**2. Definition**

Patients in ICU comprise heterogeneous population, and around 30 different definitions were used to describe acute deterioration of renal failure, which both caused difficulties to interpret studies and to discuss the conclusions. Since even small changes in renal function makes prominent differences, it was suggested that re-evaluation of the definition of AKI was mandatory. For the consensus of the definition and improvement of the quality of studies on AKI, Acute Dialysis Quality Initiative (ADQI) group was established. They recommended the term of AKI instead of ARF, since spectrum of AKI is broader and covers different degrees of severity of the disease. In 2002, for a uniform definition of AKI, they described three categories for severity (Risk of ARF, Injury of the kidney, and Failure of kidney function) and two classes for kidney outcome (Loss of kidney function and ESRD), which is called shortly RIFLE criteria(Table1) (Bellomo, 2004). Since AKI is not a stable condition, definition based on changes in function seems to be more suitable. Two measures to reflect kidney function have been recommended for the purpose of definition of AKI; changes in serum creatinine level and urine output. Importantly, these criteria were validated based on ability to distinguish the patients with higher risk for the worst outcome (death and requirement for renal replacement therapy).

The RIFLE criteria have been validated in more then 550,000 patients worldwide. Many studies confirmed that a rise in mortality rates occurred with progressing RIFLE criteria in ICU cases (Murugan, 2011). RIFLE criteria validated in a study on two Turkish tertiary hospital ICU population. This study also showed that increasing mortality rate with increase in severity, difference in mortality rate between stages was statistically significant (χ2=15.037, p<0.001) (Yegenaga, 2010).

Later on in 2007, ADQI group with worldwide collaboration of nephrologists and critical care societies developed Acute Kidney Injury Network (AKIN), which intends to improve worldwide coordination and provokes the development of uniform standards in the AKI area. RIFLE criteria were refined by this group and reported a new AKIN classification, which classified AKI in three stages (stages 1, 2, and 3) according to the degree of damage and outcome categories Loss and ESRD were removed (Mehta, 2007). The purpose of these modifications were to include a small but important group of patients with early and mild AKI who experience a change in renal function that is greater than physiological variation but less than the 50% increase required for the RIFLE criteria. Thus, an absolute increase in serum creatinine levels of at least 0.3mg/dl in 48 hours was recommended to include in stage 1. Patients who required renal replacement therapy (RRT) are classified as stage 3 AKI, regardless of their serum creatinine levels and urine output. Nevertheless there is no common worldwide guideline about the timing to initiate of RRT, using it as criteria to describe stage 3 may cause some misunderstandings. In particular, patients with underlying chronic kidney disease (CKD) may be missed by the original RIFLE criteria but not by AKIN since it counts even smaller changes in serum creatinine level. In addition, 48-hours time

classification for the definition of AKI, the viewpoint of this subject has changed and now it is possible to be aware of patients with high risk for AKI(Bellomo 2004). Nevertheless recently several biomarkers have been introduced to diagnose AKI even before creatinine

In this review, new perspective and favorable improvement of this subject will be discussed; furthermore definition, epidemiology, risk factors, biomarkers of AKI will be evaluated.

Patients in ICU comprise heterogeneous population, and around 30 different definitions were used to describe acute deterioration of renal failure, which both caused difficulties to interpret studies and to discuss the conclusions. Since even small changes in renal function makes prominent differences, it was suggested that re-evaluation of the definition of AKI was mandatory. For the consensus of the definition and improvement of the quality of studies on AKI, Acute Dialysis Quality Initiative (ADQI) group was established. They recommended the term of AKI instead of ARF, since spectrum of AKI is broader and covers different degrees of severity of the disease. In 2002, for a uniform definition of AKI, they described three categories for severity (Risk of ARF, Injury of the kidney, and Failure of kidney function) and two classes for kidney outcome (Loss of kidney function and ESRD), which is called shortly RIFLE criteria(Table1) (Bellomo, 2004). Since AKI is not a stable condition, definition based on changes in function seems to be more suitable. Two measures to reflect kidney function have been recommended for the purpose of definition of AKI; changes in serum creatinine level and urine output. Importantly, these criteria were validated based on ability to distinguish the patients with higher risk for the worst outcome

