Mortality of elderly patients with acute kidney injury undergoing continuous renal replacement therapy: is age a risk factor?
Article information
Abstract
Background
Whether advanced age is associated with poor outcomes of elderly patients with acute kidney injury (AKI) requiring continuous renal replacement therapy (CRRT) is controversial. This study aimed to evaluate age effect and predictors for mortality in elderly AKI patients undergoing CRRT.
Methods
Data of 480 elderly AKI patients who underwent CRRT were retrospectively analyzed. Subjects were stratified into two groups according to age: younger-old (age, 65–74 years; n = 205) and older-old (age, ≥75 years; n = 275). Predictors for 28-day and 90-day mortality and age effects were analyzed using multivariable Cox regression analysis and propensity score matching.
Results
Urine output at the start of CRRT (adjusted hazard ratio [aHR], 0.99; 95% confidence interval [CI], 0.99–1.00; p = 0.04), operation (aHR, 0.53; 95% CI, 0.30–0.93; p = 0.03), and use of an intra-aortic balloon pump (aHR, 3.60; 95% CI, 1.18–10.96; p = 0.02) were predictors for 28-day mortality. Ischemic heart disease (aHR, 1.74; 95% CI, 1.02–2.98; p = 0.04) and use of a ventilator (aHR, 0.56; 95% CI, 0.36–0.89; p = 0.01) were predictors for 90-day mortality. The older-old group did not exhibit a higher risk for 28-day or 90-day mortality than the younger-old group in multivariable or propensity score-matched models.
Conclusion
Advanced age was not a risk factor for mortality among elderly AKI patients undergoing CRRT, suggesting that advanced age should not be considered for therapeutic decisions in critically ill elderly patients with AKI requiring CRRT.
Introduction
Elderly individuals conventionally defined as those aged 65 years or older constitute the fastest-growing segment of the population in developed countries [1–3]. Acute kidney injury (AKI) is a common occurrence among geriatric patients with acute illnesses. Its incidence is on the rise among the elderly. This vulnerability to AKI can be attributed to multiple comorbidities, polypharmacy, and age-related structural, functional, and hemodynamic changes in their kidneys [1,3].
Elderly AKI patients face an elevated risk of hemodynamic instability, making them more likely to require continuous renal replacement therapy (CRRT). Nevertheless, the decision to initiate CRRT in elderly individuals is complex, given that these patients might not always benefit from this aggressive, expensive, and life-sustaining therapy [3]. Although there are reports comparing outcomes between elderly and non-elderly critically ill patients with AKI [3–5], not many studies have examined differences in mortality rates based on age in a subpopulation of elderly AKI patients [3,6]. Additionally, there are conflicting results in terms of the impact of age on adverse outcomes of AKI patients [2]. Therefore, this study aimed to evaluate age effect and identify predictors of mortality in elderly AKI patients requiring CRRT.
Methods
Study subjects
This was a multicenter, retrospective cohort study based on data collected from AKI patients who underwent CRRT in intensive care units (ICUs) at three university hospitals (Seoul St. Mary’s Hospital, Yeouido St. Mary’s Hospital, and Bucheon St. Mary’s Hospital) from 2012 to 2020. Patients with end-stage kidney disease were excluded. Among 892 adults with AKI who underwent CRRT, those who were younger than 65 years old were excluded. A total of 480 patients were included in the analysis. They were stratified into two groups according to age: younger-old (age of 65–74 years) and older-old (age of ≥75 years). The reason for dividing the groups based on the age of 75 years is that the definition of “old-old” population may refer to individuals aged over 75 years [7]. In addition, when conducting propensity score matching (PSM) to control confounding factors, using the age of 75 years as the dividing point ensured that a similar number of individuals were distributed between the two groups, facilitating a balanced comparison.
In line with the principles of the Declaration of Helsinki, this study was approved by the Institutional Review Board of The Catholic University of Korea (No. XC20RIDI0198).
