Acetate-free, citrate-acidified bicarbonate dialysis improves serum calcification propensity—a preliminary study

Acetate-free, citrate-acidified bicarbonate dialysis improves serum calcification propensity—a... Abstract Background A novel in vitro test (T50 test) assesses ex vivo serum calcification propensity and predicts mortality in chronic kidney disease and haemodialysis (HD) patients. For the latter, a time-dependent decline of T50 was shown to relate to mortality. Here we assessed whether a 3-month switch to acetate-free, citrate-acidified, standard bicarbonate HD (CiaHD) sustainably improves calcification propensity. Methods T50 values were assessed in paired midweek pre-dialysis sera collected before and 3 months after CiaHD in 78 prevalent European HD patients. In all, 44 were then switched back to acetate. Partial correlation was used to study associations of changing T50 and changing covariates. Linear mixed effect models were built to assess the association of CiaHD and covariates with changing T50. Results A significant intra-individual increase of serum calcification resilience was found after 3 months on CiaHD (206  ±  56 to 242  ±  56 min; P < 0.001), but not after switching back to acetate (252  ±  63 to 243  ±  64 min; n = 44; P = 0.29). CiaHD, Δ serum phosphate and Δ albumin but not Δ ionized calcium and magnesium were the strongest determinants of changing T50. Beneath T50, only serum albumin but not phosphate changed significantly during 3 months of CiaHD. Conclusion CiaHD dialysis favourably affected calcification propensity as measured by the T50 test. Whether this treatment, beyond established phosphate-directed treatments, has the potential to sustainably tip the balance towards a more anti-calcific serum milieu needs to be further investigated. albumin, citrate, dialysis, serum calcification propensity, T50 INTRODUCTION Patients on haemodialysis (HD) experience excess cardiovascular morbidity and mortality, which is related to a disturbed calcium and phosphate homeostasis promoting vascular disease progression [1–3]. Recently an in vitro test (T50 test) was developed by Pasch and colleagues. This test time-dependently assesses the calcification propensity of human serum after the addition of supraphysiologic doses of calcium and phosphate [4, 5]. Basically the test determines the crystal formation time from amorphous primary calciprotein particles (CPP) into crystallized secondary CPP in human serum [4]. The advantage of this assay is its ability to integrate the complex interplay of pro- and anti-calcific players (e.g. calcium and phosphate versus magnesium and albumin) into a functional assay conceptually comparable to functionally testing the activated clotting time [4–6]. Shorter T50 times predict vascular stiffness progression and mortality in chronic kidney disease Stages 3 and 4 [6]. We recently found that in stable HD patients, serum calcification propensity declined over the course of 2 years, indicating a loss of calcification resilience [7]. In this post hoc analysis of 188 stable dialysis patients, the time-dependent decline of serum calcification propensity T50 was associated with overall and cardiovascular mortality [7]. However, this study left the question unanswered as to which specific measures might be associated with stabilization, i.e. the prevention of the decline of serum calcification propensity T50 [8]. Acetate is widely used as a stabilizing acidic agent within the cation concentrate (A-concentrate) in standard bicarbonate HD, hereafter referred to as acetate-acidified standard bicarbonate haemodialysis (AaHD). In Europe, this A-concentrate usually contains 3–4 mmol/L of acetate [9–12]. Whereas physiological and pre-dialysis acetate concentrations are <100 µmol/L, intradialytic concentrations increase up to 10-fold above normal and range from <100 to 222 µmol/L immediately after dialysis, depending on the metabolization status [13, 14]. This results in intra-individually varying post-dialysis acetate levels of up to 500 µmol/L. Acetate is not free of side effects, and intradialytic hypotensive episodes are less frequent in acetate-free dialysis modalities [15–17]. Further replacement of acetate with the putatively more biocompatible citric acid, which is converted to citrate, as an acidifier of the A-concentrate in bicarbonate HD (CiaHD) [9] was reported to provide better control of acidic balance [12] and to positively affect serum albumin and nutritional status in those with hypoalbuminemia [18, 19]. Other studies suggest better elimination of phosphate and less need for anticoagulation [11, 20]. As a chelator of bivalent anions, citrate could further influence net calcium mass balance and the pro-calcific environment in HD patients [21]. In this prospective analysis of stored serum samples we wanted to assess how 3-months use of an acetate-free, 1 mmol citrate-acidified A-concentrate (SelectBagCitrate, Gambro/Baxter, Hechingen, Germany) would impact on serum calcification propensity (T50 time) in HD patients. Therefore we determined the T50 time in paired pre-dialysis sera collected before and after 3 months on CiaHD in 78 eligible patients. Of these, 44 patients had been switched back to AaHD thereafter. MATERIALS AND METHODS Study protocol and study subjects This pre–post-quasi-interventional study was conducted in a subgroup of the ‘Substitution of Acetate by Citrate in Bicarbonate-Based-Hemodialysis’ study (NCT02745340). Between April and June 2016, 88 HD patients were recruited from two local dialysis units of the same dialysis centre in Munich, Germany. In all, 78 patients completed the full observation period. According to the registered study protocol, all patients started out on AaHD (SelectBagOne; 3 mmol of acetate). Dialysis Unit 1 was then switched to CiaHD (SelectBagCitrate; acidified with citric acid, which is converted to 1 mmol/L citrate) for 3 months, while Unit 2 remained on AaHD. Predicted solute concentrations resulting from the A-concentrates as provided by Gambro/Baxter can be retrieved from Supplementary data, Table S1. After 3 months and a washout of 48 h, units were switched to AaHD or CiaHD, respectively. Unit 1 was afterwards switched back to AaHD. Of the 78 patients, 44 had serum available for determination of T50 at baseline, after 3 months of CiaHD and after another 3 months back on CiaHD (Figure 1). Inclusion criteria were age  ≥18 years, HD vintage of at least 90 days on a thrice-weekly schedule with a session duration of  ≥4 h. Pregnancy, ongoing infection and lack of written and informed consent were the main exclusion criteria. The study protocol was approved by the Ethics Commission of the Klinikum rechts der Isar, Technical University Munich. It was carried out in accordance with the Declaration of Helsinki, adhering to good clinical practice. The original study was registered at ClinicalTrials.gov (NCT02745340). Written informed consent was obtained from all participants. For the derived study, no additional ethics committee approval was sought. FIGURE 1 View largeDownload slide Observation schedule and study time course. All patients had been on standard bicarbonate dialysis with an AaHD for at least 2 months. Then Unit 1 (n = 47) underwent 3 months of CiaHD. After 3 months Unit 2 (n = 31) was also switched to CiaHD, whereas Unit 1 underwent another 3 months of AaHD. Dashed lines indicate groups before versus after CiaHD dialysis for pooled data analysis. T50 was determined at the indicated time points from pre-dialysis serum samples. FIGURE 1 View largeDownload slide Observation schedule and study time course. All patients had been on standard bicarbonate dialysis with an AaHD for at least 2 months. Then Unit 1 (n = 47) underwent 3 months of CiaHD. After 3 months Unit 2 (n = 31) was also switched to CiaHD, whereas Unit 1 underwent another 3 months of AaHD. Dashed lines indicate groups before versus after CiaHD dialysis for pooled data analysis. T50 was determined at the indicated time points from pre-dialysis serum samples. Clinical data assessment Patients’ age, comorbidities and medication were continuously assessed using medical records and patient interviews at inclusion and after the CiaHD or AaHD phases (3 months). Comorbidities were recorded following Liu et al.’s adapted version of the Charlson Comorbidity Index (CCI) [22]. Body mass index (BMI) was calculated as body weight/height2 (kg/m2). All patients underwent bicarbonate dialysis with different types of synthetic membranes that remained constant during the study period. Dialysis prescription [ultrafiltration, session duration, Kt/V as a measure of dialysis efficiency, HD/haemodiafiltration (HDF), anticoagulation] was provided by the units. Medical staff were the same for both dialysis units. A total of 10 of the included patients did not participate in the present study, as they had either died (n = 5), moved (n = 1), were hospitalized for a longer period (n = 2) or reported an unwillingness to donate blood specimens at any of the required time points (n = 2). Missing data at relevant time points were Kt/V (n = 3), intact parathyroid hormone (iPTH; n = 2), C-reactive protein (CRP; n = 1) and n = 2 for ionized magnesium and calcium levels. The latter two were estimated from total corrected calcium and magnesium values [23], choosing their individual ratios of ionized calcium or magnesium levels/total calcium or magnesium values as conversion factor. Exclusion of these imputed values revealed comparable results to the presented data (not shown). Blood specimen collection and laboratory methods Serum was collected prior to a midweek dialysis session at the time points indicated (Figure 1). After 30 min at room temperature, serum was centrifuged (2000 g, 10 min), aliquoted and frozen at −80°C. Routine laboratory analysis was performed by ISO accredited laboratories. Ionized magnesium and calcium levels were determined from frozen sera using the Nova 8 Analyzer and ion-selective electrodes for calcium and magnesium (Nova Biomedical, Waltham, MA, USA). pH, base excess and standard bicarbonate values were derived from routinely available pre-dialysis venous blood gas analyses (BGAs) at the middle of each treatment period and were performed by ISO-accredited laboratories. T50 was determined as previously described by Pasch et al. [5, 24]. Frozen sera were sent to Bern on dry ice and analysed in a blinded manner. Likewise, fetuin A was assessed from 4-fold diluted sera (phosphate-buffered saline) in a blinded manner using a nephelometric assay and a polyclonal rabbit anti-human fetuin A antibody following published protocols [25]. Citrate and acetate concentrations in serum were determined from a subset of 15 patients on both CiaHD and AaHD treatment using one dimensional 1H-nuclear magnetic resonance (NMR) spectroscopy as formerly described by Prokesch et al. (Graz, Austria) [26]. Statistical analysis SPSS Statistics 23 (IBM, Armonk, NY, USA) was used for statistical analysis. For pooled data before and after CiaHD, we report mean  ±  SD, median and interquartile range (IQR) or counts and percentage of total as appropriate. The change in T50 (⁠ T50Change ⁠) was calculated as (⁠ T50before CiaHD −  T50after CiaHD ⁠)/ T50before CiaHD *100 (%). Accordingly, T50Change in 44 patients, who underwent an additional 3 months of AaHD treatment was calculated as (⁠ T50after AaHD −  T50after CiaHD ⁠)/ T50after CiaHD *100 (%). Increasing T50Change was defined by T50Change values >5.1% before versus after the respective treatment period according to the interassay coefficient for T50 standards at 260 min in the Evaluation of Cinacalcet Hydrochloride Therapy to Lower Cardiovascular Events (EVOLVE) cohort and in accordance with our previous report [7, 24]. A paired t-test was used to compare differences between absolute T50 values before and after CiaHD. Unpaired t-tests were used to compare already normalized T50Change values in 44 patients on AaHD versus CiaHD. Group differences for basal comorbidities in those not included were tested using analysis of variance, Kruskal–Wallis and Fisher’s exact test. Partial regression adjusted for phosphate was used for correlating T50, T50Change with standard laboratory parameters and their respective Δ values. For example, intra-individual Δ − albumin = albuminafter CiaHD − albuminbefore CiaHD (mmol/L). Linear mixed-effects models assessing the association of treatment time points before (CiaHD = 0) versus after citrate (CiaHD = 1), adjusted for other laboratory parameters (additional fixed effects), were built based on an unstructured covariance matrix (best model fit according to −2 log likelihood) including patients’ IDs as random effects. Regression estimates are reported per 1 SD increase of the regressor. Indirect effect coefficients were calculated according to Judd and Kenny [27]. A P-value <0.05 was considered significant. RESULTS Study design and study population This study was performed as a prospective secondary analysis of the Substitution of Acetate by Citrate in Bicarbonate-Based Hemodialysis study. Of 88 originally included patients, 78 patients were accessible for the study time course (Figure 1). The study population (n = 78) did not significantly differ from the originally included patient collective (n = 88) with respect to age, gender, basic laboratory workup and dialysis modalities (Supplementary data, Table S2). The median age of the study population was 75.1 years (IQR 57.2–80.3), with 41 (53%) being male patients on HD treatment (96%). Three patients received HDF treatment. Three (4%) black participants were included. Three (4%) patients