Analysis of the extracorporeal anticoagulation effect of modified citrate infusion during continuous haemodialysis in critically ill patients


 This study aimed to analyse the anticoagulation effect of different local infusion methods for citrate continuous haemodialysis in critically ill patients in order to identify a safe and effective citrate infusion method. Critically ill patients admitted to our hospital from April 2019 to December 2019 who underwent continuous renal replacement therapy (CRRT) using citrate for anticoagulation were divided into the conventional citrate infusion before the filter (conventional group) and citrate infusion group according to the different local citrate infusion methods (modified group). A total of 30 treatment sessions were performed for each group. the modified group patients were found to have longer mean treatment times(67.67 ± 18.69 vs 52.11 ± 24.26,p = 0.007), lower transmembrane pressures at the disconnection time from dialysis circuit(147.77 ± 66.85 vs 200.63 ± 118.66,p = 0.038), fewer citrate bag replacements(1.43 ± 0.50 vs 10.60 ± 3.19,p < 0.001), and steady ionised calcium at the venous end compared to the conventional group patients with statistically significant differences(0.35 ± 0.06 vs 0.40 ± 0.05,p = 0.006). The total calcium level was significantly higher in the conventional group patients than those of the other(2.29 ± 0.23 vs 2.19 ± 0.14,p = 0.038). Incidences of citrate accumulation and tubing coagulation were also marginally lower in the modified group. We infer that the modified local citrate infusion method can prolong the treatment time, reduce the nursing workload and the occurrence of citrate accumulation, resulting in safe and effective clinical outcome.


Introduction
Continuous renal replacement therapy (CRRT) is an essential treatment modality for patients with acute renal failure which helps remove toxins continuously from the body, effectively maintains homeostasis, and is more in line with the physiological characteristics of the body [1]. CRRT has become an important tool for the treatment of critically ill patients because of its low hemodynamic impact, safety and high e cacy [2]. Anticoagulation is an important requirement to ensure smooth delivery of CRRT. However, management of adequate anticoagulation is a challenge as it can aggravate risk of bleeding events [3][4]. Local citrate anticoagulation has several advantages like minimal impact on the patient's coagulation system, better anticoagulation effect, longer lter life, and fewer bleeding complications [5][6][7][8]. It can meet the anticoagulation needs of CRRT, and therefore has been widely used in clinical practice.
The main principle of anticoagulation by citrate is the binding to and chelating of the calcium ions, which leads to a decrease in the ionised calcium level of the blood in the extracorporeal circulation. This can prevent blood coagulation, thus prolonging the treatment time of CRRT ensuing better therapeutic outcome [4][5]. Therefore, the infusion rate of citrate is precisely controlled and the ionised calcium level in the extracorporeal circulation is monitored during treatment. However, in conventional local citrate anticoagulation for CRRT, separate infusion of citrate makes its synchronisation with CRRT poor, resulting in huge uctuations in the ionised calcium level of the extracorporeal circulation. This in turn can cause disconnection from the dialysis circuit during CRRT due to coagulation in the tubing making it di cult to achieve the therapeutic dose. To address this problem in clinical practice, we modi ed the citrate infusion method according to the therapeutic characteristics of CRRT and received satisfactory results.
We designed this study to further evaluate the effect of the improved citrate anticoagulation protocol by analysing the data related to 60 sessions of CRRT treatment in 24 critically ill patients who were admitted to the intensive care unit (ICU) and underwent CRRT with citrate anticoagulation from April 2019 to December 2019 in our hospital, so as to provide a reference for clinical treatment.

Study subjects
Critically ill patients who were admitted to the Department of Critical Care Medicine of the First A liated Hospital of Xi'an Jiaotong University from April 2019 to December 2019 for CRRT with citrate anticoagulation meeting the inclusion criteria were included in this study.
All methods were carried out in accordance with relevant guidelines and regulations.
All study participants provided informed consent, and This study was reviewed and approved by the 'The First A liated hospital of Xi'an Jiaotong University Ethics committee'.

Inclusion and exclusion criteria
The inclusion criteria were as follows: (1) critically ill patients admitted to the ICU and treated by CRRT for various reasons, (2) patients using local citrate for anticoagulation for CRRT, (3) patients having normal vascular access function, (4) patients aged 18-80 years. The exclusion criteria were as follows: (1) contraindications or relative contraindications for citrate anticoagulation, (2) disconnection from dialysis circuit due to the reason of vascular access (e.g., catheter dysfunction, etc.), (3) active discontinuation of CRRT due to human reasons, (4) use of other anticoagulants or anticoagulation modalities, and (5) unwillingness to sign the informed consent form.
CRRT protocol Vascular access and devices All patients were treated using a 12 Fr double-lumen central venous catheter (ARROW, Reading, PA, USA) to establish vascular access through catheterisation in the right internal jugular vein, right femoral vein, or left femoral vein. A PrismaFlex haemodialysis machine from a Swedish company as well as the corresponding disposable tubes and haemo lters were used(PrismaFlex M150).

