Safety and Feasibility of Regional Citrate Anticoagulation for Continuous Renal Replacement Therapy With Calcium‐Containing Solutions: A Randomized Controlled Trial

Calcium‐free (Ca‐free) solutions are theoretically the most ideal for regional citrate anticoagulation (RCA) in continuous renal replacement therapy (CRRT). However, the majority of medical centers in China had to make a compromise of using commercially available calcium‐containing (Ca‐containing) solutions instead of Ca‐free ones due to their scarcity. This study was designed to probe into the potential of Ca‐containing solution as a secure and efficient substitution for Ca‐free solutions.


| Introduction
Continuous renal replacement therapy (CRRT) is a blood purification technology that can continuously and slowly remove water and solutes; it has become an important treatment for critical illnesses, due to its ability to provide accurate volume control, gentle metabolic correction, and hemodynamic stability [1][2][3].Adequate anticoagulation plays a vital role in the effective implementation of CRRT.In the current era of CRRT, regional citrate anticoagulation (RCA) has become the predominant anticoagulation method recommended by the Kidney Disease Improving Global Outcomes (KDIGO) Clinical Practice Guidelines [4], which prolongs filter life and reduces hemorrhagic complications in critically ill patients [5][6][7].
The anticoagulant effect of citrate is based on binding and chelating free ionized calcium (iCa) in the extracorporeal circuit to reduce the level of iCa, where the intrinsic and extrinsic coagulation cascades are suppressed [8].Approximately 60% of the citrate-calcium complex is removed to effluent via the hemofilter because of its low molecular weight and high water solubility.At the same time, unchelated calcium also partially enters the effluent.In this situation, adequate calcium supplementation at the venous end of the circuit is necessary to maintain the physiologic iCa level in the systemic circulation [9].Calcium-free (Ca-free) solutions are ideal for RCA to maintain anticoagulant effect and reduce metabolic complications.However, commercial Ca-free solutions are unavailable in China.Artificial preparation of Ca-free dialysate requires more labor, takes longer time, and raises the risk of infection.Therefore, commercial calcium-containing (Ca-containing) solutions are often used in clinical practice to simplify the RCA process.To our knowledge, there are little published data on the application of Cafree and Ca-containing solutions in CRRT with RCA.It is still unknown whether Ca-containing solutions are safe and feasible for cardiac intensive care unit (CICU) populations.This prospective study was designed to compare the application of RCA for CRRT with the use of Ca-free and Ca-containing dialysate in CICU.

| Study Design and Population
This study was a prospective, single-center, randomized controlled trial, which has been approved by the Ethics Committee of Xiamen Cardiovascular Hospital of Xiamen University (2022-XMX-16) and registered in the Chinese Clinical Trial Registry (ChiCTR2100048238).The study cohort included 99 patients who were admitted to the CICU of Xiamen Cardiovascular Hospital of Xiamen University (Xiamen, China) and received CRRT from April 16, 2021, to April 16, 2022.The indications for CRRT include severe congestive heart failure, severe acid-base electrolyte imbalance, acute kidney injury (AKI), contrast use with volume overload, and end-stage renal disease intolerant to intermittent hemodialysis.Patients with Severe liver dysfunction (Child-Pugh C), refractory hypoxemia (PO2 < 60 mmHg) or hypercalcemia (serum calcium > 2.8 mmol/L) were excluded from the study.
All the patients who were eligible according to the inclusion and exclusion criteria agreed to attend this study and signed the informed consent form (ICF) indicating that they fully understand the study and their rights of confidentiality and withdrawal from the study without providing a reason.
The included participants were randomly assigned to the CVVHD Ca-free group, CVVHDF Ca-free group, or CVVHDF Ca-containing group using a computer-based random number approach (Figure 1).The randomization process was conducted by nurses not involved in the treatment of the patients.Patients with multiple hospital admissions or undergoing repeated CRRT were only analyzed for their first enrolled procedure.Basic clinical information including age, gender, past medical history, and echocardiography parameters was extracted from the medical records.Current disease status information, including acute coronary syndrome, shock, cardiac arrest, acute infection, sepsis, recent use of contrast agents, usage of vasoactive medications, mechanical ventilation, and mechanical circulatory support was obtained by direct clinical contact and relevant medical records.The serum level of urea nitrogen, creatinine, sodium, potassium, total calcium, ionized calcium, pH value, actual bicarbonate, lactate, activated partial thromboplastin time (APTT), prothrombin time (PT), thromboplastin time (TT), Ddimer, fibrinogen degradation products (FDP), fibrinogen (Fib), hematocrit (HCT), and platelet count before CRRT procedure in patients of each group were recorded and compared.The clinical characteristics and CRRT operating parameters were compared between three groups.

