Acute kidney injury: how do we define it?

Abstract

Over recent years, there has been welcome increased interest in acute kidney injury (AKI) and its association with patient outcome. The term AKI has replaced the term acute renal failure (ARF) and encompasses all types of ARF. New definitions and staging systems for AKI have been proposed, which have stimulated a multitude of different studies to evaluate their clinical utility. These recent advances need to be communicated to the wider health care community so that we are using a shared nomenclature. In 2009 the National Confidential Enquiry into Patient Outcome and Death AKI study (‘Adding Insult to Injury’) announced its findings and recommendations. The report recommended that the detection of AKI and its management should be improved. These recommendations along with the adoption of the new staging systems will potentially have an impact on clinical biochemistry departments and exert an increased demand on resources. Running in parallel with these initiatives is the quest to discover novel biomarkers to detect AKI, the development and introduction of which will require laboratory support.

Introduction

Over the last few years, there has been an increased interest in acute kidney injury (AKI) particularly in relation to its impact on patient morbidity, mortality and length of hospital stay. ^1,2 The term AKI encompasses all types of acute renal failure (ARF) and it now generates over 6000 citations when entered on PubMed. However, the term has not been adopted extensively outside of renal medicine and further education is required at undergraduate and postgraduate levels. The National Service Framework (NSF) for Renal Services has recommended that patients at risk of suffering from AKI should be identified promptly, with hospital services delivering high-quality, clinically appropriate care in partnership with specialized teams. ³ This provides us with the challenge of how to improve the quality of care for patients who develop AKI in the modern health care setting.

Now is an opportune time for the clinical community to address this challenge and strengthen collaborations between specialist health care societies. It is time to consider how to close the gap between the current care delivered and that to which the NSF aspires. In June 2009, the National Confidential Enquiry into Patient Outcome and Death (NCEPOD) AKI study announced its findings and recommendations. As with previous NCEPOD reports this will act as a major incentive to health care institutions to improve the quality of care provided. To enable us to be better prepared to respond to the study's recommendations a dialogue between different stakeholder organizations was initiated at the Acute Kidney Injury Care Initiative meeting in London in Spring 2009. The aim of this meeting was to understand how different specialists identify and manage AKI and how to develop a shared nomenclature. It is hoped that ultimately this will allow a transformational change in how AKI is managed. One of the first steps will be to determine a consensus definition for AKI.

Definitions of AKI

ARF has traditionally been recognized as a rapid decline in kidney function over hours to days with failure to regulate fluid, electrolyte and acid–base balance. More than 35 different definitions of ARF have been used in the literature. This has hampered our ability to characterize the true incidence of the disease, assess its impact and make progress with clinical and scientific research. The proposals for a universal definition and classification system for AKI is the result of a collaborative effort by both nephrologists and intensive care specialists in the form of the Acute Dialysis Quality Initiative (ADQI) and the Acute Kidney Injury Network (AKIN). ^4,5 In 2002 the ADQI proposed the term AKI to represent the entire spectrum of ARF, preferring the term injury to failure, as it was felt this more accurately reflects the earlier degrees of injury that may occur prior to failure of the kidney. Simultaneously the ADQI proposed the RIFLE staging system (Table 1), which includes three levels of progressive kidney dysfunction, risk, injury and failure and two outcomes, loss of function and end stage kidney disease. Categorization into a particular stage is dependent upon increases in serum creatinine concentration from baseline values within a one week time interval or reductions in urine output. The staging criteria (serum creatinine or urine output) selected is that which corresponds with the higher stage of AKI.

Table 1

RIFLE and AKIN definition and classification schemes for AKI

	Serum creatinine criteria	Urine output criteria
RIFLE
Risk	↑SCr ≥1.5 × from baseline or ↓ GFR ≥25%	<0.5 mL/kg/h ≥ 6 h
Injury	↑SCr ≥2.0 × from baseline or ↓ GFR ≥50%	<0.5 mL/kg/h ≥12 h
Failure	↑SCr ≥3.0 × from baseline or ↓ GFR ≥75% or an acute ↑ SCr ≥44 µmol/L from baseline SCr ≥354 μmol/L	<0.3 mL/kg/h ≥24 h or anuria ≥12 h
Loss	Complete loss of kidney function >4 weeks
End-stage	End-stage kidney disease >3 months
AKIN
Stage 1	↑SCr ≥1.5 × or ↑ SCr ≥26 μmol/L from baseline	<0.5 mL/kg/h ≥ 6 h
Stage 2	↑SCr ≥2.0 × from baseline	<0.5 mL/kg/h ≥12 h
Stage 3	↑SCr ≥3.0 × from baseline or an acute ↑ SCr ≥44 μmol/L from baseline SCr ≥354 μmol/L or initiated on RRT (irrespective of stage at time of initiation)	<0.3 mL/kg/h ≥24 h or anuria ≥12 h