Tube and lter priming and basic parameter settings
The tubes and lters were routinely primed using 2,000 ml saline with 6,250 U of heparin sodium. The primed heparin saline was drained from the tubes and lters at the time of connection to the CRRT circuit. Continuous venous-venous hemo ltration(CVVH) mode was used in both groups, with a blood ow rate of 180 ml/min and a replacement uid volume of 2,000 ml/h.

Replacement uid formulation
The formulations of the replacement uids are listed below. Fluid A: 3,000 ml of uid containing 2,250 ml of saline and 750 ml sterilised water for injection (glucosefree replacement uid), or 2,250 ml saline, 500 ml sterilised water for injection and 250 ml glucose (5%) (glucose-containing replacement uid). The choice was based on the patient's blood glucose levels.
Fluid B: 5% sodium bicarbonate was administered from the outlet end during CRRT treatment, starting at 100 ml/h and adjusted during treatment according to the acid-base levels.
Fluid C: 2.4 ml of 25% magnesium sulphate and 10 ml of 10% potassium chloride, were added to Fluid A before CRRT and adjusted during treatment according to the monitored electrolyte level.
Fluid A + Fluid B + Fluid C made up the complete replacement uid formulation during CRRT.

Sodium citrate and calcium gluconate
We used blood-preserving solution ( ) (Batch number: 19052071, 3% sodium citrate)during CRRT (speci cation: 500 ml/bag; ingredients: 22.0 g of sodium citrate, 8 g of citrate, and 24.5 g of glucose per 1,000 ml of blood-preserving solution (I)). Citrate was introduced into the blood from the arterial end (outlet end) of the CRRT tubing at an initial rate of 240 ml/h, while calcium gluconate (10%) was administered through the venous end (inlet end) by a pump at a rate of 20 ml/h, and both were synchronised with the treatment. The ionised calcium level in the CRRT tubing before supplementation with 10% calcium gluconate (venous iCa) and the ionised calcium level in the peripheral arteries (arterial iCa) were monitored during treatment. The infusion rates of citrate and calcium gluconate were adjusted according to the arterial and venous iCa levels to maintain venous iCa at 0.25-0.35 mmol/L and arterial iCa at 1.0-1.2 mmol/L to achieve relatively good anticoagulation effect [5].

Grouping and treatment protocols
Patients who met the inclusion criteria during the study period were randomised at each CRRT session according to odd and even numbers into the conventional citrate infusion group (conventional group) and modi ed citrate infusion group (modi ed group) respectively.
Patients in the two groups were treated identically except for differences in the modes of citrate and 5% sodium bicarbonate administrations and the setting of CRRT treatment parameters.
For conventional citrate infusion group, 5% sodium bicarbonate was placed in the pre-blood pump (PBP) position, and the reconstituted replacement uid was placed in the pre-dilution and post-dilution positions on the three scales of the CRRT machine, according to the manufacturer's settings. Citrate was pumped continuously by a separate infusion pump before the lter. The pre-dilution was set to 50% during treatment, and the rest of the CRRT parameters were maintained as described above. In case of modi ed citrate infusion group, citrate (reconstituted to 3,000 ml/bag) was placed in the PBP position of the CRRT machine and the corresponding initial rate on the CRRT machine was set to 240 ml/h (i.e., the infusion rate of citrate was 240 ml/h). Citrate was administered continuously before the lter when blood was withdrawn from the body. 5% sodium bicarbonate was placed in the pre-dilution position with the corresponding CRRT parameter of pre-dilution set to 5% (the replacement uid rate being set to 2,000 ml/h, and pre-dilution of 5% being 100 ml/h, such that the initial rate of 5% sodium bicarbonate was 100 ml/h). The reconstituted replacement uid was placed in the post-dilution position with post-dilution set to 95%. The rest of the parameters were xed as above during treatment. Data collection Basic information such as sex, age and underlying disease, Acute Physiology and Chronic Health Evaluation II (APACHE II) score for each patient in the two groups at the start of CRRT were noted. Further, duration of each CRRT session, transmembrane pressure at the time of disconnection from dialysis circuit of each session, catheterisation sites, arterial and venous ionised calcium levels during treatment, number of citrate infusion bag replacements, tubing and lter coagulation, as well as relevant biochemical parameters before and after each treatment session, including hepatic function, renal function, coagulation, electrolytes, blood gas analysis, etc., were evaluated.