| CRRT Equipment and Solutions
Vascular access was established by insertion of a 12-Fr duallumen catheter (YIXINDA SCW, 12F, 16/20 cm, China) in the internal jugular vein or femoral vein.Some patients with long-term regular dialysis used semipermanent hemodialysis catheters or dialysis fistulas.We performed CVVHD and post-dilution CVVHDF using FRESENIUS acute dialysis and extracorporeal blood therapy machine (model: multiFiltrate with integrated Ci-Ca module, Germany), with FRESENIUS continuous veno-venous hemodialysis local anticoagulation tubing system (model: MultiFiltrate Kit Ci-Ca CVVHD 1000, Germany), which include extracorporeal circulation tube (mul-tiFiltrate Cassette Ci-Ca, Germany), dialysis tube (dialysate system multiFiltrate, Germany), and hollow fiber hemodialysis filter (Ultraflux AV1000S, Germany).When using CVVHDF mode, an additional set of hemofiltration tubes (Substitute system multiFiltrate, Germany) was required.
The 4% sodium citrate (200 mL: 8.0 g/bag, Sichuan NANGEER Biotechnology Co., Ltd., China) was infused at the arterial site of the circuit to maintain the effect of anticoagulation.Calcium was replaced via the venous end using a solution of 5% calcium gluconate, which was comprised of 250 mL of 10% calcium chloride added to 250 mL of 0.9% saline.The concentration was recommended by the instructions of the renal replacement machine.We used commercially available replacement solution (4000 mL/bag, Qingshan Likang, Pharmaceutical Co., Ltd., Chengdu, China) as A solution and 5% sodium bicarbonate solution as B solution, which composed Ca-containing solutions.When using 4000 mL A solution combined with 250 mL of B solution, the concentrations of each component were as follows: Na + 141.00 mmol/L, Cl − 110.00 mmol/L, Mg 2+ 0.75 mmol/L, Ca 2+ 1.50 mmol/L, HCO 3 − 35.00 mmol/L, glucose 10.00 mmol/L.The Ca-free dialysate was prepared by nurses with 2500 mL of 0.9% sodium chloride, 1000 mL of sterile water for injection, 15 mL of 10% sodium chloride, 10 mL of 50% glucose, 3 mL of 25% magnesium sulfate, and 130 mL of 5% sodium bicarbonate by nurses, in which the concentrations of each component were as follows: Na + 133.42 mmol/L, Cl − 112.27 mmol/L, Mg 2+ 0.83 mmol/L, HCO 3 − 21.15 mmol/L, glucose 6.90 mmol/L, osmotic pressure 275 mOsm/L.An appropriate amount of 10% potassium chloride was added to the Ca-free and Ca-containing solutions according to the patient's serum potassium level to maintain the serum potassium level at 4.0-4.5 mmol/L.

| Study Endpoints
The primary endpoint was the incidence of metabolic complications including citrate accumulation and metabolic alkalosis.The secondary endpoints included premature termination of treatment, thrombus of filter, and bubble trap after the process.Citrate accumulation was defined as the worst total to ionized calcium ratio during the procedure higher than 2.5 [10].The criterion for metabolic alkalosis was the occurrence of actual bicarbonate >27 mmol/L during CRRT.The grading criteria for filter thrombus were as follows: Grade 0: no residual clot; Grade 1: the residual clots occupy about 1/3 of the filter area; Grade 2: the residual clots occupy more than 1/2 of the filter area; Grade 3: all filter was clogged completely with residual clots.The thrombus level of the bubble trap was defined as four grades: Grade 0: no residual clot; Grade 1: blood clot in the arterial or venous bubble trap; Grade 2: blood clot in both arterial and venous bubble trap; Grade 3: blood clot completely blocked the arterial or venous bubble trap.

