Introduction
Orthotopic liver transplantation (OLT) is associated with significant changes in intravascular volume and complex acid-base perturbations. In this setting, intravenous (IV) fluids play a critical role in maintaining acid-base balance, intravascular volume, and end-organ perfusion [1]. Plasma-LyteTM (Baxter Healthcare) has been described as a ‘physiological’ replacement fluid and is frequently used for fluid resuscitation and maintenance in the critical care settings, perioperative management, and in patients undergoing OLT [2]. Plasma-LyteTM, however, is an alkalizing solution with a strong ion difference (SID) of 49 mEq/L, greater than in normal human plasma. Moreover, it does not contain calcium, enhancing blood transfusions compatibility, and does not contain lactate, that requires hepatic metabolism, a potential problem with OLT [3–6]. For these reasons, Plasma-LyteTM is often the preferred crystalloid in patients undergoing OLT.
Plasma-LyteTM, however, contains acetate and gluconate as buffers that are partly dependent on the liver for their metabolism. In patients with liver failure, this can lead to hyperacetatemia, a state associated with myocardial depression and hypotension [7]. Hyperacetatemia can exacerbate the vasodilatory states associated with OLT and hence, acetate is no longer used as a hemodialysis buffer because of cardiovascular instability [8,9].
OLT commonly involves large fluid volumes and numerous blood transfusions. Thus, the use of an alternative to Plasma-LyteTM that contains no acetate, gluconate, or calcium may be desirable. One such physiological solution would be a simple buffered bicarbonate solution with physiochemical properties similar to human plasma. However, because bicarbonate is unstable and gradually dissociates to carbon dioxide, it has not been commercially feasible to prepare a stable solution that can be packaged in polyvinyl chloride bags and stored at room temperature.
However, a fluid-manufacturing company (Adcock Ingram) has now been able to produce a stable bicarbonate buffered solution by creating a double-wrap such that an additional plastic cover is overlaid around the typical polyvinyl chloride bags. This prevents the escape of bicarbonate as carbon dioxide and thus maintains its stability.
Despite its widespread use in many countries, there is a notable lack of clinical research evaluating buffered bicarbonate solution in real-world settings. While buffered bicarbonate fluid has been approved for clinical use by the United States Food and Drug Administration, it is not currently approved by the Therapeutic Goods Administration, the regulatory agency of the Australian Government responsible for ensuring the safety, quality, and efficacy of medicines in Australia.
To date, no studies have compared buffered bicarbonate solution to Plasma-LyteTM in adult patients undergoing OLT. We hypothesized that a buffered bicarbonate solution would be non-inferior to Plasma-LyteTM for preventing the development of metabolic acidosis in adult patients undergoing OLT, as reflected by the standard base excess (SBE) measured 5 minutes post-reperfusion of the donor liver. To address this question, we performed a randomized, open-label, non-inferiority study to compare the physiological effects of these two solutions in adult patients undergoing OLT, with the aim of establishing whether larger trials could be justified.
Materials and Methods
Ethics and clinical trial registration
Ethics approval was obtained from Austin Health Human Research Ethics Committee (No. HREC/53250/Austin-2019) on 5 August 2019 and the trial was prospectively registered with the Australian New Zealand Clinical Trials Registry (No. ACTRN12619001227189) on 5 September 2019. Written informed consent was obtained from all participants. The first and last participants were enrolled on 26 February 2021 and 1 March 2022, respectively. The date of the final participant follow-up was 31 July 2022. The study was conducted at a university affiliated tertiary health service in Melbourne, Australia, that performs 80–100 OLT per annum.
Patients
The inclusion criteria were adult patients (> 18 years of age) who were undergoing primary or a revision of a deceased donor OLT, including donation after brain death or donation after cardiac death. Pediatric patients, pregnant individuals, patients undergoing combined multi-organ (kidney or small bowel) transplants, recipients with fulminant liver failure, those with end-stage renal disease (serum creatinine > 300 mmol/L), or those requiring planned intraoperative veno-veno bypass and hemofiltration were excluded from this study. The transplant unit does not perform adult living-related liver transplantation.
Intervention
Eligible patients were allocated randomly to receive either a bicarbonate-buffered solution (Adcock Ingram Inc.) or Plasma-LyteTM (Baxter HealthcareTM) intraoperatively. The bicarbonate-buffered solution was supplied in VIAFLEX® (Baxter International Inc) plastic bag containers that were produced from a uniquely formulated form of polyvinyl chloride. The clinician responsible for delivering the intervention intraoperatively was not blinded to the group allocation. The physicochemical properties of the two solutions are presented in Table 1.
