Incidence and Clinical Outcomes of Critically Ill Patients with Augmented Renal Clearance: A Retrospective Cohort

Augmented renal clearance (ARC) is a clinical phenomenon of glomerular hyperfiltration typically found in critically ill populations, which leads to enhanced elimination of solutes and medications [1]. Although there is no standardized definition for ARC, most studies define it as a creatinine clearance (CrCl) greater than 130 mL/min/1.73 m² [2]. However, the pathophysiological mechanisms of ARC are not well understood. The primary theory of its occurrence relates to changes in cardiovascular dynamics due to nonspecific immune system activation and release of inflammatory mediators [3, 4]. Increased cardiac output and decreased vascular resistance may result in increased renal blood flow and glomerular filtration, especially in patients with systemic inflammatory response syndrome, sepsis, burns, trauma, or major surgery [3]. Another proposed theory suggests that ARC may result from increased glomerular filtration capacity through the renal functional reserve [3, 5].

The incidence rate of ARC in intensive care unit (ICU) patients with normal serum creatinine values has been reported to be approximately 30–65% across multiple observational studies that vary based on the risk of the studied population [3]. High-risk populations, including those with early sepsis, traumatic brain injury, subarachnoid hemorrhage, burns, and hematologic malignancies, have increased occurrence rates ranging from 50 to 85% [5, 6]. However, the incidence among a general ICU population is less elucidated.

Pharmacokinetic changes in renally eliminated medications are a potential concern. These patients often receive medications with time-dependent activity and/or narrow therapeutic indices, which may include beta-lactam antibiotics, vancomycin, anticoagulants, or antiepileptic medications [3, 5, 7‒10]. In pharmacokinetic studies, enhanced elimination has been demonstrated to result in lower serum concentrations of renally eliminated medications in ARC, potentially resulting in adverse events, development of antimicrobial-resistant organisms, or treatment failure [5, 11]. However, clinical outcome data of patients with ARC in the general ICU population are scarce. There are no established guidelines on dosing adjustments for renally cleared medications in patients at risk or suspected of having ARC. Information regarding the onset, duration, and severity of ARC during ICU admission is limited, which may improve the identification and management of suspected patients.

While ARC may alter pharmacokinetics and serum concentrations of medications, the impact on clinical outcomes is not well known. The decrease in serum concentrations may potentially reduce antimicrobial and anticoagulant exposure, leading to undertreatment of severe infections and potential risk for venous thromboembolism. The impact on development of acute kidney injury (AKI) has not been established, but it is plausible to decrease the risk of AKI for patients with ARC. However, this is largely unknown. Due to the paucity of literature on the clinical outcomes associated with ARC, this study aimed to determine the incidence and clinical outcomes of ARC in a broad, critically ill population.

This single-center, retrospective cohort study was performed at a 532-bed, 65-ICU bed academic tertiary care hospital. This study was approved by the Cleveland Clinic Akron General Institutional Research Review Board (IRRB 19025). The requirement for written informed consent was waived. The study procedures were conducted in accordance with the ethical standards of the IRRB and the Helsinki Declaration of 1975. The study was conducted and reported following the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) criteria checklist [12].

The source population was obtained from an institutional query of the electronic medical record, which included all patients admitted to any ICU between July 1, 2018, and June 30, 2019. Patients were excluded if the documentation indicated (1) a preexisting chronic kidney disease (CKD); (2) transfer to the ICU from the floor; (3) admission for cardiovascular surgery, including coronary artery bypass grafting, vascular, or valvular surgery; (4) the development of AKI requiring renal replacement therapy within the index hospitalization; (5) an ICU length of stay (LOS) of less than 72 h; (6) a preexisting renal transplantation; or (7) age less than 18 years. These criteria were selected in order to include patients with a similar likelihood of developing ARC at baseline, remove potential confounders, and primarily focus on adult patients. The study groups were assigned based on the presence or absence of ARC. The ARC group was defined as CrCl ≥130 mL/min/1.73 m² for at least 1 day during the initial 7 days of the ICU encounter. The non-ARC group consisted of patients with all CrCl values <130 mL/min/1.73 m² during the initial 7-day period [13].

