Erythropoiesis-stimulating agent responsiveness and hemoglobin variability is associated with fat tissue index in hemodialysis patients with darbepoetin-alfa treatment: a prospective observational cohort study
Article information
Abstract
Background
Although the introduction of erythropoietin-stimulating agents (ESAs) has led to better clinical outcomes in patients undergoing hemodialysis (HD), fluctuations in hemoglobin (Hb) levels, known as Hb variability, are frequently observed. However, only a few studies have evaluated the association between Hb variability and nutritional status in patients undergoing HD.
Methods
In this prospective study conducted between March 1, 2020, and June 1, 2022, we included 109 patients aged over 20 years undergoing HD and receiving darbepoetin. We checked the average NESP (darbepoetin-alfa; Kyowa Kirin Korea Co., Ltd.) dose weekly and nutritional parameters such as body mass index (BMI), fat tissue index (FTI), and lean tissue index obtained by body composition monitoring. Additionally, the ESA resistance index (ERI) and the coefficient of variation of Hb (Hb-CV) were evaluated.
Results
In this study, the mean age of the patients was 64.0 ± 11.9 years, and 55.0% were male. Mean Hb was 10.7 ± 1.3 g/dL. Patients were categorized into three groups according to the ERI or Hb-CV tertiles. The highest ERI tertile was associated with lower Hb levels, BMI, and FTI. The highest Hb-CV tertile was associated with lower BMI and FTI. In multiple linear regression analysis, FTI was negatively associated with ERI (β = –0.218, p = 0.01) and Hb-CV (β = –0.181, p = 0.04).
Conclusion
These findings suggest that FTI is negatively associated with ERI and Hb-CV, and that ESAs responsiveness and Hb variability are associated with FTI in patients undergoing HD with darbepoetin treatment.
Introduction
Anemia is a major comorbidity of end-stage renal disease (ESRD) undergoing hemodialysis (HD) [1]. As renal function deteriorates, erythropoietin production decreases, and hemoglobin (Hb) levels gradually decline. In addition, iron deficiency worsens, making it difficult to treat anemia in patients undergoing HD. Anemia leads to a poor prognosis owing to cardiovascular complications and increased mortality [2]. Although the efficacy of erythropoiesis-stimulating agents (ESAs) is well established, fluctuations in Hb levels, known as Hb variability, are well observed during anemia treatment [3], and Hb variability is related to the difficulty of anemia treatment. Additionally, it is difficult to maintain a patient’s Hb level within a narrow optimal range; factors that can affect Hb variability include loss of physiological control over erythropoiesis, chronic inflammation, secondary hyperparathyroidism, iron deficiency, inadequate dialysis, and malnutrition. Because of these factors, it is difficult to maintain Hb within the target range in HD patients. Previous studies showed that only 30% of patients fell within this range at any given time because Hb level fluctuations result in frequent under- and overshooting of the target level [4,5]. Other studies have shown that only 5.0% to 6.5% of patients undergoing HD could maintain target Hb levels (11–12 g/dL) [6,7].
In Korea, the target Hb level is 10–11 g/dL according to the reimbursement guidelines of the Health Insurance Review & Assessment Service (HIRA). A strict approach to ESA dose adjustment based on monthly Hb levels was implemented across all HD centers. The Hb levels in our country may be controlled within a narrow range [8]. Therefore, in the real world, it is difficult to stably maintain the Hb level within the target range.
Patients undergoing HD with stable target Hb levels have a lower risk of adverse events than those without stable Hb levels [9]. The response to ESAs in patients undergoing HD varies individually. Additionally, recent studies have identified an association between higher fluctuations in Hb variability and cardiovascular mortality [10,11], as well as all-cause mortality in patients undergoing HD [12]. Malnutrition is a recognized risk factor in the general population and in patients undergoing HD. However, only a few studies have evaluated the association between Hb variability and nutritional status in patients undergoing HD. A previous observational study reported that body mass index (BMI) may determine ESA response, with better responses observed in patients with higher BMIs [13] but, the mechanisms linking Hb variability and BMI values are unclear. The nutritional status of patients undergoing HD is usually assessed using the BMI, malnutrition-inflammation score, and serum albumin levels. Body composition monitoring (BCM) analysis is an accurate and noninvasive instrument for distinguishing between excessive and insufficient hydration levels and can accurately assess nutritional status [14]. However, few long-term prospective studies regarding the relationship between Hb variability and nutritional parameters have been conducted [15]. Therefore, we aimed to evaluate the Hb variability according to ESA responsiveness and body composition using BCM.
