This was the subject of Question 12 from the first paper of 2012. It was answered to a satisfactory standard by only 2% of the candidates.
The RIFLE system
The classification system separates renal failure into 5 categories:
|GFR criteria||Urine output criteria|
|Creatinine x 1.5||u/o < 0.5ml/kg/hr x 6 hrs|
|Creatinine x 2||u/o < 0.5ml/kg/hr x 12 hrs|
|Creatinine x 3||u/o < 0.3ml/kg/hr x 24 hrs|
|Complete loss of function > 4 weeks|
|Complete loss of function > 3 months|
The Acute Dialysis Quality Initiative formulated this classification to fill a certain vacuum in renal failure research. The paper was published by Ronco Bellomo and Kellum in 2004. Good dissections are available from Life In The Fast Lane and from the ADQI people. The point of this system was to predict mortality and hospital stay from easily measured parameters, so as to be able to compare the efficacy of interventions.
Critique of the RIFLE classification
Lopes et al have published "a critical and comprehensive review" of the RIFLE and AKIN criteria in 2012. This summary owes much to their article.
Advantages of the RIFLE system:
- RIFLE has been validated as a means of prognostic stratification of acute kidney injury. i.e. as the patient deteriorates down the RIFLE classes, there is a stepwise increase in their mortality. RIFLE classes are strongly associated with increased lengths of stay, RRT requirement, renal function recovery and discharge from hospital to a care facility.
- It is successful in its originally intended purpose, which is to act as a definition and stratification of AKI severity.
Criticisms of the RIFLE system:
- The system classifies renal failure once it has already become established. Some might say that this is too late. One might wish for a system which detects renal failure before its onset, and stratifies risk in that manner.
- The criteria are unbalanced in outcome, but equal in weighting. The creatinine (GFR) criteria and the urine criteria are weighed the same, but have different associations with mortality. For instance, patients stratified into the "Risk" class by the creatinine criteria are more severely ill than those stratified into the same class by urine output criteria.
- Urine criteria rely on impractical measurements: The system requires 6th hour and 12th hour urine measurements, which are not routinely collected. This makes retrospective data collection difficult.
- Urine output criteria are altered by diuretics, DI, etc.- i.e. they may not be a brilliantly accurate reflection of renal function.
- Urine output measurements may be inaccurate. A mildly unwell "Risk" class patient may have no catheter, and may be voiding in an uncontrolled manner all over the hospital bathroom or bedsheets.
- The criteria for creatinine rely on the availability of baseline values, which are rarely available. One can make educated guesses as to what the creatinine should be, but these are yet to be validated.
- Nobody seems to agree on what the definiton of "baseline" is. Is it the creatinine on admission? Is it the last creatinine measured by the GP? Is it the predicted value, based on age and weight? Confusion reigns.
- Creatinine levels may be affected by factors which do not affect outcome, such as mild dehydration.
- The rate of creatinine rise is unaccounted for. A sudden doubling of creatinine may be more troubling than a gradual rise over the course of a week.
- The "Risk" criteria for creatinine may not be sensitive enough: in one study patients with a mild creatinine elevation still had increased mortality, but did not qualify for the "Risk" category.
- The criteria require true GFR, rather than an estimated GFR: using the eGFR is inherently invalid, because the MDRD equation used to calculate eGFR is valid only in steady-state conditions (and you with your failing kidneys are by definition not in a steady state).
- The criteria make no attempt to distinguish between different aetiologies; the differences may have powerful prognostic implications (eg. renal impairement due to dehydration vs. lupus nephritis).