Data collection
Baseline demographic data including age, sex, body mass index (BMI), and cause of AKI were collected. BMI was calculated as the patient’s weight in kilograms divided by the square of height in meters (kg/m2). All comorbidities of each patient and biochemical data were collected based on medical chart review. The presence of each disease was defined based on the description in the medical record. Biochemical data included blood levels of hemoglobin, blood urea nitrogen (BUN), creatinine, albumin, bilirubin, alanine aminotransferase (ALT), prothrombin time-international normalized ratio (PT-INR), sodium, potassium, chloride, calcium, phosphorous, magnesium, and uric acid. To assess disease severity at the time of CRRT initiation, data regarding mean blood pressure (MBP), the use of ventilator and vasopressor, and urine output were reviewed. In addition, data on the operation (any surgery or procedure performed during the hospitalization period), use of mechanical cardiac support (intra-aortic balloon pump [IABP], left ventricular assist device, and extracorporeal membrane oxygenation), and medications including renin-angiotensin system (RAS) blockers, diuretics, and statins were collected.
Patient outcomes
Outcomes of this study were short-term mortality and long-term mortality. Short-term mortality was defined as 28-day mortality after CRRT initiation. Long-term mortality was defined as 90-day mortality after CRRT initiation.
Statistical analysis
Continuous variables are presented as mean ± standard deviation and categorical variables are presented as numbers and percentages. Analyses of short- and long-term survival rates for different age cohorts were performed using the log-rank test. Results are presented as a Kaplan-Meier plot. To identify predictive factors for 28-day and 90-day mortality, Cox regression analyses were performed. Variables included in equations were chosen based on the results of univariable analyses if the parameter demonstrated an association with 28-day or 90-day mortality (p < 0.30). The following factors were adjusted in the multivariable Cox regression analysis for short-term mortality: age group, BMI, comorbidities (including hepatobiliary disease, liver cirrhosis, and cancer), ventilator support, urine output at the start of CRRT, any type of operation, cardiac surgery, usage of IABP, and blood levels of bilirubin, PT-INR, sodium, and chloride at the time of CRRT initiation. For long-term mortality, adjusted factors were as follows: age group, comorbidities (including hepatobiliary disease, cancer, and ischemic heart disease), MBP, usage of a ventilator, urine output at the start of CRRT, any type of operation, liver transplantation, and levels of BUN, creatinine, bilirubin, ALT, PT-INR, potassium, chloride, and phosphorous at the time of CRRT initiation. Subgroup Cox regression analysis for 28-day and 90-day mortality was done after excluding patients who received liver transplantation. PSM was additionally used in Cox regression analysis to increase the precision of the estimated effect without increasing bias resulting from the presence of variables potentially associated with survival (confounding factors). Propensity scores were estimated using multiple logistic regression analysis with adjustments for sex, comorbidities including diabetes mellitus, cancer, hypertension, cardiac surgery, and levels of BUN, ALT, bilirubin, and potassium at the time of CRRT initiation. After calculating propensity scores, patients in the younger-old and older-old groups with similar propensity scores were matched at a 1:1 ratio using the nearest-neighbor method with a 0.1-caliper width. Stratified Cox regression analysis was done after PSM.
All statistical tests were conducted using a two-tailed 95% confidence interval (CI). A p-value of <0.05 was considered statistically significant. All descriptive and survival analyses were performed using R version 4.3.1 software program (R Foundation for Statistical Computing).