Data were extracted from the electronic medical record by a data analyst, and the remaining data were collected by manual chart review. Trauma-related variables were extracted from the institutional trauma registry. Data were collected by two trained investigators in data collection using a standardized electronic data collection form [14]. Two additional investigators reviewed and validated the data for inconsistencies, missing data, or confirmation of outliers. The collected data included baseline demographic data, anthropometric data including admission height, weight, body mass index (BMI), body surface area, Charlson Comorbidity Index [15], presence of mechanical ventilation (MV), vasopressor use for at least 4 h, severity of illness scores, and CrCl data. Patients were categorized based on one or more reasons for ICU admission, including neurological, operative, nonoperative, trauma, sepsis, cardiovascular, or other reasons. The severity of illness scores included Acute Physiology, Age, Chronic Health Evaluation (APACHE) III, Sequential Organ Failure Assessment (SOFA), modified SOFA [6], ARC [6], and Augmented Renal Clearance in Trauma Intensive Care (ARCTIC) [16]. Daily serum creatinine values were collected, and CrCl estimations using the Cockcroft-Gault equation were calculated for each day in the ICU for up to 7 days [17, 18]. Patients with multiple serum creatinine values for the day had the highest value utilized for the estimate. Online supplementary Table 1 (for all online suppl. material, see https://doi.org/10.1159/000542881) depicts the formulas and pertinent equations used to determine the daily CrCl values. All calculations for the weights and CrCl formulas were built into the data collection tool and validated to ensure accuracy and minimize errors.

The primary outcome was the incidence of ARC. The severity of ARC was categorized as either mild (CrCl 130–170 mL/min/1.73 m²), moderate (CrCl 171–240 mL/min/1.73 m²), or severe (CrCl >240 mL/min/1.73 m²). Secondary outcomes included duration of invasive MV, 28-day alive and ventilator-free status, mortality (inhospital and 30-day), LOS (ICU and hospital), and 30-day readmission rate. Additional outcomes included the incidence rates of venous thromboembolism (during hospitalization and 90 days after discharge), AKI during hospitalization, antimicrobial days, and development of multidrug-resistant (MDR) pathogens (during hospitalization and 90 days after discharge). AKI was defined as an absolute increase in serum creatinine by more than 0.3 mg/dL within 48 h or an increase in serum creatinine by at least 50% from baseline [19]. An MDR pathogen was defined as non-susceptible to at least one agent in three or more antimicrobial categories tested for susceptibility based on the institution’s policy, excluding antimicrobials to which the pathogen is inherently non-susceptible [20]. Subgroup analyses were performed for trauma, operative, nonoperative, and neurologic injury populations.

All eligible consecutive patients within the study timeframe were included in the final analysis. The number of patients with missing data for each variable is shown in the tables. Categorical data are presented as frequencies with proportions and were analyzed using the chi-square or Fisher’s exact test, as appropriate. Continuous variables were reported as means with standard deviations or medians with interquartile ranges (IQRs) and analyzed by Student’s t test or Mann-Whitney U test, based on the normality of data distribution as evaluated by Shapiro-Wilk’s test. Subgroup analysis of ARC severity utilized ANOVA or nonparametric equivalent for analysis of continuous variables across three groups. An a priori value of 0.05 was set to determine statistical significance. Statistical analyses were performed using IBM SPSS (version 24.0; IBM Corp., Armonk, NY, USA).

Propensity score analyses were performed using logistic regression with group as the outcome and confounding variables of age, BMI, sex, and reasons for ICU care as the predictors. Boxplots were used to gauge the degree of overlap and distributional similarity of propensity scores between the ARC and non-ARC groups. Average treatment effect weights were then calculated from the inverse of the propensity score. We evaluated the mean standardized differences of covariates between the groups for both the original and weighted data. We then compared outcome levels for the weighted data using linear regression, with and without log transformation and Rao-Scott chi-square tests. Analyses were performed using SAS^® Software (version 9.4; Cary, NC), and a significance level of 0.05 was assumed for all tests.

A total of 1,445 patients were admitted to the ICU, and 561 (39%) patients were included in the study. The most common reasons for exclusion from the study were baseline CKD, ICU admission from the floor, and admission for cardiovascular surgery (Fig. 1).