Methods
Ethics statement
The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of Myongji Hospital, Hanyang University College of Medicine in Goyang, Republic of Korea (No. MJH 2020-03-004-025). Written informed consent was obtained from all patients prior to enrollment.
Study design and population
We performed a prospective, open-label, observational trial to evaluate the association between darbepoetin-alpha and nutritional status in adult patients undergoing HD. Patients with ESRD undergoing maintenance HD were enrolled. This study was conducted between March 1, 2020, and June 1, 2022, and included patients from Myongji Hospital in Korea.
The enrollment criteria for patients were: 1) adult patients aged over 20 years, undergoing HD for more than 3 months, 2) patients received darbepoetin-alfa (NESP, Kyowa Kirin Korea Co., Ltd.; dose-titration, maintenance, or discontinuation) to achieve the target Hb level (10–11 g/dL) over 4 weeks before enrollment and during the study period, and 3) patients receiving outpatient dialysis care (including less than 4 weeks of hospitalization in total during the study period). The exclusion criteria comprised: 1) patients with acute infection, malignancy, intact parathyroid hormone (PTH) >500 pg/mL during the study period, or Kt/Vurea <1.2 during the study period; 2) those receiving HD via a catheter; or 3) patients unable to undergo measurement by bioelectrical impedance analysis such as those using a cardiac pacemaker.
Dosing schedule
The target Hb level was set at 10 to 11 g/dL according to the HIRA reimbursement guidelines. The route of NESP administration was intravenous in all patients enrolled in this study. Dose adjustments of NESP were conducted according to the monthly Hb level measurements in our facility. Therefore, each study participant had more than 24 monthly Hb data points.
Determination of hemoglobin indices for hemoglobin variability and erythropoiesis-stimulating agent responsiveness
We used several Hb indices, including the conventional standard deviation of Hb (Hb-SD); coefficient of variation of Hb (Hb-CV), that is, the ratio of the SD to the mean Hb; range of Hb (Hb-Ran), that is, the difference between maximum and minimum values; minimum value of Hb (Hb-Min), maximum value of Hb (Hb-Max), average value of Hb (Hb-Avg), and median value of Hb (Hb-Med) during the study period in this study [16,17]. All Hb indices were calculated using monthly Hb levels spanning at least 24 months. Additionally, we used the ESA resistance index (ERI) to assess ESAs responsiveness: ERI = (average weekly NESP dose/body weight)/average Hb level [18]. During the study period, we measured Hb levels at 1-month intervals to calculate Hb variability. The NESP dose, ferric sucrose dose, ERI, Hb-SD, Hb-CV, Hb-Ran, Hb-Min, Hb-Max, Hb-Avg, and Hb-Med values used in this study were time-averaged, while other laboratory values were considered as baseline measures.
Body composition monitoring analysis
The nutritional status of the patients was assessed using a BCM (Fresenius Medical Care Deutschland GmbH). The BCM measures the body resistance and reactance after applying low-strength alternating electric currents at 50 different frequencies, ranging between 5 and 1,000 kHz. Based on the measured resistance and reactance data, extracellular volume, intracellular volume, and total body water were determined using the approach described by Moissl et al. [19]. Lean tissue mass, fat tissue mass, and overhydration were calculated automatically using the BCM software according to a three-compartment model. The lean tissue index (LTI) and fat tissue index (FTI) were determined by fat and lean tissue mass adjusted for body surface (kg/m2) [20]. BCM measurements were performed three times (at baseline, 12 months, and 24 months) in this study. Measurements were taken before the onset of the dialysis session at mid-week with four conventional electrodes placed on the patient in the supine position: two on the hand and two on the foot contralateral to the vascular access. The parameters obtained using BCM included BMI, FTI (kg/m2), LTI (kg/m2), body cell mass index (BCMI), and phase angle (PhA) which is the most potent predictor of malnutrition and a useful predictor of mortality [14,16,21,22]. The BMI, FTI, LTI, BCMI, and PhA were time-averaged.
