​While we wait for more data on the impact of renal biomarkers on diagnostic and therapeutic strategies in AKI, we are stuck with creatinine as our primary laboratory marker of kidney injury. Although the Acute Kidney Injury Network criteria recognise the importance of minor elevations in creatinine (a rise of 0.3 mg/dl or 25.4 micromol/L from baseline is enough to diagnose Stage 1 AKI), there is often little mention of the impact of fluid balance on our ability to diagnose and prognosticate in AKI.

Creatinine is distributed in total body water, and cumulative positive fluid balances can be expected to increase total body water. This will have the effect of diluting creatinine and providing a falsely lowered serum concentration. Similarly, states of total body water depletion will lead to an apparent rise in creatinine concentration. This merits the question of whether there is any utility in adjusting the serum creatinine to take into account changes in total body water secondary to fluid balance (presupposing accurate intake/output measurement).

Using data from the FACT trial (Fluid and Catheter Treatment Trial, a RCT of conservative against liberal fluid management and CVP against PA catheter guidance in ARDS), Liu et al adjusted creatinine levels to take into account the fluid balance of the patient. They found that there were two "hidden" populations of patient.

Firstly, there were patients who had biochemical AKI (based on creatinine) before adjustment for fluid balance, whose creatinine was not elevated after the adjustment. These patients were presumably in a negative fluid balance compared to their baseline creatinine point, and would have been falsely labelled as STAGE 1 AKI.

Secondly, there were patient who had normal creatinine prior to adjustment, but an elevated creatinine and AKI after adjustment. These patients were in a significant positive fluid balance compared to their baseline position.

Interestingly, following adjustment for fluid balance, AKI occurred more frequently in the liberal fluid group than in the conservative group (66% against 58%, p=0.007). In addition, patients who were considered to have normal creatinine prior to adjustment, but after adjustment were classed as AKI, had worse outcomes than patients who remained normal pre- and post-adjustment, suggesting that this was indeed a relevant phenomenon. The group who met AKI criteria before adjustment but not after it had similar outcomes to those who never scored for AKI. 

This has big implications for both clinical practice and research. We know that even small degrees of AKI are associated with worse outcome. For example, we know that patients with a normal creatinine but elevated renal biomarkers have worse outcome than patients with no biomarker rise. We also know that even small creatinine elevations are associated with worse outcome. This paper suggests strongly that we have within our ICU's a population of patients who are at significantly increased risk of poor outcome, yet are invisible to us by virtue of TBW expansion. As a result, we may not be directing therapy as aggressively as we might if the warning signal of creatinine elevation was evident.

So what is the message here. Firstly, we need to be conscious of our patients fluid balance in a consistent and accurate manner. Secondly, we need to use that knowledge to recheck our creatinine adjusting for the fluid balance of the patient. There is an ever increasing body of literature suggesting a link between positive fluid balance and worsening outcomes in AKI and several other conditions. From the viewpoint of the kidneys, persistent positive balance leads to tissue fluid overload, renal interstitial edema, renal venous congestion, impaired renal perfusion, increased intra-abdominal pressure, all reducing GFR and ultimately injuring tubular cells.

So it is time to make fluid balance a central feature in our ICU rounds, and to familiarise ourselves with the data.

You can read Liu et al's article in Critical Care Medicine 2011; 39: 2665-2671.