The study population was 55% male (n = 308), with a median [IQR] age of 64 [53–74] years, median [IQR] baseline serum creatinine of 0.9 mg/dL [0.7–1.0], and primarily comprised a nonoperative (i.e., medical) population (n = 396, 71%). Patients in the ARC group were more likely to be younger (median [IQR] 48 [31–57] vs. 69 [60–77] years, p < 0.001), male (68% vs. 51%, p < 0.001), have lower BMI (median [IQR] 26 [21–31] vs. 28 [24–33] kg/m², p < 0.001), lower admission baseline Scr (median [IQR] 0.7 [0.5–0.9] vs. 0.9 [0.7–1.1] mg/dL, p < 0.001), and lower Charlson Comorbidity Index (median [IQR] 2 [1–4] vs. 4 [3–6], p < 0.001). The ARC group had a lower severity of illness with lower APACHE III scores (median [IQR] 40 [26–62] vs. 53 [41–68] mg/dL, p < 0.001), but no difference in SOFA or modified SOFA scores. The ARC group had higher utilization of invasive MV (65% vs. 46%, p < 0.001). The ARC group had higher risk assessments (Table 1). There were lower serum creatinine levels and higher CrCl values between the groups for all seven ICU days (online suppl. Table 2). Figure 2 illustrates the difference between the average CrCl values and the ARC group, demonstrating a gradual upward trend until day seven.

The primary outcome of incidence of ARC was 25% (n = 139/561). In the ARC group, the severity was primarily mild (n = 87, 63%) or moderate (n = 36, 26%). While the ARC group had a higher rate of MV, there was no difference in MV duration between the groups. The ARC group had a longer ICU stay (median [IQR] 6 [4–9] vs. 5 [4–8] days, p < 0.001) and a lower rate of inhospital (6% vs. 12%, p = 0.041) and 30-day mortality (10% vs. 21%, p = 0.005). There was a lower rate of AKI (13% vs. 21%, p = 0.027) but a higher rate of developing an MDR pathogen (22% vs. 9%, p < 0.001) in the ARC group (Table 2). After applying the PS analysis, the inhospital and 30-day mortality, AKI, and development of an MDR pathogen during the hospitalization remained statistically significant (online suppl. Table 3). Among the subgroups, the incidence of ARC was found to be 32%, 31%, 32%, and 24% in the trauma, neurological, operative, and nonoperative patients, respectively. Within the ARC severity subgroups, there were no significant differences in the outcomes, except for a higher rate of MDR pathogen development as severity increased (17% vs. 19% vs. 56%, p = 0.002 for mild, moderate, and severe, respectively) (Table 3). There were no differences in the clinical outcomes for the trauma (online suppl. Table 4), neurological (online suppl. Table 6), or operative (online suppl. Table 8) subgroups. However, the trauma patients with ARC had a higher duration of MV after PS weighting (online suppl. Table 5). The operative patients with ARC had a lower rate of AKI and higher rate of developing an MDR during the hospitalization (online suppl. Table 9). Within the nonoperative subgroup, the ARC group had a higher rate of MV and lower rate of ventilator-free days. However, they had lower 30-day mortality and AKI, but a higher rate of developing an MDR pathogen (online suppl. Table 10). This was consistent with the PS weighted analysis (online suppl. Table 11).

In this retrospective cohort study, the overall incidence of ARC in the general ICU population was 25%. Within the ARC group, the CrCl values consistently increased over the 7-day course. Most patients with ARC exhibited mild to moderate severity. Although the ARC group had a longer duration of MV, ICU LOS, and the development of MDR pathogens, they demonstrated a lower rate of mortality and AKI. Most patients likely had a lower ARC risk, given the high number of nonoperative (i.e., medical) patients. This study is novel because ARC incidence has not been elucidated in broader ICU populations and impact on clinical outcomes is lacking. Additionally, the association of ARC severity and clinical outcomes, such as development of MDR pathogens and AKI, has not been previously described.