| Initial Parameters and Adjustment
The treatment modes, solution types, initial parameters, and adjustment strategies used in each group were shown in Table 1.The filtration fraction (FF) was calculated when setting parameters for patients in the CVVHDF Ca-free group and the CVVHDF Ca-containing group.Filtration fraction is the ratio of ultrafiltration rate to plasma flow, which should be maintained less than 25% according to the recommendation of the Acute Disease Quality Initiative (ADQI) [11].In order to keep the Ca-containing group running as smoothly as possible, we reduced the dialysate rate to lower the calcium exchange between the blood flow and dialysate on the one hand and increased the blood flow rate to reduce the filtration fraction on the other.When the FF was too high, we firstly adjusted the substitution fluid speed, reducing it by 200 mL each time, and then reduce the ultrafiltration speed if necessary.The substitution fluid speed had a minimum setting of 600 mL.
Five minutes, 1 h, 2 h, 4 h, and 6 h after the initiation of CRRT, peripheral (pre-filter) and post-filter blood gas analysis were conducted to determine the level of pre-filter iCa and post-filter iCa concentrations.The calcium solution and citrate flow rates were adjusted to keep peripheral blood iCa of 1.0-1.2mmol/L and post-filter levels iCa values between 0.25 and 0.35 mmol/L to ensure the safety and effectiveness of RCA.The adjustment protocol of calcium solution and sodium citrate was based on the recommendations of the machine instructions.After reaching the stable concentration, blood gas analysis was evaluated every 4 h.If there were other conditions such as acid-base imbalance, or electrolyte imbalance, the frequency of blood gas testing was increased depending on the actual clinical state of the patients.The transmembrane pressure of the filter was recorded every 4 h.
When citrate accumulation occurred, the net citrate load administered to the patient had to be quickly reduced, which can be accomplished in the following three ways: (1) reducing the blood flow rate to decrease intake through blood flow-citrate coupling; (2) increasing dialysate rate to improve clearance; or (3) adjusting targeted iCa values in the circuit to 0.4-0.5 mmol/L.If the citrate accumulation was difficult to correct by aforementioned methods, an alternative anticoagulation strategy should be taken into consideration.The correction procedures of metabolic alkalosis for various groups were detailed in Table 1.Conversely, metabolic acidosis was rectified by reducing dialysate velocity or increasing blood flow (decreasing dialysate/ blood velocity ratio).

| Statistical Analysis
The present study was exploratory because a sample size large enough to provide an adequate power to conduct a formal noninferiority trial would have made the research impracticable.We decided that a descriptive comparison employing 33 patients per arm would produce clinically significant information under the assumption that there was no difference in the absolute rates of metabolic complications between the three groups.Normally distributed continuous data were expressed as mean ± standard deviation (SD), whereas nonnormally distributed continuous data were presented as medians with interquartile ranges (IQR).Categorical variables were summarized as counts and percentages.One-way ANOVA or Kruskal-Wallis test was used to compare baseline characteristics between the three groups according to the type of data.Chi-square test and Fisher's exact test were employed in the analysis of categorical variables.We considered p < 0.05 as statistically significant.Analyses were performed using SPSS (Version 27.0) for Windows (IBM SPSS Software, NY, USA).

| Results
The study cohort included 99 patients who were randomly assigned in a 1:1:1 ratio to the CVVHD Ca-free, CVVHDF Ca-free, and CVVHDF Ca-containing groups.The baseline characteristics, including demographics, original cardiovascular disease, past medical history, comorbidities, current situation, indication for CRRT, and vascular access were well balanced among the three groups (Table 2).The mean age was 66.69 ± 14.74 years.The majority of patients were male (63.6%) and had a preadmission history of coronary heart disease (70.7%) and end-stage renal disease (60.6%).There were no overall differences in the use of vasoactive drugs, mechanical ventilation, and mechanical circulatory support.The main clinical indications for CRRT were congestive heart failure (31.3%) and dialysis dependent (35.4%).The most frequent vascular access was a temporary femoral vein catheter (75.8%).
The clinical variables at the time of CRRT initiation were outlined in Table 3.The vital signs, laboratory data, and echocardiographic parameters did not differ significantly among the three groups except that the ionized calcium level of the CVVHDF Ca-containing group was lower than CVVHDF Ca-free group (1.06 ± 0.08 mmol/L vs. 1.12 ± 0.09 mmol/L, p = 0.044).The average blood urea nitrogen and creatinine at the time of CRRT initiation for the overall population were 24.06 ± 10.10 mmol/L and 540.89 ± 301.43 umol/L.The CRRT operating parameters were presented in Table 4.There were no overall differences in the machine runtime, therapy duration, and ultrafiltration volume between the three groups.Although the transmembrane pressure of the filter in 4 h was comparable between the three groups, the transmembrane pressures of CVVHD Ca-free group were lower than another two groups in 8 and 12 h (20 mmHg vs. 30 mmHg vs. 40 mmHg and 30 mmHg vs. 40 mmHg vs. 40 mmHg, p < 0.001 for both).
The incidence of premature termination did not significantly differ among the three groups no matter due to patient factors or circuit clotting (18.2% vs. 9.1% vs. 9.1%).The thrombus level of the filter and bubble trap was similar in the three groups (p > 0.05 for all; Table 4).

