% (reference) and serum ferritin levels as <50 , 50–100 , 100–200 and >200 ng/ml (reference).
Outcome measures
The main outcome measures were CVD events, including fatal or non-fatal myocardial infarction, CHF, angina pectoris, arrhythmia, aortic dissection, cerebrovascular disorder and peripheral artery disease, observed during the study period {median follow-up duration 4.12 years [interquartile range (IQR) 3.3–9.8]}. These events were identified at each facility and adjudicated by an independent cardiac events evaluation committee.
Statistical analysis
Binary variables are expressed as frequencies and percentages for each category. Continuous variables are presented as mean [standard deviation (SD)] or median (IQR). The incidence of outcome measures was compared among the four TSAT categories. Event-free survival for CVD and CHF was calculated using the Kaplan–Meier method and compared among the four categories of TSAT and serum ferritin levels. Between-group differences in the incidence of events were examined using the logrank trend test.
Fine–Gray subdistribution hazard models were employed to examine the association between serum TSAT or ferritin levels and time to events with death unrelated to CVD and the transition to renal replacement therapy (RRT)—i.e. developing end-stage kidney disease (ESKD)—considered competing risks. All models were stratified by facilities and adjusted for age, sex, body mass index (BMI), smoking status, systolic blood pressure (SBP), diabetes mellitus, history of any CVD, eGFR, urinary albumin:creatinine ratio (UACR), haemoglobin, serum albumin level, serum calcium level, serum phosphorus level, intact parathyroid hormone (iPTH) levels, erythropoietin-stimulating agents (ESAs), iron supplementation and the use of renin–angiotensin system (RAS) inhibitors, β-blockers and calcium channel blockers (CCBs). We also performed sensitivity analysis by adding C-reactive protein (CRP) as a covariate.
Multiple imputations were used to account for missing data. The covariates included in our imputation models were baseline age; sex; eGFR; BMI; SBP and diastolic blood pressure (DBP); log-transformed UACR; haemoglobin; TSAT; log-transformed ferritin; serum albumin level; serum calcium level; serum phosphate level; log-transformed PTH; history of diabetes mellitus and CVD; smoking status; use of RAS inhibitors, β-blockers and CCBs; ESAs; iron supplementation and outcome variables. We conducted chained equations with five imputations using logistic regression for categorical variables and linear regression for continuous variables and then combined the results across the five imputed datasets using Rubin's formula [16].
Considering the possibility that the association between each of the iron indices (TSAT, ferritin, TIBC, and Fe) and outcomes may be non-linear, we used a restricted cubic spline regression model. We modelled TSAT using restricted cubic spline analysis with four knots (at 20%, 30%, 40% and 50%) and analysed using the multivariable Cox proportional hazards model after adjusting for age, sex, BMI, smoking status, SBP, diabetes mellitus, history of CVD, eGFR, UACR, haemoglobin, serum albumin level, serum calcium level, serum phosphorus level, iPTH level, ESAs, iron supplementation and use of RAS inhibitors, β-blockers and CCBs. Thresholds for ferritin, TIBC and Fe were set at 100 ng/ml, 200 µg/dl and 13 µmol/L (73 µg/dl), respectively [17].
We also applied the multivariable fractional polynomial interaction (MFPI) approach to investigate whether TSAT levels were an effect modifier of the association between iron supplementation and the outcomes [18]. All statistical tests were two-sided and statistical significance was set at P < .05. All statistical analyses were performed using Stata/MP version 17 (StataCorp, College Station, TX, USA).
Ethical considerations
The CKD-JAC study was approved by each facility's institutional review board. This study was conducted following the tenets of the Declaration of Helsinki and any subsequent amendments or equivalent ethical standards. Each participant provided informed consent in written form according to institutional review board and institutional requirements.
RESULTS
Study participants
Fig. 1 shows the selection process of the participants analysed using the CKD-JAC study data. The total number of participants recruited for the CKD-JAC was 3087, of whom 121 were excluded; 2966 participants were ultimately enrolled. Since serum TSAT and serum ferritin levels were the primary exposures to be tested, we excluded 1550 participants without these values. Data from 1416 participants were included in the final analysis.
Figure 1:
Selection process of the analysed participants.
Table 1 summarises the baseline characteristics of the participants according to their serum TSAT levels (<20, 20–30, 30–40 and ≥40%). The mean age of the participants was 61 years (SD 11), 62.0% were male, 41.7% had diabetes mellitus and 27.5% had CVD. Further, mean eGFR was 29.0 ml/min/1.73 m2 (SD 12), mean UACR was 533 mg/g Cr (IQR 123–1406) and mean SBP was 131 mmHg (SD 19). Their mean haemoglobin and serum ferritin levels were 12.0 g/dl (SD 1.9) and 101 ng/ml (IQR 51–183), respectively. Overall, 16.5% of the participants were administered ESAs. Iron was prescribed to 9.7% of all participants, with >86% administered orally.