Statistical analysis
All relevant data collected were analysed using SPSS 19.0 statistical software. The relevant enumeration data were expressed as frequencies and percentages and compared using X 2 test or Fisher's exact probability test. Measurement data that satis ed normal distribution were expressed as x ± s, and intergroup comparison was performed using independent samples t-test or one-way ANOVA. p < 0.05 was adopted as the level of statistically signi cant difference for all data.

General information of patients
Twenty-four patients were selected for the study who met all the inclusion criteria. Out of 24 patients, 17 were male and seven were female, with a mean age of 62.04 ± 18.71 years. Each group was treated for 30 sessions. The conventional group had 10 patients whereas the modi ed group had 14 patients. There was no statistical difference between the two groups in terms of sex, age, hypertension and diabetes mellitus status, or APACHE II score and sequential organ failure assessment (SOFA) score at the time of ICU admission. The general parameters were comparable between the two groups. Table 1 Analysis of patient sources Most of the patients enrolled for the study were from other departments of this hospital, mainly the emergency department and other medical departments, with only one patient from the surgical department, and one patient transferred from another ICU in each of the two groups. Few patients were transferred from other hospitals. However, there was no statistical difference in the source of patients between the two groups (p > 0.05). Table 2 Principal diagnosis of included patients All the patients included in the study were admitted mainly because of acute kidney injury, multi-organ dysfunction, sepsis or septic shock, severe pneumonia etc. Two patients in each group had acute kidney injury as a complication after cardiopulmonary resuscitation, and one patient in each group had an incidence of poisoning. There was no statistical difference in patient diagnosis at the time of admission between the two groups (p > 0.05). Table 3 Comparison of clinical data between the two groups  Table 4 CRRT Vascular Access Selection In each group of 30 treatments, the right femoral vein vascular access was the main one, followed by the left femoral vein and the right internal jugular vein vascular access. There was no statistically signi cant difference between the two groups in the choice of vascular access (p>0.05). Table 5 Comparison of biochemical indicators between the two groups of patients before and after treatment There was no signi cant difference in clinical related indicators between the two groups of patients before treatment (p>0.05). After treatment, the total calcium level of the traditional group was higher, the difference was statistically signi cant (p=0.038), and the difference in other related indicators was not statistically signi cant (p>0.05). Table 6, Table 7 Usage time of cardiopulmonary bypass 96h after the start of CRRT treatment was used as the observation period, and the Kaplan-Meier method was used as the lter usage time curve. The graph shows that the lter usage time is longer in improved citrate infusion than the traditional citrate infusion methods (p = 0.002). Figure 1

