Contrast-Induced Nephropathy: Signal or Noise?

John Mandrola, MD


March 07, 2017

Medical science done well deserves attention. When it upends expert opinion and questions entrenched quality measures, it delivers humility. And humility surely enhances wisdom.

Contrast medium is used in millions of diagnostic and interventional procedures, and contrast-induced nephropathy (CIN) is a feared adverse outcome. Current guidelines recommend intravenous (IV) hydration for the prevention of CIN in patients with impaired baseline renal function and the presence of risk factors for kidney injury.

Expert Opinion and the Absence of Evidence

How contrast dye injures the kidney is not clear, nor is it known how or if IV hydration protects the kidney from this damage. In the absence of evidence, therefore, recommendations for periprocedural hydration are based on expert opinion.

Although hundreds of studies have compared CIN prevention strategies, only a handful have included a control group receiving no prophylaxis. None of these trials have compared IV hydration with no hydration in the high-risk population targeted by the guidelines.

Researchers at the Maastricht University Medical Center in the Netherlands decided to test the value of IV hydration to prevent CIN. Also on trial was the expertness of expert opinion.

The AMACING trial was elegant because it was simple[1]. Investigators enrolled a wide swath of high-risk patients who were to have contrast and assigned them to hydration or no hydration. Then they tallied easy-to-measure outcomes, such as a 25% rise in creatinine in the days after the procedure and major adverse events.

They found that in the two well-matched groups of approximately 330 patients for whom guidelines recommend prophylactic hydration, rates of CIN were nearly identical—2.7% in the hydration arm vs 2.6% without hydration. The findings were consistent in the prespecified subgroups.

Adverse outcomes due to hydration were observed in 18 (5.5%) of 328 patients, including 13 cases of symptomatic heart failure. No patient in either group required dialysis or ICU admission.


Given the provocative nature of the results, it seems best to start with the limits of the study and its criticisms.

First is the size of the study and the chance that the nearly identical rates of CIN occurred by chance. On Twitter, cardiology trialist Dr Gregg Stone (Columbia University, NY) pointed out that investigators enrolled less than half the planned patients and event rates were low. I agree that we must entertain the tiny possibility of a statistical fluke. This is where replication would help.

Another critique is that higher-risk patients may still benefit from IV hydration. That is also true. The investigators excluded patients with an estimated glomerular filtration rate (eGFR) <30 mL/min/1.73 m2 for safety reasons. This represented only 0.5% of patients referred for contrast procedures in this single-center study. Translating research to practice requires attention to exclusion criteria of studies.

In an accompanying editorial, three authors took exception to the noninferiority trial design.[2] Their argument does not persuade me. I'm no statistician, but primary-outcome rates of 2.7% and 2.6% are essentially the same and 18 vs 0 adverse events are lopsided in favor of no hydration. That's the beauty of elegant simple experiments: you don't need fancy statistics or composite end points to explain the results.

The strengths of AMACING outnumber its weaknesses. A principal strength is its relevance to clinical practice. Investigators enrolled patients who were similar to the population for which guidelines recommend hydration. Three-quarters of the cohort had cardiovascular disease, nearly half were on NSAIDs, and almost one-third had diabetes. And this representative sample underwent varied types of contrast procedures; nearly half had intra-arterial contrast.

The trialists also mirrored good clinical practice by using minimum volume, prewarmed, low-osmolar, monomer, nonionic contrast material. That may lessen the chance of CIN (lower event rates) and thus reduce the benefit-to-harm ratio of hydration. But that's the point of pragmatic and contemporary research. IV hydration may have been useful in another era.

Rethinking CIN

Finally, the most provocative aspect of AMACING is how it prompts us to reexamine the very existence of CIN. Perhaps hydration does not prevent CIN because our way of thinking about CIN is flawed. Nephrology thought-leader Dr Joel Topf (Oakland University, Michigan) describes his change of position on CIN "from being mostly true positives with a small amount of noise to mostly noise with only a few cases of true contrast nephropathy."[3]

To support the notion that CIN could be incidental acute kidney injury, Dr Topf first argues that the definition of CIN as a 25% increase in creatinine is an "admirably" sensitive but "wildly nonspecific" surrogate. He then cites large observational studies that show that after adjustment for risk factors of kidney injury, the incidence of CIN is not different from non–contrast-induced acute kidney injury.[4,5]Putting these ideas together, Topf poses the hypothetical that if we were to exclude all cases of presumed CIN that could have an alternative reason for kidney injury, "I suspect contrast nephropathy would look more like an idiopathic condition."

I began by saying we should pay attention to medical science that delivers humility. Results of the AMACING study force us to 1) be suspicious of expert opinion, 2) object to quality measures not backed by randomized trial data, and 3) reconsider the existence of an entire disease entity (CIN), and in doing so, think about how our brains can trick us into seeing signal when there is mostly noise.

It is said that a good experiment raises more questions than it answers. This one surely does that.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.