Acute kidney injury (AKI) is a clinical disorder that has an extremely high mortality rate, and the cases that are induced by septicemia or multiorgan failure syndrome show poor outcomes. Although the mortality rate of AKI has been reduced since the use of renal replacement therapy (RRT) for patients with AKI, its mortality rate is still reported to be as high as 40~70%. Intermittent hemodialysis (IDH) has the disadvantage in that it cannot successfully treat AKI due to the hypotension that occurs during dialysis. Since continuous renal replacement therapy (CRRT) has the advantage of removing the solute slowly and continuously, it has been primarily used for treating AKI patients with an unstable hemodynamic status or the AKI is associated with multiorgan failure syndrome1-4). However, although CRRT has many advantages, previous studies have reported that CRRT has not reduced the mortality rate any more than intermittent RRT5, 6). Although AKI itself may contribute to its high mortality rate, there is the possibility that RRT can further increase the mortality rate.

Thus, the present study was conducted to determine the important factors that are related to the mortality rate of patients who undergo CRRT.

A total of 125 patients received RRT. The mean age of the patients was 61.4 (14.3) years. There were 55 males (44.0%) and 70 females (56.0%). Medical diseases (n=91, 72.8%) were the most common cause of AKI, followed by surgical diseases (n=29, 23.2%) and trauma (n=5, 4.0%). Oliguria occurred in 84 (67.2%) patients and septicemia occurred in 48 (38.4%) patients who underwent RRT. The mean SAPS II score was 60.5±16.3 (Table 1).

The most common indication for CRRT was hemodynamic instability (n=88, 70.4%), followed by cardiogenic shock (n=23, 18.4%) and acute respiratory failure (n=13, 10.4%). The mean time lapse between admission and the start of RRT was 7.3±11.1 days and the mean duration of RRT was 4.8±4.5 days. The mean amount of ultrafiltrate was 956.8±180.7 mL/h and heparin was used in 29 (23.2%) patients. Complications occurred in 33 (23.2%) patients after RRT. The use of filters 2 or more times per day (n=26, 22.0%) was the most common complication, followed by bleeding (n=3, 2.5%) and catheter-related infections (n=3, 2.5%).

Of the patients who underwent CRRT, 50 (40%) survived. Table 2 shows the clinical characteristics, laboratory data and changes after RRT between the survivor (n=50) and nonsurvivor (n=75) groups. There was no significant differences in the causes of AKI between the 2 groups (p=0.875), including septicemia, hepatic failure, congestive heart failure and acute respiratory distress syndrome. The mean SAPS II score was higher for the nonsurvivor group (62.2±16.2) than for the survivor group (58.0±16.4), but the difference was statistically insignificant. Septicemia occurred in 19 (38%) patients of the survivor group and in 29 (38.7%) patients of the nonsurvivor group, but the difference was statistically insignificant. The mean ABP before the start of RRT was significantly lower for the nonsurvivor group than for the survivor group (73.0±17.4 mm Hg vs. 82.9±19.0 mm Hg, respectively; p<0.05). The mean ABP measured the day after RRT was significantly lower for the nonsurvivor group than for the survivor group (69.2±24.6 mm Hg vs 86.7±23.7 mm Hg, respectively; p<0.05). The serum albumin level was significantly lower for the nonsurvivor group than for the survivor group (2.7±0.6 g/dL vs. 3.0±0.5 g/dL, respectively; p<0.05). The serum C-reactive protein level was higher for the nonsurvivor group (14.6±9.7 mg/dL) than for the survivor group (12.1±10.8 mg/dL), but the difference was statistically insignificant.

There were no significant differences between the 2 groups for the blood flow rate, the amount of ultrafiltrate, the duration of dialysis and the use of heparin. The mean time lapse between admission and the start of RRT was 8.4±12.8 days for the nonsurvivor group and 5.8±7.7 days for the survivor group, but the difference was statistically insignificant. There was no significant difference for the changes of the serum levels of creatinine and urea after RRT between the 2 groups, but the serum bicarbonate level was significantly lower for the nonsurvivor group than for the survivor group (16.4±5.4 mmol/L vs. 21.4±3.5 mmol/L, respectively; p<0.05).

On the multivariate analysis of the contributory factors for the mortality rate, the serum albumin level and the mean ABP values measured before and the day after RRT and the serum bicarbonate level measured the day after RRT were significantly lower. However, on logistic regression analysis, the mean ABP and the serum bicarbonate level were the important factors for the mortality rate (Table 3).

For the stage 3 subgroup, the mortality rate, the SAPS II scores and the serum levels of urea and creatinine were significantly higher for the stage 3 subgroup, and the serum levels of albumin and bicarbonate tended to be lower (Table 4).

The mortality rate was 90.9% for the stage 1 subgroup, 44.8% for the stage 2 subgroup and 29.5% for the stage 3 subgroup, and so the mortality rate was lowest for the stage 3 subgroup (Figure).