Table 1:
Participants’ baseline characteristics by serum TSAT levels.
| Characteristics | Patients, n | Total (N = 1416) | <20 (n = 249) | 20–30 (n = 497) | 30–40 (n = 397) | >40 (n = 273) | P-value |
|---|---|---|---|---|---|---|---|
| Age (years), mean (SD) | 1416 | 61 (11) | 61 (12) | 62 (11) | 62 (11) | 59 (13) | .043 |
| Male, n (%) | 1416 | 878 (62.0) | 130 (52.2) | 299 (60.2) | 275 (69.3) | 174 (63.7) | <.001 |
| Diabetes mellitus, n (%) | 1416 | 591 (41.7) | 104 (41.8) | 212 (42.7) | 180 (45.3) | 95 (34.8) | .053 |
| History of any CVD, n (%) | 1416 | 390 (27.5) | 79 (31.7) | 144 (29.0) | 103 (25.9) | 64 (23.4) | .14 |
| Smoker, n (%) | 1226 | .18 | |||||
| Never smoker, n (%) | 651 (53.1) | 125 (59.8) | 219 (51.0) | 184 (52.9) | 123 (51.2) | ||
| Active or ex-smoker, n (%) | 575 (46.9) | 84 (40.2) | 210 (49.0) | 164 (47.1) | 117 (48.8) | ||
| BMI (kg/m2) , mean (SD) | 1305 | 23.6 (3.8) | 23.1 (3.7) | 24.0 (3.8) | 23.7 (3.6) | 23.0 (3.9) | .002 |
| eGFR (ml/min/1.73 m2), mean (SD) | 1416 | 29 (12) | 29 (12) | 29 (13) | 29 (12) | 27 (13) | .32 |
| CKD stage, n (%) | 1416 | .18 | |||||
| 3a | 155 (10.9) | 23 (9.2) | 62 (12.5) | 45 (11.3) | 25 (9.2) | ||
| 3b | 470 (33.2) | 92 (36.9) | 161 (32.4) | 133 (33.5) | 84 (30.8) | ||
| 4 | 554 (39.1) | 99 (39.8) | 184 (37.0) | 165 (41.6) | 106 (38.8) | ||
| 5 | 237 (16.7) | 35 (14.1) | 90 (18.1) | 54 (13.6) | 58 (21.2) | ||
| UACR (mg/g), median (IQR) | 1324 | 533 (123–1406) | 373 (60–1113) | 598 (171–1456) | 574 (123–1399) | 513 (107–1407) | .046 |
| UACR category, n (%) | 1324 | .056 | |||||
| <30 | 148 (11.2) | 35 (15.4) | 43 (9.3) | 45 (12.0) | 25 (9.6) | ||
| 30–300 | 360 (27.2) | 70 (30.8) | 119 (25.7) | 93 (24.9) | 78 (30.0) | ||
| >300 | 816 (61.6) | 122 (53.7) | 301 (65.0) | 236 (63.1) | 157 (60.4) | ||
| SBP (mmHg), mean (SD) | 1398 | 131 (19) | 131 (19) | 132 (19) | 131 (18) | 129 (18) | .17 |
| DBP (mmHg), mean (SD) | 1396 | 75 (12) | 75 (12) | 75 (12) | 75 (12) | 75 (12) | .93 |
| Haemoglobin (g/dl), mean (SD) | 1412 | 12.0 (1.9) | 11.4 (1.8) | 12.0 (1.8) | 12.3 (1.8) | 12.0 (2.0) | <.001 |
| TSAT, n (%) | 1416 | 30.5 (12.1) | 14.8 (3.9) | 25.1 (2.8) | 34.5 (2.8) | 48.9 (9.0) | <.001 |
| Serum ferritin (ng/ml), median (IQR) | 1416 | 101 (51–183) | 51 (20–109) | 89 (51–158) | 122 (66–211) | 150 (87–241) | <.001 |
| Albumin (mg/dl), mean (SD) | 1394 | 4.0 (0.4) | 4.0 (0.4) | 4.0 (0.4) | 4.0 (0.4) | 3.9 (0.5) | .003 |
| CRP (mg/dl), median (IQR) | 1373 | 0.1 (0.0–0.2) | 0.2 (0.1–0.5) | 0.1 (0.0–0.2) | 0.1 (0.0–0.2) | 0.1 (0.0–0.1) | <.001 |
| Serum calcium (mg/dl), mean (SD) | 1373 | 9.0 (0.5) | 9.0 (0.5) | 9.0 (0.4) | 9.0 (0.4) | 9.0 (0.5) | .61 |
| Serum phosphate (mg/d), mean (SD) | 1393 | 3.6 (0.7) | 3.6 (0.7) | 3.6 (0.7) | 3.5 (0.7) | 3.6 (0.8) | .22 |
| iPTH (pg/ml), median (IQR) | 1336 | 82 (53–128) | 83 (52–134) | 88 (55–133) | 73 (51–117) | 82 (52–126) | .047 |
| Antihypertensive agents, n (%) | 1416 | 1304 (92.1) | 224 (90.0) | 456 (91.8) | 368 (92.7) | 256 (93.8) | .41 |
| ACEI/ARB | 1416 | 1171 (82.7) | 200 (80.3) | 404 (81.3) | 337 (84.9) | 230 (84.2) | .33 |
| β-blockers | 1416 | 299 (21.1) | 57 (22.9) | 100 (20.1) | 78 (19.6) | 64 (23.4) | .54 |
| CCBs | 1416 | 802 (56.6) | 141 (56.6) | 299 (60.2) | 215 (54.2) | 147 (53.8) | .22 |
| ESA | 1416 | 234 (16.5) | 39 (15.7) | 60 (12.1) | 57 (14.4) | 78 (28.6) | <.001 |
| Iron | 1416 | 138 (9.7) | 28 (11.2) | 42 (8.5) | 24 (6.0) | 44 (16.1) | <.001 |
| Oral iron | 1416 | 121 (8.5) | 23 (9.2) | 38 (7.6) | 22 (5.5) | 38 (13.9) | .002 |
| Intravenous iron | 1416 | 18 (1.3) | 5 (2.0) | 4 (0.8) | 2 (0.5) | 7 (2.6) | .059 |
