2+ Year Member
Jan 13, 2015
Resident [Any Field]
A few months ago, I learned about the updated ACR recommendations regarding contrast-induced AKI (see below). However, when I attempt to convince my attendings/coresidents to be more liberal with their use of contrast I'm usually met with an "oh, precious" look and polite rejection. Has anyone had better luck? If so, any suggestions?

From the manual on contrast media:
"Unfortunately, very few published studies have a suitable control group to permit the separation of contrast-induced nephropathy from post-contrast acute kidney injury. At the current time, it is the position of the ACR Committee on Drugs and Contrast Media that CIN is a real, albeit rare, entity. Published studies on CIN have been heavily contaminated by bias and conflation."

"There is no agreed-upon threshold of serum creatinine elevation or eGFR declination beyond which the risk of CIN is considered so great that intravascular iodinated contrast medium should never be administered. At the current time, there is very little evidence that IV iodinated contrast material is an independent risk factor for AKI in patients with eGFR >30. Therefore, if a threshold for CIN risk is used at all, 30 seems to be the one with the greatest level of evidence."

...and a related article:
"AKI temporally related to contrast material administration, a common occurrence in hospitalized patients, is often blamed on contrast material and not on one of potentially many other coexistent factors."

"Almost all prospective and retrospective studies investigating the nephrotoxic potential of iodinated contrast media have failed to disentangle contrast-induced AKI from post-contrast AKI, leading to confusing and sometimes uninterpretable results."

"...a series of recent large-scale (>10,000 patients each) propensity-adjusted studies have assessed the risk of costrast-induced AK. Each of these studies shows with excellent power that the per-patient risk of contrast-induced AKI after IV contrast administration is either rare or non-existent for patients with a stable eGFR of >45."

"However, the studies reached different conclusions for patients with an eGFR <45. Specifically, those with an eGFR of 30-44 were determined to be at either borderline increased risk or no risk, and those with an eGFR <30 were determined to be at either substantially increased risk or no risk."

"Assuming the worst-case point estimates among these studies, the number needed to harm would compute to 39 contrast administrations for one case of contrast-induced AKI in patients with an eGFR of 30-44 and to 6 contrast administrations for one case of contrast-induced AKI in patients with an eGFR of <30."


Destination darkroom
10+ Year Member
Jun 25, 2006
The potassium mines
Fellow [Any Field]
CIN is one of those things where you are trying to protect against a worst-case scenario, even one that is purely hypothetical. At my institution, we routinely give iodinated contrast for patients with GFR in the 30s to 40s. Often, people will instruct the techs to either bolus some saline or encourage PO fluids if they are an outpatient, but there is no evidence to support this as anything other than making you feel like you did something. Similarly, there is no good evidence that reducing the amount of contrast injected helps, but people will often do this for no good reason. Sometimes, contrast isn't necessary to answer the clinical question. Many times, the clinical question may be more appropriately studied with MRI or ultrasound, in which case giving iodinated contrast would be a moot point and not worth the potential risk. A lot of it probably has to do with modern contrast agents being less osmotically active than older ones, but much of the classic teaching is by people who were trained using old contrast agents. Dialysis patients are easy because as long as they are scheduled for HD, you can inject with impunity.

CIN is akin to NSF with gadolinium-based contrast agents, though the latter is probably even more of a unicorn since there are basically no documented cases of NSF in recent years.


Knowledge worker
2+ Year Member
Feb 10, 2016
Resident [Any Field]
More than one time I've seen someone order a noncontrast scan that would be suboptimal for the clinical question (eg, locate source of hemoptysis, many abdominal CTs) just because "we don't want to box the kidneys" of a patient with slightly elevated creatinine. The way people speak about it is like CIN is almost a given, rather than a worst-case hypothetical. People think of a number needed to harm of 3 rather than 39.