Randomization
Randomization was performed using computer software with a random number generator. Two treatment groups were used (buffered bicarbonate solution and Plasma-LyteTM) using consecutive block sizes of eight and four to generate a total list of 52 numbers. Concealment of the randomization sequence was in sequentially numbered, opaque, sealed envelopes. Each envelope contained the treatment allocation for the participant. Envelopes were opened sequentially as participants were enrolled.
Standardization of fluid therapy
The only crystalloid fluids administered intravenously to patients were the allocated trial fluid. All patients received Albumex® 20% (200 g/L;) (CSL Behring) as the only form of colloid fluid replacement. The components of Albumex® 20% are human albumin (200 g/L), sodium (48–100 mmol/L), and octanoate (32 mmol/L). The volumes of crystalloid, colloid, and blood product administration to counterbalance continuous blood loss were at the discretion of the anesthesiologists.
Fluid intervention followed a protocol designed to standardize patient care. Cell salvage was utilized for all patients to reduce the need for allogeneic blood transfusions. Processed red blood cells were re-infused into the patient via a Belmont® Rapid Infuser RI-2 (Belmont Medical Technologies) to facilitate rapid and precise control of fluid delivery. All donor organs were preserved using Belzer University of Wisconsin (UW)® Cold Storage Solution (Preservation Solutions, Inc.) at temperatures between 2°C and 6°C (36°F and 43°F). Belzer UW® Cold Storage Solution consists of sodium (29 mEq/L) and potassium (125 Eq/L) with a pH of approximately 7.4 at a temperature of 20°C (estimated osmolarity of 320 mosm/L). During caval anastomosis, all donor livers were flushed with two liters of trial solution at a standard temperature of 10°C. During the neo-hepatic phase, additional colloid and crystalloid fluids were administered to counterbalance blood loss, again at the discretion of the anesthesiologists.
A standardized protocol of thrombo-elastography and conventional laboratory coagulation tests facilitated point-of-care assessment of the quality of hemostasis and guided the administration of fresh frozen plasma, platelet, and cryoprecipitate transfusions. The use of allogeneic blood was in accordance with Australian patient blood management guidelines [10]. Electrolyte and metabolic disturbances were managed using standard intraoperative protocols. Administration of 8.4% bicarbonate was permitted as needed to treat severe acidemia (pH < 7.2), or the management of life-threatening hyperkalemia.
Outcomes
The primary outcome was the development of metabolic acidosis, as reflected by the SBE that was measured 5 minutes after reperfusion of the donor liver. The non-inferiority margin was defined as a mean difference between the groups in SBE of −2.5 mEq/L. Secondary outcomes included measurements of the SBE, bicarbonate, pH, lactate levels, sodium-chloride difference, and SID at the following eight prespecified time points: baseline (preoperatively), 30 minutes after clamping of the portal vein (i.e., the anhepatic phase), 5 minutes pre-reperfusion, 5 minutes post-reperfusion, one-hour post-reperfusion, at the completion of the surgical procedure (skin closure), on admission to the intensive care unit (ICU), and at 24 hours postoperatively. The strong ion gap (SIG) was measured at three prespecified time points, namely baseline (preoperatively), on admission to the ICU, and 24 hours postoperatively.
Data on ICU length of stay (LOS), hospital LOS, post-operative complications, and 30-day mortality were also collected. Postoperative complications were defined and classified according to the European Perioperative Clinical Outcome definitions [11] and graded according to the Clavien-Dindo classification [12]. Acute renal failure was defined by the Risk, Injury, Failure, Loss of kidney function, and End-stage kidney disease (RIFLE) classification [13].
Measurement of primary and secondary outcomes
Clinical chemistry analyses were conducted at 37°C. SBE, pH, lactate, and routine electrolytes in arterial blood were measured using an ABL 800 Blood Gas Analyzer® from Radiometer. The analyzer has a fully automated micromode that eliminates the possibility of user-induced bias or loss of accuracy when working with very small samples. Additionally, it has an interference-protected lactate analysis feature. SBE was computed utilizing the Van Slyke equation and bicarbonate concentration was determined using the Henderson-Hasselbalch equation [14]. The SID and SIG were determined using Stewart’s theory [14,15] in previously documented formulas [16] with adjustments to account for the influence of weak acids.
Other data collected
Preoperative data included patient demographics, anthropometric measurements, hematological and biochemical blood test results, Model for End-Stage Liver Disease (MELD) scores, and indications for the liver transplant procedure. Intraoperative data collected included volumes of administration for the trial fluid, Albumex 20%®, blood, and blood products. Postoperative data included time to tracheal extubation, sedation scores, requirement for vasoactive drugs, and type and volume of postoperative IV fluids administered.
