Renal angina index predicts fluid overload in critically ill children: an observational cohort study

Background Fluid overload and acute kidney injury are common and associated with poor outcomes among critically ill children. The prodrome of renal angina stratifies patients by risk for severe acute kidney injury, but the predictive discrimination for fluid overload is unknown. Methods Post-hoc analysis of patients admitted to a tertiary care pediatric intensive care unit (PICU). The primary outcome was the performance of renal angina fulfillment on day of ICU admission to predict fluid overload ≥15% on Day 3. Results 77/139 children (55%) fulfilled renal angina (RA+). After adjusting for covariates, RA+ was associated with increased odds of fluid overload on Day 3 (adjusted odds ratio (aOR) 5.1, 95% CI 1.23–21.2, p = 0.025, versus RA-). RA- resulted in a 90% negative predictive value for fluid overload on Day 3. Median fluid overload was significantly higher in RA+ patients with severe acute kidney injury compared to RA+ patients without severe acute kidney injury (% fluid overload on Day 3: 8.8% vs. 0.73%, p = 0.002). Conclusion Among critically ill children, fulfillment of renal angina was associated with increased odds of fluid overload versus the absence of renal angina and a higher fluid overload among patients who developed acute kidney injury. Renal angina directed risk classification may identify patients at highest risk for fluid accumulation. Expanded study in larger populations is warranted. Supplementary Information The online version contains supplementary material available at 10.1186/s12882-021-02540-6.


Background
Fluid overload (FO) and acute kidney injury (AKI) are common and both are associated with poor outcomes among critically ill children [1][2][3][4][5]. Over 25% of all children admitted to a pediatric intensive care unit (PICU) experience AKI and severe AKI is independently associated with a 12% mortality rate [5]. Similarly, FO occurs in over one-third of critically ill children and it is consistently associated with increased mortality, mechanical ventilation duration, hospital and intensive care unit (ICU) length of stay [1,2]. Furthermore, emerging evidence suggests concurrent FO and AKI worsens outcomes significantly [6].
There is no definitive therapy for AKI after it occurs [7]. In the absence of proven restorative therapy, prevention of AKI has been identified as a priority for management of at risk critically ill patients [7]. Identification of these patients appears to be possible using the renal angina (RA) prodrome for AKI risk stratification [8,9]. The Renal Angina Index (RAI) is a validated measurement of RA that combines patient-specific risk factors as well as early signs of renal dysfunction [9,10].

Open Access
*Correspondence: smgorga@med.umich.edu 1 Department of Pediatrics, University of Michigan Medical School, 1500 E. Medical Center Drive, F-6890, Ann Arbor, MI 48109, USA Full list of author information is available at the end of the article ( Figure 1) Scored at 12 h into a hospitalization course, the RAI can predict the development of severe AKI on Day 3 of hospitalization with a high negative predictive value (NPV) [9][10][11][12]. Utilizing the RAI early in a hospitalization to identify patients at risk for severe AKI provides an opportunity for the development of targeted, patient-centered care to prevent AKI.
Similar to AKI, FO after Day 3 of hospitalization is associated with poor outcomes [13][14][15][16]. Further, higher peak FO and the rate of accumulation of FO is associated with considerable mortality and morbidity [1,2]. Randomized and observational trial data demonstrate consistent benefit of targeting neutral fluid balances in the ICU and clinicians report management aligned with these evidence-based practices [17][18][19][20][21]. Despite this, evidence suggests that in daily practice, clinicians fail to recognize nearly one-third of patients who have significant FO and patients continue to receive unrecognized levels of fluid administration [19,22,23]. This delayed recognition likely contributes to worse outcomes [14,19]. Currently, there are no instruments to assist clinicians' identification of patients at risk for the development of significant FO.
While the RAI can identify patients who are at high risk for AKI, its ability to identify patients who are at risk for FO is unknown. We therefore investigated the predictive efficacy of the RAI to identify significant FO among critically ill children after the initial resuscitative period. We hypothesize that RA would be associated with higher FO states on Day 3 after ICU admission.

Study population
A prospective, observational convenience cohort of patients admitted to a PICU at a tertiary care children's hospital from March 2017 through August 2018 was evaluated, as part of a larger study of clinician identification and prediction of multiple organ dysfunction syndrome in children [24]. All patients included in the parent study period had an expected length of stay greater than 2 days at admission determined by a member of the care team as previously described. For our current study, patients were required to have a serum creatinine measurement within 12 h of ICU admission as well as complete fluid balance data through Day 3 after ICU admission. Patients were also required to have the complete data necessary to determine RAI, including transplant, mechanical ventilation, and vasoactive medication status. Exclusion criteria were death before Day 3, history of stage 5 chronic kidney disease (i.e., estimated glomerular filtration rate [eGFR] < 15 mL/min per 1.73m 2 or on maintenance dialysis), kidney transplantation in the last 90 days, incomplete kidney injury information, or patients admitted during the study period but were not part of the parent study. All methods were carried out in accordance with relevant ethics guidelines and regulations and approved by the University of Michigan. The Institutional Review Board (IRB) at the University of Michigan approved this study and because standard of care was not altered in any way, written informed consent was waived for patients.