All models were stratified by facilities and adjusted for potential confounders as follows: age, sex, BMI, smoking status, SBP, diabetes mellitus, history of any CVD, eGFR, UACR, haemoglobin, serum albumin, serum calcium, serum phosphorus, iPTH, ESAs, iron supplementation, RAS inhibitors, β-blockers and CCBs.
Supplementary Table S1 shows a comparison of the baseline characteristics between participants with and without values for serum iron indices. Participants with values for serum iron indices were more likely to present with diabetes mellitus and a CVD history. The incident rates of CVD and CHF for both groups are presented in Supplementary Table S2. CVD and CHF incident rates (per 1000 person-years) in participants with available iron parameters were higher than those in participants without available iron parameters (26.5 versus 18.4 and 12.0 versus 5.8, respectively).
Incidence of outcome measures
Supplementary Fig. S1 shows the incidence rates of CVD and CHF events based on serum TSAT levels. The overall incidence rates of CVD and CHF in the analysed participants were 26.7 and 12.0 events/1000 person-year, respectively. Participants with TSAT <20% had the highest incidence rates of CVD and CHF (33.9 and 16.5 events/1000 person-year, respectively). Conversely, the incidence rates of CVD and CHF were lowest in those with TSAT >40% (17.2 and 6.2 events/1000 person-year, respectively).
Event-free survival curves for CVD and CHF
We calculated the event-free survival curves for each outcome (CVD and CHF) according to TSAT and serum ferritin categories (Fig. 2). We observed significant differences in the estimated probability of developing CVD (Fig. 2A) or CHF (Fig. 2B) events using the Kaplan–Meier method with the logrank trend test in the TSAT categories. However, among the serum ferritin categories, there was no significant difference in either CVD (Fig. 2C) or CHF (Fig. 2D) event probability.
Figure 2:
Event-free survival curves for CVD/CHF by serum iron indices categories.
TSAT and serum ferritin levels and the risk of CVD and CHF
Fig. 3 shows the risk of CVD and CHF according to TSAT and serum ferritin levels. Compared with patients in the TSAT >40% category, those in the TSAT <20% category demonstrated an increased risk of CVD events {subdistribution hazard ratio [SHR] 2.13 [95% confidence interval (CI) 1.24–3.66]} and CHF events [SHR 2.42 (95% CI 1.02–5.74)], respectively. The risk of CVD and CHF events was not significantly increased in the TSAT categories of 20–30% and 30–40% compared with the risk in the TSAT >40% category. There was no association between serum ferritin levels and the risk of developing CVD or CHF.
Figure 3:
TSAT or serum ferritin levels and CVD or CHF risk.
Fine–Gray subdistribution hazard models were employed to examine the association between serum TSAT or ferritin levels and time to event with the consideration of death not due to CVD and dialysis initiation, i.e. developing ESKD as competing risks. All models were stratified by facilities and adjusted for potential confounders as follows: age, sex, BMI, smoking status, SBP, diabetes mellitus, history of any CVD, eGFR, UACR, haemoglobin, serum albumin, serum calcium, serum phosphorus, iPTH, ESAs, iron supplementation, RAS inhibitors, β-blockers and CCBs.
Sensitivity analyses for the association between the serum TSAT levels and risk of CVD and CHF using multiple imputations
Table 2 presents the sensitivity analysis results for the association between TSAT and CVD and CHF risk with CRP, an inflammatory marker, as a covariate (Model 4). By adding log-transformed CRP, the point estimate of the SHR for CVD and CHF was slightly reduced in the TSAT <20% category compared with that in the TSAT >40% category. The results of the other models with multiple imputations showed the robustness of the analysis (Table 2).