| Discussion
RCA was the recommended anticoagulation modality for CRRT, which lead to a decrease in hemorrhagic complications and filter thromboses [12].The traditional protocols of RCA required the use of Ca-free solutions to optimize the anticoagulation effect within the circuit [13].The absence of commercial Ca-free solutions increased the labor intensive and the risk of iatrogenic errors and thus the process of traditional RCA-CRRT.
To solve this dilemma, Zhang et al. developed a simplified RCAbased continuous veno-venous hemofiltration (CVVH) protocol that utilizes a commercial calcium-containing replacement solution, which has been proven effective and safe and obviates the need for continuous intravenous calcium infusion [14].It has been also confirmed that CVVH using a Ca-containing replacement solution was a simplified and effective implementation for CRRT [15,16].With the deepening understanding of the importance of FF, it was not difficult to find that the FF of CVVH was significantly higher than that of CVVHD and CVVHDF under the premise of the same dose, which may lead to the shortening of filter life [17].In the era of RCA, the anticoagulation effect of citrate and the low filtration fraction cannot be guaranteed at the same time in CVVH mode.Under comprehensive evaluation of the FF and the solute removal ability of CRRT, CVVHDF and citrate anticoagulation may be the best partners [18,19].Previous studies have shown that using commercial Ca-containing solution was a simplified and efficient method of RCA for CVVHDF [20,21].Rhee et al. demonstrated that RCA-CRRT with Ca-containing solutions was feasible and safe in critically ill patients including those with sepsis and liver disease [22].Furthermore, a recent randomized controlled trial showed that RCA-based CVVHDF with Ca-containing solution had a similar circuit lifespan, hospital mortality and kidney outcome with Ca-free solution [23].This randomized study indicated that RCA-CRRT using Ca-containing solutions had a comparable safety and feasibility to Ca-free solutions in CICU populations.
The majority of patients included in this study had normal consciousness and end-stage renal disease history (CKD stage 5), in which the prolonged period of renal replacement therapy was unnecessary and intolerant.Among the indications for CRRT, correction of congestive heart failure and dialysis dependence accounted for the majority, where the target CRRT time was brief.Due to the short duration of therapies, we did not routinely measure the filter efficacy.Alternatively, we assessed the clearance of small molecules in vivo by examining creatinine and blood urea nitrogen levels before and after treatment, and we found that the clearance of creatinine and blood urea nitrogen did not differ significantly among three groups, which indicated that the application of Ca-containing solutions did not reduce the solute removal efficiency.
We found that the consumption of total solution was highest in the CVVHDF Ca-free group, which caused by the setting of a higher therapeutic dose.However, the efficiency of solute clearance and occurrence of metabolic complications under such settings were no better than another two groups, which suggested that the protocol of Ca-free CVVHDF mode in our study can be further optimized in terms of economic benefits.Moreover, the consumption of sodium citrate in the Ca-containing CVVHDF mode was similar to that in the Ca-free CVVHDF mode and higher than that in the Ca-free CVVHD, where the requirement of calcium gluconate in the Ca-containing CVVHDF group was least between three groups.As a consequence, even if the use of calcium-containing dialysate increased the consumption of sodium citrate, it did not increase the total economic burden due to decreased consumption of calcium gluconate.
The basic principle of RCA is to reduce the level of ionized calcium in the extracorporeal circuit via infusion of citrate, where ionized calcium and citrate form citrate-calcium complexes (CCC) [24].A large portion (up to 60%) of CCC was removed through a filter with the remaining amount entering the patient blood compartment for metabolism [10].Metabolism of citrate mostly took place in the liver, kidney, and muscles fitting into the Krebs (citric acid) cycle [25].Metabolism of 1 mol of citrate in the Krebs cycle will generate 3 mol of bicarbonate.An elevated citrate load might result in metabolic alkalosis in which the capacity to metabolize citrate was normal [26].If citrate administration exceeded the body's capacity, CCC tended to accumulate, generating a mild acidosis.The recurrence of high anion gap metabolic acidosis increased the need for calcium substitution, and the trend for a decreased ionized calcium level should be considered the warning sign of citrate accumulation [10,12].In severe citrate