Results
Baseline characteristics of the study population
In the study population, 42.7% (n = 205) belonged to the younger-old group and 57.3% (n = 275) belonged to the older-old group. As shown in Fig. 1, patients aged 79 years were the most prevalent among the study population. The oldest patient was 99 years old. Table 1 summarizes baseline demographics and laboratory data of the study population. Sepsis was the most common cause of AKI, followed by cardiogenic, postoperative, hypovolemic, hepatorenal, and nephrotoxic causes. The mean age was 69.3 ± 2.8 years in the younger-old group and 80.9 ± 4.6 years in the older-old group. There were more male patients in the older-old group than in the younger-old group. There was no significant difference in BMI, prevalence of comorbidities, disease severity indices, proportion of operation and mechanical cardiac support, or medications between the younger-old group and the older-old group. Laboratory data showed that the older-old group had higher BUN but lower bilirubin and ALT levels at the start of CRRT than the younger-old group. The older-old group showed a tendency to have a higher proportion of those with high potassium levels than the younger-old group. There were no significant differences in blood levels of hemoglobin, creatinine, albumin, PT-INR, sodium, chloride, phosphorous, calcium, magnesium, or uric acid at the start of CRRT between the two groups. After 1:1 PSM, the two groups demonstrated a well-balanced distribution across all baseline characteristics. This includes comorbidities, disease severity, operation, utilization of mechanical cardiac support, medications, and biochemical data. The 1:1 PSM model resulted in 170 patients in each group, with a male ratio of 63.5% in both groups. Table 1 presents the specifics of the parameters both before and after the application of PSM.
Comparison of 28-day and 90-day survival rates between the younger-old group and the older-old group
Fig. 2 shows the survival rates of the two groups using the Kaplan-Meier method. There was no significant difference in the 28-day survival rate (log-rank p = 0.09) (Fig. 2A) or 90-day survival rates (log-rank p = 0.07) (Fig. 2B) between the younger-old group and the older-old group.
Predictors for 28-day mortality
Univariable and multivariable analyses were performed to identify independent prognostic factors for 28-day mortality (Table 2). In univariable analysis, hepatobiliary disease, urine output at the start of CRRT, any type of operation, usage of IABP, and serum chloride level were significantly associated with 28-day mortality (all p < 0.05). Multivariable Cox regression analysis showed that urine output at the start of CRRT (adjusted hazard ratio [HR], 0.99; 95% CI, 0.99–1.00; p = 0.04), any type of operation (adjusted HR, 0.53; 95% CI, 0.30–0.93; p = 0.03), and usage of IABP (adjusted HR, 3.60; 95% CI, 1.18–11.00; p = 0.02) were independent prognostic factors for 28-day mortality.
After excluding patients who underwent liver transplantation, a subgroup analysis was conducted. In univariable analysis, factors such as hepatobiliary disease, liver cirrhosis, urine output at the start of CRRT, any type of operation, usage of IABP, and serum bilirubin and chloride level were found to be associated with 28-day mortality (all p < 0.05). Multivariable Cox regression analysis showed that the use of IABP (adjusted HR, 3.95; 95% CI, 1.29–12.12; p = 0.02) was an independent prognostic factor for 28-day mortality (Supplementary Table 1, available online).
Predictors for 90-day mortality
Univariable and multivariable analyses were performed to assess independent prognostic factors for 90-day mortality (Table 3). In univariable analysis, hepatobiliary disease, cancer, usage of ventilator, any type of operation, and serum phosphorus level were significantly associated with 90-day mortality (all p < 0.05). Multivariable Cox regression analysis showed that older age (adjusted HR, 0.65; 95% CI, 0.44–0.96; p = 0.03), usage of ventilator (adjusted HR, 0.61; 95% CI, 0.43–0.88; p = 0.007) and liver transplantation (adjusted HR, 0.07; 95% CI, 0.01–0.64; p = 0.02) were significantly associated with a decreased risk of 90-day mortality, whereas ischemic heart disease (adjusted HR, 1.77; 95% CI, 1.04–3.01; p = 0.04) was significantly associated with an increased risk of 90-day mortality.
