Outcomes of Intermittent Hemodialysis versus Continuous Kidney Replacement Therapy in Hemodynamically Stable Patients with Acute Kidney Injury: A Prospective, Observational, Multicenter Study

Kidney replacement therapy (KRT) for acute kidney injury (AKI) has evolved considerably over the years, making extracorporeal therapy safer and easier [1]. The current practice is to utilize continuous kidney replacement therapy (CKRT) in hemodynamically unstable critically ill patients and in patients with increased intracranial pressure and to utilize intermittent hemodialysis (IHD) in clinically stable patients [1‒3]. In the absence of clear indications, there is no strong evidence demonstrating a difference in mortality between CKRT and IHD; however, the impact of dialysis modality on kidney recovery has not yet been resolved [1, 2]. We report the results of a multicenter observational study that compared the outcomes of CKRT with those of IHD in AKI patients who were clinically and hemodynamically stable and fit to receive IHD but were instead offered CKRT. CKRT was used instead of IHD in the recruited patients because of the lack of access to water treatment units (i.e., reverse osmosis units) in these hospitals’ wards. Such patients could not be transferred to the inpatient dialysis ward or the dialysis center because of the patients’ condition (e.g., in the cardiac care unit), isolation status, pain, family objection to transfer, or initiation of dialysis after closure of the inpatient services at night. However, there were no differences between the participating hospitals in the quality of nephrology care, such as the number of healthcare professionals or case volume, since they are all under the Ministry of Health. We evaluated whether CKRT had any additional advantage over IHD in this clinically stable group of patients with AKI.

This was a prospective observational study performed in seven hospitals for the Ministry of Health (MoH) of Kuwait from January 1 to December 31, 2021. The study recruited all cases of AKI in adult inpatients above the age of 18 years with native kidneys and an estimated baseline glomerular filtration (eGFR) rate above 10 mL/min who required either IHD or CKRT but did not require inotropic support or mechanical ventilation (i.e., hemodynamically and clinically stable). Kidney transplant recipients, dialysis patients, chronic kidney disease (CKD) patients not on dialysis but with eGFR <10 mL/min (or even higher if they had preemptive hemodialysis access), and patients who died within 24 h of consultation were excluded. Baseline eGFR was the eGFR upon admission before the onset of AKI in cases where AKI developed during admission, or eGFR prior to admission if AKI was diagnosed upon hospital admission. We excluded 2 patients with missing baseline eGFR (one from each group) and 1 patient with missing 30-day eGFR because the patient had left against the medical advice and did not return for follow-up. The decision on when to initiate and terminate dialysis, as well as which modality to use, was left to the treating team. This study was approved by the Joint Committee on Medical and Scientific Research of the Ministry of Health and Kuwait University. Patients’ verbal consent, or in the case of an invalid or a patient younger than 21, the verbal consent was obtained from the next of kin or legal guardian. This study was not a clinical trial; therefore, it was not registered. Patient demographics, clinical profiles, and data on the dialytic and non-dialytic management of AKI were collected. Patients were followed up for 30 days (or less if recovery or death occurred earlier), even after discharge, to record the kidney and patient outcomes. Kidney Disease: Improving Global Outcome (KDIGO) definitions and staging of AKI [3] and for patients with preexisting CKD, the KDIGO definition of CKD [4] were used. The cohort was divided into two subgroups according to the dialysis modality: group 1, IHD, and group 2, CKRT. We defined complete recovery of kidney function at 30 days as a return to within 5 mL/min of the baseline eGFR and a lack of kidney recovery at 30 days as no change from the nadir eGFR. Partial recovery was defined as eGFR between 5 mL/min above nadir and 5 mL/min below baseline at 30 days.

For statistical analysis, the means ± standard deviations (SDs) were used for continuous variables, and numbers/percentages were used for categorical variables for descriptive statistics of the study population. Student’s t tests, Mann-Whitney Wilcoxon tests, or Kruskal-Wallis tests were used for contingency table analysis, as appropriate, whereas Pearson’s chi-square tests or Fisher’s exact tests were used to compare categorical variables. We used univariable logistic regression analysis to evaluate the possible risk factors associated with two main outcomes: mortality and lack of kidney recovery. Backward stepwise multiple logistic regression analysis (p value exit criterion = 0.05) was also performed on the entire cohort to evaluate the possible independent risk factors for mortality and lack of kidney recovery. A p value of <0.05 was considered statistically significant. STATA statistical software version 17 (StataCorp LLC, College Station, TX, USA) was used for the statistical analysis.

