Reduction Rate of Uric Acid in Blood during Continuous Renal Replacement Therapy for Acute Kidney Injury: A Multicenter Retrospective Observational Study

Renal replacement therapy (RRT) is essential for managing patients with severe acute kidney injury (AKI) and end-stage kidney disease (ESRD). RRT is useful in treating life-threatening conditions in patients with AKI and enables long-term survival for patients with ESRD lacking residual renal function. The optimal therapeutic doses for AKI and ESRD have been extensively investigated. Several large clinical trials have shown that increasing the dose of continuous RRT (CRRT) beyond 20–25 mL/kg/h does not provide additional benefits [1]. For patients with ESRD, the optimal dose is mostly assessed using the parameter Kt/V, which represents the clearance of urea. The widely accepted target for hemodialysis is a single-pool Kt/V of at least 1.2 [2]. The efficacy of RRT and the clearance of uremic toxins are clinically evaluated by measuring small solutes in blood, such as blood urea nitrogen (BUN) and serum creatinine concentration (sCre).

Uric acid (UA) is a product of the metabolic breakdown of purine nucleotides, which are essential components of DNA and RNA. In humans, UA is mostly produced in the liver and excreted through the kidneys and gastrointestinal tract. Blood concentration of UA increases with renal dysfunction, similar to BUN and sCre. The molecular size of UA (186 Da) is comparable to that of BUN (60 Da) and creatinine (113 Da), allowing UA to be eliminated by CRRT with almost equal removal efficacy under identical conditions, including blood flow rate, dialysate flow rate, and ultrafiltration rate with the same filter size. BUN and sCre are known to be affected by factors other than renal excretion. In critically ill patients being treated in the ICU, gastrointestinal bleeding, increased catabolism, muscle atrophy, and decreased protein intake may affect BUN and sCre levels. However, UA has not been used for the clinical evaluation of the efficacy of RRT or uremic toxin accumulation.

This study aimed to investigate whether the reduction rates of three different small solutes – urea, creatinine, and UA – differ under the same CRRT conditions. Measuring these solutes in a single patient undergoing CRRT can provide insights into whether different reduction rates in blood concentration are influenced by factors other than removal by RRT. Furthermore, the study evaluated which small solute reduction rate would be useful for predicting the outcomes of patients with AKI who were treated with CRRT. Although several guidelines recommend a CRRT dose of 20–25 mL/kg/h for AKI [1, 3], the reduction of solutes in blood and the accumulation of uremic substances have not been considered so far. Therefore, this study evaluated a possible predictive performance of small solute reductions for the outcome of death.

The CHANGE study is a multicenter retrospective observational study conducted in 18 tertiary hospitals across Japan (online suppl. Table 1; for all online suppl. material, see https://doi.org/10.1159/000542329). It enrolled adult ICU patients who received CRRT for severe AKI or ESRD complications within 1 year before institutional Ethical Committee approval in each hospital. The decision to initiate CRRT was made by onsite physicians. Diagnosis of AKI and its treatment with CRRT followed the KDIGO AKI guidelines (2012) [4] and the Surviving Sepsis Campaign guidelines (2016 [5] and 2021 [6]). The CHANGE study adhered to the principles of the Declaration of Helsinki. The study protocol was approved by the Ethical Research Review Board of the University of Tokyo (registration number: 2021424NI) as the centralized review under the Ethical Guidelines for Medical and Biological Research Involving Human Subjects by the Ministry of Health, Labour and Welfare, Japan, and subsequently by the ethical boards of all participating hospitals. The requirement for obtaining written informed consent was waived owing to the retrospective nature of the study.

This study utilized the data of limited patients registered in the CHANGE study, as described below. Each onsite healthcare provider collected data on clinical characteristics (demographic data, past medical history, and clinical laboratory results), information on the CRRT procedure, and patient outcomes. For blood tests, the pre-CRRT test was defined as being conducted within 6 h before CRRT initiation, the 1st test was defined as the routine daily test on the next day of CRRT initiation, and the 2nd test was defined as the routine daily test on the 3rd day. Patients with missing data for BUN, sCre, UA, and urine volume, as well as those for whom CRRT was discontinued before the 2nd blood test, were excluded. The primary outcomes of this study were the reduction rates of BUN, sCre, and UA in each patient, and the secondary outcomes were death during CRRT or within 7 days after CRRT termination.