received phosphate binders while on CiaHD, four (5%) patients received increased doses of calcium acetate, three (4%) patients received increased doses of other phosphate binders and only one patient was prescribed an increased dose of active vitamin D. For detailed baseline descriptions at baseline after 3 months see Table 1. Table 1 Characteristics of the study population at study baseline (before CiaHD) and after CiaHD Parameter Before CiaHD After 3 months of CiaHD Age (years), median (IQR) 75.1 (57.2–80.3) +0.25 Gender (males), n (%) 41 (53) – BMI (kg/m2), mean ± SD 26.2 ± 5.1 – White patients, n (%) 75 (96) – Upper arm circumference (cm), mean ± SD 28.7 ± 3.6 28.7 ± 3.5 Adapted CCI (0–21), median (IQR) 3 (1–6) 4 (2–7)b History of MI, n (%) 15 (19) 15 (19) Hypertensive heart disease, n (%) 40 (51) 41 (53) Diabetes, n (%) 26 (33) 28 (36) COPD, n (%) 7 (9) 9 (11) Cancer, n (%) 30 (39) 33 (42) CHD, n (%) 26 (33) 27 (35) PAOD, n (%) 11 (14) 13 (16) CerVD, n (%) 13 (16) 13 (16) CHF, n (%) 8 (10) 8 (10) Atrial fibrillation, n (%) 17 (22) 19 (24) GIT bleeding, n (%) 5 (6) 5 (6) Smoking (at the moment), n (%) 8 (10) 8 (10) HD not HDF, n (%) 75 (96) 75 (96) HD vintage (months), median (IQR) 36 (17–56) +0.25 Bicarbonate setting, median (IQR) 32 (30–33) 32 (30–33) Kt/Va, mean ± SD 1.2 ± 0.3 1.2 ± 0.3 Haemoglobin (g/dL), mean ± SD 11.9 ± 1.3 11.6 ± 1.2 Ionized magnesium (mmol/L)a, mean ± SD 0.5 ± 0.1 0.4 ± 0.1b Ionized calcium (mmol/L)a, mean ± SD 1.2 ± 0.2 1.1 ± 0.2b Phosphate (mmol/L), mean ± SD 1.7 ± 0.5 1.8 ± 0.5 Albumin (g/L)a, mean ± SD 38 ± 4 40 ± 5b Fetuin A (g/L), mean ± SD 0.37 ± 0.11 0.36 ± 0.08 iPTH (pg/mL)a, median (IQR) 267 (134–439) 254 (134–484) CRP (mg/L)a, mean ± SD 10.4 ± 21.8 9 ± 12.3 New PB while on CiaHD n (%) 3 (4) – Increase of calcium containing PB while on CiaHD, n (%) 4 (5) – Increase of other PB while on CiaHD, n (%) 3 (4) – Increase of active vitamin D while on CiaHD, n (%) 1 (1) – Dialysis centre (1 not 2), n (%) 47 (60) – Parameter Before CiaHD After 3 months of CiaHD Age (years), median (IQR) 75.1 (57.2–80.3) +0.25 Gender (males), n (%) 41 (53) – BMI (kg/m2), mean ± SD 26.2 ± 5.1 – White patients, n (%) 75 (96) – Upper arm circumference (cm), mean ± SD 28.7 ± 3.6 28.7 ± 3.5 Adapted CCI (0–21), median (IQR) 3 (1–6) 4 (2–7)b History of MI, n (%) 15 (19) 15 (19) Hypertensive heart disease, n (%) 40 (51) 41 (53) Diabetes, n (%) 26 (33) 28 (36) COPD, n (%) 7 (9) 9 (11) Cancer, n (%) 30 (39) 33 (42) CHD, n (%) 26 (33) 27 (35) PAOD, n (%) 11 (14) 13 (16) CerVD, n (%) 13 (16) 13 (16) CHF, n (%) 8 (10) 8 (10) Atrial fibrillation, n (%) 17 (22) 19 (24) GIT bleeding, n (%) 5 (6) 5 (6) Smoking (at the moment), n (%) 8 (10) 8 (10) HD not HDF, n (%) 75 (96) 75 (96) HD vintage (months), median (IQR) 36 (17–56) +0.25 Bicarbonate setting, median (IQR) 32 (30–33) 32 (30–33) Kt/Va, mean ± SD 1.2 ± 0.3 1.2 ± 0.3 Haemoglobin (g/dL), mean ± SD 11.9 ± 1.3 11.6 ± 1.2 Ionized magnesium (mmol/L)a, mean ± SD 0.5 ± 0.1 0.4 ± 0.1b Ionized calcium (mmol/L)a, mean ± SD 1.2 ± 0.2 1.1 ± 0.2b Phosphate (mmol/L), mean ± SD 1.7 ± 0.5 1.8 ± 0.5 Albumin (g/L)a, mean ± SD 38 ± 4 40 ± 5b Fetuin A (g/L), mean ± SD 0.37 ± 0.11 0.36 ± 0.08 iPTH (pg/mL)a, median (IQR) 267 (134–439) 254 (134–484) CRP (mg/L)a, mean ± SD 10.4 ± 21.8 9 ± 12.3 New PB while on CiaHD n (%) 3 (4) – Increase of calcium containing PB while on CiaHD, n (%) 4 (5) – Increase of other PB while on CiaHD, n (%) 3 (4) – Increase of active vitamin D while on CiaHD, n (%) 1 (1) – Dialysis centre (1 not 2), n (%) 47 (60) – Demographics, comorbidities and basic laboratory values from pooled analysis (both dialysis units) are reported before and after CiaHD. CerVD, cerebral vascular disease; CHD, coronary heart disease; CHF, congestive heart failure; COPD, chronic obstructive pulmonary disease; CRP, C-reactive protein; GIT, gastrointestinal; iPTH, intact parathyroid hormone; MI, myocardial infarction; PAOD, peripheral arterial occlusive disease; PB, phosphate binders. a Missing values: 1–3 missing values, due to missed quartal assessment. b Paired samples t-test and Wilcoxon signed-rank test were used for comparing before versus after citrate P < 0.05. Table 1 Characteristics of the study population at study baseline (before CiaHD) and after CiaHD Parameter Before CiaHD After 3 months of CiaHD Age (years), median (IQR) 75.1 (57.2–80.3) +0.25 Gender (males), n (%) 41 (53) – BMI (kg/m2), mean ± SD 26.2 ± 5.1 – White patients, n (%) 75 (96) – Upper arm circumference (cm), mean ± SD 28.7 ± 3.6 28.7 ± 3.5 Adapted CCI (0–21), median (IQR) 3 (1–6) 4 (2–7)b History of MI, n (%) 15 (19) 15 (19) Hypertensive heart disease, n (%) 40 (51) 41 (53) Diabetes, n (%) 26 (33) 28 (36) COPD, n (%) 7 (9) 9 (11) Cancer, n (%) 30 (39) 33 (42) CHD, n (%) 26 (33) 27 (35) PAOD, n (%) 11 (14) 13 (16) CerVD, n (%) 13 (16) 13 (16) CHF, n (%) 8 (10) 8 (10) Atrial fibrillation, n (%) 17 (22) 19 (24) GIT bleeding, n (%) 5 (6) 5 (6) Smoking (at the moment), n (%) 8 (10) 8 (10) HD not HDF, n (%) 75 (96) 75 (96) HD vintage (months), median (IQR) 36 (17–56) +0.25 Bicarbonate setting, median (IQR) 32 (30–33) 32 (30–33) Kt/Va, mean ± SD 1.2 ± 0.3 1.2 ± 0.3 Haemoglobin (g/dL), mean ± SD 11.9 ± 1.3 11.6 ± 1.2 Ionized magnesium (mmol/L)a, mean ± SD 0.5 ± 0.1 0.4 ± 0.1b Ionized calcium (mmol/L)a, mean ± SD 1.2 ± 0.2 1.1 ± 0.2b Phosphate (mmol/L), mean ± SD 1.7 ± 0.5 1.8 ± 0.5 Albumin (g/L)a, mean ± SD 38 ± 4 40 ± 5b Fetuin A (g/L), mean ± SD 0.37 ± 0.11 0.36 ± 0.08 iPTH (pg/mL)a, median (IQR) 267 (134–439) 254 (134–484) CRP (mg/L)a, mean ± SD 10.4 ± 21.8 9 ± 12.3 New PB while on CiaHD n (%) 3 (4) – Increase of calcium containing PB while on CiaHD, n (%) 4 (5) – Increase of other PB while on CiaHD, n (%) 3 (4) – Increase of active vitamin D while on CiaHD, n (%) 1 (1) – Dialysis centre (1 not 2), n (%) 47 (60) – Parameter Before CiaHD After 3 months of CiaHD Age (years), median (IQR) 75.1 (57.2–80.3) +0.25 Gender (males), n (%) 41 (53) – BMI (kg/m2), mean ± SD 26.2 ± 5.1 – White patients, n (%) 75 (96) – Upper arm circumference (cm), mean ± SD 28.7 ± 3.6 28.7 ± 3.5 Adapted CCI (0–21), median (IQR) 3 (1–6) 4 (2–7)b History of MI, n (%) 15 (19) 15 (19) Hypertensive heart disease, n (%) 40 (51) 41 (53) Diabetes, n (%) 26 (33) 28 (36) COPD, n (%) 7 (9) 9 (11) Cancer, n (%) 30 (39) 33 (42) CHD, n (%) 26 (33) 27 (35) PAOD, n (%) 11 (14) 13 (16) CerVD, n (%) 13 (16) 13 (16) CHF, n (%) 8 (10) 8 (10) Atrial fibrillation, n (%) 17 (22) 19 (24) GIT bleeding, n (%) 5 (6) 5 (6) Smoking (at the moment), n (%) 8 (10) 8 (10) HD not HDF, n (%) 75 (96) 75 (96) HD vintage (months), median (IQR) 36 (17–56) +0.25 Bicarbonate setting, median (IQR) 32 (30–33) 32 (30–33) Kt/Va, mean ± SD 1.2 ± 0.3 1.2 ± 0.3 Haemoglobin (g/dL), mean ± SD 11.9 ± 1.3 11.6 ± 1.2 Ionized magnesium (mmol/L)a, mean ± SD 0.5 ± 0.1 0.4 ± 0.1b Ionized calcium (mmol/L)a, mean ± SD 1.2 ± 0.2 1.1 ± 0.2b Phosphate (mmol/L), mean ± SD 1.7 ± 0.5 1.8 ± 0.5 Albumin (g/L)a, mean ± SD 38 ± 4 40 ± 5b Fetuin A (g/L), mean ± SD 0.37 ± 0.11 0.36 ± 0.08 iPTH (pg/mL)a, median (IQR) 267 (134–439) 254 (134–484) CRP (mg/L)a, mean ± SD 10.4 ± 21.8 9 ± 12.3 New PB while on CiaHD n (%) 3 (4) – Increase of calcium containing PB while on CiaHD, n (%) 4 (5) – Increase of other PB while on CiaHD, n (%) 3 (4) – Increase of active vitamin D while on CiaHD, n (%) 1 (1) – Dialysis centre (1 not 2), n (%) 47 (60) – Demographics, comorbidities and basic laboratory values from pooled analysis (both dialysis units) are reported before and after CiaHD. CerVD, cerebral vascular disease; CHD, coronary heart disease; CHF, congestive heart failure; COPD, chronic obstructive pulmonary disease; CRP, C-reactive protein; GIT, gastrointestinal; iPTH, intact parathyroid hormone; MI, myocardial infarction; PAOD, peripheral arterial occlusive disease; PB, phosphate binders. a Missing values: 1–3 missing values, due to missed quartal assessment. b Paired samples t-test and Wilcoxon signed-rank test were used for comparing before versus after citrate P < 0.05. Serum calcification propensity (T50) increases after 3 months on acetate-free/citrate-acidified A-concentrate Individual T50 times before (⁠ T50before CiaHD ⁠) and after (⁠ T50after CiaHD ⁠) 3 months of CiaHD treatment were normally distributed (Supplementary data, Figure S1). We had previously reported a longitudinal decline of T50 in stable HD patients on standard dialysis regimes [7]. In our present study, the mean T50before CiaHD was 206  ±  56 min and increased significantly to a T50after CiaHD of 242  ±  65 min after 3 months of intermittent CiaHD treatment (P <0.001; Figure 2A). Further, when testing for intra-individual variation, T50before CiaHD versus T50after CiaHD increased significantly after 3 months of CiaHD (Figure 2B; P <0.001). In accordance, T50Change [= (⁠ T50after CiaHD −  T50before CiaHD ⁠)/ T50after CiaHD] exceeded +5.1% (derived from the interassay coefficient for T50 standards at 260 min as reported previously [7, 24]) in 57 patients on CiaHD, whereas only 21 showed stable or declining T50Change (Figure 2C). FIGURE 2 View largeDownload slide T50 values and their change during the study course. (A) Increasing absolute pre-dialysis T50 values after 3 months of CiaHD in 78 patients and not significantly altered levels after an additional 3 months on an AaHD in 44 patients eligible for this analysis (Figure 1). Paired-samples t-test was used for comparison. (B) Intra-individual T50Change values for 78 patients before versus after 3 months of citrate. Blue rhombuses indicate stable or declining T50 (⁠ T50Change ≤5.1%). Green rhombuses indicate increasing (⁠ T50Change >5.1%). T50Change values were sorted in ascending order. (C) Mean values and 95% confidence intervals of T50Change observed during CiaHD versus T50Change back on AaHD thereafter. Independent samples t-test was used for comparing the latter. FIGURE 2 View largeDownload slide T50 values and their change during the study course. (A) Increasing absolute pre-dialysis T50 values after 3 months of CiaHD in 78 patients and not significantly altered levels after an additional 3 months on an AaHD in 44 patients eligible for this analysis (Figure 1). Paired-samples t-test was used for comparison. (B) Intra-individual T50Change values for 78 patients before versus after 3 months of citrate. Blue rhombuses indicate stable or declining T50 (⁠ T50Change ≤5.1%). Green rhombuses indicate increasing (⁠ T50Change >5.1%). T50Change values were sorted in ascending order. (C) Mean values and 95% confidence intervals of T50Change observed during CiaHD versus T50Change back on AaHD thereafter. Independent samples t-test was used for comparing the latter. A total of 44 patients had additionally undergone another 3 months of AaHD after 3 months on CiaHD (Figure 1). No further increase in T50 time was observed in this period. The mean T50after CiaHD for this subgroup was 252  ± 63 min and 243 ± 64 min (P = 0.29) after these additional 3 months back on AaHD, indicating that the beneficial effects observed on citrate dialysate had not persisted (Figure 2B). In accordance with this, T50Change (+28 ± 28%; n = 44) within the first 3 months on citrate was significantly different from T50Change (−1 ± 28%; n = 44) observed thereafter (Figure 2D; P < 0.001). Individual values for T50Change while on CiaHD and AaHD thereafter are depicted in Supplementary data, Figure S2. Albumin partially mediates increasing T50 during acetate-free, citrate-acidified bicarbonate HD Aside from changing T50 after 3 months of CiaHD, we observed significantly decreasing levels for ionized serum magnesium (0.5 ± 0.1 to 0.4 ± 0.1 mmol/L; P <0.001) and calcium levels (1.2 ± 0.2 to 1.1 ± 0.2 mmol/L; P = 0.01), whereas serum albumin levels had significantly increased (38 ± 4 to 40 ± 5 g/L, P = 0.001). Serum phosphate levels did not significantly change (1.7 ± 0.4  to 1.8 ± 0.4 mmol/L; P = 0.47) nor did fetuin A levels (0.37 ± 0.11 to 0.36 ± 0.08 g/L). For these and other parameters, including routine venous pre-dialysis BGA, see Supplementary data, Figure S3. Similar results were obtained when testing the influence of CiaHD (1 = yes) on these parameters in linear mixed models, including patient IDs as random effects and these variables as dependents (Supplementary data, Table S3). It should be noted that the mentioned changes were probably not directly conveyed due to persistence of citrate in the circulation until the next dialysis session. In line with its reported half-life [28, 29], we did not find significant differences in serum citrate or acetate levels in pre-dialysis samples from 15 patients on CiaHD versus AaHD treatment (Supplementary data, Figure S4). Since both phosphate (rbefore CiaHD = −0.44, P < 0.001; rafter CiaHD = −0.50, P < 0.001) and fetuin A (rbefore CiaHD = −0.33, P = 0.003; rafter CiaHD = 0.31, P = 0.005) levels did not significantly change on CiaHD (‘treatment’) but are known determinants of T50 and T50Change ⁠, we ran partial regression adjusted for phosphate and fetuin A or Δ phosphate and Δ fetuin A, respectively, to screen for parameters that associated with absolute T50 values and changed in parallel to T50Change but were also significantly related to CiaHD. Herein, only albumin (rbefore CiaHD = 0.38, P = 0.001; rafter CiaHD = 0.29, P < 0.01), Δ albumin (r = 0.42, P < 0.001) but not ionized magnesium, calcium nor their respective Δs were significantly related with T50Change ⁠. For more correlations see Table 2. Interestingly, in univariate analysis per tertiles of Δ phosphate, the observed association of T50Change with Δ albumin was strongest in those with increasing or stable phosphate serum levels, indicating that decreasing serum