Table 1 shows basic characteristics of the cohort. The total number of hemodynamically stable AKI patients on no inotropic support or mechanical ventilation was 229 (age: 59.9 years; males: 60.3%; Kuwaitis: 53.7%; baseline eGFR: 56 mL/min). CKRT was utilized in 72.9% of the cohort. There was no statistically significant difference between the two groups (IHD vs. CKRT) in terms of age, nationality, baseline eGFR, prevalence of major comorbidities (diabetes, hypertension, cardiovascular disease), or cause of AKI, although obesity was more common in the IHD group. Volume depletion was a common precipitating factor for AKI in both groups; however, renin-angiotensin-aldosterone system inhibitors (RAASi) were more significant precipitating factors for AKI in the IHD group.

The details of non-dialytic and dialytic management of patients are presented in Table 2. The intensive care unit (ICU) contributed 21% of the cases (which was greater but not significantly different in the CKRT group). The majority of patients received intravenous (IV) fluids prior to dialysis initiation, given the frequency of volume depletion, but there was no statistically significant difference in receiving IV fluids between the two groups; however, more patients in the IHD group received diuretics (62.9% vs. 43.1% for CKRT, p = 0.008). Dialysis catheter was inserted in the right internal jugular vein in 60.9% of patients (76.9% of CKRT patients, 54.3% of IHD patients, p = 0.005), followed by the femoral veins in 33% of patients (17.3% of CKRT patients and 39.4% of IHD, p = 0.004). Extracorporeal membrane oxygenation was used in only 2 patients in the CKRT group.

The details of 30-day kidney and patient outcomes have been listed in Table 3. Among the total cohort, 21.8% died, with no statistically significant difference between the two groups (16.1% for IHD vs. 24% for CKRT, p = 0.2). Partial recovery was greater in the IHD group (30.8% vs. 17.3%, p = 0.046), and complete recovery was greater in the CKRT group (33.1% vs. 13.5%, p = 0.009). The eGFR at 30 days was 53.2 mL/min, with no statistically significant difference between the two groups, although it was numerically greater for CKRT (44.6 for IHD vs. 56.1 for CKRT, p = 0.15).

Obesity, gastrointestinal/liver disorders, and the use of RAASi were statistically significant factors associated with increased mortality in our cohort after adjustment in Table 4. Table 5 shows that hypertension was associated with reduced chances of recovery. Although RAASi were significantly more common as precipitating factors for AKI in the IHD group, they did not impede the recovery.

Patients with severe AKI usually require KRT in the form of IHD, extended HD (slow low-efficiency dialysis), CKRT, or peritoneal dialysis [2]. Current evidence and KDIGO guidelines recommend the utilization of CKRT in hemodynamically unstable critically ill patients and in patients with increased intracranial pressure [2, 3]. Since there is no evidence demonstrating a difference in mortality between these modalities of KRT in the absence of clear indications supporting a specific modality, IHD can therefore be safely used in clinically stable patients [1]. However, the impact of the dialysis modality on kidney recovery has not yet been resolved [1, 2]. A meta-analysis of 110 studies conducted worldwide reported the AKI-related mortality rate in adults was 24%; however, the mortality rates for AKI stage 3 and AKI requiring dialysis in that metanalysis were 47.8% and 49.4%, respectively [5]. The mortality of adult non-ICU hospitalized patients with dialysis-requiring and non-dialysis-requiring AKI reported in the literature was 10–20% [6]. These results suggest that the mortality rate of 21.8% in dialysis-requiring AKI in our study was similar to what has been reported in the literature. Our multicenter observational study showed no difference in 30-day mortality between the CKRT and IHD groups. This is consistent with the findings reported in the Hemodiafe study [7], which reported similar survival rates between CVVH and IHD at 60 days, and in the CONVINT trial, which reported no difference in in-hospital mortality or 30-day mortality [8]. Similar conclusions were drawn from the Cochrane Collaboration meta-analysis [9], which included 1,550 patients from 15 RCTs and reported no difference in mortality between CKRT and IHD. In contrast, Mehta et al. [10] reported higher mortality with CKRT than with IHD. However, there were significant differences in the baseline characteristics between the two groups in that study, which may explain their results.