Other data collection
All demographic and clinical data were retrieved from patients’ electronic medical records. Age, sex, height, body weight, the presence of diabetes, HD duration, and various laboratory data were collected. Laboratory data included total iron-binding capacity (TIBC), transferrin saturation (TS), as well as levels of serum albumin, iron, ferritin, calcium, phosphorus, triglyceride, total cholesterol, high-density lipoprotein (HDL)-cholesterol, low-density lipoprotein (LDL)-cholesterol, intact PTH and highly sensitive C-reactive protein (hs-CRP). Baseline values were the laboratory parameters used in the analysis. The NESP dosages were collected during the entire study period and calculated as µg per week.
Statistical analysis
All normally distributed numerical variables were expressed as the mean ± SD, whereas variables with skewed distributions were expressed as the median and interquartile range. Patients were categorized into three groups based on the ERI or Hb-CV tertiles, and the differences between the groups were determined using an analysis of variance for continuous variables or the chi-squared test for categorical data. Multivariate linear regression analysis was used to assess the combined impact of the FTI values adjusted for variables that demonstrated significance in the univariate analysis or were clinically important. Additionally, multicollinearity was confirmed for the variables included in the multivariate analysis. These included age, sex, presence of diabetes mellitus, triglyceride level, intravenous iron replacement dose, Hb-CV, and ERI. Statistical significance was defined as p-values less than 0.05. All statistical analyses were performed using IBM SPSS version 23.0 (IBM Corp.).
Results
Baseline clinical characteristics of the study population
In this study, we enrolled 109 patients undergoing HD. The mean age of the patients was 64.0 ± 11.9 years, and 55.0% were male. Dialysis vintage was 54.9 ± 46.8 months. Follow-up duration was 24.1 ± 4.6 months. Comorbidities included hypertension (82.6%), diabetes mellitus (60.6%), and coronary artery disease (22.9%). The mean Hb level was 10.7 ± 1.3 g/dL, and serum albumin level was 3.9 ± 0.3 g/dL. The mean iron, TIBC, TS, and ferritin levels were 74.7 ± 31.8 µg/dL, 256.8 ± 44.9 mg/dL, 29.7% ± 13.4%, and 215.6 ± 209.0 ng/dL, respectively. Total cholesterol, LDL cholesterol, HDL cholesterol, and triglyceride were 139.2 ± 36.1 mg/dL, 72.1 ± 30.3 mg/dL, 44.5 ± 12.5 mg/dL, and 110.8 ± 52.9 mg/dL, respectively. Serum calcium, phosphorus, intact PTH, and hs-CRP levels were 8.4 ± 0.6 mg/dL, 4.7 ± 1.5 mg/dL, 284.3 ± 168.9 mg/dL, and 0.55 ±1.14 mg/dL, respectively. The dry weight, pre-dialysis weight, and overhydration were 61.8 ± 11.7 kg, 62.8 ± 11.5 kg, and 13.7% ± 8.1%, respectively. The BMI, LTI, FTI, BCMI, and PhA were 27.79 ± 3.63 kg/m2, 13.32 ± 3.34 kg/m2, 9.68 ± 4.00 kg/m2, 7.35 ± 2.38 kg/m2, and 4.44º ± 0.91º, respectively (Table 1).
Comparisons of hemoglobin and nutrition index by erythropoiesis-stimulating agent resistance index
We categorized the patients into tertiles according to the ERI. The average ERI was 0.02 ± 0.01, 0.04 ± 0.01, and 0.07 ± 0.03 in ERI-T1, ERI-T2, and ERI-T3, respectively. When comparing the three groups, the ERI-T3 group had the lowest Hb level (10.4 ± 1.2 g/dL, p = 0.04), lowest TIBC (229.0 ± 32.5 µg/dL, p < 0.001), and highest serum ferritin level (307.9 ± 237.6 ng/mL, p = 0.005). Administration of NESP dose was highest in ERI-T3 (36.8 ± 17.5 µg/wk, p < 0.001). The Hb-CV was higher in ERI-T3 (0.10 ± 0.03, p = 0.02). Hb-Min (8.0 ± 0.9 g/dL, p < 0.001) and Hb-Med (10.4 ± 0.4 g/dL, p < 0.001) were significantly lower in the ERI-T3 group (Table 2).
When comparing the three groups, the ERI-T3 group had the lower PhA value (4.06º ± 0.83º, p = 0.04), lowest BMI (22.4 ± 3.2 kg/m2, p = 0.001), and lowest FTI (7.7 ± 4.1 kg/m2, p = 0.046). Overhydration was highest in the ERI-T3 group (17.5% ± 5.8%, p = 0.01). No significant differences were observed in other nutritional parameters between the groups (Table 3).