phosphate levels might have antagonized the effects of albumin in a few patients with respect to T50Change (Figure 3). Table 2 Partial correlation matrix of absolute and Δ values of determinants of T50 and T50Change Correlation matrix T50 before CiaHD T50 after CiaHD T50Change before versus after (r; P-value) (r; P-value) (r; P-value) Phosphate (before/after CiaHD/Δ)  −0.44; <0.001  -0.50; <0.001 −0.48; <0.001 Fetuin A (before/after CiaHD/Δ)  0.33; 0.003  0.48; <0.001 −0.12; 0.28 Albumin (before/after CiaHD/Δ)a  0.38; 0.001  0.29; <0.01 0.42, <0.001 Ca2+(before/after CiaHD/Δ)a  0.04; 0.76  0.12, 0.30 −0.07, 0.51 Mg2+(before/after CiaHD/Δ)a  0.03; 0.82 0.16, 0.17 −0.03, 0.83 Correlation matrix T50 before CiaHD T50 after CiaHD T50Change before versus after (r; P-value) (r; P-value) (r; P-value) Phosphate (before/after CiaHD/Δ)  −0.44; <0.001  -0.50; <0.001 −0.48; <0.001 Fetuin A (before/after CiaHD/Δ)  0.33; 0.003  0.48; <0.001 −0.12; 0.28 Albumin (before/after CiaHD/Δ)a  0.38; 0.001  0.29; <0.01 0.42, <0.001 Ca2+(before/after CiaHD/Δ)a  0.04; 0.76  0.12, 0.30 −0.07, 0.51 Mg2+(before/after CiaHD/Δ)a  0.03; 0.82 0.16, 0.17 −0.03, 0.83 Partial correlations were adjusted for phosphate and fetuin A levels or their respective Δ values. Correlations were assessed before CiaHD and after CiaHD. In addition, we assessed the correlation of post-pre CiaHD values with T50Change ⁠. a Adjusted for phosphate and fetuin A. Table 2 Partial correlation matrix of absolute and Δ values of determinants of T50 and T50Change Correlation matrix T50 before CiaHD T50 after CiaHD T50Change before versus after (r; P-value) (r; P-value) (r; P-value) Phosphate (before/after CiaHD/Δ)  −0.44; <0.001  -0.50; <0.001 −0.48; <0.001 Fetuin A (before/after CiaHD/Δ)  0.33; 0.003  0.48; <0.001 −0.12; 0.28 Albumin (before/after CiaHD/Δ)a  0.38; 0.001  0.29; <0.01 0.42, <0.001 Ca2+(before/after CiaHD/Δ)a  0.04; 0.76  0.12, 0.30 −0.07, 0.51 Mg2+(before/after CiaHD/Δ)a  0.03; 0.82 0.16, 0.17 −0.03, 0.83 Correlation matrix T50 before CiaHD T50 after CiaHD T50Change before versus after (r; P-value) (r; P-value) (r; P-value) Phosphate (before/after CiaHD/Δ)  −0.44; <0.001  -0.50; <0.001 −0.48; <0.001 Fetuin A (before/after CiaHD/Δ)  0.33; 0.003  0.48; <0.001 −0.12; 0.28 Albumin (before/after CiaHD/Δ)a  0.38; 0.001  0.29; <0.01 0.42, <0.001 Ca2+(before/after CiaHD/Δ)a  0.04; 0.76  0.12, 0.30 −0.07, 0.51 Mg2+(before/after CiaHD/Δ)a  0.03; 0.82 0.16, 0.17 −0.03, 0.83 Partial correlations were adjusted for phosphate and fetuin A levels or their respective Δ values. Correlations were assessed before CiaHD and after CiaHD. In addition, we assessed the correlation of post-pre CiaHD values with T50Change ⁠. a Adjusted for phosphate and fetuin A. FIGURE 3 View largeDownload slide Scatter plots depicting the association of T50Change and Δ albumin per tertiles of Δ-phosphate (Δ albumin = albuminafter CiaHD − albuminbefore CiaHD). Pearson correlation was performed within these subgroups, since alterations in phosphate have been shown to be strong determinants of serum calcification propensity. Results were confirmed by plotting residuals adjusted for changing serum phosphate (data not shown). FIGURE 3 View largeDownload slide Scatter plots depicting the association of T50Change and Δ albumin per tertiles of Δ-phosphate (Δ albumin = albuminafter CiaHD − albuminbefore CiaHD). Pearson correlation was performed within these subgroups, since alterations in phosphate have been shown to be strong determinants of serum calcification propensity. Results were confirmed by plotting residuals adjusted for changing serum phosphate (data not shown). In line with these results, phosphate, fetuin A levels per se, albumin levels and CiaHD (1 = yes) were significant univariate regressors of the outcome T50 in linear mixed-effects models (not shown) and remained so after adjustment for phosphate and fetuin A (Table 3) and in the full model taking other laboratory parameters into account (Table 3). The best model fit (−2log likelihood = −284.725, df = 8) was obtained when only including CiaHD; with phosphate, fetuin A and albumin as fixed effects and subject ID as a random effect (Table 3). Table 3 Linear mixed models: effect quantification and mediation analysis Parameter SD Dependent = z(T50) adjusted for phosphate and fetuin A Dependent = z(T50) full model Dependent = z(T50) best fit model → estimate (95% CI) P-value → estimate (95% CI) P-value → estimate (95% CI) P-value CiaHD (=1) – 0.55 (0.34–0.77) <0.001 0.50 (0.27–0.74) <0.001 0.43 (0.23–0.63) <0.001 Phosphate (mmol/L) 0.5 −0.43 (−0.57 to −0.30) <0.001 −0.53 (−0.66 to −0.39) <0.001 −0.52 (−0.65 to −0.39) <0.001 Albumin (g/L) 4.0 0.29 (0.15–0.43) <0.001 0.27 (−0.13–0.41) <0.001 0.29 (0.15–0.43) <0.001 Fetuin A (g/L) 0.1 0.37 (0.23–0.40) <0.001 0.22 (0.10–0.34) 0.001 0.21 (0.09–0.33) 0.001 Mg2+ (mmol/L) 0.1 −0.07 (−0.07–0.21) 0.31 0.07 (−0.10–0.23) 0.41 – – Ca2+ (mmol/L) 0.2 −0.08 (−0.06–0.22) 0.27 0.02 (−0.15–0.19) 0.81 - – iPTH (pg/mL) 344 0.05 (−0.10–0.19) 0.52 0.02 (−0.13–0.16) 0.84 – – Hb (g/dL) 1.2 −0.01 (−0.11–0.13) 0.92 −0.02 (−0.13–0.10) 0.77 - – Dialysate HCO3−(mmol/L) 2 0.06 (−0.03–0.14) 0.18 0.06 (−0.03–0.14) 0.21 – – Parameter SD Dependent = z(T50) adjusted for phosphate and fetuin A Dependent = z(T50) full model Dependent = z(T50) best fit model → estimate (95% CI) P-value → estimate (95% CI) P-value → estimate (95% CI) P-value CiaHD (=1) – 0.55 (0.34–0.77) <0.001 0.50 (0.27–0.74) <0.001 0.43 (0.23–0.63) <0.001 Phosphate (mmol/L) 0.5 −0.43 (−0.57 to −0.30) <0.001 −0.53 (−0.66 to −0.39) <0.001 −0.52 (−0.65 to −0.39) <0.001 Albumin (g/L) 4.0 0.29 (0.15–0.43) <0.001 0.27 (−0.13–0.41) <0.001 0.29 (0.15–0.43) <0.001 Fetuin A (g/L) 0.1 0.37 (0.23–0.40) <0.001 0.22 (0.10–0.34) 0.001 0.21 (0.09–0.33) 0.001 Mg2+ (mmol/L) 0.1 −0.07 (−0.07–0.21) 0.31 0.07 (−0.10–0.23) 0.41 – – Ca2+ (mmol/L) 0.2 −0.08 (−0.06–0.22) 0.27 0.02 (−0.15–0.19) 0.81 - – iPTH (pg/mL) 344 0.05 (−0.10–0.19) 0.52 0.02 (−0.13–0.16) 0.84 – – Hb (g/dL) 1.2 −0.01 (−0.11–0.13) 0.92 −0.02 (−0.13–0.10) 0.77 - – Dialysate HCO3−(mmol/L) 2 0.06 (−0.03–0.14) 0.18 0.06 (−0.03–0.14) 0.21 – – Linear mixed model regressions for CiaHD (yes = 1) and changing covariates on T50 (outcome) were conducted for each covariate. First column reports overall SD for predictors. Second column reports single correlations after adjustment for varying serum phosphate levels. Further, the associations for these covariates were assessed in a full model including all these variables (third column) and after elimination of non-significant predictors (fourth column) for mediation analysis. All metric parameters were entered as centred variables. Bold font indicates estimates used for calculation of the indirect mediation effect: CiaHD →albumin→T50. Inclusion of pre-dialysis pH data from routine midterm BGA did not significantly affect the results depicted above (data not shown). Table 3 Linear mixed models: effect quantification and mediation analysis Parameter SD Dependent = z(T50) adjusted for phosphate and fetuin A Dependent = z(T50) full model Dependent = z(T50) best fit model → estimate (95% CI) P-value → estimate (95% CI) P-value → estimate (95% CI) P-value CiaHD (=1) – 0.55 (0.34–0.77) <0.001 0.50 (0.27–0.74) <0.001 0.43 (0.23–0.63) <0.001 Phosphate (mmol/L) 0.5 −0.43 (−0.57 to −0.30) <0.001 −0.53 (−0.66 to −0.39) <0.001 −0.52 (−0.65 to −0.39) <0.001 Albumin (g/L) 4.0 0.29 (0.15–0.43) <0.001 0.27 (−0.13–0.41) <0.001 0.29 (0.15–0.43) <0.001 Fetuin A (g/L) 0.1 0.37 (0.23–0.40) <0.001 0.22 (0.10–0.34) 0.001 0.21 (0.09–0.33) 0.001 Mg2+ (mmol/L) 0.1 −0.07 (−0.07–0.21) 0.31 0.07 (−0.10–0.23) 0.41 – – Ca2+ (mmol/L) 0.2 −0.08 (−0.06–0.22) 0.27 0.02 (−0.15–0.19) 0.81 - – iPTH (pg/mL) 344 0.05 (−0.10–0.19) 0.52 0.02 (−0.13–0.16) 0.84 – – Hb (g/dL) 1.2 −0.01 (−0.11–0.13) 0.92 −0.02 (−0.13–0.10) 0.77 - – Dialysate HCO3−(mmol/L) 2 0.06 (−0.03–0.14) 0.18 0.06 (−0.03–0.14) 0.21 – – Parameter SD Dependent = z(T50) adjusted for phosphate and fetuin A Dependent = z(T50) full model Dependent = z(T50) best fit model → estimate (95% CI) P-value → estimate (95% CI) P-value → estimate (95% CI) P-value CiaHD (=1) – 0.55 (0.34–0.77) <0.001 0.50 (0.27–0.74) <0.001 0.43 (0.23–0.63) <0.001 Phosphate (mmol/L) 0.5 −0.43 (−0.57 to −0.30) <0.001 −0.53 (−0.66 to −0.39) <0.001 −0.52 (−0.65 to −0.39) <0.001 Albumin (g/L) 4.0 0.29 (0.15–0.43) <0.001 0.27 (−0.13–0.41) <0.001 0.29 (0.15–0.43) <0.001 Fetuin A (g/L) 0.1 0.37 (0.23–0.40) <0.001 0.22 (0.10–0.34) 0.001 0.21 (0.09–0.33) 0.001 Mg2+ (mmol/L) 0.1 −0.07 (−0.07–0.21) 0.31 0.07 (−0.10–0.23) 0.41 – – Ca2+ (mmol/L) 0.2 −0.08 (−0.06–0.22) 0.27 0.02 (−0.15–0.19) 0.81 - – iPTH (pg/mL) 344 0.05 (−0.10–0.19) 0.52 0.02 (−0.13–0.16) 0.84 – – Hb (g/dL) 1.2 −0.01 (−0.11–0.13) 0.92 −0.02 (−0.13–0.10) 0.77 - – Dialysate HCO3−(mmol/L) 2 0.06 (−0.03–0.14) 0.18 0.06 (−0.03–0.14) 0.21 – – Linear mixed model regressions for CiaHD (yes = 1) and changing covariates on T50 (outcome) were conducted for each covariate. First column reports overall SD for predictors. Second column reports single correlations after adjustment for varying serum phosphate levels. Further, the associations for these covariates were assessed in a full model including all these variables (third column) and after elimination of non-significant predictors (fourth column) for mediation analysis. All metric parameters were entered as centred variables. Bold font indicates estimates used for calculation of the indirect mediation effect: CiaHD →albumin→T50. Inclusion of pre-dialysis pH data from routine midterm BGA did not significantly affect the results depicted above (data not shown). Since only albumin changed significantly during 3 months of CiaHD and remained independently associated with the outcome T50, but the effect was reduced by addition of CiaHD (‘treatment’) to the model, some of the effect observed during CiaHD might have been mediated by increasing albumin levels. The indirect effect estimate (Judd and Kenny coefficient) reflecting ‘CiaHD (1 = after CiaHD treatment) → albumin → T50’ was 0.55 − 0.43 = 0.12 (Table 3), meaning that ∼22% (0.12/0.55) of the beneficial T50-modifying effect observed during 3 months of CiaHD can be mathematically attributed to increasing serum albumin after adjusting for variations in serum phosphate and fetuin A levels (both unrelated to CiaHD treatment). DISCUSSION In this pre–post-quasi-interventional study assessing the influence of an acetate-free, citrate-based (1 mmol/L) A-concentrate, we found a significant improvement of serum calcification propensity T50 after switching to CiaHD for 3 months. In line with this, no further increase of T50 was observed in 44 patients after switching back to AaHD. The calculated T50Change values were significantly different for these treatment phases. As a natural liver substrate, citrate has a reported half-life of 5–60 min [28, 29]. We did find comparable levels of citrate and acetate in pre-dialysis sera in a sample of patients after switching treatment modalities. However, our data indicate that CiaHD solutions might still sustainably alter the calcification resilience in HD patients upon long-term use, although citrate per se has already been metabolized by the next session. Although intra-individual phosphate and fetuin A levels were associated with changing T50 in accordance with previous reports [7, 30], overall phosphate and fetuin A levels remained stable for 3 months, indicating that the beneficial effect observed during CiaHD were not mediated by alteration in these parameters. Likewise, although we detected decreased ionized serum magnesium and calcium levels after 3 months of CiaHD, these factors were not independently associated with changing T50. Instead, we observed increasing albumin levels after the treatment phase, which has also been observed in other studies assessing the impact of midterm (3 months) acetate-free HD or HDF treatment [18, 19]. Further, long-term acetate-free dialysis was shown to reduce post-dialysis malaise [17] and improve the creatinine index, a nutritional marker associated with survival in HD patients [19, 31]. In contrast, no variation of serum albumin was seen in Grundstrom et al.’s short-term (6 weeks) randomized cross-over trial, demonstrating safety and biocompatibility of CiaHD [9]. Nevertheless, in our study population, albumin was the only identified factor that was significantly associated with both the treatment modality and the outcome (changing T50). In this context, albumin seems to mediate a minor part of the beneficial treatment effect observed during CiaHD, which could route back to improved well-being and better nutritional status in these patients. However, no changes in BMI and upper arm circumference were observed during treatment (not shown), maybe related to the short observation times. An alternative hypothesis would be that citrate could contribute to the dissolution of primary CPP during dialysis. This effect could also explain the increased availability of albumin and the improvement of the T50 test. Yet, although we replicated the known association of fetuin A with T50, we did not find increasing serum levels of this anti-calcific protein. Unfortunately, we did not have data on pre-dialysis serum bicarbonate at the time of T50 assessment since improved acid-balance on CiaHD might as well have mediated another part of the observed effect. Yet, published data remain conflicting on the effect of citrate-acidified A-concentrates on pre-dialysis acidosis [11, 12]. In this context it should be noted that pre-dialysis pH, base excess and bicarbonate from routine venous BGA at the middle of the treatment phases