Statistical methodology
For the primary endpoint, the sample size was determined to evaluate the non-inferiority of bicarbonate-buffered solution based on findings from a previous pilot study of patients undergoing OLT at our institution [16]. Based on this primary outcome, with a one-sided significance level (P value) of 0.025, a power (1−beta) of 80%, a standard deviation (SD) of 3.16, and a non-inferiority margin of −2.5 mEq/L, the required sample size was 26 patients in each arm. Non-inferiority was defined as a mean difference between the groups of the SBE of less than −2.5 mEq/L. While a deviation from baseline SBE > −2.5 mEq/L is unlikely to have clinically meaningful effects on patient outcomes in patients undergoing liver transplantation, the clinical and scientific rationale for setting the non-inferiority margin to −2.5 mEg/L was to ensure that any observed effects of the buffered bicarbonate solution fell within an extremely safe range. Further, an absolute value of −2.5 mEq/L for SBE is within one and a half of SD, which is a commonly used standards to delineate the smallest differences of clinical relevance.
For the analysis of the primary outcome, a quantile regression method was applied to confirm the non-inferiority using R Statistical SoftwareTM (v4.1.2; R Core Team 2021). Compared with the classical method to assess non-inferiority, the quantile regression method produces reciprocal P values. That is, non-inferiority is established with a P value < 0.025 by quantile regression. The regression model was estimated with quantile regression for the median difference of each group. To ensure the robustness and reliability of the estimated coefficients, standard errors were derived from 1,000 bootstrap samples. A Z-score was then calculated using the difference between the regression coefficients and the predefined non-inferiority margin of −2.5. The one-sided P value, derived from the Z-score, evaluates whether the observed median difference is significantly worse than the non-inferiority margin. Non-inferiority is established if this P value is less than 0.025, indicating a less than 2.5% chance that such a median difference would occur if the actual difference were worse than −2.5. Furthermore, the lower boundary of the 95% CI for the coefficient was calculated from the bootstrapped standard error and estimated coefficient.
For the analysis of all the secondary outcomes, two-sided tests were used to investigate for differences between the two solutions. Continuous variables were assessed for normality using Shapiro’s test or visualized using Q-Q plots. Descriptive statistics were presented as mean ± standard deviations (SD), median (Q1, Q3), or frequencies (percentages), as appropriate. Repeated measures analysis of variance (RM-ANOVA) was conducted to test for differences between prespecified acid-base variables over time. RM-ANOVA was conducted as a one-repeated and one-between-factors design; the sphericity assumption was evaluated using Mauchly’s test, and the correction was applied to the interpretations of the results if the sphericity assumption was violated.
For all secondary outcomes, statistical significance was defined as a two-sided P value of < 0.05. The study was reported according to the guidelines for reporting non-inferiority and equivalence randomized trials, which is an extension of the Consolidated Standards of Reporting Trials (CONSORT Statement) [17]. All secondary analyses were performed using R Statistical Software (v4.1.2; R Core Team 2021). All figures were generated using GraphPad Prism version 10.0; Mac GraphPad Software).
Results
Baseline characteristics and intraoperative data
Of the 64 consecutive patients undergoing OLT screened for this study, 52 fulfilled the inclusion criteria (Fig. 1). There were no deviations from the study protocol, and all patients received the allocated treatments. Patients’ demographics, anthropometric measurements, preoperative blood results, indications for transplantation, and transplant characteristics are presented in Table 2. The intraoperative data, including fluid administration, fluid output and blood loss, and use of vasoactive medications are summarized in Table 3. There were no reported adverse events or safety concerns associated with the use of Plasma-LyteTM or bicarbonate solution.
Primary outcome
The median (Q1, Q3) SBE at 5 minutes post-reperfusion was −4.857 (−6.231, −3.565) mEq/L in patients in the bicarbonate-buffered solution group and −4.749 (−7.574, −2.963) mEq/L in those in the Plasma-LyteTM group. The estimated median difference by quantile regression was −0.043 mEq/L (95% CI −1.988 to 1.902 mEq/L; one sided P = 0.015).
Secondary outcomes
As shown in Fig. 2, using ANOVA, there were no statistically significant differences in any of these acid-base variables between the two groups at the specified time points; SBE: F (7, 350) = 0.656, P = 0.709; bicarbonate: F (7, 350) = 1.216, P = 0.293; pH: F (7, 350) = 0.465, P = 0.860); and lactate: F (7, 350) = 0.464, P = 0.860. Similarly, no statistically significant differences were detected in the comparisons of SID or sodium-chloride difference: F (7, 350) = 1.227, P = 0.287, or SIG: F (2, 100) = 2.340, P = 0.102] (Fig. 3). There were no significant differences in the duration of ICU or hospital LOS, complications, or 30-day mortality (Table 4). Fluids and vasopressors administered in the first 24 hours after ICU admission are presented in the Supplementary File 1.