The RIFLE criteria have been validated in more then 550,000 patients worldwide. Many studies confirmed that a rise in mortality rates occurred with progressing RIFLE criteria in ICU cases (Murugan, 2011). RIFLE criteria validated in a study on two Turkish tertiary hospital ICU population. This study also showed that increasing mortality rate with increase in severity, difference in mortality rate between stages was statistically significant

Later on in 2007, ADQI group with worldwide collaboration of nephrologists and critical care societies developed Acute Kidney Injury Network (AKIN), which intends to improve worldwide coordination and provokes the development of uniform standards in the AKI area. RIFLE criteria were refined by this group and reported a new AKIN classification, which classified AKI in three stages (stages 1, 2, and 3) according to the degree of damage and outcome categories Loss and ESRD were removed (Mehta, 2007). The purpose of these modifications were to include a small but important group of patients with early and mild AKI who experience a change in renal function that is greater than physiological variation but less than the 50% increase required for the RIFLE criteria. Thus, an absolute increase in serum creatinine levels of at least 0.3mg/dl in 48 hours was recommended to include in stage 1. Patients who required renal replacement therapy (RRT) are classified as stage 3 AKI, regardless of their serum creatinine levels and urine output. Nevertheless there is no common worldwide guideline about the timing to initiate of RRT, using it as criteria to describe stage 3 may cause some misunderstandings. In particular, patients with underlying chronic kidney disease (CKD) may be missed by the original RIFLE criteria but not by AKIN since it counts even smaller changes in serum creatinine level. In addition, 48-hours time

starts to increase(Waiker, 2008).

(death and requirement for renal replacement therapy).

(χ2=15.037, p<0.001) (Yegenaga, 2010).

**2. Definition** 


Table 1. Two definitions of AKI

interval for the diagnosis of AKI was introduced to ensure that the process is acute. Unlike RIFLE, in AKIN criteria there has been an attempt to resolve some easily reversible causes of azotemia (for example, volume depletion and urinary obstruction). AKIN criteria declare that diagnosis based on the urine criterion alone will require exclusion of urinary tract obstructions and other easily reversible causes of decreased urine output.

*Determination of baseline renal function:* Since AKI was defined as rapid decline in renal function from baseline levels, it is important how to measure baseline renal function. The baseline serum creatinine level has been recommended for use as a marker to reflect the renal function. The baseline serum creatinine value has been estimated in various ways, such as the serum creatinine level on hospital admission, the minimum creatinine level during the hospital stay, the serum creatinine value estimated from the MDRD calculation (assuming estimated GFR=75ml/min/1,73m2) or the lowest value among these (Ricci,

Evaluation of Acute Kidney Injury in Intensive Care Unit 177

2004). This study revealed that, etiology of AKI was ATN (acute tubular necrosis) in 50% of cases with no specific determined cause. The next most common etiologies included nephrotoxin administration (26%), cardiac disease (20%) including myocardial infarction, cardiogenic shock, and congestive heart failure, ATN from hypotension (12%), sepsis (19%), unresolved prerenal factors (16%) and liver disease (11%). They found that predictors of mortality using proportional hazards Cox regression at the day of diagnosis of ARF were age, BUN, liver failure. At the day of consultation; age, Log urine output, creatinine (<2mg/dl), BUN, liver failure, ARDS (acute respiratory distress syndrome) and platelet count were related to mortality. When the day renal replacement therapy started, the predictive factors for survival were found as age, platelet count, liver failure, sepsis or septic shock. Oliguria, sepsis, respiratory failure, and hepatic failure would be

The largest cohort with most participant of AKI to date was the Beginning and Ending Supportive Therapy for Kidney (BEST) study (Uchino, 2005). Out of 29.269 critically ill patients, there were 1.738 (5.7%) patients with AKI. The most common cause of AKI was septic shock (47.5 %), followed by major surgery (34%), cardiogenic shock (27%), hypovolemia (26%) and nephrotoxin administration (19%). In-hospital mortality varied from 50.5% to 76.8% between centers. A multivariate logistic regression model to identify independent correlates of in-hospital mortality yielded several previously identified risk factors, including delayed diagnosis of AKI, age, sepsis, and known disease severity score that included BUN and urine output, which are consistent with the