Table 2:
Sensitivity analyses for the association of TSAT with CVD and CHF risks using multiple imputations.
| Unadjusted | Model 1 | Model 2 | Model 3 | Model 4 | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Outcome | TSAT category | SHR (95% CI) | P-value | SHR (95% CI) | P-value | SHR (95% CI) | P-value | SHR (95% CI) | P-value | SHR (95% CI) | P-value |
| CVD | |||||||||||
| <20 | 2.31 (1.38–3.86) | .002 | 2.13 (1.26–3.77) | .005 | 2.13 (1.23–3.69) | .007 | 2.13 (1.24–3.66) | .006 | 2.05 (1.19–3.55) | .010 | |
| 20–30 | 1.7 (1.04–2.76) | .033 | 1.46 (0.88–2.43) | .15 | 1.42 (0.85–2.39) | .18 | 1.41 (0.84–2.35) | .19 | 1.4 (0.84–2.33) | .20 | |
| 30–40 | 1.88 (1.15–3.08) | .012 | 1.62 (0.97–2.70) | .066 | 1.60 (0.96–2.69) | .074 | 1.58 (0.94–2.65) | .082 | 1.59 (0.95–2.67) | .080 | |
| >40 | Reference | Reference | Reference | Reference | Reference | ||||||
| CHF | |||||||||||
| <20 | 3.07 (1.36–6.92) | .007 | 2.56 (1.07–6.08) | .034 | 2.56 (1.09–6.05) | .031 | 2.42 (1.02–5.74) | .045 | 2.39 (0.99–5.78) | .053 | |
| 20–30 | 2.65 (1.23–5.68) | .012 | 2.16 (0.97–4.80) | .058 | 2.17 (0.98–4.81) | .057 | 1.99 (0.88–4.50) | .10 | 1.98 (0.88–4.48) | .10 | |
| 30–40 | 2.07 (0.93–4.63) | .075 | 1.77 (0.78–4.04) | .17 | 1.77 (0.77–4.05) | .18 | 1.63 (0.71–3.76) | .25 | 1.64 (0.71–3.78) | .25 | |
| >40 | Reference | Reference | Reference | Reference | Reference | ||||||
Model 1 was adjusted for age, sex, haemoglobin, eGFR, log-transformed UACR, smoking status, BMI, serum albumin, SBP, diabetes mellitus, any CVD history, serum phosphate, corrected serum calcium and log-transformed PTH.
Model 2 was adjusted for Model 1 + ACEIs/ARBs, β-blockers and CCBs.
Model 3 was adjusted for Model 2 + iron supplementation and use of ESAs.
Model 4 was adjusted for Model 3 + log-transformed CRP.
Analysis of the association between serum TSAT levels and risk of CVD and CHF using a restricted cubic spline regression model
Supplementary Fig. S2 shows the results of a restricted cubic spline regression model that considers the possibility that the association between TSAT level and CVD and CHF risk is non-linear. Four knots were set at the prespecified cut-off points of TSAT levels of 20%, 30%, 40% and 50%. The vertical dashed lines include the prespecified cut-off points of the continuous predictors; the horizontal dashed line at the HR of 1 denotes equivalent treatment effects. As shown in this figure, the risk of CVD and CHF was lowest at a TSAT of 40% and increased below a TSAT of 20%. An increasing trend in the risk of CVD and CHF was also observed above a TSAT of 50%, although this result was not statistically significant.
Effect of iron supplementation on the risk of CVD by TSAT levels using the MFPI approach
Fig. 4 shows the effect of iron supplementation on CVD risk reflected by serum TSAT levels using the MFPI approach. The vertical dashed lines include prespecified cut-off points of TSAT (20% and 40%); the horizontal line at the log(HR) of 0 denotes equivalent treatment effects. Thus a treatment effect function parallel to the horizontal line indicates no treatment interaction. Log(HRs) <0 indicate that patients who received iron supplementation had a lower risk of developing CVD events than those who did not receive iron supplementation. This MFPI analysis suggested greater risk reduction for CVD by iron supplementation in patients with pre-dialysis CKD having lower TSAT levels (P for interaction = .042).
Figure 4:
Effect of iron supplementation on CVD risk by TSAT levels.
TSAT levels and the risk of CVD stratified by haemoglobin categories
We performed a stratified analysis of the association between TSAT levels and the risk of CVD by haemoglobin categories (Supplementary Table S3). The interaction test for haemoglobin and TSAT was not statistically significant (P = .072).
Serum Fe level and CVD risk using a restricted cubic spline regression model
Supplementary Fig. S3 shows the association between serum Fe levels and CVD risk using a restricted cubic spline regression model. We set a threshold for Fe at 13 μmol/l (73 μg/dl) based on the criteria of the ‘definition of iron deficiency’ [17], and a trend towards increased CVD risk was observed at Fe levels below this threshold.
