Comparison of Outcomes From Intermittent Hemodialysis and Continuous Renal Replacement Therapy With Systemic or Regional Anticoagulation in 128 Patients in a Polish Intensive Care Unit

Introduction

Acute kidney injury (AKI) remains one of the most common complications among the organ dysfunctions recognized in critically ill patients treated in intensive care units (ICUs). It is diagnosed based on an abrupt loss of kidney function, characterized by an increase in serum creatinine, blood urea nitrogen, and other waste products levels, along with a decrease in urine output [1]. The most up-to-date criteria for the diagnosis of AKI are based on the Kidney Disease Improving Global Outcome (KDIGO) guidelines [1,2]. Over 50% of ICU patients develop AKI, which often results in longer hospital stays and increased risks of chronic kidney disease (CKD) and mortality [1,3]. In fact, mortality in ICU patients diagnosed with AKI may be as high as 80%, while 3% to 19% of patients require long-term dialysis treatment after hospital discharge [1,4,5]. AKI can be managed conservatively or with the use of renal replacement therapies (RRT). AKI is not the sole indication for initiating RRT; other well-established indications include fluid overload unresponsive to conservative management, severe refractory hyperkalemia, refractory metabolic acidosis, severe or symptomatic uremia (eg, uremic encephalopathy or pericarditis), and intoxications with certain dialyzable toxins (eg, lithium, ethylene glycol) [6]. RRT are intended to replace the kidney in the situation of renal failure, with the modalities comprising intermittent and continuous procedures. Several studies were conducted in ICUs worldwide to assess the effect of introducing renal replacement therapy into ICUs on long-term outcomes such as mortality and long-term dialysis dependency of the patients discharged [7–9]. That study results failed to conclusively demonstrate that type of RRT affected the outcomes of mortality or long-term dialysis dependency [10,11], but the data did not include patients from Poland. Due to the differences in the severity of Polish ICU patients and another organization of the health care system Poland, as compared to those in other countries of Europe and worldwide, the data obtained cannot be extrapolated without prior analysis [12–15]. Health status and severity of disease were assessed using the Acute Physiology and Chronic Health Evaluation II (APACHE II) scoring system upon admission to the ICU of the Regional Specialist Hospital (RSH) in Olsztyn, Poland. Compared with reports from other countries, the severity of patients admitted to the ICU RSH in Olsztyn over the years 2003–2011 differed significantly [16–23]. The APACHE II scores obtained by the patients in Olsztyn were considerably higher, indicating a greater severity of illness. Alongside with the markedly higher APACHE II score on admission to Polish ICUs, there is a higher mortality rate. The available figures show, however, that Polish ICU mortality rate is lower than the estimates based on APACHE II score suggest [24]. The demographic differences mentioned above formed the basis for the analysis of mortality and long-term dialysis dependency in patients treated in a Polish ICU who received RRT. Additionally, to date, no single study has compared the outcomes of patients assigned to cohorts according to both the type of RRT and the type of anticoagulation used. Studies have been conducted comparing intermittent with continuous modalities, while separate investigations have examined only continuous techniques with distinguished regional and systemic anticoagulation. Therefore, we decided to compare 3 methods and types of anticoagulation within a single study in a Polish ICU. This retrospective study of 128 patients admitted to an ICU aimed to compare outcomes from 3 treatment approaches: intermittent hemodialysis (IHD), continuous renal replacement therapy (CRRT) with 2 types of heparin, and CRRT with citrate anticoagulation. Another aim was to determine which factors contribute to an increase or decrease in mortality rate in patients receiving a particular type of renal replacement technique and anticoagulation.

Material and Methods

ETHICS APPROVAL AND PATIENT CONSENT:

The study was approved by the Bioethics Committee at Warmia and Mazury Medical Chamber (Ref. No. WMIL-KB/14/2022). Informed consent was not required due to the retrospective design of the study, which was based solely on review of patients’ medical records without any direct contact with the patients, and the anonymized data analysis.