remained unaffected by treatment modality (CiaHD versus AaHD) in a subgroup of our patients with available data. Further, as the solutions for performing the T50 assay are strongly buffered with 50 mmol HEPES, an impact of the actual serum pH on T50 assay results is unlikely [4]. Thus, although we are unable to rule out improved pre-dialysis blood pH as a mediator of the observed variation of T50 with absolute certainty, we consider a major modulation of T50 by pre-dialysis pH unlikely in our study. It is worth mentioning that citrate as a chelator of bivalent ions has similar affinity for ionized magnesium as for calcium [32] and might thereby not only negatively affect calcium [11, 21], but also magnesium mass balance. Whereas the former effect is certainly desirable, magnesium probably has a more anti-calcifying effect and can modulate serum calcification propensity (increasing T50) in vitro [5, 33, 34]. In our study, the effects of lower ionized calcium and magnesium levels might have eliminated each other with respect to T50. Since we observed similar pre-dialytic citrate levels on both CiaHD and AaHD, the concept of direct complexing of bivalent cations by citrate does not help us in explaining our observations. Instead, we rather consider the changes in ionized calcium and magnesium a phenomenon of complexing to rather long-lasting molecules or higher regeneration of anti-calcific buffer proteins (e.g. albumin). This hypothesis also explains why the modulation of T50 lasts until the next dialysis session although citrate has already been metabolized. Nevertheless, we are unable to provide definite evidence for these ideas. Despite the decent number of patients within a longitudinal design, this post hoc study has several limitations. Patients had not been blinded to the treatment. The pre–post (non-randomized controlled trial) design did not include a placebo-controlled arm, so this effect cannot be excluded with certainty. Further, the relatively short study period and the overall number of patients did not permit inferences regarding mortality. Moreover, we only assessed basic routine laboratory parameters. Differential elimination of uraemic toxins or an alteration of the anion gap’s composition on CiaHD versus AaHD might as well have influenced T50 in addition to albumin. The study protocol allowed regular alterations in medications during periods including phosphate binders, which per se remained unaltered during treatment but might have masked correlations of naturally weaker determinants of T50 and outcome during CiaHD. Although no further increase of T50 was observed in 44 patients and T50Change differed significantly between periods, it cannot be ruled out with absolute certainty that time per se, instead of CiaHD treatment, could have been the relevant determinant. After all, this scenario appears unlikely, as T50 has been shown to time-dependently decrease in HD patients on standard bicarbonate AaHD regimes [7]. Lastly, the exact mechanisms underlying increasing albumin levels and T50 after 3 months of CiaHD dialysis remain elusive and the post hoc character of this study only allows the generation of hypotheses that need elaboration and proof from further studies including clinical endpoints. Intriguingly, hints at improved survival due to acetate-free dialysis regimes have been reported from registries [35]. We recently reported an association of worsening calcification propensity with all-cause and cardiovascular mortality in a subgroup of patients from the observational ISAR cohort [7]. Following up on these findings, the present study indicates that aside from strict phosphate control, the putative risk factor T50 can be therapeutically modified by acetate-free dialysis regimes. ACKNOWLEDGEMENTS Special thanks are to be given to R.B. and staff for making this study possible. Isabelle Gsponer and Sandra Haderer must be thanked for technical support. FUNDING G.L. and C.S. received investigator-initiated research grant support from Baxter for the mother study. G.L. was supported by a research scholarship from the Else Kröner-Fresenius-Stiftung for physician scientists. U.H. reports grants from Baxter during the conduct of the study. This work was supported by the German Research Foundation (DFG) and the Technical University of Munich within the funding programme Open Access Publishing. This work was supported by the President’s International Fellowship Initiative of The Chinese Academy of Science (2015VBB045 to T.M.), the National Natural Science Foundation of China (31450110423 to T.M.), the Austrian Science Fund (FWF: P28854 to T.M.), the Austrian Research Promotion Agency (FFG: 864690 to T.M.), the Integrative Metabolism Research Center Graz, the Austrian infrastructure program 2016/2017, BioTechMed/Graz and the OMICS Center Graz. AUTHORS’ CONTRIBUTIONS All listed authors have contributed sufficiently to the project. G.L. and C.S. were responsible for the study design. G.L., C.M., B.H. and S.W. were responsible for statistical analysis. R.B. recruited all patients and provided access to patients’ data and secured protocol adherence. J.C-L., A.S., C.S., V.B., S.W., J.A., S.K. and M.C.B. contributed to patients’ data acquisition and entry. C.C.M. assessed ionized calcium and magnesium values. G.L. drafted the manuscript, which was reviewed and approved by all authors. S.A., U.H. and C.S. were responsible for the database and ethics approval. A.P. evaluated T50 in a blinded fashion and contributed to the manuscript. T.M. and S.S. performed NMR-based metabolomics technically and were responsible for the quantification of metabolites during the revision process. CONFLICT OF INTEREST STATEMENT A.P. is an employee and stockholder of Calciscon. 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 – 1305 Google Scholar Crossref Search ADS PubMed 2 Foley RN , Parfrey PS , Sarnak MJ. Epidemiology of cardiovascular disease in chronic renal disease . J Am Soc Nephrol 1998 ; 9 : S16 – S23 Google Scholar PubMed 3 Stokes JB. Consequences of frequent hemodialysis: comparison to conventional hemodialysis and transplantation . Trans Am Clin Climatol Assoc 2011 ; 122 : 124 – 136 Google Scholar PubMed 4 Pasch A. Novel assessments of systemic calcification propensity . Curr Opin Nephrol Hypertens 2016 ; 25 : 278 – 284 Google Scholar Crossref Search ADS PubMed 5 Pasch A , Farese S , Graber S et al. Nanoparticle-based test measures overall propensity for calcification in serum . J Am Soc Nephrol 2012 ; 23 : 1744 – 1752 Google Scholar Crossref Search ADS PubMed 6 Smith ER , Ford ML , Tomlinson LA et al. Serum calcification propensity predicts all-cause mortality in predialysis CKD . J Am Soc Nephrol 2014 ; 25 : 339 – 348 Google Scholar Crossref Search ADS PubMed 7 Lorenz G , Steubl D , Kemmner S et al. Worsening calcification propensity precedes all-cause and cardiovascular mortality in haemodialyzed patients . Sci Rep 2017 ; 7 : 13368 Google Scholar Crossref Search ADS PubMed 8 Pasch A , Covic A , Marzell B et al. Serum calcification propensity of HD patients is therapeutically improved by a calcium acetate/magnesium carbonate containing phosphate binder [abstract]. In: 51st Congress of the European-Renal-Association(ERA)/European-Dialysis-and-Transplant-Association (EDTA), Amsterdam, Netherlands, 31 May–03 Jun 2014. Oxford: Oxford University Press. 9 Grundstrom G , Christensson A , Alquist M et al. Replacement of acetate with citrate in dialysis fluid: a randomized clinical trial of short term safety and fluid biocompatibility . BMC Nephrol 2013 ; 14 : 216 Google Scholar Crossref Search ADS PubMed 10 Panesar M , Shah N , Vaqar S et al. Changes in serum bicarbonate levels caused by acetate-containing bicarbonate-buffered hemodialysis solution: an observational prospective cohort study . Ther Apher Dial 2017 ; 21 : 157 – 165 Google Scholar Crossref Search ADS PubMed 11 Gabutti L , Lucchini B , Marone C et al. Citrate- vs. acetate-based dialysate in bicarbonate haemodialysis: consequences on haemodynamics, coagulation, acid-base status, and electrolytes . BMC Nephrol 2009 ; 10 : 7 – 7 Google Scholar Crossref Search ADS PubMed 12 De Sequera Ortiz P , Albalate Ramón M , Pérez-García R et al. Acute effect of citrate bath on postdialysis alkalaemia . Nefrologia 2015 ; 35 : 164 – 171 Google Scholar Crossref Search ADS PubMed 13 Wolever TM , Fernandes J , Rao AV. Serum acetate: propionate ratio is related to serum cholesterol in men but not women . J Nutr 1996 ; 126 : 2790 – 2797 Google Scholar PubMed 14 Fournier G , Potier J , Thebaud HE et al. Substitution of acetic acid for hydrochloric acid in the bicarbonate buffered dialysate . Artif Organs 1998 ; 22 : 608 – 613 Google Scholar Crossref Search ADS PubMed 15 Tessitore N , Santoro A , Panzetta GO et al. Acetate-free biofiltration reduces intradialytic hypotension: a European multicenter randomized controlled trial . Blood Purif 2012 ; 34 : 354 – 363 Google Scholar Crossref Search ADS PubMed 16 Verzetti G , Navino C , Bolzani R et al. Acetate-free biofiltration versus bicarbonate haemodialysis in the treatment of patients with diabetic nephropathy: a cross-over multicentric study . Nephrol Dial Transplant 1998 ; 13 : 955 – 961 Google Scholar Crossref Search ADS PubMed 17 Daimon S , Dan K , Kawano M. Comparison of acetate-free citrate hemodialysis and bicarbonate hemodialysis regarding the effect of intra-dialysis hypotension and post-dialysis malaise . Ther Apher Dial 2011 ; 15 : 460 – 465 Google Scholar Crossref Search ADS PubMed 18 Kuragano T , Kida A , Furuta M et al. Effects of acetate-free citrate-containing dialysate on metabolic acidosis, anemia, and malnutrition in hemodialysis patients . Artif Organs 2012 ; 36 : 282 – 290 Google Scholar Crossref Search ADS PubMed 19 Matsuyama K , Tomo T , Kadota J. Acetate-free blood purification can impact improved nutritional status in hemodialysis patients . J Artif Organs 2011 ; 14 : 112 – 119 Google Scholar Crossref Search ADS PubMed 20 Richtrova P , Mares J , Kielberger L et al. Citrate-buffered dialysis solution (Citrasate) allows avoidance of anticoagulation during intermittent hemodiafiltration-at the cost of decreased performance and systemic biocompatibility . Artif Organs 2017 ; 41 : 759 – 766 Google Scholar Crossref Search ADS PubMed 21 Perez-Garcia R , Albalate M , Sequera P et al. Calcium mass balance with citrate dialysate is lower than with acetate . Nefrologia 2017 ; 37 : 109 – 110 Google Scholar Crossref Search ADS PubMed 22 Liu J , Huang Z , Gilbertson DT et al. An improved comorbidity index for outcome analyses among dialysis patients . Kidney Int 2010 ; 77 : 141 – 151 Google Scholar Crossref Search ADS PubMed 23 Jain A , Bhayana S , Vlasschaert M et al. A formula to predict corrected calcium in haemodialysis patients . Nephrol Dial Transplant 2008 ; 23 : 2884 – 2888 Google Scholar Crossref Search ADS PubMed 24 Pasch A , Block GA , Bachtler M et al. Blood calcification propensity, cardiovascular events, and survival in patients receiving hemodialysis in the EVOLVE Trial . Clin J Am Soc Nephrol 2017 ; 12 : 315 – 322 Google Scholar Crossref Search ADS PubMed 25 Ix JH , Shlipak MG , Brandenburg VM et al. Association between human fetuin-A and the metabolic syndrome: data from the Heart and Soul Study . Circulation 2006 ; 113 : 1760 – 1767 Google Scholar Crossref Search ADS PubMed 26 Prokesch A , Graef FA , Madl T et al. Liver p53 is stabilized upon starvation and required for amino acid catabolism and gluconeogenesis . FASEB J 2017 ; 31 : 732 – 742 Google Scholar Crossref Search ADS PubMed 27 Judd CM , Kenny DA. Process analysis: estimating mediation in treatment evaluations . Eval Rev 1981 ; 5 : 602 – 619 Google Scholar Crossref Search ADS 28 Kozik-Jaromin J , Nier V , Heemann U et al. Citrate pharmacokinetics and calcium levels during high-flux dialysis with regional citrate anticoagulation . Nephrol Dial Transplant 2009 ; 24 : 2244 – 2251 Google Scholar Crossref Search ADS PubMed 29 Davenport A , Tolwani A. Citrate anticoagulation for continuous renal replacement therapy (CRRT) in patients with acute kidney injury admitted to the intensive care unit . NDT Plus 2009 ; 2 : 439 – 447 Google Scholar PubMed 30 Dekker M , Pasch A , van der Sande F et al. High-flux hemodialysis and high-volume hemodiafiltration improve serum calcification propensity . PLoS One 2016 ; 11 : e0151508 Google Scholar Crossref Search ADS PubMed 31 Desmeules S , Levesque R , Jaussent I et al. Creatinine index and lean body mass are excellent predictors of long-term survival in haemodiafiltration patients . Nephrol Dial Transplant 2004 ; 19 : 1182 – 1189 Google Scholar Crossref Search ADS PubMed 32 Lee G , Arepally GM. Anticoagulation techniques in apheresis: from heparin to citrate and beyond . J Clin Apheresis 2012 ; 27 : 117 – 125 Google Scholar Crossref Search ADS PubMed 33 Sakaguchi Y , Hamano T , Isaka Y. Effects of magnesium on the phosphate toxicity in chronic kidney disease: time for intervention studies . Nutrients 2017 ; 9 : 112 Google Scholar Crossref Search ADS 34 Sakaguchi Y , Hamano T , Nakano C et al. Association between density of coronary artery calcification and serum magnesium levels among patients with chronic kidney disease . PLoS One 2016 ; 11 : e0163673 Google Scholar Crossref Search ADS PubMed 35 Mercadal L , Franck JE , Metzger M et al. Improved survival associated with acetate-free haemodialysis in elderly: a registry-based study . Nephrol Dial Transplant 2015 ; 30 : 1560 – 1568 Google Scholar Crossref Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nephrology Dialysis Transplantation Oxford University Press

Loading next page...