Serum TIBC level and CVD risk using a restricted cubic spline regression model
An association between TIBC levels and CVD risk was not found using a restricted cubic spline regression model (Supplementary Fig. S4).
Serum ferritin level and CVD risk using a restricted cubic spline regression model
No association between ferritin levels and CVD risk was observed in a restricted cubic spline regression model (Supplementary Fig. S5).
Comparison of CVD incident rates stratified by more detailed TSAT categories
We compared CVD incident rates with more detailed categories of TSAT levels (<15%, 15–20%, 20–30%, 30–40%, 40–50% and >50%) in Supplementary Table S4. In the CKD-JAC data, the CVD incidence rate was not the highest in patients with a TSAT <15%.
DISCUSSION
We examined the association between serum iron indices and CVD incidence using data from the CKD-JAC study, a multicentre prospective cohort of Japanese patients with pre-dialysis CKD. The results suggested an association between increased CVD risk in patients with pre-dialysis CKD and low TSAT levels (<20%). In contrast, no association was observed between ferritin levels and CVD risk. Furthermore, the association between low TSAT levels and increased CVD risk in patients with pre-dialysis CKD was robust in sensitivity analyses performed with different regression models, multiple imputations and in the non-linear regression analysis. Interaction analysis showed that the reduction in CVD risk with iron supplementation in patients with pre-dialysis CKD was greater in patients with lower TSAT levels.
Differences in the serum levels of iron markers and practice patterns of iron supplementation in patients with pre-dialysis CKD have been observed across countries. The Chronic Kidney Disease Outcomes and Practice Patterns Study (CKDOPPS) [12] conducted on patients with pre-dialysis CKD in France, Germany, Brazil and the USA reported that 38.7% of patients had TSAT <20%. In our study, the percentage of patients with TSAT <20% was 17.6%, indicating that Japanese patients with pre-dialysis CKD were less likely to be iron deficient than those in other countries. In the CKDOPPS [10], ≈40% of patients with pre-dialysis CKD were administered iron, whereas in our study (CKD-JAC), <10% were administered iron. In our study, among patients with TSAT <20%, 10% and 2% of them received oral and intravenous iron, respectively. Another CKDOPPS [19] reported that 26% of patients received oral iron and 6% received intravenous iron among patients with TSAT <20%.
It is biologically plausible that iron deficiency interferes with myocardial function and affects cardiovascular outcomes independent of the haemoglobin level [20]. The negative effects of intramyocardial iron deficiency have been confirmed in animal studies using haemojuvelin knockout mice [21] and in human cardiomyocytes from patients with heart failure [22]. Iron deficiency increases platelet aggregation [23] and thrombosis risk [24]. Our results showed that an iron deficiency status—indicated by TSAT <20%—increased the risk of subsequent CVD in patients with pre-dialysis CKD. In contrast, serum ferritin levels were not associated with CVD risk in patients with pre-dialysis CKD. Serum ferritin levels are reportedly influenced by inflammation and are not suitable iron deficiency markers [25]. An observational study involving 975 patients with pre-dialysis CKD showed that TSAT <10% was associated with adverse outcomes, including CVD, independent of serum ferritin levels [26]. Another recent observational study in a veteran cohort [27] found that the group with low TSAT (<16%) had a higher risk of major adverse cardiovascular events (MACEs) than the control group (TSAT 16–28%), consistent with our results. Similar findings have been reported in recent years in a CKDOPPS [12]. In the previous cohort study of 5145 patients with pre-dialysis CKD, the low TSAT group (TSAT <15%) had a higher risk of MACEs than the control group (TSAT 26–35%). The study also showed that the lowest risk of MACEs was associated with a TSAT of 40% in a spline regression analysis. This threshold for TSAT is consistent with the results of our analysis. These results suggest the need to meet the iron demand adequately in patients with pre-dialysis CKD. The most recent observational study of outpatients with CHF showed that low serum iron levels were associated with an increased risk of all-cause mortality but not with the risk of CVD-associated death [28]. We also examined the association between serum Fe levels and CVD risk with a multivariate analysis using a spline model (Supplementary Fig. S3). There was increased CVD risk at Fe levels below the threshold of 13 μmol/L (73 μg/dl) based on the criteria of the ‘definition of iron deficiency’ [17].
Among patients with pre-dialysis CKD, our results suggest greater CVD risk reduction with iron supplementation in patients with lower TSAT. The possible underlying mechanisms include inhibition of both platelet aggregation and vascular calcification via iron supplementation [29, 30]. To date, the association of intravenous iron levels with reduced CVD risk has been shown in a randomised controlled trial (RCT), systematic reviews and meta-analyses conducted in patients with CHF [31–33]. These results are consistent with the findings of a previous study that reported the amelioration of myocardial dysfunction due to intramyocardial iron deficiency in humans with iron supplementation [22]. There is a lack of evidence regarding oral iron agents, and a recent RCT examined the clinical efficacy of ferric citrate in patients with pre-dialysis CKD [34]. The results of this RCT showed that the ferric citrate group had a lower risk of a composite endpoint (death, RRT or renal transplantation) than the usual care group.