Comparisons of hemoglobin and nutrition index by coefficient of variation of hemoglobin
The patients were categorized into tertiles according to their Hb-CV. The average Hb-CV was 0.07 ± 0.01, 0.09 ± 0.01, and 0.12 ± 0.03 in Hb-CV-T1, Hb-CV-T2, and Hb-CV-T3, respectively. When comparing the three groups, the Hb-SD, Hb-Max, and Hb-Ran were significantly higher in Hb-CV-T3 (1.28 ± 0.30, p < 0.001; 13.7 ± 1.4, p < 0.001; 5.7 ± 1.3, p < 0.001). Hb-Min was significantly lower in the Hb-CV-T3 group (8.1 ± 1.0, p < 0.001) (Table 4).
When comparing the three groups, the Hb-CV-T3 group had the lowest BMI (22.6 ± 2.6 kg/m2, p = 0.003) and lowest FTI (7.3 ± 3.6 kg/m2, p = 0.002) with significantly higher overhydration (17.0 ± 5.7 %, p = 0.03) (Table 5).
Relationship between body mass index, fat tissue index, erythropoiesis-stimulating agent resistance index, and coefficient of variation of hemoglobin
We evaluated the relationship among BMI-ERI, BMI-Hb-CV, FTI-ERI, and FTI-Hb-CV. In this study, BMI showed a significantly negative correlation with ERI (R2 = 0.076) and Hb-CV (R2 = 0.062). Also, FTI was significantly negatively correlated with ERI (R2 = 0.037) and Hb-CV (R2 = 0.068) (Fig. 1).

Scatter plot demonstrating the relationship between BMI, FTI, ERI, and Hb-CV.
(A) Relationship between BMI and ERI. (B) Relationship between BMI and Hb-CV. (C) Relationship between FTI and ERI. (D) Relationship between FTI and Hb-CV.
BMI, body mass index; ERI, erythropoietin-stimulating agent resistance index; FTI, fat tissue index; Hb-CV, coefficient of variation of hemoglobin.
Correlations between clinical and biochemical variables and fat tissue index
Age, female sex, presence of diabetes mellitus, and triglyceride were positively correlated with FTI (β = 0.203, p = 0.04; β = 0.376, p < 0.001; β = 0.191, p = 0.049; and β = 0.227, p = 0.02, respectively). Hb-CV and ERI were negatively correlated with FTI (β = –0.268, p = 0.005 and β = –0.193, p = 0.046) (Table 6). In multiple linear regression analysis, FTI was negatively associated with Hb-CV, and ERI (β = –0.185, p = 0.04 and β = –0.216, p = 0.04, respectively), whereas FTI was positively associated with age and female sex (β = 0.185, p = 0.03 and β = 0.394, p < 0.001, respectively).
Discussion
This study demonstrated the relationship between Hb variability and body composition in patients undergoing HD, using BCM. Our findings showed that high ERI was associated with lower ferritin levels, BMI, FTI, and increased overhydration. Similarly, high Hb-CV was associated with lower BMI, FTI, and increased overhydration. However, no association was observed among ERI, Hb-CV, and LTI in this study. Finally, the FTI exhibited a negative association with ERI and Hb-CV.
The treatment of anemia is of paramount importance in patients undergoing HD as it affects their quality of life, morbidity, and mortality [23,24]. Furthermore, observed Hb variability during anemia treatment is independently associated with quality of life, infectious events, and cardiovascular mortality [10,11,25]. Kalantar-Zadeh and Aronoff [17] reported that the factors influencing Hb variability include drug-related factors, patient characteristics, iron storage, infection, and inflammation. In addition, the half-life of ESA also affects Hb variability and Portolés et al. [26] have reported that long-acting darbepoetin achieves greater Hb stability than short-acting ESA. Notably, BMI has been shown to correlate negatively with weekly ESA dose and ERI showing Hb variability, suggesting that obesity has a protective effect against anemia in patients undergoing dialysis [13]. However, the specific components (body fat mass, lean mass, or hydration) associated with Hb variability remain controversial. In a study using BCM, Chiang et al. [20] reported that ERI was negatively correlated with BMI and LTI, but not with FTI, suggesting LTI as a significant predictor of ERI. However, in our study, lean mass had no statistical correlation with ERI, but fat mass showed a significant negative correlation with ERI. Whereas Chiang et al.’s study [20] was a cross-sectional observational study, our study was a prospective study. As a strength of our study, Hb variability was observed over 24.1 months using a single formulation of ESA, and we adjusted the iron supplement doses administered to patients undergoing HD. Therefore, this study exclusively evaluated the effect of nutritional status such as FTI on Hb variability.