While there is no standardized method or estimation for CrCl in the ICU, the CG equation is commonly utilized based on Federal Drug Administration recommendations for medication dosing [21]. Although the CG equation has higher accuracy and precision for ICU patients without ARC or underlying CKD, this equation may be biased and inaccurate in patients with ARC and has been found to primarily underestimate the CrCl in this setting [4, 22‒31]. This may have underestimated the incidence of ARC in the present study. Other equations, such as the modified CG, 4-variable and 6-variable Modified Diet in Renal Disease, and Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI), have been shown to result in inaccurate, biased, and poor correlation in the ARC setting [4, 6, 23‒25, 27‒29]. Additionally, newer equations, such as the 2021 CKD-EPI, with and without cystatin C, have not been rigorously tested or validated in patients with ARC or in the ICU setting [32]. Additionally, these formulas have been validated for the staging or prognosis of CKD. Patients with preexisting CKD were excluded from the study, limiting the use of these formulas in this study. While previous studies have not found a reliable estimate for CrCl in patients with ARC, the CG equation has been found to be the most accurate option relative to other estimates of CrCl but still underestimates the actual CrCl [27, 29].

The incidence of ARC in the ICU has been highly variable in the literature. In high-risk ICU populations with neurological, trauma, and burn injuries, the incidence of ARC has been reported to be 65–100% [6, 7, 33, 34]. In lower risk populations such as those with febrile neutropenia or sepsis, this has been shown to range from 16 to 56% [27, 34, 35]. We found a 25% occurrence of ARC in this study, which likely represents a combination of high- and low-risk patients with various diagnoses. This incidence was comparable to that reported in previously published studies that included a primary medical population [36, 37]. Additionally, we observed an increase in the ARC rate over the 7-day period, which was similar to the findings of Egea et al. [36].

Although the characteristics, risk factors, and diagnostic criteria for ARC have been evaluated in the literature, very little outcome data exist. Patients with ARC had a longer duration of MV, presence of MDR pathogens, and ICU LOS. Inhospital mortality, 30-day mortality, and AKI were observed to be lower in patients with ARC. Unfortunately, there are few comparative data on the implications of ARC in critically ill populations, which serves as an area of investigation for future studies. Similar to this study, Egea et al. [36] reported a lower rate of 30-day mortality within the ARC population than in the non-ARC population (16.4% vs. 35.1%, p < 0.01).

Of great concern are the implications of ARC for renally eliminated medications, particularly for antimicrobials. Patients with sepsis represented approximately 20% of the study population, with a higher percentage of patients in the ARC group. The development of MDR pathogens was greater in patients with ARC and occurred more frequently during the index hospitalization. Additionally, there was a higher rate of MDR pathogens for higher ARC severity as compared to mild ARC severity. Given that prior studies have demonstrated lower serum concentrations of renally eliminated antimicrobials in ARC, the development of MDR pathogens is plausible. In previous studies, the presence of ARC was associated with increased rates of therapeutic antimicrobial failure, defined as an impaired clinical response requiring alternative antimicrobial therapy, and development of MDR pathogens [5, 10, 38, 39]. While our study population did not evaluate antimicrobial serum concentrations or treatment failure, there was a significant association with the development of MDR pathogens in this study population. Claus et al. [38] reported a higher rate of therapeutic failure and development of antimicrobial resistance among patients with ARC. It is possible that lower serum concentrations of antimicrobials, as previously described in pharmacokinetic studies, may have played a role in the development of MDR pathogens; however, pharmacokinetic data could not be retrieved retrospectively to confirm medication concentrations as a contributing factor in this study population [5, 7‒9, 40]. Altered pharmacokinetics in patients at high risk for ARC should be considered. However, further research should be conducted to determine whether these patients require aggressive initial antimicrobial therapy for the appropriate treatment of infection.

This study aimed to explore the knowledge gap regarding the clinical outcomes of patients with ARC. One of the assessed outcomes in our population was mortality, and the ARC group had lower rates of inhospital and 30-day mortality. Although previous literature describes a lower severity of illness associated with ARC [3], data on its impact on mortality are lacking. However, this aligns with the severity of illness scores in the ARC group, which had lower APACHE III scores. However, the inhospital mortality rates of both groups were below the expected rates based on the APACHE III scores [41]. Adverse outcomes associated with ARC include increased time on MV and ICU LOS, both of which are unique findings. Although the reason for this occurrence is unknown and warrants further exploration, the populations identified in this study as having increased incidence rates of ARC (neurological injury, trauma, and sepsis) could have influenced these identified outcome rates. Further exploration of the outcomes in patients with ARC is warranted to expand the findings of this study.