STUDY DESIGN AND DATA COLLECTION:

The study involved a retrospective review of the medical histories of patients treated in the Department of Anesthesiology and Intensive Care of The RSH in Olsztyn between January 2003 and December 2011. The analysis identified patients receiving any type of renal replacement therapy with different types of anticoagulation. Initially, 200 patients were considered. The predefined inclusion and exclusion criteria were applied. As a result, 128 patients were enrolled in the study. We retrospectively analyzed the medical histories of 128 patients, including laboratory test results and doctors and nurses notes. Subsequently, patients were classified into 3 groups according to the RRT and anticoagulation type used. The division was made to compare 3 treatment approaches and their influence on mortality and long-term dialysis dependency: cohort I consisted of patients who received intermittent hemodialysis (the IHD cohort), cohort II consisted of patients who received CRRT with heparin anticoagulation (unfractionated and low-molecular-weight heparin) (the HEP cohort), and cohort III consisted of patients who received CRRT with citrate anticoagulation (the CITR cohort).

DATA EVALUATION AND CLINICAL FOLLOW-UP:

After the division, the 3 cohorts were compared for homogeneity, including parameters such as age, sex, APACHE II score, the presence of arterial hypertension, type 1 or 2 diabetes, surgical intervention, and the time of initiating renal replacement therapy. The main outcomes were: ICU mortality, in-hospital mortality (every in-hospital death after ICU discharge), 90-day mortality measured from the day of research enrollment, long-term mortality (ie, at the end of the observation period [31/12/2020]), and long-term dialysis dependency after hospital discharge. Study enrollment was on the day of RRT introduction, and the observation period ended December 31, 2020. The duration of RRT use was not recorded, but the mean follow-up duration was 720.5 days. The shortest follow-up was observed in the CITR group (509.9 days), while the longest was in the IHD group (1002.2 days). Based on the medical records available, the date of death was recorded for each patient enrolled (during ICU stay or after discharge). The missing 90-day and long-term mortality data after hospital discharge were obtained from the Registry Office in Olsztyn, in accordance with the General Data Protection Regulation. Coded information on long-term dialysis therapy used after hospital discharge was obtained from the National Health Fund, the Department for Warmia and Mazury. The available data were analyzed for mortality, including early and long-term mortality, within and among the 3 cohorts.

STUDY INCLUSION CRITERIA:

The patients were included in the study if they received renal replacement therapy during ICU hospitalization and met at least 1 of the following criteria: the minimum serum creatinine concentration of ≥4 mg/dL, a 2-fold rise in the creatinine concentration comparing to the value on ICU admission, oliguria (defined as a urine output of less than 0.5 mL/kg/h for at least 6 hours), serum urea concentration ≥150 mg/dL, hyperkalemia ≥6 mmol/L, acidosis with blood pH ≤7.2, overhydration refractory to conservative treatment (diuretics administration), pulmonary edema, age >16. Patients were enrolled in the study if they met inclusion criteria at any time during ICU stay before implementing RRT.

As defined by the study inclusion criteria, 63 patients met only 1 criterion for initiation of RRT. The remaining patients fulfilled 2 or more criteria. AKI was diagnosed in 72 patients, of whom 52 had this as the sole indication for RRT. The remaining patients had additional indications for RRT.

STUDY EXCLUSION CRITERIA:

Patients were excluded if they met at least 1 of the following criteria: history of advanced chronic kidney disease, long-term dialysis dependency prior to ICU admission, history of kidney transplantation, ICU stay <24 hours, change from one renal replacement technique or anticoagulation to another during ICU stay.

TYPES OF RENAL REPLACEMENT THERAPY AND ANTICOAGULATION:

The patients included in the study underwent RRT in the ICU, or in the hospital dialysis unit if receiving IHD. IHD cohort patients received the intermittent technique, which causes significant changes in body fluid osmolality and thus affects circulatory stability. The HEP cohort and CITR cohort patients were treated with the continuous technique but with 2 different types of anticoagulation. Continuous techniques are less intensive, with a distinctly slower dialysate flow, which mitigates the adverse effects of the procedure on the circulatory system, while ensuring optimal blood purification [25]. During RRT, the patient’s blood is introduced into extracorporeal circuit, made entirely of synthetic material, which activates the coagulation system. This complication can be prevented by using appropriate anticoagulation. There are 2 types of anticoagulation used in RRT: systemic and regional. Unfractionated heparin (UFH) is one of the substances widely applied to obtain systemic anticoagulation. Low-molecular-weight heparin also can be used. Their use, however, is limited by their adverse effects, of which bleeding is the most common and serious complication [26,27]. Regional anticoagulation is usually achieved with sodium citrate administered by an infusion into the blood in the extracorporeal unit immediately after its collection from the patient. This results in the binding of ionized calcium (factor IV), which is an essential cofactor involved in the blood clotting process. Immediately before the blood from the circuit is returned to the patient, the calcium concentration is normalized with an infusion, to prevent clotting disorders or other complications of hypocalcemia. This virtually eliminates the risk of bleeding and makes this form of anticoagulation suitable for patients in whom heparin is contraindicated [28]. The decision regarding the type of the RRT and anticoagulation to be used was made by the doctor in charge and was guided by the availability of a particular technique, to select the optimal solution for each patient. There was only 1 type of RRT and anticoagulation used for each patient included in the study. Patients who received various types of either RRT or anticoagulation were excluded from the study. The procedures were performed with a double-lumen dialysis catheter placed in the central vein. IHD in the dialysis center was conducted using a range of machines and filters: Elisio 150H, 170H, 190H, 210H by Nipro (Belgium); Polyflux 150L, 170L, 210L, 170H, 190H by Baxter (Gambro Dialysatoren GmbH, Germany); and the FX classix dialyzer (Fresenius Medical Care, Germany). CRRT was conducted in the ICU using multiFiltrate dialysis machines by Fresenius (Fresenius Medical Care, Germany). When CRRT was performed with heparin anticoagulation, either unfractionated or low-molecular-weight heparin was used. We used filters with surface area of 1.8 m2 (Ultraflux AV 1000S, Fresenius Polysulfone membrane, Fresenius Medical Care, Germany) and 1.4 m2 (Ultraflux AV 600S, Fresenius Polysulfone membrane, Fresenius Medical Care, Germany). The procedures conducted with this type of anticoagulation were: CVVHDF with pre- and post-dilution, and HV-CVVH and CVVH with pre- and post-dilution. MultiBic Fresenius bags with adjusted concentration of potassium were used. CRRT with citrate anticoagulation was performed with filters with a surface area of 1.8 m2 (Ultraflux AV 1000S, Fresenius Polysulfone membrane, Fresenius Medical Care, Germany). CVVHD was conducted with citrate anticoagulation. CiCa Dialysate Plus Fresenius bags with adjusted concentration of potassium and no calcium were used.

STATISTICAL ANALYSIS:

Statistical analysis was conducted using TIBCO software (2017), Statistica data analysis software version 13 (StatSoft, Cracow, Poland; http://statistica.io), and MS Excel with Office 360.

Categorical measurements are presented as numbers and percentages. Quantitative results are presented as mean values with standard deviation; additionally, median and minimal and maximal values were included. The normality of data was assessed using the Shapiro-Wilk test, separately for the results obtained for each of the groups compared.

The comparison of the results for the 2 cohorts was conducted using the t test when the distribution of variables was close to normal, while the Mann-Whitney U test was used when the variables did not have normal distribution. For 3 and more independent groups, the comparisons were made using ANOVA or the Kruskal-Wallis test, when the variables were normally distributed or not, respectively. Mortality data in 3 patient cohorts were obtained using Pearson’s chi-squared test. Because of the limited sample size, the statistical analysis of data from patients discharged from the hospital was performed using the chi-square test with Yates’ continuity correction.

Odds ratios (univariable and multivariable) for death by 31/12/2020 as well as 90-day mortality and in-hospital death were estimated using logistic regression analysis. Univariate variables with P<0.1 were included in the multivariable model. A P value of <0.05 was considered to be statistically significant.

Results

PATIENTS CHARACTERISTICS INCLUDING SEX AND AGE:

The analysis of sex distribution was conducted for each of the 3 cohorts. Among all patients, 75% were men and 25% women. The cohorts were not homogenous regarding the variable under analysis. Most patients in each cohort were men; in the IHD cohort there were more men than in the HEP or CITR cohorts; this finding was statistically significant (P=0.024). Additionally, patients discharged from the hospital, both those requiring and those not requiring dialysis, were analyzed for sex distribution. In both groups, most patients were men. No statistically significant difference was found between the groups (P=0.925).

Patients aged 60–69 years old were the largest group, accounting for 32.03% of all patients. The youngest patient was 16 years old and the oldest was 89. The mean age was 59.53 years and the median age was 61 years. The age distribution in the 3 cohorts was not significantly different (P=0.099). Statistical analysis of patient age was conducted separately for patients who required chronic dialysis treatment and those who did not receive dialysis after discharge. The mean age in both groups was comparable (52 vs 53), with no statistically significant difference found (P=0.620).