Data collection
Patients' medical records were reviewed for creatinine measurements, past medical history, total fluid input and total fluid output. Hourly urine output was recorded by PICU nursing staff per unit protocol. In an effort to minimize false positive identification of oliguria, in patients without urethral catheters, the total output of mixed stool Fig. 1 The renal angina index. Adapted from Basu, et al. [10]. Renal angina index is calculated by multiplying the Risk Strata score by the Injury Strata score. The score assigned in the Injury Strata is the highest score based on either the serum creatinine or FO accumulation. The score is calculated at 12 h after ICU admission. FO: Fluid overload, by percentage and urine was used to define urine output as previously published [25]. Additionally, if the patient had an unmeasurable output, it was defined as a normal amount for age by default. To further protect against false positive identification of oliguria in patients without urethral catheters, total urine output was divided over the period of time between instances of output, allowing for normal sleep periods of > 6 h with no urine output. This protected false identification of stage 1 AKI by Kidney Diseases: Improving Global Outcomes (KDIGO) [26].

Definitions
Baseline creatinine was measured based on the lowest measured creatinine in the previous 3 months prior to hospitalization [27][28][29]. If the patient was hospitalized for the last 3 months prior to PICU admission, the lowest creatinine measurement more than 1 month prior to PICU admission was used. If this measurement was not available, baseline creatinine was imputed using age dependent calculations with the assumption of a glomerular filtration rate (GFR) of 120 ml/min/1.73m 2 , as previously published [29]. We utilized KDIGO staging criteria to define and classify AKI [26]. The worst stage achieved by serum creatinine or urine output was used to classify kidney injury stage. Severe acute kidney injury was defined as stage 2 or worse.
The renal angina index was determined after 12 h on the day of ICU admission as previously described, combining the risk and injury scores [9]. (Figure 1) A RAI score of 8 or more is considered fulfilment of renal angina (RA+) based on derivation and validation studies [9,10].
Hourly FO was determined using intake and output. Because percent FO is part of the definition of RAI, cumulative percent FO started at 12 h after ICU admission, after RAI fulfilment. All subjects were normalized to 0% FO after RAI calculation and FO was calculated thereafter. Cumulative FO was then calculated every 12 h through 96 h according to previously published definitions [14]:

Outcomes
The primary outcome was the amount of FO at Day 3 after ICU admission. Due to anticipated nonparametric data, FO as an outcome for regression models was dichotomized at 15% in order to harmonize multiple previously published definitions, accepted published FO data, and recent consensus statements regarding FO in various critically ill populations of children [1,14,30]. Secondary outcomes included differences in FO at 12 h increments through Day 3, as this is seen as clinically relevant time points for assessment and intervention after the period of resuscitation [7].

Percent FO =
Sum of daily Fluid Liters in − Fluid Liters out x 100

ICU admission weight kg
Furthermore, assessment of the predictive characteristics of RAI for severe AKI were explored. Given that we investigated the predictive characteristics of RAI for both AKI and FO, and these states can exist together or separately, patients were further characterized into 4 discrete phenotypes at Day 3 based on the presence or absence of FO and AKI: 1) FO+/AKI+, 2) FO−/AKI+,3) FO+/ AKI-, and 4) FO−/AKI- [6]. Finally, we also investigated included the association of FO with ICU length of stay, hospital length of stay, renal replacement therapy utilization, and mortality at 30 days.

Statistical analysis
Statistical analysis was performed using Stata 16 (Stata-Corp, LLC, College Station, TX, USA). Categorical data are presented as number and percentages and compared by χ 2 or Fisher's exact test. Continuous data are presented as medians and interquartile ranges (IQR), as we anticipated non-normally distributed data, and compared using the Mann-Whitney test. Repeated admissions were included and considered as a separate, discrete risk of developing AKI. We used bivariate and multivariate logistic models to correct for significant covariate effects and identify independent associations with outcomes. All bivariate associations carrying associations with p < 0.15 were included as multivariate model terms. A p value ≤0.05 was considered significant.

Acute kidney injury
Severe AKI developed in 33/139 (23.7%) children (  Table 1). RA+ was also associated with increased renal replacement therapy (RRT) use, but not with mortality, ICU, or hospital length of stay (Table 1).