Discussion
The application of CRRT is no longer limited to replacement therapy for patients with renal insu ciency but has also been extended to the treatment of critically ill patients with severe sepsis, severe trauma, burns, severe acute pancreatitis etc., and has played an irreplaceable role in reducing mortality in critically ill patients [9,[10][11]. This method requires drawing blood from the patient. Moreover, the treatment duration is long, and coagulation in the tubing and lter reduces the service life of the lter, resulting in treatment interruption, increased treatment costs, increased health care workload, and blood loss of patients [12][13]. Therefore, anticoagulation is a key component to ensure the smooth performance of CRRT. However, this requirement can be a challenge, as many critically ill patients are at increased risk of bleeding.
The safety and e cacy of anticoagulation in CRRT have been the focus of clinical researchers.
The application of local citrate anticoagulation in CRRT for critically ill patients has received much attention in recent years because of its unique advantages. The principle of citrate anticoagulation is mainly to affect the level of ionised calcium in blood and thus intervene in the coagulation process [14]. After pumping citrate through the arterial end of the CRRT tubing, the ionised calcium in blood binds to citrate radical to form calcium citrate, so that the ionised calcium concentration in extracorporeal blood is reduced and the conversion of prothrombin to thrombin is affected. It also affects other coagulation links, thus maintaining the post-lter calcium ion concentration at 0.25-0.35 mmol/L for optimal anticoagulation effect in the extracorporeal circulation [5]. After entering the body, citrate group participates in the tricarboxylic acid cycle in the liver, muscle tissue and renal cortex, and is quickly metabolised to bicarbonate radical ions without any residue [4,14]. Supplementation with su cient ionised calcium to maintain its blood concentration at 1.00-1.20 mmol/L before the blood returns to the body through the venous end will restore the coagulation process in the body to normal. As a result, there will be no systemic anticoagulant effects, which in turn can reduce the incidence of bleeding complications [5,[15][16], and avoid the adverse effects associated with low calcium. Previous studies have con rmed that citrate is safe and effective for haemodialysis in patients with a high-risk of bleeding [6,[17][18][19]. Local citrate anticoagulation was also reported to prolong the duration of CRRT and reduce the cost of treatment [20][21][22][23].
There are two main infusion methods commonly used when citrate is applied for anticoagulation in CRRT: one is priming citrate anticoagulation, in which citrate is added to the replacement uid so that it becomes a part of the replacement uid [24]. The advantage of this method is that electrolyte disturbances are less likely to occur, although it is not suitable for precise adjustment. The other mode is synchronous local citrate anticoagulation [25], in which citrate is infused separately from the replacement uid and the infusion rate of citrate is adjusted according to the monitored ionised calcium and acid-base concentrations. The advantage of this method is that the infusion rate of citrate can be adjusted in real time, thus achieving a better anticoagulation effect. In clinical implementation, citrate is pumped from the outlet end of CRRT via a separate infusion pump. There are clinical situations where the actual infusion rate of the infusion pump is not consistent with the intended infusion rate, which may result in variations in the actual amount of citrate administered over a certain period of time. Reduced citrate concentration will lead to inadequate anticoagulation, tubing coagulation, reduced treatment time, acidosis etc. whereas increased citrate levels can lead to a subsequent risk of citrate accumulation, alkalosis etc. We have implemented a modi ed citrate infusion method in clinical practice by placing the citrate solution directly in the original position with 5% sodium bicarbonate on the CRRT machine, setting the pump speed of the machine(Grouping and treatment protocols), and administering it when the blood is just drawn out of the body, which is closer to the outlet end than the conventional infusion position. This can result in effective calcium chelation in blood reducing the risk of coagulation, and allowing for longer treatment duration and lower transmembrane pressure at the time of disconnection from CRRT circuit. We further inferred that the modi ed citrate infusion method can stabilise the citrate infusion rate, reduce problems associated with unstable citrate infusion rates, and allow more adequate chelation of blood calcium. We therefore monitored venous end ionised calcium levels of patients. The levels were found to be closer to the values required for anticoagulation with minimum uctuations. A higher level of ionised calcium at the venous end and a greater range of uctuations imply an increased chance of coagulation.
The major complications of citrate anticoagulation include citrate accumulation, hypocalcaemia, hypernatraemia and metabolic acid-base disorders [26]. Citrate affects systemic coagulation, and there have been few reports of bleeding induced by citrate anticoagulation [27]. The ideal state of anticoagulation in CRRT is to achieve a good anticoagulation effect without associated complications. We included 60 treatment sessions in each of the study groups and none of the patients had hypocalcaemia or hypernatraemia, due to close monitoring and adjustment during our treatment. The occurrence of citrate accumulation is related to various factors such as the patient's hepatic function, shock, as well as the infusion rate and dose of citrate [28]. Citrate accumulation was not observed in the patients of modi ed infusion group, while three episodes of citrate accumulation occurred in the conventional infusion group. Although the difference in incidences of citrate accumulation between the two groups was not statistically signi cant, the values were marginally lower in the modi ed infusion group. This can be related to accurate infusion rate of citrate in the modi ed infusion group. In addition, the total calcium monitored after treatment was higher in the conventional infusion group, and the elevated total calcium may mean a greater possibility of citrate accumulation. There was no statistically signi cant difference in acid-base status, coagulation, or hepatic function indicators between the two groups after treatment.
We currently use citrate (3% sodium citrate) as a blood-preserving solution. Due to the immediate anticoagulant effect of citrate, in the conventional way of local citrate infusion, nurses have to change the citrate infusion bag about once every 2 h. But with high clinical workload, increased number of critically ill patients, and heavy workload on nurses, frequent changing of infusion bag undoubtedly poses a challenge [29]. Absence of citrate infusion due to delay in changing of the infusion bag can increase the risk of coagulation. Moreover, while infusing citrate in the conventional way, nurses have to include the amount of citrate in the calculation of uid volume, which also increases the clinical workload of nurses and causes di culties in uid management. In contrast, the modi ed local citrate infusion method adds citrate to a three-litre bag, which is placed on a scale same as the CRRT replacement uid. The number of infusion bag replacements and hence time without citrate-free infusion is reduced. It also avoids inclusion of citrate volume in the uid volume calculations, thus reducing the di culty in uid management. Nurses changed the citrate infusion bags much less frequently for the patients in the modi ed citrate compared to those in conventional group, and the differences between the two groups was statistically signi cant.
Thus, the modi ed local citrate infusion method improves lter service life, prolongs treatment duration, achieves a signi cantly better anticoagulant effect than conventional citrate infusion, has a lower risk of citrate accumulation, does not increase metabolic complications, reduces the burden of nursing operations, and increases patient safety.