 
/lp/ou_press/acetate-free-citrate-acidified-bicarbonate-dialysis-improves-serum-0PichugaJc
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved.
ISSN
0931-0509
eISSN
1460-2385
D.O.I.
10.1093/ndt/gfy134
Publisher site
See Article on Publisher Site

Abstract

Abstract Background A novel in vitro test (T50 test) assesses ex vivo serum calcification propensity and predicts mortality in chronic kidney disease and haemodialysis (HD) patients. For the latter, a time-dependent decline of T50 was shown to relate to mortality. Here we assessed whether a 3-month switch to acetate-free, citrate-acidified, standard bicarbonate HD (CiaHD) sustainably improves calcification propensity. Methods T50 values were assessed in paired midweek pre-dialysis sera collected before and 3 months after CiaHD in 78 prevalent European HD patients. In all, 44 were then switched back to acetate. Partial correlation was used to study associations of changing T50 and changing covariates. Linear mixed effect models were built to assess the association of CiaHD and covariates with changing T50. Results A significant intra-individual increase of serum calcification resilience was found after 3 months on CiaHD (206  ±  56 to 242  ±  56 min; P < 0.001), but not after switching back to acetate (252  ±  63 to 243  ±  64 min; n = 44; P = 0.29). CiaHD, Δ serum phosphate and Δ albumin but not Δ ionized calcium and magnesium were the strongest determinants of changing T50. Beneath T50, only serum albumin but not phosphate changed significantly during 3 months of CiaHD. Conclusion CiaHD dialysis favourably affected calcification propensity as measured by the T50 test. Whether this treatment, beyond established phosphate-directed treatments, has the potential to sustainably tip the balance towards a more anti-calcific serum milieu needs to be further investigated. albumin, citrate, dialysis, serum calcification propensity, T50 INTRODUCTION Patients on haemodialysis (HD) experience excess cardiovascular morbidity and mortality, which is related to a disturbed calcium and phosphate homeostasis promoting vascular disease progression [1–3]. Recently an in vitro test (T50 test) was developed by Pasch and colleagues. This test time-dependently assesses the calcification propensity of human serum after the addition of supraphysiologic doses of calcium and phosphate [4, 5]. Basically the test determines the crystal formation time from amorphous primary calciprotein particles (CPP) into crystallized secondary CPP in human serum [4]. The advantage of this assay is its ability to integrate the complex interplay of pro- and anti-calcific players (e.g. calcium and phosphate versus magnesium and albumin) into a functional assay conceptually comparable to functionally testing the activated clotting time [4–6]. Shorter T50 times predict vascular stiffness progression and mortality in chronic kidney disease Stages 3 and 4 [6]. We recently found that in stable HD patients, serum calcification propensity declined over the course of 2 years, indicating a loss of calcification resilience [7]. In this post hoc analysis of 188 stable dialysis patients, the time-dependent decline of serum calcification propensity T50 was associated with overall and cardiovascular mortality [7]. However, this study left the question unanswered as to which specific measures might be associated with stabilization, i.e. the prevention of the decline of serum calcification propensity T50 [8]. Acetate is widely used as a stabilizing acidic agent within the cation concentrate (A-concentrate) in standard bicarbonate HD, hereafter referred to as acetate-acidified standard bicarbonate haemodialysis (AaHD). In Europe, this A-concentrate usually contains 3–4 mmol/L of acetate [9–12]. Whereas physiological and pre-dialysis acetate concentrations are <100 µmol/L, intradialytic concentrations increase up to 10-fold above normal and range from <100 to 222 µmol/L immediately after dialysis, depending on the metabolization status [13, 14]. This results in intra-individually varying post-dialysis acetate levels of up to 500 µmol/L. Acetate is not free of side effects, and intradialytic hypotensive episodes are less frequent in acetate-free dialysis modalities [15–17]. Further replacement of acetate with the putatively more biocompatible citric acid, which is converted to citrate, as an acidifier of the A-concentrate in bicarbonate HD (CiaHD) [9] was reported to provide better control of acidic balance [12] and to positively affect serum albumin and nutritional status in those with hypoalbuminemia [18, 19]. Other studies suggest better elimination of phosphate and less need for anticoagulation [11, 20]. As a chelator of bivalent anions, citrate could further influence net calcium mass balance and the pro-calcific environment in HD patients [21]. In this prospective analysis of stored serum samples we wanted to assess how 3-months use of an acetate-free, 1 mmol citrate-acidified A-concentrate (SelectBagCitrate, Gambro/Baxter, Hechingen, Germany) would impact on serum calcification propensity (T50 time) in HD patients. Therefore we determined the T50 time in paired pre-dialysis sera collected before and after 3 months on CiaHD in 78 eligible patients. Of these, 44 patients had been switched back to AaHD thereafter. MATERIALS AND METHODS Study protocol and study subjects This pre–post-quasi-interventional study was conducted in a subgroup of the ‘Substitution of Acetate by Citrate in Bicarbonate-Based-Hemodialysis’ study (NCT02745340). Between April and June 2016, 88 HD patients were recruited from two local dialysis units of the same dialysis centre in Munich, Germany. In all, 78 patients completed the full observation period. According to the registered study protocol, all patients started out on AaHD (SelectBagOne; 3 mmol of acetate). Dialysis Unit 1 was then switched to CiaHD (SelectBagCitrate; acidified with citric acid, which is converted to 1 mmol/L citrate) for 3 months, while Unit 2 remained on AaHD. Predicted solute concentrations resulting from the A-concentrates as provided by Gambro/Baxter can be retrieved from Supplementary data, Table S1. After 3 months and a washout of 48 h, units were switched to AaHD or CiaHD, respectively. Unit 1 was afterwards switched back to AaHD. Of the 78 patients, 44 had serum available for determination of T50 at baseline, after 3 months of CiaHD and after another 3 months back on CiaHD (Figure 1). Inclusion criteria were age  ≥18 years, HD vintage of at least 90 days on a thrice-weekly schedule with a session duration of  ≥4 h. Pregnancy, ongoing infection and lack of written and informed consent were the main exclusion criteria. The study protocol was approved by the Ethics Commission of the Klinikum rechts der Isar, Technical University Munich. It was carried out in accordance with the Declaration of Helsinki, adhering to good clinical practice. The original study was registered at ClinicalTrials.gov (NCT02745340). Written informed consent was obtained from all participants. For the derived study, no additional ethics committee approval was sought. FIGURE 1 View largeDownload slide Observation schedule and study time course. All patients had been on standard bicarbonate dialysis with an AaHD for at least 2 months. Then Unit 1 (n = 47) underwent 3 months of CiaHD. After 3 months Unit 2 (n = 31) was also switched to CiaHD, whereas Unit 1 underwent another 3 months of AaHD. Dashed lines indicate groups before versus after CiaHD dialysis for pooled data analysis. T50 was determined at the indicated time points from pre-dialysis serum samples. FIGURE 1 View largeDownload slide Observation schedule and study time course. All patients had been on standard bicarbonate dialysis with an AaHD for at least 2 months. Then Unit 1 (n = 47) underwent 3 months of CiaHD. After 3 months Unit 2 (n = 31) was also switched to CiaHD, whereas Unit 1 underwent another 3 months of AaHD. Dashed lines indicate groups before versus after CiaHD dialysis for pooled data analysis. T50 was determined at the indicated time points from pre-dialysis serum samples. Clinical data assessment Patients’ age, comorbidities and medication were continuously assessed using medical records and patient interviews at inclusion and after the CiaHD or AaHD phases (3 months). Comorbidities were recorded following Liu et al.’s adapted version of the Charlson Comorbidity Index (CCI) [22]. Body mass index (BMI) was calculated as body weight/height2 (kg/m2). All patients underwent bicarbonate dialysis with different types of synthetic membranes that remained constant during the study period. Dialysis prescription [ultrafiltration, session duration, Kt/V as a measure of dialysis efficiency, HD/haemodiafiltration (HDF), anticoagulation] was provided by the units. Medical staff were the same for both dialysis units. A total of 10 of the included patients did not participate in the present study, as they had either died (n = 5), moved (n = 1), were hospitalized for a longer period (n = 2) or reported an unwillingness to donate blood specimens at any of the required time points (n = 2). Missing data at relevant time points were Kt/V (n = 3), intact parathyroid hormone (iPTH; n = 2), C-reactive protein (CRP; n = 1) and n = 2 for ionized magnesium and calcium levels. The latter two were estimated from total corrected calcium and magnesium values [23], choosing their individual ratios of ionized calcium or magnesium levels/total calcium or magnesium values as conversion factor. Exclusion of these imputed values revealed comparable results to the presented data (not shown). Blood specimen collection and laboratory methods Serum was collected prior to a midweek dialysis session at the time points indicated (Figure 1). After 30 min at room temperature, serum was centrifuged (2000 g, 10 min), aliquoted and frozen at −80°C. Routine laboratory analysis was performed by ISO accredited laboratories. Ionized magnesium and calcium levels were determined from frozen sera using the Nova 8 Analyzer and ion-selective electrodes for calcium and magnesium (Nova Biomedical, Waltham, MA, USA). pH, base excess and standard bicarbonate values were derived from routinely available pre-dialysis venous blood gas analyses (BGAs) at the middle of each treatment period and were performed by ISO-accredited laboratories. T50 was determined as previously described by Pasch et al. [5, 24]. Frozen sera were sent to Bern on dry ice and analysed in a blinded manner. Likewise, fetuin A was assessed from 4-fold diluted sera (phosphate-buffered saline) in a blinded manner using a nephelometric assay and a polyclonal rabbit anti-human fetuin A antibody following published protocols [25]. Citrate and acetate concentrations in serum were determined from a subset of 15 patients on both CiaHD and AaHD treatment using one dimensional 1H-nuclear magnetic resonance (NMR) spectroscopy as formerly described by Prokesch et al. (Graz, Austria) [26]. Statistical analysis SPSS Statistics 23 (IBM, Armonk, NY, USA) was used for statistical analysis. For pooled data before and after CiaHD, we report mean  ±  SD, median and interquartile range (IQR) or counts and percentage of total as appropriate. The change in T50 (⁠ T50Change ⁠) was calculated as (⁠ T50before CiaHD −  T50after CiaHD ⁠)/ T50before CiaHD *100 (%). Accordingly, T50Change in 44 patients, who underwent an additional 3 months of AaHD treatment was calculated as (⁠ T50after AaHD −  T50after CiaHD ⁠)/ T50after CiaHD *100 (%). Increasing T50Change was defined by T50Change values >5.1% before versus after the respective treatment period according to the interassay coefficient for T50 standards at 260 min in the Evaluation of Cinacalcet Hydrochloride Therapy to Lower Cardiovascular Events (EVOLVE) cohort and in accordance with our previous report [7, 24]. A paired t-test was used to compare differences between absolute T50 values before and after CiaHD. Unpaired t-tests were used to compare already normalized T50Change values in 44 patients on AaHD versus CiaHD. Group differences for basal comorbidities in those not included were tested using analysis of variance, Kruskal–Wallis and Fisher’s exact test. Partial regression adjusted for phosphate was used for correlating T50, T50Change with standard laboratory parameters and their respective Δ values. For example, intra-individual Δ − albumin = albuminafter CiaHD − albuminbefore CiaHD (mmol/L). Linear mixed-effects models assessing the association of treatment time points before (CiaHD = 0) versus after citrate (CiaHD = 1), adjusted for other laboratory parameters (additional fixed effects), were built based on an unstructured covariance matrix (best model fit according to −2 log likelihood) including patients’ IDs as random effects. Regression estimates are reported per 1 SD increase of the regressor. Indirect effect coefficients were calculated according to Judd and Kenny [27]. A P-value <0.05 was considered significant. RESULTS Study design and study population This study was performed as a prospective secondary analysis of the Substitution of Acetate by Citrate in Bicarbonate-Based Hemodialysis study. Of 88 originally included patients, 78 patients were accessible for the study time course (Figure 1). The study population (n = 78) did not significantly differ from the originally included patient collective (n = 88) with respect to age, gender, basic laboratory workup and dialysis modalities (Supplementary data, Table S2). The median age of the study population was 75.1 years (IQR 57.2–80.3), with 41 (53%) being male patients on HD treatment (96%). Three patients received HDF treatment. Three (4%) black participants were included. Three (4%) patients received phosphate binders while on CiaHD, four (5%) patients received increased doses of calcium acetate, three (4%) patients received increased doses of other phosphate binders and only one patient was prescribed an increased dose of active vitamin D. For detailed baseline descriptions at baseline after 3 months see Table 1. Table 1 Characteristics of the study population at study baseline (before CiaHD) and after CiaHD Parameter Before CiaHD After 3 months of CiaHD Age (years), median (IQR) 75.1 (57.2–80.3) +0.25 Gender (males), n (%) 41 (53) – BMI (kg/m2), mean ± SD 26.2 ± 5.1 – White patients, n (%) 75 (96) – Upper arm circumference (cm), mean ± SD 28.7 ± 3.6 28.7 ± 3.5 Adapted CCI (0–21), median (IQR) 3 (1–6) 4 (2–7)b History of MI, n (%) 15 (19) 15 (19) Hypertensive heart disease, n (%) 40 (51) 41 (53) Diabetes, n (%) 26 (33) 28 (36) COPD, n (%) 7 (9) 9 (11) Cancer, n (%) 30 (39) 33 (42) CHD, n (%) 26 (33) 27 (35) PAOD, n (%) 11 (14) 13 (16) CerVD, n (%) 13 (16) 13 (16) CHF, n (%) 8 (10) 8 (10) Atrial fibrillation, n (%) 17 (22) 19 (24) GIT bleeding, n (%) 5 (6) 5 (6) Smoking (at the moment), n (%) 8 (10) 8 (10) HD not HDF, n (%) 75 (96) 75 (96) HD vintage (months), median (IQR) 36 (17–56) +0.25 Bicarbonate setting, median (IQR) 32 (30–33) 32 (30–33) Kt/Va, mean ± SD 1.2 ± 0.3 1.2 ± 0.3 Haemoglobin (g/dL), mean ± SD 11.9 ± 1.3 11.6 ± 1.2 Ionized magnesium (mmol/L)a, mean ± SD 0.5 ± 0.1 0.4 ± 0.1b Ionized calcium (mmol/L)a, mean ± SD 1.2 ± 0.2 1.1 ± 0.2b Phosphate (mmol/L), mean ± SD 1.7 ± 0.5 1.8 ± 0.5 Albumin (g/L)a, mean ± SD 38 ± 4 40 ± 5b Fetuin A (g/L), mean ± SD 0.37 ± 0.11 0.36 ± 0.08 iPTH (pg/mL)a, median (IQR) 267 (134–439) 254 (134–484) CRP (mg/L)a, mean ± SD 10.4 ± 21.8 9 ± 12.3 New PB while on CiaHD n (%) 3 (4) – Increase of calcium containing PB while on CiaHD, n (%) 4 (5) – Increase of other PB while on CiaHD, n (%) 3 (4) – Increase of active vitamin D while on CiaHD, n (%) 1 (1) – Dialysis centre (1 not 2), n (%) 47 (60) – Parameter Before CiaHD After 3 months of CiaHD Age (years), median (IQR) 75.1 (57.2–80.3) +0.25 Gender (males), n (%) 41 (53) – BMI (kg/m2), mean ± SD 26.2 ± 5.1 – White patients, n (%) 75 (96) – Upper arm circumference (cm), mean ± SD 28.7 ± 3.6 28.7 ± 3.5 Adapted CCI (0–21), median (IQR) 3 (1–6) 4 (2–7)b History of MI, n (%) 15 (19) 15 (19) Hypertensive heart disease, n (%) 40 (51) 41 (53) Diabetes, n (%) 26 (33) 28 (36) COPD, n (%) 7 (9) 9 (11) Cancer, n (%) 30 (39) 33 (42) CHD, n (%) 26 (33) 27 (35) PAOD, n (%) 11 (14) 13 (16) CerVD, n (%) 13 (16) 13 (16) CHF, n (%) 8 (10) 8 (10) Atrial fibrillation, n (%) 17 (22) 19 (24) GIT bleeding, n (%) 5 (6) 5 (6) Smoking (at the moment), n (%) 8 (10) 8 (10) HD not HDF, n (%) 75 (96) 75 (96) HD vintage (months), median (IQR) 36 (17–56) +0.25 Bicarbonate setting, median (IQR) 32 (30–33) 32 (30–33) Kt/Va, mean ± SD 1.2 ± 0.3 1.2 ± 0.3 Haemoglobin (g/dL), mean ± SD 11.9 ± 1.3 11.6 ± 1.2 Ionized magnesium (mmol/L)a, mean ± SD 0.5 ± 0.1 0.4 ± 0.1b Ionized calcium (mmol/L)a, mean ± SD 1.2 ± 0.2 1.1 ± 0.2b Phosphate (mmol/L), mean ± SD 1.7 ± 0.5 1.8 ± 0.5 Albumin (g/L)a, mean ± SD 38 ± 4 40 ± 5b Fetuin A (g/L), mean ± SD 0.37 ± 0.11 0.36 ± 0.08 iPTH (pg/mL)a, median (IQR) 267 (134–439) 254 (134–484) CRP (mg/L)a, mean ± SD 10.4 ± 21.8 9 ± 12.3 New PB while on CiaHD n (%) 3 (4) – Increase of calcium containing PB while on CiaHD, n (%) 4 (5) – Increase of other PB while on CiaHD, n (%) 3 (4) – Increase of active vitamin D while on CiaHD, n (%) 1 (1) – Dialysis centre (1 not 2), n (%) 47 (60) – Demographics, comorbidities and basic laboratory values from pooled analysis (both dialysis units) are reported before and after CiaHD. CerVD, cerebral vascular disease; CHD, coronary heart disease; CHF, congestive heart failure; COPD, chronic obstructive pulmonary disease; CRP, C-reactive protein; GIT, gastrointestinal; iPTH, intact parathyroid hormone; MI, myocardial infarction; PAOD, peripheral arterial occlusive disease; PB, phosphate binders. a Missing values: 1–3 missing values, due to missed quartal assessment. b Paired samples t-test and Wilcoxon signed-rank test were used for comparing before versus after citrate P < 0.05. Table 1 Characteristics of the study population at study baseline (before CiaHD) and after CiaHD Parameter Before CiaHD After 3 months of CiaHD Age (years), median (IQR) 75.1 (57.2–80.3) +0.25 Gender (males), n (%) 41 (53) – BMI (kg/m2), mean ± SD 26.2 ± 5.1 – White patients, n (%) 75 (96) – Upper arm circumference (cm), mean ± SD 28.7 ± 3.6 28.7 ± 3.5 Adapted CCI (0–21), median (IQR) 3 (1–6) 4 (2–7)b History of MI, n (%) 15 (19) 15 (19) Hypertensive heart disease, n (%) 40 (51) 41 (53) Diabetes, n (%) 26 (33) 28 (36) COPD, n (%) 7 (9) 9 (11) Cancer, n (%) 30 (39) 33 (42) CHD, n (%) 26 (33) 27 (35) PAOD, n (%) 11 (14) 13 (16) CerVD, n (%) 13 (16) 13 (16) CHF, n (%) 8 (10) 8 (10) Atrial fibrillation, n (%) 17 (22) 19 (24) GIT bleeding, n (%) 5 (6) 5 (6) Smoking (at the moment), n (%) 8 (10) 8 (10) HD not HDF, n (%) 75 (96) 75 (96) HD vintage (months), median (IQR) 36 (17–56) +0.25 Bicarbonate setting, median (IQR) 32 (30–33) 32 (30–33) Kt/Va, mean ± SD 1.2 ± 0.3 1.2 ± 0.3 Haemoglobin (g/dL), mean ± SD 11.9 ± 1.3 11.6 ± 1.2 Ionized magnesium (mmol/L)a, mean ± SD 0.5 ± 0.1 0.4 ± 0.1b Ionized calcium (mmol/L)a, mean ± SD 1.2 ± 0.2 1.1 ± 0.2b Phosphate (mmol/L), mean ± SD 1.7 ± 0.5 1.8 ± 0.5 Albumin (g/L)a, mean ± SD 38 ± 4 40 ± 5b Fetuin A (g/L), mean ± SD 0.37 ± 0.11 0.36 ± 0.08 iPTH (pg/mL)a, median (IQR) 267 (134–439) 254 (134–484) CRP (mg/L)a, mean ± SD 10.4 ± 21.8 9 ± 12.3 New PB while on CiaHD n (%) 3 (4) – Increase of calcium containing PB while on CiaHD, n (%) 4 (5) – Increase of other PB while on CiaHD, n (%) 3 (4) – Increase of active vitamin D while on CiaHD, n (%) 1 (1) – Dialysis centre (1 not 2), n (%) 47 (60) – Parameter Before CiaHD After 3 months of CiaHD Age (years), median (IQR) 75.1 (57.2–80.3) +0.25 Gender (males), n (%) 41 (53) – BMI (kg/m2), mean ± SD 26.2 ± 5.1 – White patients, n (%) 75 (96) – Upper arm circumference (cm), mean ± SD 28.7 ± 3.6 28.7 ± 3.5 Adapted CCI (0–21), median (IQR) 3 (1–6) 4 (2–7)b History of MI, n (%) 15 (19) 15 (19) Hypertensive heart disease, n (%) 40 (51) 41 (53) Diabetes, n (%) 26 (33) 28 (36) COPD, n (%) 7 (9) 9 (11) Cancer, n (%) 30 (39) 33 (42) CHD, n (%) 26 (33) 27 (35) PAOD, n (%) 11 (14) 13 (16) CerVD, n (%) 13 (16) 13 (16) CHF, n (%) 8 (10) 8 (10) Atrial fibrillation, n (%) 17 (22) 19 (24) GIT bleeding, n (%) 5 (6) 5 (6) Smoking (at the moment), n (%) 8 (10) 8 (10) HD not HDF, n (%) 75 (96) 75 (96) HD vintage (months), median (IQR) 36 (17–56) +0.25 Bicarbonate setting, median (IQR) 32 (30–33) 32 (30–33) Kt/Va, mean ± SD 1.2 ± 0.3 1.2 ± 0.3 Haemoglobin (g/dL), mean ± SD 11.9 ± 1.3 11.6 ± 1.2 Ionized magnesium (mmol/L)a, mean ± SD 0.5 ± 0.1 0.4 ± 0.1b Ionized calcium (mmol/L)a, mean ± SD 1.2 ± 0.2 1.1 ± 0.2b Phosphate (mmol/L), mean ± SD 1.7 ± 0.5 1.8 ± 0.5 Albumin (g/L)a, mean ± SD 38 ± 4 40 ± 5b Fetuin A (g/L), mean ± SD 0.37 ± 0.11 0.36 ± 0.08 iPTH (pg/mL)a, median (IQR) 267 (134–439) 254 (134–484) CRP (mg/L)a, mean ± SD 10.4 ± 21.8 9 ± 12.3 New PB while on CiaHD n (%) 3 (4) – Increase of calcium containing PB while on CiaHD, n (%) 4 (5) – Increase of other PB while on CiaHD, n (%) 3 (4) – Increase of active vitamin D while on CiaHD, n (%) 1 (1) – Dialysis centre (1 not 2), n (%) 47 (60) – Demographics, comorbidities and basic laboratory values from pooled analysis (both dialysis units) are reported before and after CiaHD. CerVD, cerebral vascular disease; CHD, coronary heart disease; CHF, congestive heart failure; COPD, chronic obstructive pulmonary disease; CRP, C-reactive protein; GIT, gastrointestinal; iPTH, intact parathyroid hormone; MI, myocardial infarction; PAOD, peripheral arterial occlusive disease; PB, phosphate binders. a Missing values: 1–3 missing values, due to missed quartal assessment. b Paired samples t-test and Wilcoxon signed-rank test were used for comparing before versus after citrate P < 0.05. Serum calcification propensity (T50) increases after 3 months on acetate-free/citrate-acidified A-concentrate Individual T50 times before (⁠ T50before CiaHD ⁠) and after (⁠ T50after CiaHD ⁠) 3 months of CiaHD treatment were normally distributed (Supplementary data, Figure S1). We had previously reported a longitudinal decline of T50 in stable HD patients on standard dialysis regimes [7]. In our present study, the mean T50before CiaHD was 206  ±  56 min and increased significantly to a T50after CiaHD of 242  ±  65 min after 3 months of intermittent CiaHD treatment (P <0.001; Figure 2A). Further, when testing for intra-individual variation, T50before CiaHD versus T50after CiaHD increased significantly after 3 months of CiaHD (Figure 2B; P <0.001). In accordance, T50Change [= (⁠ T50after CiaHD −  T50before CiaHD ⁠)/ T50after CiaHD] exceeded +5.1% (derived from the interassay coefficient for T50 standards at 260 min as reported previously [7, 24]) in 57 patients on CiaHD, whereas only 21 showed stable or declining T50Change (Figure 2C). FIGURE 2 View largeDownload slide T50 values and their change during the study course. (A) Increasing absolute pre-dialysis T50 values after 3 months of CiaHD in 78 patients and not significantly altered levels after an additional 3 months on an AaHD in 44 patients eligible for this analysis (Figure 1). Paired-samples t-test was used for comparison. (B) Intra-individual T50Change values for 78 patients before versus after 3 months of citrate. Blue rhombuses indicate stable or declining T50 (⁠ T50Change ≤5.1%). Green rhombuses indicate increasing (⁠ T50Change >5.1%). T50Change values were sorted in ascending order. (C) Mean values and 95% confidence intervals of T50Change observed during CiaHD versus T50Change back on AaHD thereafter. Independent samples t-test was used for comparing the latter. FIGURE 2 View largeDownload slide T50 values and their change during the study course. (A) Increasing absolute pre-dialysis T50 values after 3 months of CiaHD in 78 patients and not significantly altered levels after an additional 3 months on an AaHD in 44 patients eligible for this analysis (Figure 1). Paired-samples t-test was used for comparison. (B) Intra-individual T50Change values for 78 patients before versus after 3 months of citrate. Blue rhombuses indicate stable or declining T50 (⁠ T50Change ≤5.1%). Green rhombuses indicate increasing (⁠ T50Change >5.1%). T50Change values were sorted in ascending order. (C) Mean values and 95% confidence intervals of T50Change observed during CiaHD versus T50Change back on AaHD thereafter. Independent samples t-test was used for comparing the latter. A total of 44 patients had additionally undergone another 3 months of AaHD after 3 months on CiaHD (Figure 1). No further increase in T50 time was observed in this period. The mean T50after CiaHD for this subgroup was 252  ± 63 min and 243 ± 64 min (P = 0.29) after these additional 3 months back on AaHD, indicating that the beneficial effects observed on citrate dialysate had not persisted (Figure 2B). In accordance with this, T50Change (+28 ± 28%; n = 44) within the first 3 months on citrate was significantly different from T50Change (−1 ± 28%; n = 44) observed thereafter (Figure 2D; P < 0.001). Individual values for T50Change while on CiaHD and AaHD thereafter are depicted in Supplementary data, Figure S2. Albumin partially mediates increasing T50 during acetate-free, citrate-acidified bicarbonate HD Aside from changing T50 after 3 months of CiaHD, we observed significantly decreasing levels for ionized serum magnesium (0.5 ± 0.1 to 0.4 ± 0.1 mmol/L; P <0.001) and calcium levels (1.2 ± 0.2 to 1.1 ± 0.2 mmol/L; P = 0.01), whereas serum albumin levels had significantly increased (38 ± 4 to 40 ± 5 g/L, P = 0.001). Serum phosphate levels did not significantly change (1.7 ± 0.4  to 1.8 ± 0.4 mmol/L; P = 0.47) nor did fetuin A levels (0.37 ± 0.11 to 0.36 ± 0.08 g/L). For these and other parameters, including routine venous pre-dialysis BGA, see Supplementary data, Figure S3. Similar results were obtained when testing the influence of CiaHD (1 = yes) on these parameters in linear mixed models, including patient IDs as random effects and these variables as dependents (Supplementary data, Table S3). It should be noted that the mentioned changes were probably not directly conveyed due to persistence of citrate in the circulation until the next dialysis session. In line with its reported half-life [28, 29], we did not find significant differences in serum citrate or acetate levels in pre-dialysis samples from 15 patients on CiaHD versus AaHD treatment (Supplementary data, Figure S4). Since both phosphate (rbefore CiaHD = −0.44, P < 0.001; rafter CiaHD = −0.50, P < 0.001) and fetuin A (rbefore CiaHD = −0.33, P = 0.003; rafter CiaHD = 0.31, P = 0.005) levels did not significantly change on CiaHD (‘treatment’) but are known determinants of T50 and T50Change ⁠, we ran partial regression adjusted for phosphate and fetuin A or Δ phosphate and Δ fetuin A, respectively, to screen for parameters that associated with absolute T50 values and changed in parallel to T50Change but were also significantly related to CiaHD. Herein, only albumin (rbefore CiaHD = 0.38, P = 0.001; rafter CiaHD = 0.29, P < 0.01), Δ albumin (r = 0.42, P < 0.001) but not ionized magnesium, calcium nor their respective Δs were significantly related with T50Change ⁠. For more correlations see Table 2. Interestingly, in univariate analysis per tertiles of Δ phosphate, the observed association of T50Change with Δ albumin was strongest in those with increasing or stable phosphate serum levels, indicating that decreasing serum