This study has several limitations. First, this was a retrospective study that relied on partial manual chart review. Missing or incomplete data for specific fields may present errors, although this was minimized by the majority of data provided by electronic databases, utilization of a standardized data collection tool, and provision of a quality assurance review of the data for completeness and accuracy. The amount of missing data was found to be low as highlighted in the tables. Factors affecting the generalizability of the results included a relatively small sample size and ICU population with broad heterogeneity. This study used the CG equation to estimate CrCl, which has several limitations for the ARC population [3, 22]. While a measured urinary CrCl collected over 8–24 h would be more accurate and ideal for determining ARC [7], these are not routinely performed due to challenges and impracticality, especially in lower risk patients. While the CG equation may be potentially inaccurate or biased in critically ill populations, there is currently no acceptable standard for estimating CrCl in the ICU setting. However, no CrCl equation has been rigorously tested or validated for this population, including the CKD-EPI with and without cystatin C. Additionally, the CG equation was selected because alterations in medication dosing were primarily based on CrCl values from the CG equation [21]. Another limitation is the controversy regarding the true threshold and units of the ARC, which are not well defined [2, 3]. However, the majority of studies utilized a threshold of at least 130, but there is variation between mL/min or normalized to body surface area with mL/min/1.73 m². Normalization for body surface area provided the most data and would provide the best comparison to previous literature; however, this may also have impacted the incidence rates of ARC. Lastly, there was a potential for confounding and bias. Given the study design and the multiple secondary outcomes evaluated, further statistical analyses to account for potential confounding factors were not performed.

This study sought to describe the incidence of ARC in the general ICU population, which is applicable and generalizable to multiple ICU populations. The implication of this study in clinical practice is that the incidence of ARC is potentially lower when applied across a broad range of ICU populations. Although this study attempted to evaluate the clinical outcomes of ARC and non-ARC patients, these results should be hypothesis generating and guide future research. Future studies should evaluate the physiologic mechanism for a lower risk of AKI in patients with ARC as compared to non-ARC patients. Additionally, further research evaluating the implications of ARC-related outcomes in sepsis populations should be conducted. While our study found that ARC was associated with higher rates of MDR pathogens, future studies should evaluate the association between antimicrobial serum concentrations and development of MDR pathogens.

This study identified the incidence of ARC in a broad general ICU population of 25%. There were multiple significant differences in clinical outcomes between the ARC and non-ARC populations. The association between ARC and the development of MDR pathogens and the potential mortality benefits should be explored in future research.

We would like to thank Taylor Tharp, DO, for the peer review support.

This study was approved by the Cleveland Clinic Akron General Institutional Research Review Board (IRRB 19025). The requirement for written informed consent was waived. The study procedures were conducted in accordance with the ethical standards of the IRRB and the Helsinki Declaration of 1975.

The authors have no conflicts of interest to declare.

The authors report no sources of funding for this study.

Drs. Cucci and Mullen had full access to all the data in the study and took responsibility for the integrity of the data and the accuracy of the data analysis. Literature search: Cucci, Radhakrishnan, and Patel. Acquisition, analysis, or interpretation of data, drafting of the manuscript, and critical revision of the manuscript for important intellectual content: Cucci, Mullen, Radhakrishnan, Syed, Patel, and Vazquez. Statistical analysis: Mullen. Administrative, technical, or material support and supervision: Cucci and Mullen.

The authors confirm that data supporting the findings of this study are available within the article and its supplementary materials. Due to the nature of the research, the raw data are not available for privacy/legal reasons and are within compliance of the Institution’s Investigational Review Board approvals. Preliminary results: Yeshwanter Radhakrishnan, Ali Syed; Jaimini S. Patel, Chanda L. Mullen, Michaelia D. Cucci. Incidence and clinical outcomes of critically ill patients with augmented renal clearance (Abstract #1135). Crit Care Med 2021;49(1):569.