COMORBIDITIES AND CLINICAL STATUS:

Of all the patients studied, 29.69% had a history of arterial hypertension, while 14% had a history of type 1 or 2 diabetes, and 69.5% underwent a surgical procedure during their hospital stay. Most patients (92%) received pressor amines while in the ICU. The results are presented in Table 1. Patient data were also analyzed for the association between a chronic disease and long-term dialysis after discharge (Table 2).

ASSESSMENT OF THE SEVERITY OF PATIENT STATUS USING APACHE II SCORING SYSTEM:

The severity of patient clinical status at ICU admission was assessed using the APACHE II scoring system. Its mean value in patients undergoing dialysis during hospitalization was 32.5, while the median was 33.5. When divided into groups according to their APACHE II score, patients with scores over 30 were the largest group (61.7%), with 35–40 points being the most prevalent.

The highest mean score (35.4) was achieved by the patients from the IHD cohort, significantly (P=0.039) higher than in the HEP and CITR cohorts (30.5 and 31.8, respectively). Post hoc (NIR) analysis comparing the IHD vs HEP cohorts showed a significant difference in APACHE II scores (P=0.014). Another comparison involved patients from the IHD vs those from the HEP and CITR cohorts combined, showing a significantly higher mean APACHE II score in IHD the group (P=0.009). Patients who did not receive long-term dialysis after hospital discharge had a mean APACHE II score of 27, versus 31.6 in the other group, with no statistically significant difference (P=0.246).

TIME OF INITIATING RENAL REPLACEMENT THERAPY AND ICU LENGTH OF STAY:

The mean time of initiating the RRT was day 4 of ICU stay. RRT was usually started within the first 5 days of ICU stay, with a mean duration of 19.7 days and a median of 12.5 days. For each of the 3 cohorts, the time of initiating RRT was analyzed. In the IHD cohort, the mean time of RRT initiation was day 5, in the HEP cohort it was day 6, and in the CITR cohort it was day 3, with no statistically significant difference found between the cohorts (P=0.273). Likewise, no statistically significant difference was found among cohorts in length of ICU stay. The mean duration of ICU stay in the IHD cohort was 19.2 days, in the HEP cohort it was 18.5 days, and in the CITR cohort it was 21.1 days (P=0.957). The mean values for the day of starting RRT were determined for patients not requiring dialysis after hospital discharge and for patients who received dialyses after hospital discharge, showing that RRT was initiated on day 7 (mean value) in the former, and on day 2 in the latter group. No statistically significant difference was found between the groups (P=0.095). The mean ICU stay in the first group was 25 days, as compared to 33 in the other group. The difference between the groups was not statistically significant (P= 0.303).

THE LENGTH OF OBSERVATION PERIOD AND MORTALITY:

The observation of patients began with initiating RRT and finished on 31/12/2020. The ICU mortality rate was 69.5%, while in-hospital mortality, including ICU deaths and in-hospital deaths after ICU discharge, was 76.56%. The 90-day mortality rate, calculated from the time of patient enrollment into the study was 75.78%, while the long-term mortality rate of all patients was 88.28%. The mortality data for all 3 cohorts were subject to analysis and the results are shown in Table 3. No statistically significant difference was found among the 3 cohorts in long-term mortality at the end of the observation period (31/12/2020) (P=0.958). Additionally, there was an analysis performed comparing the mortality in IHD cohort and in HEP and CITR cohorts combined, aiming to identify differences in mortality rates between the intermittent and continuous techniques. This analysis of IHD cohort versus the HEP+CITR cohorts failed to demonstrate statistically significant differences in mortality rates (ICU death P=0.839, in-hospital death P=0.785, 90-day mortality P=0.680).

FACTORS AFFECTING LONG-TERM MORTALITY BY 31/12/2020:

The available data were analyzed to investigate the effect of particular factors on long-term mortality in the 3 cohorts. Tables 4 and 5 show the results of univariable and multivariable logistic regression analyses conducted for the factors influencing in-hospital, 90-day, and long-term mortality.