phosphate levels might have antagonized the effects of albumin in a few patients with respect to T50Change (Figure 3). Table 2 Partial correlation matrix of absolute and Δ values of determinants of T50 and T50Change Correlation matrix T50 before CiaHD T50 after CiaHD T50Change before versus after (r; P-value) (r; P-value) (r; P-value) Phosphate (before/after CiaHD/Δ)  −0.44; <0.001  -0.50; <0.001 −0.48; <0.001 Fetuin A (before/after CiaHD/Δ)  0.33; 0.003  0.48; <0.001 −0.12; 0.28 Albumin (before/after CiaHD/Δ)a  0.38; 0.001  0.29; <0.01 0.42, <0.001 Ca2+(before/after CiaHD/Δ)a  0.04; 0.76  0.12, 0.30 −0.07, 0.51 Mg2+(before/after CiaHD/Δ)a  0.03; 0.82 0.16, 0.17 −0.03, 0.83 Correlation matrix T50 before CiaHD T50 after CiaHD T50Change before versus after (r; P-value) (r; P-value) (r; P-value) Phosphate (before/after CiaHD/Δ)  −0.44; <0.001  -0.50; <0.001 −0.48; <0.001 Fetuin A (before/after CiaHD/Δ)  0.33; 0.003  0.48; <0.001 −0.12; 0.28 Albumin (before/after CiaHD/Δ)a  0.38; 0.001  0.29; <0.01 0.42, <0.001 Ca2+(before/after CiaHD/Δ)a  0.04; 0.76  0.12, 0.30 −0.07, 0.51 Mg2+(before/after CiaHD/Δ)a  0.03; 0.82 0.16, 0.17 −0.03, 0.83 Partial correlations were adjusted for phosphate and fetuin A levels or their respective Δ values. Correlations were assessed before CiaHD and after CiaHD. In addition, we assessed the correlation of post-pre CiaHD values with T50Change ⁠. a Adjusted for phosphate and fetuin A. Table 2 Partial correlation matrix of absolute and Δ values of determinants of T50 and T50Change Correlation matrix T50 before CiaHD T50 after CiaHD T50Change before versus after (r; P-value) (r; P-value) (r; P-value) Phosphate (before/after CiaHD/Δ)  −0.44; <0.001  -0.50; <0.001 −0.48; <0.001 Fetuin A (before/after CiaHD/Δ)  0.33; 0.003  0.48; <0.001 −0.12; 0.28 Albumin (before/after CiaHD/Δ)a  0.38; 0.001  0.29; <0.01 0.42, <0.001 Ca2+(before/after CiaHD/Δ)a  0.04; 0.76  0.12, 0.30 −0.07, 0.51 Mg2+(before/after CiaHD/Δ)a  0.03; 0.82 0.16, 0.17 −0.03, 0.83 Correlation matrix T50 before CiaHD T50 after CiaHD T50Change before versus after (r; P-value) (r; P-value) (r; P-value) Phosphate (before/after CiaHD/Δ)  −0.44; <0.001  -0.50; <0.001 −0.48; <0.001 Fetuin A (before/after CiaHD/Δ)  0.33; 0.003  0.48; <0.001 −0.12; 0.28 Albumin (before/after CiaHD/Δ)a  0.38; 0.001  0.29; <0.01 0.42, <0.001 Ca2+(before/after CiaHD/Δ)a  0.04; 0.76  0.12, 0.30 −0.07, 0.51 Mg2+(before/after CiaHD/Δ)a  0.03; 0.82 0.16, 0.17 −0.03, 0.83 Partial correlations were adjusted for phosphate and fetuin A levels or their respective Δ values. Correlations were assessed before CiaHD and after CiaHD. In addition, we assessed the correlation of post-pre CiaHD values with T50Change ⁠. a Adjusted for phosphate and fetuin A. FIGURE 3 View largeDownload slide Scatter plots depicting the association of T50Change and Δ albumin per tertiles of Δ-phosphate (Δ albumin = albuminafter CiaHD − albuminbefore CiaHD). Pearson correlation was performed within these subgroups, since alterations in phosphate have been shown to be strong determinants of serum calcification propensity. Results were confirmed by plotting residuals adjusted for changing serum phosphate (data not shown). FIGURE 3 View largeDownload slide Scatter plots depicting the association of T50Change and Δ albumin per tertiles of Δ-phosphate (Δ albumin = albuminafter CiaHD − albuminbefore CiaHD). Pearson correlation was performed within these subgroups, since alterations in phosphate have been shown to be strong determinants of serum calcification propensity. Results were confirmed by plotting residuals adjusted for changing serum phosphate (data not shown). In line with these results, phosphate, fetuin A levels per se, albumin levels and CiaHD (1 = yes) were significant univariate regressors of the outcome T50 in linear mixed-effects models (not shown) and remained so after adjustment for phosphate and fetuin A (Table 3) and in the full model taking other laboratory parameters into account (Table 3). The best model fit (−2log likelihood = −284.725, df = 8) was obtained when only including CiaHD; with phosphate, fetuin A and albumin as fixed effects and subject ID as a random effect (Table 3). Table 3 Linear mixed models: effect quantification and mediation analysis Parameter SD Dependent = z(T50) adjusted for phosphate and fetuin A Dependent = z(T50) full model Dependent = z(T50) best fit model → estimate (95% CI) P-value → estimate (95% CI) P-value → estimate (95% CI) P-value CiaHD (=1) – 0.55 (0.34–0.77) <0.001 0.50 (0.27–0.74) <0.001 0.43 (0.23–0.63) <0.001 Phosphate (mmol/L) 0.5 −0.43 (−0.57 to −0.30) <0.001 −0.53 (−0.66 to −0.39) <0.001 −0.52 (−0.65 to −0.39) <0.001 Albumin (g/L) 4.0 0.29 (0.15–0.43) <0.001 0.27 (−0.13–0.41) <0.001 0.29 (0.15–0.43) <0.001 Fetuin A (g/L) 0.1 0.37 (0.23–0.40) <0.001 0.22 (0.10–0.34) 0.001 0.21 (0.09–0.33) 0.001 Mg2+ (mmol/L) 0.1 −0.07 (−0.07–0.21) 0.31 0.07 (−0.10–0.23) 0.41 – – Ca2+ (mmol/L) 0.2 −0.08 (−0.06–0.22) 0.27 0.02 (−0.15–0.19) 0.81 - – iPTH (pg/mL) 344 0.05 (−0.10–0.19) 0.52 0.02 (−0.13–0.16) 0.84 – – Hb (g/dL) 1.2 −0.01 (−0.11–0.13) 0.92 −0.02 (−0.13–0.10) 0.77 - – Dialysate HCO3−(mmol/L) 2 0.06 (−0.03–0.14) 0.18 0.06 (−0.03–0.14) 0.21 – – Parameter SD Dependent = z(T50) adjusted for phosphate and fetuin A Dependent = z(T50) full model Dependent = z(T50) best fit model → estimate (95% CI) P-value → estimate (95% CI) P-value → estimate (95% CI) P-value CiaHD (=1) – 0.55 (0.34–0.77) <0.001 0.50 (0.27–0.74) <0.001 0.43 (0.23–0.63) <0.001 Phosphate (mmol/L) 0.5 −0.43 (−0.57 to −0.30) <0.001 −0.53 (−0.66 to −0.39) <0.001 −0.52 (−0.65 to −0.39) <0.001 Albumin (g/L) 4.0 0.29 (0.15–0.43) <0.001 0.27 (−0.13–0.41) <0.001 0.29 (0.15–0.43) <0.001 Fetuin A (g/L) 0.1 0.37 (0.23–0.40) <0.001 0.22 (0.10–0.34) 0.001 0.21 (0.09–0.33) 0.001 Mg2+ (mmol/L) 0.1 −0.07 (−0.07–0.21) 0.31 0.07 (−0.10–0.23) 0.41 – – Ca2+ (mmol/L) 0.2 −0.08 (−0.06–0.22) 0.27 0.02 (−0.15–0.19) 0.81 - – iPTH (pg/mL) 344 0.05 (−0.10–0.19) 0.52 0.02 (−0.13–0.16) 0.84 – – Hb (g/dL) 1.2 −0.01 (−0.11–0.13) 0.92 −0.02 (−0.13–0.10) 0.77 - – Dialysate HCO3−(mmol/L) 2 0.06 (−0.03–0.14) 0.18 0.06 (−0.03–0.14) 0.21 – – Linear mixed model regressions for CiaHD (yes = 1) and changing covariates on T50 (outcome) were conducted for each covariate. First column reports overall SD for predictors. Second column reports single correlations after adjustment for varying serum phosphate levels. Further, the associations for these covariates were assessed in a full model including all these variables (third column) and after elimination of non-significant predictors (fourth column) for mediation analysis. All metric parameters were entered as centred variables. Bold font indicates estimates used for calculation of the indirect mediation effect: CiaHD →albumin→T50. Inclusion of pre-dialysis pH data from routine midterm BGA did not significantly affect the results depicted above (data not shown). Table 3 Linear mixed models: effect quantification and mediation analysis Parameter SD Dependent = z(T50) adjusted for phosphate and fetuin A Dependent = z(T50) full model Dependent = z(T50) best fit model → estimate (95% CI) P-value → estimate (95% CI) P-value → estimate (95% CI) P-value CiaHD (=1) – 0.55 (0.34–0.77) <0.001 0.50 (0.27–0.74) <0.001 0.43 (0.23–0.63) <0.001 Phosphate (mmol/L) 0.5 −0.43 (−0.57 to −0.30) <0.001 −0.53 (−0.66 to −0.39) <0.001 −0.52 (−0.65 to −0.39) <0.001 Albumin (g/L) 4.0 0.29 (0.15–0.43) <0.001 0.27 (−0.13–0.41) <0.001 0.29 (0.15–0.43) <0.001 Fetuin A (g/L) 0.1 0.37 (0.23–0.40) <0.001 0.22 (0.10–0.34) 0.001 0.21 (0.09–0.33) 0.001 Mg2+ (mmol/L) 0.1 −0.07 (−0.07–0.21) 0.31 0.07 (−0.10–0.23) 0.41 – – Ca2+ (mmol/L) 0.2 −0.08 (−0.06–0.22) 0.27 0.02 (−0.15–0.19) 0.81 - – iPTH (pg/mL) 344 0.05 (−0.10–0.19) 0.52 0.02 (−0.13–0.16) 0.84 – – Hb (g/dL) 1.2 −0.01 (−0.11–0.13) 0.92 −0.02 (−0.13–0.10) 0.77 - – Dialysate HCO3−(mmol/L) 2 0.06 (−0.03–0.14) 0.18 0.06 (−0.03–0.14) 0.21 – – Parameter SD Dependent = z(T50) adjusted for phosphate and fetuin A Dependent = z(T50) full model Dependent = z(T50) best fit model → estimate (95% CI) P-value → estimate (95% CI) P-value → estimate (95% CI) P-value CiaHD (=1) – 0.55 (0.34–0.77) <0.001 0.50 (0.27–0.74) <0.001 0.43 (0.23–0.63) <0.001 Phosphate (mmol/L) 0.5 −0.43 (−0.57 to −0.30) <0.001 −0.53 (−0.66 to −0.39) <0.001 −0.52 (−0.65 to −0.39) <0.001 Albumin (g/L) 4.0 0.29 (0.15–0.43) <0.001 0.27 (−0.13–0.41) <0.001 0.29 (0.15–0.43) <0.001 Fetuin A (g/L) 0.1 0.37 (0.23–0.40) <0.001 0.22 (0.10–0.34) 0.001 0.21 (0.09–0.33) 0.001 Mg2+ (mmol/L) 0.1 −0.07 (−0.07–0.21) 0.31 0.07 (−0.10–0.23) 0.41 – – Ca2+ (mmol/L) 0.2 −0.08 (−0.06–0.22) 0.27 0.02 (−0.15–0.19) 0.81 - – iPTH (pg/mL) 344 0.05 (−0.10–0.19) 0.52 0.02 (−0.13–0.16) 0.84 – – Hb (g/dL) 1.2 −0.01 (−0.11–0.13) 0.92 −0.02 (−0.13–0.10) 0.77 - – Dialysate HCO3−(mmol/L) 2 0.06 (−0.03–0.14) 0.18 0.06 (−0.03–0.14) 0.21 – – Linear mixed model regressions for CiaHD (yes = 1) and changing covariates on T50 (outcome) were conducted for each covariate. First column reports overall SD for predictors. Second column reports single correlations after adjustment for varying serum phosphate levels. Further, the associations for these covariates were assessed in a full model including all these variables (third column) and after elimination of non-significant predictors (fourth column) for mediation analysis. All metric parameters were entered as centred variables. Bold font indicates estimates used for calculation of the indirect mediation effect: CiaHD →albumin→T50. Inclusion of pre-dialysis pH data from routine midterm BGA did not significantly affect the results depicted above (data not shown). Since only albumin changed significantly during 3 months of CiaHD and remained independently associated with the outcome T50, but the effect was reduced by addition of CiaHD (‘treatment’) to the model, some of the effect observed during CiaHD might have been mediated by increasing albumin levels. The indirect effect estimate (Judd and Kenny coefficient) reflecting ‘CiaHD (1 = after CiaHD treatment) → albumin → T50’ was 0.55 − 0.43 = 0.12 (Table 3), meaning that ∼22% (0.12/0.55) of the beneficial T50-modifying effect observed during 3 months of CiaHD can be mathematically attributed to increasing serum albumin after adjusting for variations in serum phosphate and fetuin A levels (both unrelated to CiaHD treatment). DISCUSSION In this pre–post-quasi-interventional study assessing the influence of an acetate-free, citrate-based (1 mmol/L) A-concentrate, we found a significant improvement of serum calcification propensity T50 after switching to CiaHD for 3 months. In line with this, no further increase of T50 was observed in 44 patients after switching back to AaHD. The calculated T50Change values were significantly different for these treatment phases. As a natural liver substrate, citrate has a reported half-life of 5–60 min [28, 29]. We did find comparable levels of citrate and acetate in pre-dialysis sera in a sample of patients after switching treatment modalities. However, our data indicate that CiaHD solutions might still sustainably alter the calcification resilience in HD patients upon long-term use, although citrate per se has already been metabolized by the next session. Although intra-individual phosphate and fetuin A levels were associated with changing T50 in accordance with previous reports [7, 30], overall phosphate and fetuin A levels remained stable for 3 months, indicating that the beneficial effect observed during CiaHD were not mediated by alteration in these parameters. Likewise, although we detected decreased ionized serum magnesium and calcium levels after 3 months of CiaHD, these factors were not independently associated with changing T50. Instead, we observed increasing albumin levels after the treatment phase, which has also been observed in other studies assessing the impact of midterm (3 months) acetate-free HD or HDF treatment [18, 19]. Further, long-term acetate-free dialysis was shown to reduce post-dialysis malaise [17] and improve the creatinine index, a nutritional marker associated with survival in HD patients [19, 31]. In contrast, no variation of serum albumin was seen in Grundstrom et al.’s short-term (6 weeks) randomized cross-over trial, demonstrating safety and biocompatibility of CiaHD [9]. Nevertheless, in our study population, albumin was the only identified factor that was significantly associated with both the treatment modality and the outcome (changing T50). In this context, albumin seems to mediate a minor part of the beneficial treatment effect observed during CiaHD, which could route back to improved well-being and better nutritional status in these patients. However, no changes in BMI and upper arm circumference were observed during treatment (not shown), maybe related to the short observation times. An alternative hypothesis would be that citrate could contribute to the dissolution of primary CPP during dialysis. This effect could also explain the increased availability of albumin and the improvement of the T50 test. Yet, although we replicated the known association of fetuin A with T50, we did not find increasing serum levels of this anti-calcific protein. Unfortunately, we did not have data on pre-dialysis serum bicarbonate at the time of T50 assessment since improved acid-balance on CiaHD might as well have mediated another part of the observed effect. Yet, published data remain conflicting on the effect of citrate-acidified A-concentrates on pre-dialysis acidosis [11, 12]. In this context it should be noted that pre-dialysis pH, base excess and bicarbonate from routine venous BGA at the middle of the treatment phases remained unaffected by treatment