Introduction
Acute kidney injury (AKI) is a critical factor in increasing the mortality of critically ill patients admitted to the intensive care unit (ICU) [
1‐
5]. Continuous renal replacement therapy (CRRT) is a rescue measure for patients with both unstable vital signs and severe AKI. The number of AKI cases requiring CRRT has increased to more than 150,000 in the United States over the past few decades [
6]. Despite advances in CRRT technology, the patient outcomes of CRRT due to AKI are still worse [
4,
5,
7‐
10]. Although guidelines exist for CRRT implementation [
11‐
13], CRRT-related complications can occur, and initiating CRRT does not always guarantee a survival advantage, which indicates the importance of an individualized approach [
14‐
17].
Metabolic acidosis is an important feature in severe AKI [
7,
18]. This pathologic condition is attributable to decreased excretion of nonvolatile acids via urination and decreased renal synthesis of bicarbonates [
19]. CRRT can successfully control metabolic acidosis, and exogenous bicarbonates may be added during CRRT. Nevertheless, incomplete correction frequently occurs, perhaps because of less correction by CRRT or high production of acids by patient aggravation. Previous studies have investigated the correlation between pH and mortality in patients receiving CRRT [
7,
18], and initial pH on CRRT was related to subsequent high mortality.
To date, no studies have considered the acidosis trend and its relationship with outcomes after starting CRRT. Herein, we aimed to address this issue by clustering the first 24-hour trajectories of pH on CRRT, and identified that certain trajectory groups had high mortality outcomes despite adjusting multiple variables or matching propensity scores. Furthermore, the pH trends determined the subsequent trend of systemic inflammation evaluated with high sensitivity C-reactive protein (CRP) levels, which might contribute to mortality differences.
Discussion
Since AKI patients who require CRRT are in critical condition, clinicians need to consider the patient status, including vital signs, biochemical results, imaging tests, and medical history. Among these, acidosis is one that reflects hemodynamics, respiration, and tissue oxygenation at once [
23,
24]. Herein, acidosis trajectories were associated with subsequent mortality and inflammatory trends after starting CRRT. The relationship between acidosis trajectories and outcomes was independent of other variables; thus, monitoring both initial and subsequent trends of acidosis is important in this patient subset.
The detrimental effect of acidosis on patient outcome has been documented, particularly in chronic kidney disease [
25‐
27]. There is a consensus that metabolic acidosis leads to insulin resistance, breakdown in skeletal muscle mass, and cardiovascular complications in addition to progression of kidney disease [
28‐
30]. Observational studies have shown that metabolic acidosis is associated with risks of doubling of serum creatinine and all-cause mortality compared to the counterpart normal status [
25,
29,
31,
32]. Small sample-sized clinical trials on alkali supplementation and dietary intervention have demonstrated the beneficial effect of correcting metabolic acidosis on preserving kidney function in patients with chronic kidney disease [
33‐
35].
Similar to chronic kidney disease, metabolic acidosis may be associated with poorer outcomes in AKI patients [
36‐
39]. Acidosis would be both the cause and result of AKI [
40]. Despite the complex relationship between acidosis and AKI, a linear relationship was observed between baseline pH and outcome in most studies [
7,
18]. Here, pH trajectories after starting CRRT were diverse, and some cases could not be recovered from acidosis despite the same protocol on CRRT. Accordingly, the relationship between acidosis trajectories and mortality outcomes was prominent, which indicates that the acidosis trend, in addition to initial acid-base status, is also important for determining the patient outcomes.
The process of acidosis begins with the formation of free radicals, which leads to oxidative stress and results in endothelial dysfunction; finally, cytokines are released that make it difficult to maintain an appropriate blood pressure level [
41‐
44]. Furthermore, acidosis causes malnutrition, which is primarily related to poor survival in patients with septic shock [
45‐
47]. This pathophysiology of metabolic acidosis in AKI is similar to that of CKD, where persistent renal acidosis leads to cardiovascular complications, bone mineral disease, and CKD progression [
25‐
27,
29,
30,
48,
49].
The present study has strengths, such as no missing values and concrete statistical analyses. Nonetheless, there are certain limitations to be discussed. First, because the study design was retrospective in nature, the results could not determine causality between acidosis trajectories and outcomes. Selection bias and residual confounding factors might exist, although we used matching methods to overcome them. Second, other pH-related biochemical parameters, such as pCO2, anion gap, and lactic acid, were not traced, making it difficult to interpret the independent relationship with outcomes. Nevertheless, we collected information such as baseline pCO2, anion gap, and lactate to differentiate the early causes of acidosis in each pH cluster. Finally, time-varying ICU cares, including shifts in catheter site, changes in CRRT settings, and inotropic dosage were not considered in the analysis.
In conclusion, acidosis trajectories determine subsequent worse outcomes, such as high mortality and systemic inflammatory response in patients starting CRRT due to AKI. Accordingly, precise monitoring of acidosis on CRRT may be helpful to predict patient outcomes. However, randomized clinical trial is needed to determine whether pH correction during CRRT improves the survival rate in patients with specific pH groups and to overcome the limitation of retrospective nature. Future trials will address other clinical outcomes, such as renal recovery and ventilator weaning. Hopefully, the present results will be a conceptual rationale for clinical trials with acidosis correction.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.