Although our findings remain controversial, the following studies have shown similar results. Kotanko et al. [27] reported that higher absolute total and subcutaneous adipose tissue in patients undergoing HD was associated with lower ESA doses and lower ESA resistance. Another study conducted by Vega et al. [28] which used BIS to measure body composition in patients undergoing maintenance HD showed that higher fat tissue was associated with a better response to ESA. However, no association was found between ERI and LTI [28]. In addition, Lee et al. [15] studied the relationship between body composition and ERI using a multifrequency bioelectrical impedance analysis device in 123 HD patients. In a previous study, patients with a smaller fat mass had higher ERI, and ERI was negatively correlated with visceral fat area. A previous study by Lee et al. [15] reported that the simple value of visceral fat area was related to ERI. However, in this study, we used the FTI, which is determined by fat tissue mass adjusted for body surface area. Thus, our result showing that FTI was negatively correlated with ERI has a strength compared with past studies in that it used quantified fat tissue mass. In this prospective study, we analyzed Hb variability in patients undergoing HD over a longer duration than in previous studies, providing a more accurate evaluation of the relationship between the nutritional status and Hb variability. Despite the differences in study duration, similar results were obtained in both our study and Vega et al.’s study [28].
Feret et al. [29] reported that low-fat mass was associated with a higher ERI, suggesting a role of leptin in this mechanism. Adipocyte-derived leptin has demonstrated an erythropoiesis-stimulating effect by reducing the pro-inflammatory effects of adipose tissue and enhancing its anti-inflammatory effects. In an interventional study by Hung et al. [30], high-calorie intake in patients undergoing HD, leading to hyperleptinemia, markedly improved hematopoiesis. This is most likely why not muscle mass, but visceral fat and fat mass in general, constitute a factor associated with erythropoietin sensitivity [31]. Although we did not evaluate leptin levels in this study, we found that a higher FTI was associated with a lower ERI and lower Hb variability in patients undergoing HD, irrespective of age, sex, and the presence of diabetes mellitus.
Moreover, overhydration was associated with higher ERI and Hb-CV in this study. In malnourished patients, the loss of lean and fat mass components results in excess extracellular and intracellular water, indicating overhydration [32]. A high degree of overhydration indicates malnutrition, and previous studies have shown that Hb variation is high in patients with overhydration [33]. The results of this study are consistent with those of previous research.
Our study had a few limitations. First, this was a single-center study with a relatively small sample size, which could have resulted in information bias. However, the 2-year prospective follow-up minimized the potential risk of missing or incorrect information. Second, this study included only Korean patients undergoing HD, so the ability to generalize the results to other populations may be limited. Third, this study used only a single erythropoietin agent, limiting the ability to suggest a correlation between other erythropoietin drugs, Hb variability, and nutritional status. Fourth, because the number of patients enrolled in this study was small, there may be heterogeneity between groups. So, the ERI-T1 group may have included relatively healthy patients. It cannot be ruled out that this may have influenced the research results.
In conclusion, we investigated the relationship between nutritional status and Hb variability in patients undergoing HD and receiving NESP treatment and found that FTI was negatively associated with ERI and Hb-CV. Therefore, the protective effects of body composition on ESA responsiveness and Hb variability are thought to be associated with fat tissue.
Notes
Conflicts of interest
All authors have no conflicts of interest to declare.
Funding
This work was supported by Kyowa Kirin Korea Co., Ltd. The authors would like to extend their gratitude to Kyowa Kirin Korea Co., Ltd.
Data sharing statement
The data presented in this study are available from the corresponding author upon reasonable request.
Authors’ contributions
Conceptualization, Methodology: DJO, DHK
Data curation: HY, DJO
Formal analysis: HY, DJO, DHK
Funding acquisition: DJO
Writing–original draft: DHK
Writing–review & editing: HY, DJO, DHK
All authors have read and agreed to the published version of the manuscript.