Only one criterion (serum creatinine, GFR or urine output) has to be fulfilled to qualify for a specific category. Baseline serum creatinine is considered to be within one week for RIFLE and within 48 h for AKIN

SCr, serum creatinine, GFR, glomerular filtration rate, AKIN, Acute Kidney Injury Network, RRT, renal replacement therapy

Subsequent studies have demonstrated that relatively small increments in serum creatinine concentration are associated with a significant increase in patient mortality. ^1,2 This prompted further refinements to the definition and classification system by the AKIN in 2005. ⁵ The AKIN modified the RIFLE staging system to produce an interim staging system to allow additional data to be gathered and research initiatives to be proposed (Table 1). The modifications included using a smaller change in serum creatinine of ≥26 μmol/L (0.3 mg/dL) to define the presence of AKI thereby attempting to increase the sensitivity of the criteria. This change is based on work from Chertow et al., ¹ who found that patients in the USA with an increase in serum creatinine of just 26 μmol/L had a 70% higher multivariable adjusted odds of death compared with patients with little or no change in serum creatinine concentration. Further changes included the introduction of a time constraint of 48 h between serum creatinine results and classifying any patient who received renal replacement therapy as stage 3 AKI. The 48 h time constraint was based upon data demonstrating that a poor outcome was associated with small increases in serum creatinine occurring within 24–48 h. ⁶ The AKIN group also proposed that the diagnostic criteria could not be applied until the patient's volume status had been optimized and urinary tract obstruction had been excluded. This has led to a certain degree of debate as to whether prerenal and obstructive aetiologies should be included or excluded from the staging criteria.

There are inherent difficulties with both of the newly proposed definitions as they rely on changes in either serum creatinine concentration or urine output, both of which are recognized as poor biomarkers. Serum creatinine concentration does not accurately reflect the true glomerular filtration rate (GFR) in a patient who is not in steady state. In the early stages of severe AKI, serum creatinine may not be significantly increased despite a marked reduction in GFR due to there having been insufficient time for creatinine to accumulate in the blood pool. Once a patient has commenced on renal replacement therapy, serum creatinine becomes less useful as a marker of kidney injury as it is removed. The accurate measurement of urine output is generally confined to patients with urinary catheters and is modified by the use of diuretics. It must also be remembered that AKI can be oliguric (<400 mL/24 h) or non-oliguric, so that increases in serum creatinine concentration can occur despite an adequate urine output. Another contentious issue is deciding what constitutes the baseline serum creatinine concentration and what value to accept when calculating the stage of AKI. Some studies have used the Modification of Diet in Renal Disease Study equation to estimate a baseline serum creatinine assuming a ‘lower limit of normal’ baseline GFR of 75 mL/min/1.73 m² when a true baseline value has not been available.⁷

More recently one of the members of the AKIN group has studied and published work on creatinine kinetics and proposed a further definition of AKI. ⁸ This group simulated creatinine kinetics following AKI in the setting of normal baseline kidney function and progressively worsening stages of chronic kidney disease (CKD). They demonstrated that the percentage changes in serum creatinine after severe AKI were highly dependent on baseline kidney function. In contrast to this the absolute increase in serum creatinine was nearly identical across the spectrum of baseline kidney function. The group has therefore proposed a definition and staging system similar to that of the AKIN but incorporating absolute changes (rather than percentage changes) in serum creatinine over a 24–48 h time period. It is anticipated that there will be further modifications to the definition as further data become available. There remains an obvious need for a universal definition to allow improved patient care and clinical research.

Epidemiology of AKI

At present the true incidence of AKI is difficult to assess because of the absence of a universal definition and standardized registry. As previously discussed the incidence of AKI is dependent upon the definition used and the cohort studied. Historically, studies of AKI have estimated an incidence of 3–7% in hospitalized patients. ⁹ More recently a study using the RIFLE staging system identified that 18% of patients admitted to a large urban medical centre had evidence of AKI. ¹⁰ The incidence of AKI on the intensive care unit has been more clearly characterized due to established data collection systems such as the Intensive Care National Audit and Research Centre in the UK and has been estimated at 25–30%. ⁹

There has previously been an emphasis in the literature on AKI in the critically ill patient. The recent Scottish Audit of Surgical Mortality report identified a 32% incidence of AKI in patients dying following surgery. ¹¹ To date over 20 different studies using the RIFLE staging system have demonstrated a wide ranging incidence of AKI dependent upon the population case-mix. ⁹ Further studies have compared the AKIN staging system with the RIFLE staging system demonstrating little difference between the systems in overall incidence and outcomes of AKI. ⁷ A consensus definition will provide us with the opportunity to assess the incidence and impact of AKI in different geographical areas around the world. It is essential to clarify the epidemiology of AKI to allow us to target resources to appropriate areas.