Baseline characteristics measured as continuous or categorical variables were summarized. Categorical data are presented as percentages. Normally and non-normally distributed variables are summarized as mean (standard deviation) and median (interquartile range [IQR]), respectively. The chi-square test or Fisher’s exact test were used to compare categorical data. Continuous variables were compared using Welch’s t test or the Mann-Whitney U test, depending on the distribution of the data. Multiple comparisons were made using the Steel-Dwass method on the 1st test and the 2nd blood test.

The prediction of mortality and the determination of the optimal cutoff value by solute reduction rates were investigated using receiver operating characteristic (ROC) curve analysis with Youden’s index. The impact of small solute reduction rates on mortality was evaluated by multivariable logistic regression analysis.

For our study population, there were missing values for body mass index and sequential organ failure assessment (SOFA) score at the start of CRRT, which may have affected the results. To address this, we replaced each missing value with a set of substituted plausible values through multiple imputation [7]. We created 20 completely filled-in datasets with 10 iterations per dataset using the multiple imputation by chained equations method [8]. The following covariates were used for imputation: age, sex, use of ventilation, and death during CRRT or within 7 days. After imputation, effect estimates were determined, and the results were pooled using Rubin’s rules [9].

We conducted sensitivity analyses to ensure the robustness of our findings. First, we performed multivariable logistic regression analysis particularly among patients with oliguria, defined as urine output under 200 mL per 24 h. Data were analyzed using R version 4.3.3 software (R Foundation for Statistical Computing, Vienna, Austria). A two-sided p value of <0.05 was considered statistically significant.

During the observational period (spanning from April 1, 2021, to March 31, 2022), 1,100 patients were initially enrolled in the CHANGE study. Among these, 651 patients were excluded because of undergoing maintenance dialysis for ESRD (n = 281) and having early termination of CRRT within 3 days (n = 81) or missing data (n = 570). In addition, 5 patients were excluded due to missing urine volume data. Finally, 163 patients were included in the analysis (shown in Fig. 1). Table 1 describes the baseline characteristics of the patients. Their median age was 64.0 (IQR 49.0–74.5) years, and they were predominantly male (63.8%). The in-hospital mortality rate for this cohort was 23.9%. In total, 2 (0.8%) and 43 (26.4%) patients had missing data for body mass index and SOFA scores, respectively. Table 2 provides a summary of the detailed prescription and CRRT settings. The median duration of CRRT was 7.5 (IQR 4.0–12.8) days.

Figure 2a shows the reduction rates of the three solutes. The reduction rates of UA were significantly greater than those of BUN and sCre (UA 34.8 [23.1–47.5]%, BUN 19.6 [4.9–34.4]%, sCre 21.7 [7.0–36.4]%, p < 0.05 for the 1st test; UA 57.9 [46.0–72.8]%, BUN 37.6 [9.6–55.3]%, sCre 37.1 [9.7–53.1]% p < 0.05 for the 2nd test). When the analysis was limited to patients with oliguria (urine output <200 mL per 24 h), similar results were observed (shown in Fig. 2b). Significant correlations between these three small solute reduction rates and the sum of dialysate flow rate and filtration rate, which determines the removal of small solutes, were observed (shown in Fig. 3).

Significant differences between survivors and nonsurvivors in the secondary outcome were observed in mechanical ventilation, oliguria, SOFA score at the start of CRRT, and the duration of CRRT (Table 3), whereas no significant differences were observed in CRRT dose (Table 4). Only the reduction rate of UA was significantly different between the survivors and nonsurvivors, while no significant difference was observed in the reduction rates of urea and sCre (Table 3). ROC analysis revealed that the reduction rates of UA were significantly associated with death during CRRT or within 7 days after the termination of CRRT (area under the ROC curve 0.62 [95% confidence interval 0.52–0.71] for the 1st test and 0.63 [0.54–0.72] for the 2nd test; shown in Fig. 4). The cutoff values for UA reduction rate were determined as 42.9% and 56.6%, respectively. We further conducted a multivariate logistic analysis (Table 5). After adjusting for age and SOFA score, a significant association between UA reduction rate and hospital mortality was observed for both the 1st and 2nd tests. Sensitivity analyses limited to patients with oliguria (urine output <200 mL/24 h) also showed significant associations between UA reduction rates and outcomes (Table 6).