Our study also found a greater rate of partial recovery of kidney function in the IHD group than in the CKRT group, whereas complete recovery of kidney function was more frequently reported in the CKRT group than in the IHD group. The difference in outcomes between partial and complete kidney recovery has not been well described in the literature. Although episodes for interdialytic hypotension were not documented, in our view this was not a factor which influenced recovery, since all patients were hemodynamically stable and did not require inotropic support, and none of the patients on IHD required switching to CKRT. Although patient outcome “death” being a competing risk for kidney outcome “kidney recovery” is a possibility, we unfortunately could not perform competing risk analysis because we did not document the number of days between consultation and kidney recovery or the number of days between consultation and death. We only reported survival versus death within that 30-day period and kidney recovery versus lack of recovery within that same period. While some observational studies have shown the advantage of CKRT in kidney recovery [11‒13], randomized controlled trials have failed to demonstrate a significant difference in the recovery of kidney function between the two modalities [7, 14, 15]. Moreover, meta-analyses of randomized controlled trials have also failed to show any advantage in terms of kidney recovery with one modality over another [16, 17]. A meta-analysis of studies on CKRT for AKI revealed that increasing age was a risk factor for mortality [18]. However, in our study, we did not find age >65 years a risk factor for mortality. Preexisting CKD did not influence mortality in our cohort. In fact, one study reported the lowest mortality rates among patients with baseline eGFRs <30 mL/min/1.73 m² [19]. Although older age has been reported to be associated with a lower likelihood of recovery [20‒22], we found the opposite result in our cohort. Sex did not influence patient or kidney outcomes in our cohort; however, female gender has been reported to be associated with lower chances of kidney recovery [20, 23], and male sex was associated with higher mortality [22, 24]. Preexisting CKD (eGFR <60 mL/min) was not associated with a lack of recovery or higher mortality in our cohort; however, it has been reported to be associated with a lower chance of recovery [25, 26]. There were variables that were predominantly more frequent among CKRT patients versus IHD patients, such as nationality and cause of death (infections and cardiovascular); however, they were statistically non-significant, most likely due to the small cohort and low power. There was no difference in the incidence of COVID-19 as a cause of AKI between the two groups, and it did not influence the outcomes in our cohort. Although the incidence of AKI is high in patients hospitalized for COVID-19, and AKI is strongly associated with COVID-19 mortality [1], we could not find a study that compares CKRT outcomes to those of IHD in COVID-19 patients with AKI requiring KRT.

Our study was observational and prone to bias (selection bias, treatment bias). It also did not control for the timing of initiation, although the practice in all participating hospitals prefers late rather than early initiation of KRT for AKI. In addition, the practice in participating hospitals was to use hemodiafiltration for both CKRT and IHD; however, we did not attempt to calculate the dose of CKRT. Some CKRT cases were not a true continuous modality in duration but rather an extended HD modality using a continuous dialysis machine. Nevertheless, this multicenter study explored the possible benefits of CKRT beyond its effect on hemodynamic stability. It also differentiated between partial and complete recovery of kidney function.

AKI patients who received either IHD or CKRT were both clinically and hemodynamically stable, had similar basic characteristics, received similar lines of management, and had similar death rates at 30 days, which is similar to findings from randomized clinical trials that showed CKRT does not improve survival in patients with AKI requiring dialysis. However, the CKRT group had higher rates of complete recovery of kidney function than did the IHD group. This finding is in accordance with the findings of published observational studies that suggest that CKRT may not improve mortality but may improve rates of recovery of kidney function.

This study was approved by the Joint Committee on Medical and Scientific Research of the Ministry of Health and the Kuwait University. Authorization number: 2019/1242 was issued on February 13, 2020.

Verbal consent was obtained, or in the case of an invalid or minor patient, verbal consent was obtained from the next of kin or legal guardian.

The authors declare that they have no competing interests.

This study was not funded.

Research idea and study design: all authors. Data acquisition: Ali AlSahow, Omar Akandari, Gmal Nessim, Bassem Mashal, Ahmad Mazroue, Alaa Abdelmoteleb, Mohamed ElAbbadi, Ali Abdelzaher, Emad Abdallah, Mohamed Abdellatif, Ziad ElHusseini, and Ahmed Abdelrady. Data analysis and interpretation and supervision or mentorship: Ali AlSahow, Omar Alkandari, Yousif Bahbahani, Anas AlYousef, Bassam AlHelal, Heba AlRajab, Ahmed AlQallaf, Monther AlSharekh, and Abdulrahman AlKandari. Statistical analysis: Ali AlSahow and Omar Alkandari. Each author contributed important intellectual content during article drafting or revision and accepts accountability for the overall work by ensuring that questions pertaining to the accuracy or integrity of any portion of the work are appropriately investigated and resolved.

Deidentified participant data as well as the study protocol can be provided by the corresponding author after approval of a reasonable request, with the publication of this paper.