A subgroup analysis after excluding patients who received liver transplantation was conducted. In univariable analysis, comorbidities such as hepatobiliary disease, cancer, usage of ventilator, serum creatinine, bilirubin, and chloride level were associated with 90-day mortality (all p < 0.05). Multivariable analysis showed that liver cirrhosis (adjusted HR, 0.29; 95% CI, 0.09–0.92; p = 0.04), ischemic heart disease (adjusted HR, 2.250; 95% CI, 1.099–4.608; p = 0.027), usage of ventilator (adjusted HR, 0.54; 95% CI, 0.35–0.83; p = 0.005) and usage of vasopressor (adjusted HR, 2.06; 95% CI, 1.26–3.36; p = 0.004) were independent prognostic factors for 90-day mortality (Supplementary Table 2, available online).
Cox proportional hazard regression analysis for mortality of the older-old group
Table 4 shows the results of the Cox proportional hazard regression analysis for the 28-day mortality of the older-old group. In univariable analysis, the risk of 28-day mortality was not higher in the older-old group than in the younger-old group (model 1). In multivariable analysis, the older-old group did not show a higher risk for 28-day mortality than the younger-old group after adjustments for age group, BMI, comorbidities (including hepatobiliary disease, liver cirrhosis, and cancer), ventilator support, urine output at the start of CRRT, any type of operation, cardiac surgery, usage of IABP, and blood levels of bilirubin, PT-INR, sodium, and chloride (model 2). After PSM (model 3), the older-old group was consistently not associated with the risk for 28-day mortality (adjusted HR, 0.92; 95% CI, 0.42–2.00; p = 0.84).
Table 5 shows the results of the Cox proportional hazard regression analysis for 90-day mortality of the older-old group. In univariable analysis, the risk of 90-day mortality was not higher in the older-old group than in the younger-old group (model 1). In multivariable analysis, the older-old group had a lower risk for 90-day mortality than the younger-old group after adjustments for comorbidities (including hepatobiliary disease, cancer, and ischemic heart disease), MBP, usage of ventilator, urine output at the start of CRRT, any type of operation, liver transplantation, and levels of BUN, creatinine, bilirubin, ALT, PT-INR, potassium, chloride, and phosphorous. However, after PSM (model 3), the older-old group was not associated with the risk for 90-day mortality (adjusted HR, 1.00; 95% CI, 0.49–2.05; p > 0.99).
Discussion
The risk of developing AKI is significantly increased in the elderly [8]. Therefore, we often need to consider CRRT for elderly patients with AKI. There are reports showing that age is a risk factor for mortality in AKI patients [3,4,9], while others have reported that age is not associated with mortality [5,6]. The purpose of this study was to examine whether age affected mortality in elderly patients with AKI undergoing CRRT. Results showed that advanced age did not increase short-term or long-term mortality in elderly AKI patients undergoing CRRT. This finding was evident after PSM. These results suggest that it is not reasonable to withhold or hesitate to start CRRT solely because of a patient’s advanced age.
Only a few studies have examined the outcomes of elderly AKI patients undergoing CRRT. There was a prospective multicenter study from South Korea that included 607 patients aged 65 years or older [10], although a detailed age distribution was not described. In a retrospective, single-center study by Rhee et al. [3], 411 patients aged 65 years or older were included, with those aged 75 years or more accounting for 44.3%. Previous studies have also focused on very elderly patients. Conroy et al. [4] have included 118 patients aged 75 years or more. Liu et al. [6] have included 41 patients aged 80 years or more. Funk et al. [5] have included 102 patients aged 80 years or more. These studies indicate that elderly AKI patients undergoing CRRT are common and that application of CRRT in the very elderly group, such as patients with ages more than 75 or 80 years, is not an uncommon finding. Similarly, in this study, the proportion of patients aged ≥75 years was higher than that of patients aged 65 to 74 years (57.3% vs. 42.7%), with patients aged 79 years being the most prevalent among the study population.