This study demonstrated varying reduction rates of three small solutes (urea, creatinine, and UA) in CRRT-treated patients with AKI. The concentrations of these solutes in blood are influenced not only by their removal via CRRT as well as renal and nonrenal excretion but also by their production rates in the body, endogenous clearance, and volume of distribution [10]. Given that blood samples from the same patient were used to measure the reduction rates of these small solutes, it is reasonable to assume no difference in CRRT removal efficiency. Thus, observed differences in reduction rates likely stem from variations in production rates, endogenous clearance (body pool), renal and nonrenal excretion, and volume of distribution [11].

Blood levels of urea nitrogen are affected by factors like protein hypercatabolism, gastrointestinal hemorrhage, and steroid administration, potentially elevating urea production in ICU patients who are frequently exposed to these conditions. Creatinine levels are widely influenced by muscle mass [12] and may decrease in sepsis conditions [13]. UA in blood is present predominantly as urate (99%), which is elevated by purine intake and increased nucleic acid metabolism [14]. Several medical conditions of rapid cell turnover and subsequent metabolism of endogenous purines are associated with an increase in purine and urate overproduction. Myelo- and lymphoproliferative disorders, hemolytic disorders, tissue hypoxia are possible mechanisms of increased UA production in ICU patients. In healthy individuals, 60% of UA is replaced daily, and blood concentrations tend to decrease when exogenous factors such as food intake are reduced [15]. Renal excretion of UA is 0.5 g/day, while the body pool is 1.2 g/day. In contrast, 100–120 g of creatinine is present in the body, 95% of which is stored in the skeletal muscle, and 2–3 g is metabolized per day, half of which is produced by the body, and the other half is obtained from the diet [16]. Compared to creatinine, a higher percentage of UA is metabolized per day relative to body stores. Therefore, among the three small solutes, UA, with its smaller body pool, is more susceptible to CRRT removal among three small solutes.

This study revealed that UA reduction rates were significantly associated with death during CRRT or within 7 days after termination of CRRT, even after adjusting for severity and age. This result suggests that sufficient reduction of UA may reflect the target condition that CRRT aims to achieve in improving patients with severe AKI. An observational study using a nationwide claims database in Japan reported an unacceptably high mortality rate of approximately 40%–50% in this patient population [17]. Several clinical trials have been conducted to determine the optimal dose of CRRT for AKI; however, CRRT dosing has typically been evaluated using effluent flow rates, expressed as total effluent volume per weight per unit of time (mL/kg/h), instead of actual solute removal or the appearance of solutes in the effluent fluid [18]. A more accurate evaluation of CRRT dosage may need to consider not only the prescribed regimens but also the reduction rate of solutes in blood, such as UA, to better reflect the efficacy of removal by CRRT. There are several possible explanations for the association of UA reduction rate with mortality. UA may be less influenced by systemic conditions such as hypercatabolism or muscle loss than urea and creatinine. As described above, smaller body pool of UA may cause rapid decrease of UA blood concentration, and therefore UA reduction rate could detect the effects of CRRT more sensitively than urea and creatinine. Because UA is recognized as a damage-associated molecular pattern [19], released from ischemic tissues and dying cells, removal of UA may perhaps be a causal factor for mortality reduction. On the other hand, the decline in UA production will also contribute to increase UA reduction rate in addition to removal by CRRT. Because UA is synthesized from xanthine via xanthine oxidase, measurement of this enzyme activity may provide another explanation.