accumulation, hypocalcemia was usually observed, potentially leading to decreased myocardial contractility, vasoplegia, and poor prognosis.Therefore, citrate accumulation was our primary concern when running RCA.It was reported that 13.3% citrate accumulation in RCA-CVVHDF with calcium-containing dialysate, and 0%-12% in RCA-CRRT with calcium-free dialysate [22,27].In a study by Zhang et al., the incidence of citrate accumulation of was higher in the patients with liver disease, ranging from 3% to 22% [28].The incidence of citrate accumulation was not rare in our experience which was detected in 17 (17.2%)patients.We were able to control citrate accumulation by the methods mentioned above so that none of the patients required interruption of the treatment and alternative anticoagulation strategy.Hence, constant monitoring must be exerted to guarantee the safe progress of RCA.
In our study, no differences were found in the incidence of metabolic complications including citrate accumulation and metabolic alkalosis among the three groups, which indicated that the usage of Ca-containing solutions did not increase the risk of metabolic complications.
It was difficult to evaluate the circuit lifespan because most participants reached the treatment goal before circuit failure.Consequently, the anticoagulant effect was evaluated through transmembrane pressure and the thrombus state of filter and bubble trap after treatment.In the current study, we found that the transmembrane pressure of CVVHDF Ca-free and CVVHDF Ca-containing groups in 8 and 12 h were comparable and higher than that of the CVVHD Ca-free group.CVVHD has a lower filtration fraction than CVVHDF at the same prescribed dose, which may lead to longer circuit longevity [29].However, the infeasibility of using Ca-containing dialysate in CVVHD with RCA in our preliminary was in line with previous studies on patients undergoing chronic hemodialysis with citrate anticoagulation at the dialysis center, which both revealed that RCA with Ca-containing dialysate resulted in significantly worse anticoagulation of dialyzer and venous bubble trap compared with Ca-free dialysate in spite of higher citrate dose [30].
For that reason, we did not perform RCA with Ca-containing solutions for CVVHD at our institution.After the procedures, we graded circuit thrombosis and discovered that there was no statistical difference in the level of thrombus in the filter and bubble trap between the three groups; meanwhile, the incidence of premature termination was similar in each group regardless of whether it was caused by patient factors or circuit clotting.There were three (one in each group) patients who was terminated prematurely due to clotting, and their filter lifespan was 4, 9, and 11 h, respectively.This demonstrated that the application of Ca-solutions did not facilitate the circuit clotting.
There were several limitations to the present study.Firstly, this was a single-center study with a small sample size, which may cause selection bias and limits the strength of our conclusions.The present exploratory sample size demonstrated the infeasibility of conducting a fully powered study according to standard calculations.Secondly, our research did not include the patients with severe liver dysfunction, refractory hypoxemia, or hypercalcemia, and the safety of calcium-containing dialysate in patients with heterogeneous conditions needs to be further evaluated.Besides, many results cannot be compared to those from previous studies due to the brief duration of target treatment and differences in assessment methods.
In conclusion, we provided an appealing alternative for RCA-CVVHDF with Ca-containing dialysate in CICU populations.
In comparison with traditional RCA with Ca-free dialysate, the utilization of Ca-containing dialysate in CVVHDF displayed a similar complication rate and anticoagulant effect.Our findings required to be confirmed in more controlled trials in heterogeneous populations for wider application.

TABLE 1 |
Initial setting and adjustment strategy of three groups.
The consumption of total solution was highest in the CVVHDF Ca-free group and lowest in the CVVHDF Ca-containing group (p < 0.001).The usage of sodium citrate was lower in the CVVHD Ca-free group (p < 0.001), while the expenditure of calcium gluconate was lower in the CVVHDF Ca-containing group (p < 0.001).The clearance of BUN and creatinine were similar in each group.

TABLE 3 |
Clinical variables at the time of CRRT initiation.

TABLE 4 |
Comparison of CRRT operating parameters.
Abbreviations: BUN, blood urea nitrogen; Cr, creatinine.a Statistical difference between CVVHDF Ca-free and CVVHDF Ca-containing group.b Statistical difference between CVVHD Ca-free and CVVHDF Ca-containing group.c Statistical difference between CVVHD Ca-free and CVVHDF Ca-free group.