Based on the previous reports; most common and complicated cause of AKI in ICU was found as sepsis and its prevalence have been reported to be 9-40% (Brivet, 1996 & Liano, 1998 & Bellomo,2008). Therefore, we prospectively evaluated 257 patients with sepsis or systemic inflammatory response syndrome who were admitted to the surgical and medical ICU during 12 months in 2001 to determine the risk factors for acute renal failure (ARF) development (Yegenaga, 2004). In this study, ARF was defined as serum creatinine level >2 mg/dL, based on this definition out of 257, twentynine (11%) patients were diagnosed as ARF. Mortality rate was 72% in patient with sepsis and ARF and in those with sepsis without ARF this rate was found as low as to 24 %. Multivariate logistic regression analysis of data showed that older age (OR:1.1, CI:1.03- 1.13) serum bilirubin >1.5 (OR:9.7, CI:1.65-60.3), higher baseline serum creatinine level (OR:1.02,CI:1.007-1.04), and higher central venous pressure (CVP) (OR:1.5,CI:1.26-1.80) were predictive for the development of ARF (Yegenaga, 2004). It was mandatory to evaluate fluid status of these patients, since it was found that CVP is an independent risk factor for the development of AKI (Van Biesen & Yegenaga, 2004). And subgroup analysis of these patients revealed that higher colloid fluid loading for the first 3 days (2037±1681vs 1116±1220, p<0.03) and lower diuresis (1347±649vs1849±916mL, p=0.005) was associated with poor outcomes and in addition interestingly the fraction of inspired oxygen (FiO2) needed to be increased significantly on the second day of sepsis in the ARF group but remained unchanged in non ARF group. These patients developed ARF despite further fluid loading and in addition, respiratory function deteriorated. Against the classical knowledge; this study brought up the idea that in critically ill patient fluid loading should be performed in cautious. It is more likely that in dehydrated patients fluid loading can prevent ARF development but in critical ill patients

consistently associated with mortality in AKI (Chertow, 2006).

previous studies.

2011). The choice of estimation technique used to obtain the baseline creatinine value has a marked effect on the prevalence of AKI, severity of disease, and mortality risk associated with various stages of AKI. When premorbid renal function is not known ADQI group has recommended back- estimation of the baseline serum creatinine level value using the MDRD formula. In a study, involving 4,863 hospitalized patents, investigators studied the performance of these three potential surrogates for the baseline serum creatinine level and they concluded that all three surrogates resulted in bidirectional misclassification of AKI. Although the use of serum creatinine level at admission found to be more specific in this study, this parameter had the lowest sensitivity (38.9%) for diagnosis of AKI (Siew, 2010).

Subgroup analysis of data from the Beginning and Ending Supportive Therapy for the Kidney (BEST) study also compared observed baseline serum creatinine levels with MDRD back estimated values to determine the RIFLE class. They found that MDRD back estimated serum creatinine value performed reasonably well for the determination of the RIFLE categories when premorbid renal function was near normal, but should be cautiously used in patients with suspected CKD (Bagshaw, 2009).

As a conclusion investigators should make every effort to find out true baseline creatinine level before using any estimated values. An appropriate baseline serum creatinine value should express normal renal function of patients, such as outpatients serum creatinine level measured within the past year (prefereably in the past 3 months). When an outpatient serum creatinine level is not available, serum creatinine level on admission to hospital or ICU could be used as a second choice. However, use of admission serum creatinine level is specific but insensitive, and may underestimate the incidence of AKI.

*Which classification is better RIFLE or AKIN?* Based on the studies to compare these two classifications; AKIN did not seem to show any improvement in the sensitivity and predictive ability of the definition and classification of AKI. Acute Physiology Score (SAPS)3 database in 2009 re-evaluated and compared the performance of the RIFLE and AKIN criteria (Joannidis, 2009). They concluded that RIFLE seems to be more sensitive for diagnosing AKI and to have a greater ability then AKIN criteria to predict the mortality. These data confirm that neither classification offers clear advantage over the other; and both systems still have some limitations. Similarly Bagshaw et al. compared AKIN classification and RIFLE criteria in 120,123 critically ill patients during the first 24 hours after admission to the ICU (Bagshaw, 2008). They concluded that AKIN classification did not improve the sensitivity and predictive ability of RIFLE criteria for the definition and classification of AKI in the first 24 hours in ICU. Based on these large database studies there is no evidence about superiority of any of these classifications. Both criteria can be used for the evaluation of critically ill patients, as long as the clinician is aware of the limitations mentioned above.