Although there is still no concordance between studies regarding the definition, treatment modalities and treatment outcomes of AKI, its mortality rate has been reduced because of the successful treatment of acute renal failure via dialysis9). Since Kramer et al.10) first described continuous arteriovenous hemofiltration (CAVH) in 1977, CRRT such as continuous venovenous hemodiafiltration (CVVHDF) has been developed along with countermeasures for correcting its shortcomings and CRRT has been widely used in clinical settings. CRRT can lead to a normal equilibrium of electrolytes and a normal acid base balance, as well as hemodynamic stability, and it may be effective for the treatment of patients with brain edema. However, despite these therapeutic effects, it is controversial whether CRRT is better than intermittent hemodialysis for intensive care11).

There have been numerous studies on the indications and adequate timing of RRT and the clinical characteristics of AKI. Although there is controversy concerning the adequate timing of RRT, it is generally recommended that RRT should be initiated before the complications of AKI occur12, 13). In the present study, RRT was performed on the patients who were admitted to the intensive care unit due to AKI associated with hemodynamic instability, brain edema, a high catabolic state and a severe excess of body fluids, and their mortality rate was 61.0%, which was similar to the 62.0% mortality rate reported by a previous study. The previous study also reported that the mortality rate of CRRT increased for the cases where the duration of this treatment was long, oliguria was present and the renal functions were reduced15). There were no significant differences in the mortality rate in terms of the blood flow rate, the amount of ultrafiltrate, the duration of RRT and the use of heparin in the present study. However, since the mean amount of ultrafiltrate was relatively low (18.7 mL/kg/min), it is thought that the survival rate should have been compared with that of the patients with a higher amount of ultrafiltrate (35 mL/kg/min). In our study, there was no significant association between the mortality rate and oliguria, the RRT outcomes or renal functions. The mortality rate tended to become high for the cases where the serum C-reactive protein was increased, the mean ABP was decreased before the start of dialysis and the serum bicarbonate level was decreased the day after dialysis. The mortality rate was significantly higher for the cases where the mean ABP and the serum bicarbonate level decreased the day after RRT. It has been reported that there is a high mortality rate when the mean ABP before RRT is low and the blood pressure is low 2 hours after RRT14). In the present study, since the patients who were undergoing RRT were hemodynamically unstable, they showed remarkable changes in blood pressure on the day of RRT; however, their blood pressure was stabilized the day after RRT. Although the blood pressures before RRT, 2 hours after RRT and the day after RRT were related to the mortality rate, only the blood pressure measured the day after RRT showed a significant correlation with the mortality rate. The blood pressure and the serum bicarbonate level were significantly correlated with the mortality rate, which indicates that rapid restoration of hemodynamic stability and recovery from metabolic acidosis may be important factors for the mortality rate for patients with AKI. Wald et al15) has reported that CRRT results in a low mortality rate through the restoration of hemodynamic stability and normal kidney functioning. The results of the present study indicate that the mortality rate is low for the cases where the mean ABP is high. This may be explained by the restoration of hemodynamic stability, which is one of the therapeutic effects of CRRT.

Since the prognosis of AKI is affected by numerous variables, its prediction and making precise comparisons of the accuracy between the prognostic factors are difficult. Various studies have compared prognostic factors with using the risk, injury, failure loss and end kidney disease (RIFLE) classification, the acute physiology and chronic health evaluation (APAHE) II scoring system, the SAPS II scoring system, the mortality prediction model (MPM), the sepsis-related organ assessment (SOFA) and the logistic organ dysfunction system (LODS)16, 17).

In particular, since a review of the literature does not yield a consensus about the definition and classification of AKI, there remain some difficulties for conducting research and improving the prognosis. In the present study, the usefulness of the SAPS II scoring system and the classification proposed in 2007 by the Acute Dialysis Quality Initiative (ADQI) group8) was evaluated for the patients who underwent CRRT. The stage 3 AKI subgroup showed the highest mortality rate (70.5%) and the highest SAPS II score (62.7±14.9), which was similar to the results of a previous study that the renal failure group according to the RIFLE classification showed the highest mortality rate (74.4%) and the highest SAPS II score (55.9±2.7)16). Based on the results that the stage 3 AKI subgroup, which was the most severe group and it exhibited the highest mortality rate, we think that the severity of AKI may affect the prognosis of AKI patients. Kresse et al18) have reported that for patients with AKI, the mortality rate increases as the severity of the underlying diseases become more severe. It has been demonstrated that the APACHE II scores sensitively reflect the mortality rate for the injury group, whereas the SAPS II scores sensitively reflect the mortality rate for the high risk and severe injury groups. For these reasons, we evaluated the severity of the patients with AKI by using the SAPS II scoring system. Although there was no significant difference of the SAPS II scores between the survivor and nonsurvivor groups, the SAPS II scores were significantly increased for the stage 3 AKI subgroup, suggesting that the SAPS II scoring system can be used along with the AKI classification system for evaluating the severity of AKI.

In conclusion, the most important prognostic factor for AKI patients may be the hemodynamic status, and especially the blood pressure or the serum bicarbonate level after undergoing CRRT. The AKI classification and the SAPS II scoring system can help evaluate the severity of AKI, which suggests that the general status of patients and the severity of AKI may be more important for the prognosis than the CRRT itself.