Haemodynamic-guided fluid administration for the prevention of contrast-induced acute kidney injury: the POSEIDON randomised controlled trial.

Brar SS, Aharonian V, Mansukhani P, Moore N, Shen AY, Jorgensen M, Dua A, Short L, Kane K.

PMID: 24856027

Accompanying Editorial

TweetChat: June 25th 9 pm EDT

For this round of NephJC, we have somewhat of a rarity in the nephrology literature - a randomized clinical trial with a positive result! The POSIEDON trial, published in the Lancet a few weeks ago, is a single-centre randomized trial which found a large reduction in acute kidney injury after cardiac catheterization, using left ventricular end-diastolic pressure to guide IV fluid administration compared to a control group.

What do we know about preventing Contrast-induced Nephropathy?
Or, as the new kids call it: ‘Contrast-Induced Acute Kidney Injury’ (CI-AKI). Many preventive measures have been tried and have fallen by the wayside, notably n-acetyl cysteine, but also lasix and mannitol. The best evidence for hydration or volume expansion is actually pretty sketchy. In the 1990s, we did have this RCT, suggesting 0.9% saline is more effective than 0.45% saline. Up next came the Merten RCT, which caused quite a bit of a splash, suggesting that not only a bicarbonate-based solution is superior to normal saline, but also for pioneering an abbreviated version of the volume expansion protocol (1 hour pre-procedure and 6 hours post). Since then, the replication of these results has run into some trouble, but another large ongoing RCT will eventually answer the bicarb versus saline question. For this week’s discussion, Brar and colleagues took a novel approach - using LVEDP to guide the volume expansion strategy.

What does LVEDP mean anyway?
LVEDP is the left ventricular end-diastolic pressure, and serves as a measure of the pre-load and thus that of circulating blood volume. It is a useful prognostic tool and has been shown to have a good correlation with clinical outcomes, especially in heart failure studies. Moreover, if a patient is undergoing cardiac catheterization, LVEDP is relatively easy to measure (though this does make the strategy difficult to use in other settings). It may indeed represent a smarter measure of the overall volume status, which is how these authors decided to use that versus a weight-based algorithm, which is standard care.

What was the study design?
This trial recruited 396 patients over 2 years, who underwent cardiac cath at the Kaiser Permanente Medical center at LA. To select patients at high risk of CI-AKI, they needed to have CKD (defined as GFR < 60) AND at least one of the following:

  • diabetes

  • history of congestive heart failure

  • hypertension

  • age > 75 years

Among the usual acceptable exclusion criteria were patients who had a changing GFR (and possible AKI ongoing). The control arm received 0.9% saline at the standard rate (3ml/kg/hour for 1 hour pre- and 1.5 ml/kg/hour for 4 hours post). The intervention arm received the same pre-procedural saline (3ml/kg/hour) but the secret sauce was what happened next - based on the LVEDP. They received

  • 5 mL/kg/h for left ventricular end-diastolic pressure lower than 13 mm Hg,

  • 3 mL/kg/h for pressure of 13–18 mm Hg, and

  • 1·5 mL/kg/h for pressure higher than 18 mm Hg.

All patients were administered the same low-osmolar non-ionic contrast medium (ioxilan). The recruits were stratified according to diabetes status and n-acetyl cysteine use. The study was single-blinded (patients, and not investigators) which is understandable as the rate of fluid infusion would be difficult to mask. The endpoint was the usual CI-AKI one (25% or 0.5 mg/dL rise in creatinine within days 1-4 post cath) and they also looked at the AKIN threshold (0.3 mg/dL in a post hoc analysis).

What did the POSIEDON trial show?
Of the 396 patients recruited, just under 40% were women, and roughly half were diabetic, with a mean GFR of ~ 48mL/min/1.73m2 (Table 1). The overall CI-AKI incidence was 11.4%, but with a striking difference of 16.3% in control versus 6.7% in the intervention arm - resulting in a RR of 0.41 with a NNT of only 11. No differences were apparent in the different subgroups or different AKI measures studied (Tables 2 and 3). Interestingly, the major adverse event rates (arguably comprising more important clinical outcomes)showed a similar  beneficial trend, which was significant at 6 months.

What did the Editors comment?
The editors are not completely convinced about this strategy - they would like it to be replicated in a higher risk population, and they seem to favour their own ‘RENALGUARD’  system - which is basically a method to maintain a high urine flow.

Disclosure: The lead author, Sam Brar, is a close personal friend and co-author along with this writer on a few papers, but there is no involvement of this writer in the paper we are discussing.