Biomarkers for AKI

The overall mortality rate among patients with AKI has remained at 50% for the past four decades despite significant advances in medical care. ¹² The use of more aggressive medical and surgical interventions in an ever ageing population has been suggested to account for the failure to improve outcomes. It has been well documented that the majority of patients who survive following AKI will recover kidney function. ¹³ However, it is now recognized that a significant proportion of these patients will have subclinical functional defects and that a smaller proportion will be at risk of developing CKD. ¹⁴ It is therefore essential that we prevent the development of AKI and have the capability for early detection when it occurs. The introduction of the new definitions of AKI has resulted in a number of epidemiological studies, which have helped identify different at-risk populations of patients.

There is now a need to translate the improved understanding of the pathophysiology of AKI into clinical advances. At present, we have very poor biomarkers (biological markers) of AKI, having to rely on serum creatinine and urine output, both of which are recognized as insensitive. By comparison development of more sensitive biomarkers such as troponin to indicate early myocardial damage has contributed to a transformational change in the way acute coronary disease is managed. Earlier detection of AKI may allow more precise targeting of therapeutic strategies, which to date have been relatively unsuccessful.

The term biomarker applies to any measurable diagnostic indicator that is used to assess the risk or the presence of disease. Desirable features of a biomarker include accuracy, reproducibility, cost-effectiveness and ease of measurement and use by the clinician. A biomarker should allow early disease detection to enable effective therapeutic intervention. If it is to be used as a diagnostic test, it should be both sensitive and specific providing a high predictive value. Novel biomarkers may be able to differentiate between the different aetiologies of AKI (e.g. that due to ischaemia, sepsis or nephrotoxins), determine the site of injury and assess response to therapy. Recently, there has been an explosion of interest in developing new biomarkers across many medical and surgical specialties to aid in the diagnosis of disease. This has been particularly helped by the advances that have been made in genomics, metabolomics and proteomics.

Progress has been made in the development of biomarkers for AKI with a flurry of publications over the last few years. ^15,16 This includes the discovery of neutrophil gelatinase-associated lipocalin (NGAL), ¹⁷ the expression of which is increased following ischaemic renal insult and released into both the blood and the urine. Clinical research in paediatric patients undergoing elective cardiac surgery has demonstrated that NGAL concentrations are increased in the urine and plasma within two hours of surgery in those children who subsequently develop kidney injury. ¹⁸ Urinary NGAL concentrations at two hours following cardiac surgery have been demonstrated to correlate strongly with the duration and severity of AKI, length of hospital stay, the need for renal replacement therapy and mortality. ¹⁹

There are a number of other potential biomarkers for AKI including interleukin 18 (IL-18), liver fatty acid binding protein (L-FABP) and kidney injury molecule-1 (KIM-1). IL-18 is a proinflammatory cytokine that has been demonstrated to be induced in the proximal tubule following ischaemic AKI and that can then be detected in the urine following cardiac surgery. ²⁰ IL-18 expression is not increased by nephrotoxins or urinary tract infections but is induced by endotoxaemia and various systemic inflammatory states. L-FABP is normally expressed in the proximal tubules and urinary concentrations have been shown to be increased in children who develop AKI following cardiac surgery. ²¹ However, measurements of urinary L-FABP can be confounded by the presence of other forms of kidney disease. KIM-1 is induced in the proximal tubules of patients with established AKI. Urinary KIM-1 concentrations have been demonstrated to be increased 12 h after cardiac surgery in children who developed AKI. ²² KIM-1 expression appears to be limited solely to injured kidney with no other systemic source identified. Its expression is also increased following exposure to nephrotoxins.

It is unlikely that a single biomarker will be able to diagnose all forms of AKI and a panel of biomarkers may be required in this complex disease. Biomarker development is in an embryonic form at present and the results of randomized controlled clinical trials should be awaited to confirm initial success. The potential importance of biomarkers has been recognized by the National Institutes of Health Research through its substantial funding of clinical research projects. This is essential in establishing a biomarker pipeline to enable the successful progression from biomarker discovery to validation in clinical studies.

Summary

AKI has a significant impact on patient morbidity, length of hospital stay and mortality. It is important that the terminology is clear and understood across the health care profession. New definitions of AKI have been proposed and this is providing useful information on the epidemiology of the disease. At present, there is no universally accepted definition but it is anticipated that any future definition will be based on the two most recently proposed definitions that are available. The Renal Association and the Association for Clinical Biochemistry are currently collaborating to determine the changes that occur in hospitalized patient's serum creatinine concentrations in a number of different centres around the UK. The advances in biomarker development may ultimately enhance the detection and management of AKI.

DECLARATIONS

Competing interests: None.

Funding: None.

Guarantor: AJPL.

Contributorship: Both authors contributed to the article in terms of ideas. AJPL edited the final version.