Our study has some limitations. First, there was a likely sampling/selection bias. The CKD-JAC was created from 17 tertiary care hospitals in Japan. Therefore it may be difficult to generalise the study results to other settings or countries. We limited our analyses to patients with recorded values of serum iron indices (TSAT and serum ferritin). As shown in Supplementary Tables S1 and S2, participants with available iron parameters had a higher prevalence of diabetes and CVD compared with participants without available iron parameters. The incident rates of CVD and CHF were also greater in this group. Therefore, since this study only included participants with available iron parameters, the results may be affected by selection bias. Second, we only measured serum iron indices at baseline. Thus we could not consider the effects of changes in iron markers due to therapeutic interventions during the observation period. Third, there could be bias due to residual confounders. We believe that the analysis in the present study covered most of the confounders presented in similar studies. However, because this was an observational study, we could not control residual bias due to unmeasurable or unknown confounders. Finally, the differences in serum iron marker levels and CVD incidence across countries, as discussed previously, can limit the generalisability of our results. However, we believe that the CKD-JAC population received less iron supplementation than those in other countries, thus allowing appropriate examination of the association between baseline serum iron indices and CVD events and the effectiveness of iron supplementation.
Our findings suggest that maintaining TSAT >20% could reduce the risk of developing CVD and CHF in patients with pre-dialysis CKD. Our study also suggests that iron supplementation reduces CVD risk in patients with lower TSAT levels. Based on the present study results, the assessment of iron deficiency status could potentially help stratify CVD risk in patients with pre-dialysis CKD. Therefore it is necessary to conduct multicentre trials in the future to evaluate the optimal target values of serum iron indices and examine the clinical benefits of iron supplementation in patients with pre-dialysis CKD.
Supplementary Material
ACKNOWLEDGEMENTS
The CKD-JAC study was supported by research funds with no restriction on publication from Kyowa Hakko Kirin. The authors wish to thank the study participants and personnel.
Contributor Information
Takeshi Hasegawa, Showa University Research Administration Center (SURAC), Showa University, Tokyo, Japan; Division of Nephrology, Department of Medicine, School of Medicine; Department of Hygiene, Public Health, and Preventive Medicine, School of Medicine, Showa University, Tokyo, Japan; Center for Innovative Research for Communities and Clinical Excellence, Fukushima Medical University, Fukushima, Japan.
Takahiro Imaizumi, Department of Advanced Medicine, Nagoya University Hospital, Nagoya, Japan; Department of Nephrology, Nagoya University Graduate School of Medicine, Nagoya, Japan.
Takayuki Hamano, Department of Nephrology, Nagoya City University Graduate School of Medicine, Nagoya, Japan; Department of Nephrology, Osaka University Graduate School of Medicine, Suita, Japan.
Kenta Murotani, Biostatistics Center, Kurume University, Kurume, Japan.
Naohiko Fujii, Medical and Research Center for Nephrology and Transplantation, Hyogo Prefectural Nishinomiya Hospital, Nishinomiya, Japan.
Hirotaka Komaba, Division of Nephrology, Endocrinology and Metabolism, Tokai University School of Medicine, Isehara, Japan.
Masahiko Ando, Department of Advanced Medicine, Nagoya University Hospital, Nagoya, Japan.
Shoichi Maruyama, Department of Nephrology, Nagoya University Graduate School of Medicine, Nagoya, Japan.
Masaomi Nangaku, Division of Nephrology and Endocrinology, University of Tokyo Hospital, Tokyo, Japan.
Kosaku Nitta, Department of Medicine, Kidney Center, Tokyo Women's Medical University, Tokyo, Japan.
Hideki Hirakata, Fukuoka Renal Clinic, Fukuoka, Japan and .
Yoshitaka Isaka, Department of Nephrology, Osaka University Graduate School of Medicine, Suita, Japan.
Takashi Wada, Division of Nephrology and Laboratory Medicine, Kanazawa University, Kanazawa, Japan.
Masafumi Fukagawa, Division of Nephrology, Endocrinology and Metabolism, Tokai University School of Medicine, Isehara, Japan.
FUNDING
This work was supported by research grants from Kyowa Kirin without publication restrictions.
AUTHORS’ CONTRIBUTIONS
T.Hasegawa, T.I., T.Hamano were responsible for the research concept and study design and protocol drafting. T.I., K.M. and M.A. were responsible for data acquisition and analysis. T.Hasegawa, T.I., T.Hamano, N.F. and H.K. were responsible for interpretation and analysis. M.N., K.N., H.H., Y.I., T.W. and M.F. were responsible for supervision. Each author contributed important intellectual content in drafting the manuscript and approved the accuracy and integrity of the final version of the manuscript.