modality (CiaHD versus AaHD) in a subgroup of our patients with available data. Further, as the solutions for performing the T50 assay are strongly buffered with 50 mmol HEPES, an impact of the actual serum pH on T50 assay results is unlikely [4]. Thus, although we are unable to rule out improved pre-dialysis blood pH as a mediator of the observed variation of T50 with absolute certainty, we consider a major modulation of T50 by pre-dialysis pH unlikely in our study. It is worth mentioning that citrate as a chelator of bivalent ions has similar affinity for ionized magnesium as for calcium [32] and might thereby not only negatively affect calcium [11, 21], but also magnesium mass balance. Whereas the former effect is certainly desirable, magnesium probably has a more anti-calcifying effect and can modulate serum calcification propensity (increasing T50) in vitro [5, 33, 34]. In our study, the effects of lower ionized calcium and magnesium levels might have eliminated each other with respect to T50. Since we observed similar pre-dialytic citrate levels on both CiaHD and AaHD, the concept of direct complexing of bivalent cations by citrate does not help us in explaining our observations. Instead, we rather consider the changes in ionized calcium and magnesium a phenomenon of complexing to rather long-lasting molecules or higher regeneration of anti-calcific buffer proteins (e.g. albumin). This hypothesis also explains why the modulation of T50 lasts until the next dialysis session although citrate has already been metabolized. Nevertheless, we are unable to provide definite evidence for these ideas. Despite the decent number of patients within a longitudinal design, this post hoc study has several limitations. Patients had not been blinded to the treatment. The pre–post (non-randomized controlled trial) design did not include a placebo-controlled arm, so this effect cannot be excluded with certainty. Further, the relatively short study period and the overall number of patients did not permit inferences regarding mortality. Moreover, we only assessed basic routine laboratory parameters. Differential elimination of uraemic toxins or an alteration of the anion gap’s composition on CiaHD versus AaHD might as well have influenced T50 in addition to albumin. The study protocol allowed regular alterations in medications during periods including phosphate binders, which per se remained unaltered during treatment but might have masked correlations of naturally weaker determinants of T50 and outcome during CiaHD. Although no further increase of T50 was observed in 44 patients and T50Change differed significantly between periods, it cannot be ruled out with absolute certainty that time per se, instead of CiaHD treatment, could have been the relevant determinant. After all, this scenario appears unlikely, as T50 has been shown to time-dependently decrease in HD patients on standard bicarbonate AaHD regimes [7]. Lastly, the exact mechanisms underlying increasing albumin levels and T50 after 3 months of CiaHD dialysis remain elusive and the post hoc character of this study only allows the generation of hypotheses that need elaboration and proof from further studies including clinical endpoints. Intriguingly, hints at improved survival due to acetate-free dialysis regimes have been reported from registries [35]. We recently reported an association of worsening calcification propensity with all-cause and cardiovascular mortality in a subgroup of patients from the observational ISAR cohort [7]. Following up on these findings, the present study indicates that aside from strict phosphate control, the putative risk factor T50 can be therapeutically modified by acetate-free dialysis regimes. ACKNOWLEDGEMENTS Special thanks are to be given to R.B. and staff for making this study possible. Isabelle Gsponer and Sandra Haderer must be thanked for technical support. FUNDING G.L. and C.S. received investigator-initiated research grant support from Baxter for the mother study. G.L. was supported by a research scholarship from the Else Kröner-Fresenius-Stiftung for physician scientists. U.H. reports grants from Baxter during the conduct of the study. This work was supported by the German Research Foundation (DFG) and the Technical University of Munich within the funding programme Open Access Publishing. This work was supported by the President’s International Fellowship Initiative of The Chinese Academy of Science (2015VBB045 to T.M.), the National Natural Science Foundation of China (31450110423 to T.M.), the Austrian Science Fund (FWF: P28854 to T.M.), the Austrian Research Promotion Agency (FFG: 864690 to T.M.), the Integrative Metabolism Research Center Graz, the Austrian infrastructure program 2016/2017, BioTechMed/Graz and the OMICS Center Graz. AUTHORS’ CONTRIBUTIONS All listed authors have contributed sufficiently to the project. G.L. and C.S. were responsible for the study design. G.L., C.M., B.H. and S.W. were responsible for statistical analysis. R.B. recruited all patients and provided access to patients’ data and secured protocol adherence. J.C-L., A.S., C.S., V.B., S.W., J.A., S.K. and M.C.B. contributed to patients’ data acquisition and entry. C.C.M. assessed ionized calcium and magnesium values. G.L. drafted the manuscript, which was reviewed and approved by all authors. S.A., U.H. and C.S. were responsible for the database and ethics approval. A.P. evaluated T50 in a blinded fashion and contributed to the manuscript. T.M. and S.S. performed NMR-based metabolomics technically and were responsible for the quantification of metabolites during the revision process. CONFLICT OF INTEREST STATEMENT A.P. is an employee and stockholder of Calciscon. 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 – 1305 Google Scholar Crossref Search ADS PubMed 2 Foley RN , Parfrey PS , Sarnak MJ. Epidemiology of cardiovascular disease in chronic renal disease . J Am Soc Nephrol 1998 ; 9 : S16 – S23 Google Scholar PubMed 3 Stokes JB. Consequences of frequent hemodialysis: comparison to conventional hemodialysis and transplantation . Trans Am Clin Climatol Assoc 2011 ; 122 : 124 – 136 Google Scholar PubMed 4 Pasch A. Novel assessments of systemic calcification propensity . Curr Opin Nephrol Hypertens 2016 ; 25 : 278 – 284 Google Scholar Crossref Search ADS PubMed 5 Pasch A , Farese S , Graber S et al. Nanoparticle-based test measures overall propensity for calcification in serum . J Am Soc Nephrol 2012 ; 23 : 1744 – 1752 Google Scholar Crossref Search ADS PubMed 6 Smith ER , Ford ML , Tomlinson LA et al. Serum calcification propensity predicts all-cause mortality in predialysis CKD . J Am Soc Nephrol 2014 ; 25 : 339 – 348 Google Scholar Crossref Search ADS PubMed 7 Lorenz G , Steubl D , Kemmner S et al. Worsening calcification propensity precedes all-cause and cardiovascular mortality in haemodialyzed patients . Sci Rep 2017 ; 7 : 13368 Google Scholar Crossref Search ADS PubMed 8 Pasch A , Covic A , Marzell B et al. Serum calcification propensity of HD patients is therapeutically improved by a calcium acetate/magnesium carbonate containing phosphate binder [abstract]. In: 51st Congress of the European-Renal-Association(ERA)/European-Dialysis-and-Transplant-Association (EDTA), Amsterdam, Netherlands, 31 May–03 Jun 2014. Oxford: Oxford University Press. 9 Grundstrom G , Christensson A , Alquist M et al. Replacement of acetate with citrate in dialysis fluid: a randomized clinical trial of short term safety and fluid biocompatibility . BMC Nephrol 2013 ; 14 : 216 Google Scholar Crossref Search ADS PubMed 10 Panesar M , Shah N , Vaqar S et al. Changes in serum bicarbonate levels caused by acetate-containing bicarbonate-buffered hemodialysis solution: an observational prospective cohort study . Ther Apher Dial 2017 ; 21 : 157 – 165 Google Scholar Crossref Search ADS PubMed 11 Gabutti L , Lucchini B , Marone C et al. Citrate- vs. acetate-based dialysate in bicarbonate haemodialysis: consequences on haemodynamics, coagulation, acid-base status, and electrolytes . BMC Nephrol 2009 ; 10 : 7 – 7 Google Scholar Crossref Search ADS PubMed 12 De Sequera Ortiz P , Albalate Ramón M , Pérez-García R et al. Acute effect of citrate bath on postdialysis alkalaemia . Nefrologia 2015 ; 35 : 164 – 171 Google Scholar Crossref Search ADS PubMed 13 Wolever TM , Fernandes J , Rao AV. Serum acetate: propionate ratio is related to serum cholesterol in men but not women . J Nutr 1996 ; 126 : 2790 – 2797 Google Scholar PubMed 14 Fournier G , Potier J , Thebaud HE et al. Substitution of acetic acid for hydrochloric acid in the bicarbonate buffered dialysate . Artif Organs 1998 ; 22 : 608 – 613 Google Scholar Crossref Search ADS PubMed 15 Tessitore N , Santoro A , Panzetta GO et al. Acetate-free biofiltration reduces intradialytic hypotension: a European multicenter randomized controlled trial . Blood Purif 2012 ; 34 : 354 – 363 Google Scholar Crossref Search ADS PubMed 16 Verzetti G , Navino C , Bolzani R et al. Acetate-free biofiltration versus bicarbonate haemodialysis in the treatment of patients with diabetic nephropathy: a cross-over multicentric study . Nephrol Dial Transplant 1998 ; 13 : 955 – 961 Google Scholar Crossref Search ADS PubMed 17 Daimon S , Dan K , Kawano M. Comparison of acetate-free citrate hemodialysis and bicarbonate hemodialysis regarding the effect of intra-dialysis hypotension and post-dialysis malaise . Ther Apher Dial 2011 ; 15 : 460 – 465 Google Scholar Crossref Search ADS PubMed 18 Kuragano T , Kida A , Furuta M et al. Effects of acetate-free citrate-containing dialysate on metabolic acidosis, anemia, and malnutrition in hemodialysis patients . Artif Organs 2012 ; 36 : 282 – 290 Google Scholar Crossref Search ADS PubMed 19 Matsuyama K , Tomo T , Kadota J. Acetate-free blood purification can impact improved nutritional status in hemodialysis patients . J Artif Organs 2011 ; 14 : 112 – 119 Google Scholar Crossref Search ADS PubMed 20 Richtrova P , Mares J , Kielberger L et al. Citrate-buffered dialysis solution (Citrasate) allows avoidance of anticoagulation during intermittent hemodiafiltration-at the cost of decreased performance and systemic biocompatibility . Artif Organs 2017 ; 41 : 759 – 766 Google Scholar Crossref Search ADS PubMed 21 Perez-Garcia R , Albalate M , Sequera P et al. Calcium mass balance with citrate dialysate is lower than with acetate . Nefrologia 2017 ; 37 : 109 – 110 Google Scholar Crossref Search ADS PubMed 22 Liu J , Huang Z , Gilbertson DT et al. An improved comorbidity index for outcome analyses among dialysis patients . Kidney Int 2010 ; 77 : 141 – 151 Google Scholar Crossref Search ADS PubMed 23 Jain A , Bhayana S , Vlasschaert M et al. A formula to predict corrected calcium in haemodialysis patients . Nephrol Dial Transplant 2008 ; 23 : 2884 – 2888 Google Scholar Crossref Search ADS PubMed 24 Pasch A , Block GA , Bachtler M et al. Blood calcification propensity, cardiovascular events, and survival in patients receiving hemodialysis in the EVOLVE Trial . Clin J Am Soc Nephrol 2017 ; 12 : 315 – 322 Google Scholar Crossref Search ADS PubMed 25 Ix JH , Shlipak MG , Brandenburg VM et al. Association between human fetuin-A and the metabolic syndrome: data from the Heart and Soul Study . Circulation 2006 ; 113 : 1760 – 1767 Google Scholar Crossref Search ADS PubMed 26 Prokesch A , Graef FA , Madl T et al. Liver p53 is stabilized upon starvation and required for amino acid catabolism and gluconeogenesis . FASEB J 2017 ; 31 : 732 – 742 Google Scholar Crossref Search ADS PubMed 27 Judd CM , Kenny DA. Process analysis: estimating mediation in treatment evaluations . Eval Rev 1981 ; 5 : 602 – 619 Google Scholar Crossref Search ADS 28 Kozik-Jaromin J , Nier V , Heemann U et al. Citrate pharmacokinetics and calcium levels during high-flux dialysis with regional citrate anticoagulation . Nephrol Dial Transplant 2009 ; 24 : 2244 – 2251 Google Scholar Crossref Search ADS PubMed 29 Davenport A , Tolwani A. Citrate anticoagulation for continuous renal replacement therapy (CRRT) in patients with acute kidney injury admitted to the intensive care unit . NDT Plus 2009 ; 2 : 439 – 447 Google Scholar PubMed 30 Dekker M , Pasch A , van der Sande F et al. High-flux hemodialysis and high-volume hemodiafiltration improve serum calcification propensity . PLoS One 2016 ; 11 : e0151508 Google Scholar Crossref Search ADS PubMed 31 Desmeules S , Levesque R , Jaussent I et al. Creatinine index and lean body mass are excellent predictors of long-term survival in haemodiafiltration patients . Nephrol Dial Transplant 2004 ; 19 : 1182 – 1189 Google Scholar Crossref Search ADS PubMed 32 Lee G , Arepally GM. Anticoagulation techniques in apheresis: from heparin to citrate and beyond . J Clin Apheresis 2012 ; 27 : 117 – 125 Google Scholar Crossref Search ADS PubMed 33 Sakaguchi Y , Hamano T , Isaka Y. Effects of magnesium on the phosphate toxicity in chronic kidney disease: time for intervention studies . Nutrients 2017 ; 9 : 112 Google Scholar Crossref Search ADS 34 Sakaguchi Y , Hamano T , Nakano C et al. Association between density of coronary artery calcification and serum magnesium levels among patients with chronic kidney disease . PLoS One 2016 ; 11 : e0163673 Google Scholar Crossref Search ADS PubMed 35 Mercadal L , Franck JE , Metzger M et al. Improved survival associated with acetate-free haemodialysis in elderly: a registry-based study . Nephrol Dial Transplant 2015 ; 30 : 1560 – 1568 Google Scholar Crossref Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Journal

Nephrology Dialysis TransplantationOxford University Press

Published: Nov 1, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off