LONG-TERM DIALYSIS DEPENDENCY AFTER HOSPITAL DISCHARGE:

Patients data were subject to statistical analysis within and between the 3 cohorts. Patients in the IHD cohort were significantly more likely to be dependent on long-term dialysis than in the 2 other cohorts combined) (P=0.001). We also analyzed the data of patients discharged from the hospital in the 3 cohorts. Owing to the small sample size, the chi-squared test with Yates correction was used. The HEP cohort patients were less likely to be dependent on long-term dialysis in comparison to other patients, but the difference was not statistically significant (P=0.064).

Another statistical analysis involving the data of patients discharged from the hospital was performed using the chi-squared test with Yates modification, due to the small sample size as above, showing the CITR cohort patients were less likely to receive long-term dialysis than those in the other 2 cohorts, but the difference was not significant (P=0.440).

Discussion

The analysis of the type of RRT modality and the anticoagulation strategy used did not demonstrate an impact on mortality, but patients treated with intermittent hemodialysis (IHD) were more likely to become dialysis-dependent after hospital discharge. Similar results with respect to mortality were obtained by Lins et al, who found no effect of intermittent and continuous renal replacement modalities on mortality or long-term dialysis dependency outcomes [9]. Similar results were obtained by Vinsonneau et al [7].

A 2008 study by Rauf et al compared IHD versus CRRT, with both regional and systemic anticoagulation used. The observation period used to determine patient mortality was 12 months. A statistically significant difference was found for in-hospital mortality rates between patients receiving CRRT versus IHD (48% vs 30%, respectively). No such difference was found during 12-month observation. Likewise, no difference was found between the cohorts in renal function recovery. The differences concerning in-hospital mortality can be attributed to the severity of patients’ condition. The patients who received CRRT were younger, but with higher APACHE II scores, which was 32.3 versus 27.9 for IHD [8]. The retrospective study by Bonnassieux et al referred to above involved patients hospitalized in 291 ICUs in France from 2010 to 2013. Its focus was the impact of intermittent and continuous RRT on mortality and renal recovery. For in-hospital mortality, similar results to those in the previously mentioned study were obtained. The difference between IHD and CRRT cohorts was 45% versus 61%, which was statistically significant. The patients subject to CRRT were also in a more severe clinical status, as assessed using SAPS II [29]. The causes of a higher in-hospital mortality in the 2 studies cited above seem to be similar and can probably be attributed to a more severe status of the patients subject to CRRT in comparison to those receiving IHD. Pointing to a particular type of RRT as the sole cause of an increased mortality rate would probably be overinterpretation, as there are numerous contributory factors involved. We found that the mean severity score for the patients receiving a particular type of RRT in our study stands in sharp contrast with the corresponding values presented in other studies. The patients undergoing IHD were in the most severe condition, with the mean APACHE II score of 35.4, while the lowest score was noted in the HEP cohort. The difference between the 2 cohorts was statistically significant. However, it had no effect on the patient mortality outcomes in particular cohorts, even at an early stage of their ICU stay, unlike in the 2 studies mentioned above. To proceed with the analysis of our study results, one might expect to see a higher ICU or in-hospital mortality in patients in the IHD cohort because of the highest APACHE II score, but the analysis brought no such finding. Therefore, the severity of patients’ condition cannot be the only explanation of the higher in-hospital mortality noted in CRRT patients in the studies discussed above. Some interesting outcomes are also presented in the study by Mehta et al. Similar to the previously cited publications, they found increased ICU and in-hospital mortality rates with CRRT versus IHD. However, better kidney recovery was seen in the CRRT cohort. Some aspects of patient severity seen in the 2 cohorts deserve attention. Patients who received CRRT were in a more severe state, as indicated by APACHE II and III scores. Thus, the lower ICU mortality rate in IHD patients may be due to their less severe clinical state, not the type of RRT used, which is also a suggestion made by the authors themselves [30].

Most studies comparing IHD and CRRT regarding their effect on mortality and long-term dialysis dependency failed to demonstrate the differences between these modalities. Some studies, particularly older ones, suggested that a greater long-term dialysis dependency was more likely with IHD versus CRRT, which is in line with the results obtained in our study. We found a statistically significant result of more frequent long-term dialysis dependency in the IHD cohort. The literature suggests that CRRT leads to better patient outcomes due to its greater hemodynamic stability. This, in consequence, would translate into maintaining optimal organ perfusion, here particularly in the kidneys. This is important because hypotension is a risk factor for AKI [31]. Hence, the recommendations issued by, among others, Kidney Disease: Improving Global Outcomes (KIDIGO), indicate CRRT as a technique dedicated to hemodynamically unstable patients [2]. In turn, the French Intensive Care Society (SRLF) recommends a balanced use of both techniques [32].