Cumulative fluid trajectories
In order to evaluate fluid trajectories among patients at risk for severe AKI by RA+, median FO% between patients who developed AKI were compared to those who did not. Median values were compared every 12 h from RAI scoring ( Fig. 3) (Supplemental Table 2). Among RA+ patients, 28/77 (36%) developed severe AKI on Day 3. Median FO% was significantly higher among those who developed severe AKI on Day 3 compared to those who did not at each timepoint starting at 36 h (4.5% vs. 2.3%, p = 0.022) through 96 h (8.8% vs. 0.73%, p = 0.002) (Fig. 3A) (Supplemental Table 3). Among RA-patients, 5 of 62 (8%) developed severe AKI at Day 3. Fluid trajectories were not statistically different at any time point through 96 h between the two groups (3.7% severe AKI vs. 3.7% no severe AKI, p = 0.77) (Fig. 3B) (Supplemental Table 4). Among RA+ patients, FO at 96 h was independently associated with CRRT use and hospital length of stay, but not ICU length of stay or mortality at 30 days (Supplemental Table 5).

Discussion
In this analysis of critically ill children, renal angina fulfillment was independently associated with the development of FO at Day 3 after ICU admission. Additionally, fluid accumulation was higher among those children who fulfilled RA criteria and developed severe AKI compared to those who did not develop severe AKI. This difference was not seen among patients who did not fulfill RA criteria, which may suggest an The RAI has the ability to discriminate critically ill children who are at risk of developing severe AKI on Day 3 [9,10]. Here, we confirm and extend these findings to fluid accumulation, with children fulfilling RAI criteria having over 5 times odds of developing FO. FO states are common among critically ill children and are associated with a large burden of resource utilization, morbidity and mortality in multiple pediatric populations [2,15,16,[31][32][33][34]. In particular, we find that all patients who developed the FO+/AKI+ phenotype were RA+. (Figure 4, Supplemental Table 6) Thus, RAI can identify patients at risk for severe AKI and may identify those at risk for fluid overload and therefore may benefit from fluid management strategies.
Continued investigation into the interplay of FO and AKI has demonstrated both independent and synergistic effects on the outcomes of critically ill children [6]. In our study, RAI was associated with FO status independently of severe AKI, further suggesting independent phenotypes within this population. The ability to predict the development of FO independent of AKI is important for clinicians at the bedside, offering an opportunity    [17,19]. With a high negative predictive value of 90% for predicting FO, RAI may provide an opportunity to identify children early in an ICU course for population enrichment for similar trials evaluating directed fluid management strategies in critically ill children.
In contrast to AKI, early FO treatment and mitigation strategies do exist and are feasible [17,19,20,[35][36][37][38]. While fluid administration is a mainstay of pediatric resuscitation and support, excessive fluid accumulation is common and often goes unrecognized with nearly one-third of patients with clinical signs and symptoms of FO ≥15% is unnoticed [22,23,39]. Major contributions to FO states tend to be from "fluid creep" as opposed to resuscitation fluids early in the ICU course and median fluid exposure exceeds weight-based maintenance on Day 3 of critical illness [22]. Here, we found that fluid accumulation was significantly different among children who developed severe AKI compared to those who did not. This finding was present as early as 36 h after ICU admission, suggesting an early opportunity to recognize and intervene before significant FO develops. Importantly, this finding was seen among patients who fulfilled RA criteria, but not among those who did not. As such, this discriminatory power of RAI may be helpful in identifying those children most at risk for excessive fluid accumulation at admission.
This study had several limitations. As a retrospective study, all findings are limited to associations as opposed to causality. Importantly, this study does not answer the cause-and-effect nature of AKI and FO. However, it does suggest the possibility to identify children early in an ICU course in a future, prospective way. Secondly, this study did not identify medications which may affect fluid balance or influence development of AKI through nephrotoxic actions. These multiple interactions should be investigated in future, prospective ways. Further, we did not find significant associations with mortality at 30 days. This finding is likely related to low overall mortality and a small cohort with inadequate power to find a difference. However, other patient-and family-centered outcomes, including dialysis receipt and hospital length of stay, were reinforced in this study. The included patients all had an expected length of stay > 48 h, which may suggest these patients were sicker and potentially bias the findings toward the development of AKI and FO. However, as the RAI was used in all patients, this inclusion factor served as a population enrichment strategy. Finally, if a patient did not have a Foley catheter in place, any mixed stool and urine output was classified as "urine output", potentially overestimating urine output and therefore underestimating the rate of oliguria. This may bias the classification of AKI to be less severe than was actually present, potentially weakening or missing important AKI related associations.

Conclusion
In this study of critically ill children at risk for AKI, we found that RA fulfillment was associated with increased odds of FO at Day 3. Furthermore, we demonstrated that fluid accumulation was different as early as 36 h after admission among children who went on to develop severe AKI at Day 3 compared to those who did not develop severe AKI. These findings suggest that RAI may identify patients at high risk of developing significant fluid overload.