This study has several limitations. First, it included a heterogeneous population from mixed ICUs with a modest sample size. This limits the generalizability of the findings. Multiple regression analysis was conducted with adjustment only by age and SOFA score because the number of the outcome was small. Underlying disease severity cannot be excluded completely. In addition, information on the cause of AKI, vasopressor use, and the presence of shock was not available. Data on longer outcomes such as mortality, dialysis dependence, and renal recovery were not also available. Future studies must validate the utility of small solute reduction rates in patients undergoing CRRT in a larger setting with sufficient clinical information and outcome measurements. Second, this study excluded patients with ESRD. The indications for CRRT in AKI and ESRD differ, and the intrinsic catabolic states and systemic severity between these two conditions may also vary. Further evaluation of patients with ESRD is necessary. Third, decisions to initiate and discontinue CRRT were made by attending physicians at each site. Fourth, the sampling points in this study were not identical for each patient owing to the retrospective nature of the analysis. Because BUN, sCre, and UA were measured with the same blood sample, the reduction rates of these solutes were equally evaluated. However, association of UA reduction rate with mortality needs to be evaluated by future prospective study. Fifth, data on other factors that could have impacts on solute reduction rates such as downtime were not available. Finally, although our findings suggest an association between high UA reduction rates and mortality, casual inference cannot be derived.

Among the three small solutes studied, UA reduction rates were significantly higher in patients with AKI treated with CRRT than those of creatinine and urea. Furthermore, UA reduction rates were significantly associated with death during CRRT or within 7 days of termination of CRRT. These results suggest that UA reduction rate serves as an indicator of effective removal of solutes in severe AKI cases, although the decline in UA production must be taken into account.

Collaborating author names from the CHANGE Study Group were as follows: Takaya Abe (Department of Urology, Iwate Medical University, Iwate, Japan); Shigeki Fujitani (Department of Emergency and Critical Care Medicine, St. Marianna University School of Medicine, Kanagawa, Japan); Noriyuki Hattori (Department of Emergency and Critical Care Medicine, Chiba University Graduate School of Medicine, Chiba, Japan); Shingo Ichiba (Department of Intensive Care Medicine Tokyo Women’s Medical University, Tokyo, Japan); Masafumi Idei (Department of Anesthesiology and Intensive Care Medicine, Yokohama City University, Kanagawa, Japan); Yasuyuki Kakihana (Emergency and Intensive Care Medicine, Kagoshima University Graduate School of Medical and Dental Sciences, Kagoshima, Japan); Mariko Miyazaki (Department of Nephrology, Rheumatology and Endocrinology, Tohoku University Graduate School of Medicine, Miyagi, Japan); Takeshi Moriguchi (Department of Emergency and Critical Care Medicine, University of Yamanashi, Graduate School of Medicine, Yamanashi, Japan); Hiroshi Morimatsu (Department of Anesthesiology and Resuscitology, Okayama University Graduate School of Medicine, Dentistry, and Pharmaceutical Sciences, Okayama, Japan); Hiromasa Nagata (Department of Anesthesiology, Keio University School of Medicine, Tokyo, Japan); Yoshifumi Ohchi (Department of Anesthesiology and Intensive Care, Faculty of Medicine, Oita University, Oita, Japan); Kasumi Satoh (Department of Emergency and Critical Care Medicine, Akita University Graduate School of Medicine, Akita, Japan); Motohiro Sekino (Department of Anesthesiology and Intensive Care Medicine, Nagasaki University Graduate School of Biomedical Sciences, Nagasaki, Japan); Yuichiro Toda (Department of Anesthesiology and Intensive Care Medicine, Kawasaki Medical School, Okayama, Japan); and Natsuko Tokuhira (Department of Anesthesiology and Intensive Care Medicine, Osaka University Graduate School of Medicine, Osaka, Japan).

This study protocol was first approved by the Ethical Research Review Board of the University of Tokyo (registration number: 2021424NI) and subsequently by the ethical boards of all other participating hospitals in the CHANGE study except Iwate Medical University. The ethical research review board of Iwate Medical University (registration number: MH2022-024) approved the CHANGE study. The requirement for obtaining written informed consent was waived owing to the retrospective nature of the study.

The authors have no conflicts of interest to declare.

This study was supported by Nipro Corporation. However, Nipro Corporation did not participate in the entire analysis or result interpretation.

Y.I., R.I., and K.D. designed the study, had full access to all the data, and take responsibility for the integrity and accuracy of the data analysis. H.N. contributed to data collection, data interpretation, and literature searches. Y.M. and O.N. supervised the study and contributed to writing the manuscript. All authors contributed to data acquisition, data analysis, and data interpretation and have reviewed and approved the final version of the manuscript.

The data that support the findings of this study are not publicly available due to their containing information that could compromise the privacy of research participants but are available from the corresponding author [K.D.] upon reasonable request ([email protected]).