RIFLE criteria for acute kidney injury

Clinical background

Before 2004 there were more than 30 different definitions of acute renal failure in the literature, which made studies impossible to compare and clinical communication difficult. The RIFLE criteria were developed by the Acute Dialysis Quality Initiative (ADQI) at a consensus conference in Vicenza in 2002 to remedy this. The aim was not to replace clinical assessment of the underlying cause, but to create a common language for severity that could be used across specialties and between centres [1].

The point of a graded classification is twofold. First, it allows early identification of patients with only mild impairment of renal function, in whom intervention may still prevent progression. Second, it provides a prognostic estimate that supports decisions on the level of monitoring and on referral to nephrology or intensive care. RIFLE was the first system to show systematically that even moderate rises in creatinine are associated with increased mortality, a finding since confirmed in more than half a million patients [3].

Applying the RIFLE criteria

RIFLE classifies acute kidney injury into three grades of severity on the basis of two criteria: the change in serum creatinine relative to baseline, and urine output. The more severe of the two criteria determines the class.

Creatinine criteria:

$$\text{Risk: } S_{cr} \geq 1{.}5 \times \text{Baseline}$$

$$\text{Injury: } S_{cr} \geq 2 \times \text{Baseline}$$

$$\text{Failure: } S_{cr} \geq 3 \times \text{Baseline} \text{ or } S_{cr} \geq 4 \text{ mg/dL}$$

Urine output criteria:

$$\text{Risk: } <0{.}5 \text{ mL/kg/h for } 6 \text{ h}$$

$$\text{Injury: } <0{.}5 \text{ mL/kg/h for } 12 \text{ h}$$

$$\text{Failure: } <0{.}3 \text{ mL/kg/h for } 24 \text{ h or anuria for } 12 \text{ h}$$

The derivation was not an empirical study population but an expert panel that systematically reviewed the available literature and formulated consensus recommendations at the second international conference of the ADQI [1]. The criteria rest on physiological principles: serum creatinine is specific to renal function and the change from baseline reflects the change in glomerular filtration, even though in non-steady-state conditions the absolute value underestimates the true fall in GFR. Urine output is less specific but can change well before biochemical signs, particularly with a prerenal cause.

Where the baseline serum creatinine is unknown, the ADQI proposed that it can be estimated with the MDRD equation assuming a normal GFR of 75 mL/min/1.73 m² [1]. This method has, however, proved problematic (see Limitations).

The two outermost categories in the acronym, Loss and ESKD, are time-based outcomes requiring the need for renal replacement therapy for more than 4 weeks and more than 3 months respectively. They are not calculated by this tool and are not relevant to acute classification.

Interpretation in practice

The RIFLE class should be interpreted as a snapshot of severity, not as a diagnosis. A patient may move up and down between classes over a single day, and the class should therefore be updated continuously.

Validation and performance

Ricci et al. carried out a systematic review of 24 studies with a total of more than 71,000 patients in whom RIFLE had been applied [2]. Pooled analysis showed a stepwise increase in the relative risk of death: Risk RR 2.40, Injury RR 4.15, Failure RR 6.37 (all p < 0.0001) compared with patients without AKI. Heterogeneity between studies was substantial, reflecting differences in patient populations, the proportion using the urine output criteria and the proportion with chronic kidney disease.

In a prospective cohort study of 190 intensive care patients in north-eastern Brazil, RIFLE was compared with the newer AKIN and KDIGO classifications [4]. The AUROC for mortality was 0.735 for RIFLE, 0.74 for AKIN and 0.733 for KDIGO, with no significant difference between the systems. Mortality rose with the RIFLE class: 17.7% without AKI, 22.6% for Risk, 24.2% for Injury and 35.5% for Failure. The odds ratio for death, adjusted in multivariable logistic regression, was 2.82 (95% CI 1.25 to 6.39) for Injury and 5.85 (95% CI 2.61 to 13.12) for Failure, while Risk did not reach significance (OR 0.56, 95% CI 0.27 to 1.12). This last finding should be interpreted with caution given the small cohort.

The 2012 KDIGO guideline summarises the evidence and notes that studies with a total of more than half a million patients from different parts of the world have confirmed that the RIFLE class correlates with both short- and long-term survival [3]. KDIGO subsequently adopted a unified definition integrating RIFLE and AKIN: AKI is defined as a rise in serum creatinine of ≥0.3 mg/dL within 48 hours, ≥1.5 times baseline within 7 days, or a urine output below 0.5 mL/kg/h for 6 hours. Stages 1 to 3 correspond to RIFLE Risk, Injury and Failure respectively, with certain modifications for children and for patients who already have a raised creatinine at presentation [3].

Limitations

Baseline creatinine. RIFLE rests on the relative change from a known baseline. In clinical practice a reliable baseline value is often lacking, particularly on acute admission. Bagshaw et al. analysed data from the BEST Kidney study, a prospective observational study in 54 intensive care units in 23 countries with 1,327 patients with severe AKI [5]. Of these, 76% had a known premorbid baseline creatinine. When the baseline was estimated with the MDRD equation instead of the observed value, 18.8% of patients were misclassified at ICU admission and 11.7% at study inclusion. In patients without chronic kidney disease, agreement was considerably better (r = 0.90, misclassification 6.6%), but in patients with suspected chronic kidney disease the MDRD method overestimated the incidence of AKI and should not be used [5].

Urine output criteria. Many validation studies have used the creatinine criteria exclusively and have not applied urine output, even though urine output can identify additional cases [3]. This applies particularly to patients with non-oliguric AKI, in whom a normal urine output can mask marked impairment of renal function. The calculator requires both criteria to be assessed and uses the more severe, which is a strength in clinical use but is not reflected in much of the published validation literature.

Chronic kidney disease. Patients with a raised baseline creatinine from chronic kidney disease may be misclassified. A patient with a baseline of 3.0 mg/dL rising to 4.5 mg/dL meets the Risk criterion (1.5×), but represents a different clinical situation from a patient with a baseline of 1.0 mg/dL rising to 1.5 mg/dL. RIFLE does not handle this adequately.

Superseded by KDIGO. RIFLE has formally been replaced by KDIGO 2012 as the international standard [3]. The KDIGO definition is broader (it includes a ≥0.3 mg/dL rise within 48 hours) and more inclusive, so that more patients are classified as having AKI. In clinical practice and in research, KDIGO is generally used today, but RIFLE retains relevance as the original system and for the interpretation of older studies.

References

1. Bellomo R, Ronco C, Kellum JA, Mehta RL, Palevsky P, the ADQI workgroup. Acute renal failure: definition, outcome measures, animal models, fluid therapy and information technology needs: the Second International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group. Crit Care 2004;8(4):R204-R212. PMID: 15312219
2. Ricci Z, Cruz D, Ronco C. The RIFLE criteria and mortality in acute kidney injury: A systematic review. Kidney Int 2008;73(5):538-546. PMID: 18160961
3. Kellum JA, Lameire N, KDIGO AKI Guideline Work Group. Diagnosis, evaluation, and management of acute kidney injury: a KDIGO summary (Part 1). Crit Care 2013;17(1):204. PMID: 23394211
4. Levi TM, de Souza SP, de Magalhães JG, et al. Comparison of the RIFLE, AKIN and KDIGO criteria to predict mortality in critically ill patients. Rev Bras Ter Intensiva 2013;25(4):290-296. PMID: 24553510
5. Bagshaw SM, Uchino S, Cruz D, et al. A comparison of observed versus estimated baseline creatinine for determination of RIFLE class in patients with acute kidney injury. Nephrol Dial Transplant 2009;24(9):2739-2744. PMID: 19349297