REPLACEMENT THERAPY IMPROVES SURVIVAL IN PATIENTS WITH ACUTE KIDNEY INJURY

Introduction/Objective: Defining renal replacement therapy (RRT) initiation in critically ill patients with acute kidney injury (AKI) has become an imperative for nephrologists and intensivists. The primary objective of this study was 28-day patient survival and the secondary objective was renal function recovery. Methods: We performed a single-center retrospective study of 385 surgical and non-surgical patients with AKI and episode of AKI in chronic kidney disease who were admitted to the Emergency Center between 2014 and 2017 and received RRT. Patients with Kidney Disease Improving Global Outcomes stage 2 AKI and/or volume overload were assigned to the „early“ group with RRT start within 24h of the diagnosis; patients with poor response to conservative treatment or evidence of clinical complications associated with AKI were assigned to the „late” RRT group. Results: Based on the retrospective analysis we found that 241 (62.6%) patients received „early“ RRT within 24h. Patients in the „early” RRT group had significantly higher survival compared to the „late“ RRT group (63.9% vs. 36.1%; p=0.001). The „early” RRT group had more patients with renal function recovery (56.8%), but without statistical significance (p=0.514). The patients who started RRT within 24 hours with the SOFA score of 1-3 were twice likely to recover renal function in relation to the patients with the SOFA score of 4 or higher (OR=2,01; 95%CI (1,37-2,95); p<0,001), while septic patients a 62% lower chance of renal function recovery in relation to non-septic patients (OR=0,38; 95%CI (0,18-0,82); p=0,013). In the „late” RRT group it was found that non-diabetic patients had 3.8 times greater chance for renal function recovery compared to diabetic patients (OR=3.53; 95% CI [1.27-9.83]; p=0.016). Conclusions: Patients with early initiation of RRT had significantly improved 28-day survival.


Introduction:
In the past decade, acute kidney injury (AKI) has become a well-recognized global occurrence that affects developed and developing countries alike, with initiatives like Saving Young Lives and the International Society for Nephrology's 0by25 Initiative aiming at reducing the economic, social and healthcare burden imposed by AKI (1). AKI frequently occurs in critically ill patients and severe AKI is associated with hospital mortality in 60% of the cases (2). Those that survive the initially high mortality rate associated with dialysis-requiring AKI, mostly become independent of renal replacement therapy (RRT) within a year, but some of them do go on to develop chronic kidney disease and even progress to end-stage renal disease (3). The fundamental principle in the treatment of AKI is to correct the underlying cause, but besides hemodynamic resuscitation and removal of nephrotoxins, we lack any established pharmacotherapy. Although drugs are tested for prevention and/or treatment of AKI, RRT appears to be our only efficacious option at the time. Thus, management of AKI is largely limited to preventing further deterioration and loss of function with the use of temporizing actions in severe cases until RRT is established (4). Interaction of RRT and the outcome of RRT, along with mechanical ventilation, vasoactive therapy and nutritional support, is one of the defined life-sustaining technologies in the current treatment of the critically ill. A recent trend suggests an increasing use of RRT in critically ill patients with AKI (5). Despite the research and growing clinical experience in dialysis, the optimal time to start RRT in a critical disease complicated with AKI is unclear (2). Heterogeneity in operational definitions of "time", "threshold" or "criteria" in individual observational data (often with variable designs and methodological qualities) have probably interfered with clear conclusions that could guide clinical practice on this issue (6). It is unclear whether a preventive/early strategy of the initiation of RRT in order to avoid complications associated with AKI leads to better patient outcomes and the use of health services, or a more conservative strategy in which RRT is started as a response to the development of complications provides better results. Neither the standard clinical parameters, nor the new biomarkers that have been introduced to clarify the definitive ideal time more precisely, nor the clinical picture have optimized patient outcomes (2). The primary objective of this study was 28-day patient survival and the secondary objective was renal function recovery.

Materials and Methods:
We performed a single-center retrospective study of 385 surgical and non- times baseline and urine output <0.5 mL/kg/h for 12 h) and/or volume overload were assigned to the early group with RRT start within 24h of the diagnosis; patients with poor response to conservative treatment or evidence of clinical complications associated with AKI were assigned to the late RRT group.
Although the condition of certain patients called for the start of RRT in the first 12 hours, they were denied this request at that time due to the organization of the team for starting RRT at weekends or at night, the unavailability of the apparatus and/or difficulty in placing the dialysis catheter. Other reasons for postponement were surgical interventions or radiological tests that were to be performed before the start of the RRT. Some patients started treatment with intermittent dialysis at the time of hemodynamic stability or the unavailability of the apparatus since "more severe" patients and/or the ones who were occasionally dialyzed with it had the need for it. Patients with an immediate RRT indication, at least one of the following conditions from the beginning were excluded: laboratory analysis at the admission urea>50 mmol/l, K>6.5 mmol/l, pH<7.15 in the context of either pure metabolic acidosis or mixed acidosis despite medical treatment; acute pulmonary edema due to fluid overload causing severe hypoxemia, as well as patients treated with conservative therapy. We analyzed: demographic data, comorbidities, laboratory and clinical data in confirmed acute kidney injury (urea, creatinine, C-reactive protein, procalcitonin, oliguria/anuria) and before continuous renal replacement therapy (CRRT) initiation (urea, creatinine, 24h diuresis (ml)); use of vasopressor therapy and mechanical ventilation; hospital length of stay The Chi-Square Test was used to compare the differences between the groups, and Cox regression analysis was used to estimate survival and recovery of renal function.