In this study, the most common cause of AKI was sepsis, both in the younger-old and older-old groups, similar to the findings of previous studies [10,11]. There was no significant difference in comorbidities, MBP, or urine output at baseline between the two groups, suggesting that hemodynamic or anuric status was not different between the two groups. The baseline BUN level was higher in the older-old group than in the younger-old group. Since BUN is influenced by a range of factors including glomerular filtration, tubular reabsorption of urea, catabolism of endogenous protein, volume status, and upper gastrointestinal bleeding [12], the older-old group might experience reduced glomerular and tubular function, more catabolic and dehydrated state, and increased vulnerability to upper gastrointestinal bleeding than the younger-old group when AKI occurs. Higher levels of bilirubin and ALT in the younger-old group than in the older-old group could be potentially attributed to the fact that those in the younger-old group were more likely to have undergone liver transplantation, although there was no significant difference in the prevalence of hepatobiliary disease or liver cirrhosis. Further investigation and analysis are needed to confirm these hypotheses and determine the reasons behind differences in bilirubin and ALT levels between the two groups. An increasing proportion of the older-old group shifting towards a high potassium level compared to the younger-old group might be associated with aging-related disturbance in renal tubular function and RAS activity [13,14]. In this study, serum creatinine levels between the older-old group and the younger-old group did not exhibit a significant difference. However, considering that serum creatinine level can be influenced by factors such as muscle mass, age, and frailty, the actual difference in serum creatinine level might not be apparent [15] and the residual renal function might have been lower in the older-old group than in the younger-old group.
In this study, the older-old group and the younger-old group did not exhibit a significant difference in 28-day or 90-day survival rate, in contrast with the results of a previous study [3]. This difference might be attributed to the fact that, in our study, there was no significant difference in comorbidities or disease severity indices between the younger-old and older-old groups, whereas there were differences in comorbidities and disease severity in the previous study [3]. We speculate that, since the older-old and young-old groups had similar baseline characteristics, survival rates after CRRT initiation were not significantly different between the two groups.
In the 28-day Cox regression analysis, age did not appear to be a significant predictive factor. However, urine output was associated with a reduced risk of 28-day death. This aligns with previous research that considered urine output as a reliable indicator for renal and multiorgan impairment, suggesting that urine output might play a role in reducing the risk of 28-day mortality. When there is a certain amount of urine output in patients who have started CRRT, it can signify a favorable prognosis. Additionally, in some respects, it can be interpreted that early initiation of CRRT in elderly AKI patients is beneficial for their outcomes. A study by Park et al. [10] supports this consideration. In that study, the mortality rate was analyzed according to urine output by categorizing elderly AKI patients undergoing CRRT based on their median 6-hour urine output immediately before CRRT initiation. Their results showed that patients with a higher urine output just before starting CRRT had better outcomes. In other words, an earlier initiation of CRRT in elderly AKI patients can contribute to improved prognosis. Similar studies suggesting the benefits of early initiation of CRRT have also been reported in AKI patients [16,17]. In this study, operation appeared to be a variable that reduced the 28-day mortality rate. Since the operation included any surgery or procedure performed during hospitalization, a proactive intervention or the presence of correctable causal factors might have led to a better prognosis. On the other hand, the use of IABP appeared to increase the risk of 28-day death. This suggested that there might have been severe cardiac conditions or multiorgan failure that was significant enough to warrant the application of IABP, which affected the short-term mortality.
In the 90-day Cox regression analysis, the use of ventilator lowered the risk of 90-day mortality. The reason was unclear. It could be speculated that the long-term survival of patients might have benefited from active respiratory support. However, in a previous study, the usage of ventilators was not significantly associated with long-term mortality [3]. Since previous research [3] and the current study did not categorize more specific details such as respiratory diseases, further research is needed to understand how different results were obtained in each study. In our study, ischemic heart disease was not a risk factor for 28-day mortality. However, it was a significant predictor for 90-day mortality. It could be assumed that more patients with ischemic heart disease had cardiac decompensation with AKI, which might have led to an increased risk for long-term mortality. Liver transplantation was also identified as a factor in reducing the risk of 90-day mortality. Previous studies have shown an improvement in the survival rates of liver transplantation over time [18,19]. Due to this enhancement, liver transplantation appeared as a factor in improving 90-day survival in this study. However, it did not appear to be a prognostic factor in 28-day mortality. The reason is uncertain; it may be associated with acute complications of liver transplantation in the early postoperative period, such as infection, bleeding, and graft dysfunction. Further research is needed to explore this respect.