Liang et al investigated the impact of the initial RRT technique applied after the diagnosis of AKI on renal function recovery. Subsequent types of RRT used during hospitalization were not taken into consideration. At 90 days following the commencement of the study, better kidney recovery was demonstrated for the CRRT than for the IHD cohort. There was no such difference found in the patient population 365 days after the study started. However, no mechanism responsible for this result was identified [33]. The study by Bonnassieux et al comparing CRRT versus IHD focused on renal recovery at hospital discharge. According to the definition used, kidney function recovery was declared if there was no RRT used 3 days prior to hospital discharge, which is a very short observation period. The superiority of CRRT over IHD in renal function recovery was demonstrated (87.2% vs 84.3%) [29]. Similar results were obtained by Wald et al and Huerfano et al [34,35]. Comparative studies of anticoagulation strategies during renal replacement therapy have thus far been confined to continuous modalities, with primary emphasis on filter lifespan [36]. Consistent with our findings, the type of anticoagulation used during CRRT has not been shown to influence mortality [37].

In our study, the criterion qualifying patients for a particular RRT was different than in most previous studies. With IHD already in use, and continuous therapy yet to be introduced, the doctor in charge had no alternative options regarding the technique. When CRRT appeared as a treatment option, it was available only with heparin anticoagulation. Another stage was the introduction of citrate anticoagulation into CRRT. It was only then that responsible physicians could choose from all the types of techniques, but it was CRRT with citrate anticoagulation that, following the existing recommendations, became the preferred technique. In our study, it is important that IHD patients had higher APACHE II scores, which means they were qualified for RRT later, when the disease was more advanced. This was not always the question of a therapeutic decision, but of limited access to the technique at that time.

Our study’s retrospective nature is a limitation. Although laboratory test records are reliable, some medical history details, such as chronic diseases or comorbidities, may be missing. This is of importance, as it was the medical history of the disease, with entries made by many doctors, that provided such information. Another limitation is the size of the cohorts. It is likely that with larger cohorts, the results of statistical analysis would be of greater power. However, extending the study period to include the preceding years would not mean having larger cohorts. This is because before 2002, renal replacement therapy was rarely used at our center, and the medical records for this period are scarce. Alternatively, extending the study to include patients treated in subsequent years would only result in an increase in CITR cohort size. A limitation to the study may also be its relatively long duration. This might have been associated with improvement in patient care and monitoring standards over the years, which in turn could have had an impact on the prognosis. Another limitation is the varying availability of RRT over the study period. The choice of a particular RRT to be instituted was made by the doctor in charge, whose decision was guided by its availability and clinical indications. However, the only option originally available at our center was IHD. It was only some years later that CRRT with heparin anticoagulation, and ultimately CRRT with citrate anticoagulation, became available. This limitation, however, can also be viewed as beneficial. The patients who received IHD were in the most severe state, and thus had the highest APACHE II scores. Had citrate anticoagulation been available, this technique would probably have been used, which would have reduced the ability to investigate the effect of IHD on prognosis in extremely severe patients. Finally, this was a single-center study. With more centers involved, the results obtained could have been more reliable.

Conclusions

We did not find a significant association between mortality and the type of renal replacement therapy (RRT) or the anticoagulation strategy. However, patients treated with intermittent hemodialysis (IHD) were more likely to be dependent on long-term dialysis. Despite numerous analyses conducted over the years aiming to resolve the question of the optimal RRT and anticoagulation in terms of effectiveness and safety, there are still some doubts remaining. It seems that developing and implementing guidelines would be a solution which would increase the safety of using extracorporeal blood purification techniques in precisely defined clinical scenarios. The implementation of such guidelines indicating an optimal therapy for a particular patient and specifying the details of a patient’s clinical status as well as laboratory test findings, may have a major impact on the prognosis. Here, it is not the type of RRT used, but the way it is consistently instituted, possibly with guidance in the form of protocols, that may influence mortality and long-term dialysis dependency. Such protocols, developed and implemented as mentioned above, should then be used to design and conduct a prospective clinical trial.