Discussion
Key findings
In this randomized study a bicarbonate-buffered solution contained in a novel soft polyvinyl chloride wrap was non-inferior to Plasma-LyteTM in maintaining SBE at 5 minutes post-reperfusion in adult patients undergoing OLT. Moreover, there were no significant differences between the two groups in any of the acid-base variables throughout surgery and for 24 hours postoperatively. Finally, our results revealed no clinical or statistical differences in the development of complications or patient mortality.
Relationship to previous studies
The bicarbonate-buffered solution used in our study is a more chloride-rich solution (chloride content 110 mmol/L) relative to Plasma-LyteTM (chloride content 98 mmol/L). However, its chloride content is similar to Hartmann’s solution (chloride content 109 mmol/L) and much less than saline (chloride content 150 mmol/L). The increased chloride load can theoretically cause adverse physiological events [18,19]. However, in this regard, our findings support those of the SPLIT (0.9% Saline versus Plasma-Lyte or Intensive Care Unit Therapy) [20], LICRA (Limiting IV Chloride to Reduce Acute Kidney Injury after Cardiac Surgery) [21], and SOLAR (Saline or Lactated Ringer’s) [22] studies, all of which showed no clinically meaningful differences in complications in patients with more chloride-rich fluids (fluid dose of 1 900 ml). A unique feature of our trial, however, is that both groups received large fluid volumes (proximately 5 000 ml per patient) of Plasma-LyteTM or bicarbonate-buffered solution and yet did not experience differences in chloride concentration or SID.
Severe hepatic dysfunction, along with the exclusion of the liver from circulation, is associated with several unique pharmacological and physiological considerations. Balanced solutions contain buffers and inorganic anions such as lactate (e.g., Hartmann’s solution) or acetate and gluconate (e.g., Plasma-LyteTM) that are normally metabolized by the liver. Exclusion of the liver during OLT surgery can result in the accumulation of these anions and lead to iatrogenic metabolic derangements such as hyperlactatemia or hyperacetatemia [23–28]. However, in our study, there were no significant differences in the SIG, suggesting that the exogenous anions (i.e., acetate and gluconate) in Plasma-LyteTM were adequately metabolized during OLT [29].
Strengths and limitations
To the best of our knowledge, this is the first randomized trial designed to compare the clinical and biochemical outcomes of large-volume infusion when comparing a bicarbonate-buffered solution to Plasma-LyteTM in patients undergoing OLT. Within the scope of a pilot study, our results suggest that volume management with a bicarbonate-buffered solution was non-inferior to Plasma-LyteTM for maintaining acid-base homeostasis in this setting.
We acknowledge several limitations. This is a pilot trial that does not have the statistical power to assess small to moderate differences in clinical outcomes. However, our intention was to assess the acid-base effect of the two fluids and to provide insight into the sample sizes that might be needed for future larger trials. Likewise, albumin and phosphate concentrations were only sampled at three time points, and not tested intraoperatively. As a result, we were unable to calculate intraoperative SIGs and extrapolate these findings to estimate the concentrations of unmeasured ions during each phase of surgery. Furthermore, acetate and gluconate levels were not measured in the present study, both of which have been identified as elevated in patients undergoing cardiac bypass surgery with Plasma-LyteTM as the priming solution [30,31]. Although surgeons, intensivists, nursing staff, data collectors, and the trial statistician were blinded to the specific trial fluid intervention received by each patient, this was not the case for the anesthesiologist. Finally, this trial was limited to a single center and conducted in adult patients only, that limits the external validity of the results to other centers and pediatric patients. However, treatment was standardized and reflected similar protocols in other transplantation centers.
In a pilot randomized trial, a bicarbonate-buffered solution contained in a novel polyvinyl chloride bag solution designed to prevent bicarbonate losses as carbon dioxide was non-inferior to Plasma-LyteTM for acid-base management during OLT. Moreover, it appeared to provide similar volume expansion effects. Finally, it carried no statistical difference in complications and clinical outcomes. Thus, our findings provide preliminary evidence that either solution can be used safely in adult patients undergoing OLT and justify further investigations of a solution that, for the first time, can be stored, comes in a soft polyvinyl chloride bag similar to other routinely used IV fluid preparation, and uses a physiological buffer. These findings could also pave the way for further research, regulatory approval, and clinical adoption of the buffered bicarbonate in Australia, potentially improving care quality and patient outcomes.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Supplement
Print