Results:
Based on the retrospective analysis we found that 241 (62,6%) patients (male 65.4%) mean age 60.6 received early RRT within 24h, and 144 (37,4%) patients (male 70.8%) mean age 63.5 received late RRT after 24h. All studied comorbidities were more prevalent in early RRT, and cardiovascular diseases were the most prevalent comorbidity in both groups of patients. Patients in the early RRT group had a higher rate of sepsis and less frequent use of mechanical ventilation and vasopressor therapy compared to the late RRT group (54,7% vs 41,7%; 58,3% vs 70,8%; 56,7% vs 75,7%). The presence of adsorptive membrane/adsorbers according to the type and the number of procedures in both groups was similar. Over 50% in both groups had a SOFA score ≥4.
Median diuresis (ml) before RRT was smaller in early group compared to the late group (150 vs 400); median urea (mmol) and creatinine (µmol) were Authors of two previously conducted meta-analyses, which included a total of 38 studies predominantly retrospective and of different quality, came to similar conclusions as ours, reporting a significant improvement in the 28-day mortality with early RRT (11,12). Other three studies have been conducted.
Two of them (the multicenter randomized controlled trial (RCT) and the retrospective study)) included septic patients and used different criteria AKI classification for starting RRT in relation to our study design, but at the same RRT initiation time, showed a significant reduction in mortality in early RRT, while the third multicenter RCT conducted three years later found that there was no difference in mortality between the early and the late onset of RRT in 224 postoperative cardiosurgical patients (13,14,15).
In fact, since 2012 the majority of the published studies have not supported the benefit of early RRT in critically ill patients. Conducting a meta-analysis of 36 predominantly retrospective studies, it was established that early initiation of RRT in critical patients did not improve survival for 28 or 30 days, nor did it reduce the length of stay in the ICU or the overall length of hospitalization. In the abovementioned studies, biochemical markers according to the Risk, Injury, Failure, Loss of kidney function, and End-stage kidney disease-RIFLE classification, Acute Kidney Injury Network-AKIN stages and time-based cutoffs were used (e.g. within the defined admission time or the development of the biochemical "start time") for defining the early start of RRT. However, the "early RRT" according to one author was the "late RRT" according to another author, which made it difficult to interpret the results. The late onset included classical indications of RRT that did not respond to conservative treatment (16). A 4 RCT meta-analysis showed similar results, except that a higher risk of catheter-related infections was reported at the early onset of RRT (17,18). A meta-analysis with six RCT that provided similar conclusions was conducted the same year, after which 4 additional RCT and 41 observational studies were included (a total of 51 studies) whose results showed that the early RRT was associated with a reduced risk of all causes of mortality, although the results were taken with caution given the variety of design studies.
In our study of the early onset of RRT, we established a better 28-day renal function recovery, although the differences between the groups were not statistically significant. The ELAIN study indicated significant benefit since the early onset of RRT in renal function recovery (53.6% vs. 38.7%, p=0.02), and the last meta-analysis also showed significant renal function recovery in 14 studies with 2570 patients (10). When we established a better renal function recovery, we tried to determine which predictors could affect this outcome. By applying multivariate Cox regression analysis in the total sample as well as in patients who started RRT early, SOFA score and sepsis stood out as significant predictors or renal function recovery. Namely, the patients with 1-3 SOFA score who started RRT early, were two times more likely to recover from renal function in comparison to the patients with SOFA score 4 and above, while septic patients had a 62% lower chance to recover their renal function compared to non-septic patients. Contrary to our results, in a large retrospective study of the critically ill with AKI, older age, heart conditions and admission to ICU were significantly linked to a lower rate of renal function recovery 60 -120 days after discharging from ICU. However, in another retrospective study, Pistolesi et al, concluded that older age, oliguria, sepsis and a higher SOFA score in 264 cardio-surgical patients with severe AKI within the first week of CRRT start were independent prognostic indicators for non-recovery of renal function. However, in the late RRT group it was found that the patients without Also, it should be noted that in studies similar to ours, the predictors of primary and secondary outcomes were less examined, because the main focus was on finding criteria for early/late start of RRT and predicting the outcomes. There are some limitations associated with our study. We performed a singlecenter retrospective study. All our patients received RRT, there was no control group (due to the limited data availability), nor the possibility that delaying RRT could provide time for renal function recovery. Unlike other studies, we did not obtain the full data on early RRT-related complications, nor the data on long-term outcomes (17,18,24).

Conclusion:
The patients who had early started RRT had significantly better 28-day survival. Further prospective research of the primary and secondary outcome predictors is necessary.
contributed to the design of the manuscript, interpreted the data, revised the article, provided intellectual content to the work and gave final approval for the manuscript to be published. T.A. provided contributions to conception of the manuscript; she interpreted the data, revised the article, provided intellectual content to the work and gave final approval for the manuscript to be published. V.S. was responsible for conception of the manuscript and she interpreted the data, revised the article, provided intellectual content to the work and gave final approval for the manuscript to be published. D.C. interpreted the data, revised the article, provided intellectual content to the work and gave final approval for the manuscript to be published. I.U. contributed to conception of the manuscript; she interpreted the data, revised the article, provided intellectual content to the work and gave final approval for the manuscript to be published.

Funding:
The autors received no financial support for the research, authorship and/or publication of this article.
Conflict of interest statement: None declared. The results presented in this paper have not been published previously in whole or part, except in abstract format.

Adittion:
Demographic and clinical data Patient survival depending on the time of RRT initiation Figure 1.
Recovery of renal function depending on the time of RRT initiation Figure 2. Legend: RFR-renal function recovery; RRT-renal replacement therapy;