DATA AVAILABILITY STATEMENT
The data underlying this article will not be made publicly available due to the privacy of the study participants. These data will be shared upon reasonable request to the corresponding author.
CONFLICT OF INTEREST STATEMENT
T.Hasegawa, H.K., M.N., K.N., Y.I. and TW have consultancy agreements with and received honoraria from Kyowa Hakko Kirin. T.I. received a research grant and honoraria from Kyowa Hakko Kirin. M.A. received a research grant from Kyowa Hakko Kirin. T.Hamano and M.F. have consultancy agreements with Kyowa Hakko Kirin and received honoraria and a research grant from Kyowa Hakko Kirin. All remaining authors have nothing to disclose.
REFERENCES
- 1. Go AS, Chertow GM, Fan D et al. Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization. N Engl J Med 2004;351:1296–305. 10.1056/NEJMoa041031 [DOI] [PubMed] [Google Scholar]
- 2. Meisinger C, Döring A, Löwel H et al. Chronic kidney disease and risk of incident myocardial infarction and all-cause and cardiovascular disease mortality in middle-aged men and women from the general population. Eur Heart J 2006;27:1245–50. 10.1093/eurheartj/ehi880 [DOI] [PubMed] [Google Scholar]
- 3. Astor BC, Muntner P, Levin A et al. Association of kidney function with anemia: the Third National Health and Nutrition Examination Survey (1988–1994). Arch Intern Med 2002;162:1401–8. 10.1001/archinte.162.12.1401 [DOI] [PubMed] [Google Scholar]
- 4. McClellan WM, Flanders WD, Langston RD et al. Anemia and renal insufficiency are independent risk factors for death among patients with congestive heart failure admitted to community hospitals: a population-based study. J Am Soc Nephrol 2002;13:1928–36. 10.1097/01.asn.0000018409.45834.fa [DOI] [PubMed] [Google Scholar]
- 5. Stancu S, Stanciu A, Zugravu A et al. Bone marrow iron, iron indices, and the response to intravenous iron in patients with non-dialysis-dependent CKD. Am J Kidney Dis 2010;55:639–47. 10.1053/j.ajkd.2009.10.043 [DOI] [PubMed] [Google Scholar]
- 6. Klip IT, Jankowska EA, Enjuanes C et al. The additive burden of iron deficiency in the cardiorenal-anaemia axis: scope of a problem and its consequences. Eur J Heart Fail 2014;16:655–62. 10.1002/ejhf.84 [DOI] [PubMed] [Google Scholar]
- 7. Babitt JL, Lin HY. Mechanisms of anemia in CKD. J Am Soc Nephrol 2012;23:1631–4. 10.1681/ASN.2011111078 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Summary of recommendation statements. Kidney Int Suppl 2012;2:283–7. 10.1038/kisup.2012.41 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. McDonagh TA, Metra M, Adamo M et al. 2021 ESC guidelines for the diagnosis and treatment of acute and chronic heart failure. Eur Heart J 2021;42:3599–726. 10.1093/eurheartj/ehab368 [DOI] [PubMed] [Google Scholar]
- 10. Wong MMY, Tu C, Li Y et al. Anemia and iron deficiency among chronic kidney disease Stages 3–5ND patients in the Chronic Kidney Disease Outcomes and Practice Patterns Study: often unmeasured, variably treated. Clin Kidney J 2020;13:613–24. 10.1093/ckj/sfz091 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Akizawa T, Makino H, Matsuo S et al. Management of anemia in chronic kidney disease patients: baseline findings from Chronic Kidney Disease Japan Cohort Study. Clin Exp Nephrol 2011;15:248–57. 10.1007/s10157-010-0396-7 [DOI] [PubMed] [Google Scholar]
- 12. Guedes M, Muenz DG, Zee J et al. Serum biomarkers of iron stores are associated with increased risk of all-cause mortality and cardiovascular events in nondialysis CKD patients, with or without anemia. J Am Soc Nephrol 2021;32:2020–30. 10.1681/ASN.2020101531 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Matsuo S, Imai E, Horio M et al. Revised equations for estimated GFR from serum creatinine in Japan. Am J Kidney Dis 2009;53:982–92. 10.1053/j.ajkd.2008.12.034 [DOI] [PubMed] [Google Scholar]
- 14. Imai E, Matsuo S, Makino H et al. Chronic Kidney Disease Japan Cohort (CKD-JAC) study: design and methods. Hypertens Res 2008;31:1101–7. 10.1291/hypres.31.1101 [DOI] [PubMed] [Google Scholar]
- 15. Yamamoto H, Nishi S, Tomo T et al. 2015 Japanese Society for Dialysis Therapy: guidelines for renal anemia in chronic kidney disease. Ren Replace Ther 2017;3:36. 10.1186/s41100-017-0114-y [DOI] [Google Scholar]
- 16. White IR, Royston P, Wood AM. Multiple imputation using chained equations: issues and guidance for practice. Stat Med 2011;30:377–99. 10.1002/sim.4067 [DOI] [PubMed] [Google Scholar]