The older-old group seemed to have a lower risk for 90-day mortality in multivariable Cox regression analysis, which might be related to a selection bias in this study. Since this was a retrospective cohort study, we tried to minimize the possible confounding effect by performing PSM. The older-old group had no significant impact of age on 28-day or 90-day mortality after PSM. This suggests that advanced age in elderly patients was neither beneficial nor harmful for survival. A previous report has shown that age is not a risk factor for poor outcomes [4]. This may be because of clinical decisions to perform CRRT in critically ill elderly patients with AKI. ICU physicians might have selected elderly patients with fewer comorbidities and greater likelihood of survival, which has been demonstrated in previous studies [20–22]. On the other hand, it also may be because elderly patients can truly benefit from intensive care, which has been demonstrated by a prospective, observational, multicenter study regarding the effects of ICU triage decisions on mortality and ICU benefit in elderly patients [23]. Its results showed that elderly patients had more ICU rejections and higher mortality than younger patients. However, the mortality benefit appeared to be greater for the elderly.
In this study, liver transplantation was performed on younger-old patients rather than older-old patients. Despite the improved survival rates associated with liver transplantation, the inevitability of potential complications such as infection, bleeding, and organ failure after surgery makes it evident that adjustments for post-liver transplantation mortality may be necessary. Therefore, we conducted a subgroup Cox regression analysis after excluding liver transplant recipients and the results consistently showed age was not a significant risk factor for 28- and 90-day mortality.
Our study has several limitations. First, this study did not include variables such as scores indicating the severity of illness because data such as APACHE II (Acute Physiology and Chronic Health Evaluation II) and SOFA (Sequential Organ Failure Assessment) score were lacking. Second, the dose of CRRT was not analyzed because those data were not collected. However, until now, there has been no randomized controlled trial demonstrating the optimal CRRT dose in elderly patients. Third, variables such as comorbidities and disease severity indices did not exhibit significant differences between the older-old and younger-old groups. This might be due to a selection bias, that is, elderly patients with more severe illness might have not been included because they could not even initiate CRRT. Fourth, resulting in due to the retrospective manner of this research, bias and several uncharted comorbidities or events could play a role in the interpretation of short- and long-term mortality in this study. Despite these limitations, the number of elderly patients in this study was not small. In addition, the use of PSM enabled control of confounding variables, which could stand as a strength of this study.
In conclusion, an older age was not a risk factor for short-term or long-term mortality in elderly patients with AKI undergoing CRRT. This supports the importance of active management and application of CRRT in critically ill elderly patients with AKI.
Supplementary Materials
Supplementary data are available at Kidney Research and Clinical Practice online (https://doi.org/10.23876/j.krcp.23.313).
Notes
Conflicts of interest
All authors have no conflicts of interest to declare.
Funding
This research was supported by a cooperative research fund from the Korean Society of Nephrology (2022) and by a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (grant number: HI23C047600).
Data sharing statement
The data presented in this study are available from the corresponding author upon reasonable request.
Authors’ contributions
Conceptualization, Methodology, Funding acquisition: HEY
Formal analysis: JHK, SHE, HWK, HEY
Investigation, Resources: JWM, ESK, EJK, HDK, BHC, SJS, CWY, HEY
Writing–original draft: JHK, HEY
Writing–review & editing: All authors
All authors read and approved the final manuscript.