- 17. Grote Beverborg N, Klip IT, Meijers WC et al. Definition of iron deficiency based on the gold standard of bone marrow iron staining in heart failure patients. Circ Heart Fail 2018;11:e004519. 10.1161/CIRCHEARTFAILURE.117.004519 [DOI] [PubMed] [Google Scholar]
- 18. Royston P, Sauerbrei W. Interaction of treatment with a continuous variable: simulation study of significance level for several methods of analysis. Stat Med 2013;32:3788–803. 10.1002/sim.5813 [DOI] [PubMed] [Google Scholar]
- 19. Lopes MB, Tu C, Zee J et al. A real-world longitudinal study of anemia management in non-dialysis-dependent chronic kidney disease patients: a multinational analysis of CKDopps. Sci Rep 2021;11:1784. 10.1038/s41598-020-79254-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Beard JL. Iron biology in immune function, muscle metabolism and neuronal functioning. J Nutr 2001;131:568S–80S. 10.1093/jn/131.2.568S [DOI] [PubMed] [Google Scholar]
- 21. Xu W, Barrientos T, Mao L et al. Lethal cardiomyopathy in mice lacking transferrin receptor in the heart. Cell Rep 2015;13:533–45. 10.1016/j.celrep.2015.09.023 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Haddad S, Wang Y, Galy B et al. Iron-regulatory proteins secure iron availability in cardiomyocytes to prevent heart failure. Eur Heart J 2017;38:362–72. 10.1093/eurheartj/ehw333 [DOI] [PubMed] [Google Scholar]
- 23. Tang X, Fang M, Cheng R et al. Iron-deficiency and estrogen are associated with ischemic stroke by up-regulating transferrin to induce hypercoagulability. Circ Res 2020;127:651–63. 10.1161/CIRCRESAHA.119.316453 [DOI] [PubMed] [Google Scholar]
- 24. Chang YL, Hung SH, Ling W et al. Association between ischemic stroke and iron-deficiency anemia: a population-based study. PLoS One 2013;8:e82952. 10.1371/journal.pone.0082952 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Ferrari P, Kulkarni H, Dheda S et al. Serum iron markers are inadequate for guiding iron repletion in chronic kidney disease. Clin J Am Soc Nephrol 2011;6:77–83. 10.2215/CJN.04190510 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Eisenga MF, Nolte IM, van der Meer P et al. Association of different iron deficiency cut-offs with adverse outcomes in chronic kidney disease. BMC Nephrol 2018;19:225. 10.1186/s12882-018-1021-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Cho ME, Hansen JL, Sauer BC et al. Heart failure hospitalization risk associated with iron status in veterans with CKD. Clin J Am Soc Nephrol 2021;16:522–31. 10.2215/CJN.15360920 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Masini G, Graham FJ, Pellicori P et al. Criteria for iron deficiency in patients with heart failure. J Am Coll Cardiol 2022;79:341–51. 10.1016/j.jacc.2021.11.039 [DOI] [PubMed] [Google Scholar]
- 29. Shovlin CL, Chamali B, Santhirapala V et al. Ischaemic strokes in patients with pulmonary arteriovenous malformations and hereditary hemorrhagic telangiectasia: associations with iron deficiency and platelets. PLoS One 2014;9:e88812. 10.1371/journal.pone.0088812 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Seto T, Hamada C, Tomino Y. Suppressive effects of iron overloading on vascular calcification in uremic rats. J Nephrol 2014;27:135–42. 10.1007/s40620-014-0046-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Ponikowski P, van Veldhuisen DJ, Comin-Colet J et al. Beneficial effects of long-term intravenous iron therapy with ferric carboxymaltose in patients with symptomatic heart failure and iron deficiency. Eur Heart J 2015;36:657–68. 10.1093/eurheartj/ehu385 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Anker SD, Kirwan BA, van Veldhuisen DJ et al. Effects of ferric carboxymaltose on hospitalisations and mortality rates in iron-deficient heart failure patients: an individual patient data meta-analysis. Eur J Heart Fail 2018;20:125–33. 10.1002/ejhf.823 [DOI] [PubMed] [Google Scholar]
- 33. Zhou X, Xu W, Xu Y et al. Iron supplementation improves cardiovascular outcomes in patients with heart failure. Am J Med 2019;132:955–63. 10.1016/j.amjmed.2019.02.018 [DOI] [PubMed] [Google Scholar]
- 34. Block GA, Block MS, Smits G et al. A pilot randomized trial of ferric citrate coordination complex for the treatment of advanced CKD. J Am Soc Nephrol 2019;30:1495–504. 10.1681/ASN.2018101016 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data underlying this article will not be made publicly available due to the privacy of the study participants. These data will be shared upon reasonable request to the corresponding author.