[JohnnyFever]

ACR needs to publish solid guidelines that we can adhere to and lean upon legally

[mario.mtz30_447]

They do:
 
[link=https://www.acr.org/-/media/ACR/Files/Clinical-Resources/Contrast_Media.pdf]https://www.acr.org/-/media/ACR/Files/Clinical-Resources/Contrast_Media.pdf[/link]
 
 

[JohnnyFever]

The last time I read that, it was a wishy-washy lit review without taking a concrete stand on anything. Is it better now?

[mario.mtz30_447]

It does seem more definitive now:
 
At the current time, it is the position of ACR Committee on Drugs and Contrast Media that CI-AKI is a real, albeit rare, entity.
 
 
 
 

[toumeray]

The ACR recommendations are fairly wishy washy still according to contrast manual. Im unable to copy paste on my phone but basically they say the risk is real and should always be considered depending on potential benefits of the test, regardless of creatinine level including normal creatinine. They dont come down hard on a threshold but say there is little evidence for it above 30 gfr so that seems like a logical threshold, but dont specify a threshold of what. Above 30 is of no concern, below 30 a potential concern? Or above 30 is of low, but real, concern, below 30 of high concern, only for acute life threatening pathology? Very little actionable guidelines in the document. Its on pages 42-43

In fairness they are wishy washy because the data is contradictory. Yes there are several large studies showing no increases risk of AKI but here is the one they cite showing a statistically significant risk for gfr<30:

This is a large scale randomized study also using the 1:1 stratification of other large studies.
[link]https://pubmed.ncbi.nlm.nih.gov/23579046/[/link]

[toumeray]

To be clear I do think the risk is overstated and data on past studies have been clearly biased and confounded. But I do think there is a real risk for patients with late stage ckd. How significant is that risk is hard to say. I think in the age of very low yield imaging we have to take into account that many of the tests that are ordered are simply of no benefit to the patient. So any risk of contrast injury, even small or potential, is often enough to justify a noncon scan, at least for me. The purpose and utility of the test is of paramount importance here

[sorana12]

I agree with BlueDeep. This article is bad science with too few patients included and flawed hypothesis. Notice its not published in a highly repsected journal

To add to BlueDeep's post, there's another study that Dr. McDonald at Mayo published around 2013 or 2015 that included > 50,000 patients (with > 100,000 scans). It showed no causal link between IV contrast administration and renal insufficiecy. The study was well set up and controlled for multuple patient factors. The power of a study with 50k patients vs this study AuntMinnie is referencing is laughable.

Here's a link to the Mayo study: [link]https://pubmed.ncbi.nlm.nih.gov/23360742/[/link]

When I teach medical students, I tell them that IV contrast is very unlikely to cause kidney injury. Like any medication, it can happen.

[JohnnyFever]

But until the ACR, RSNA, ARRS make some definitive statement, we’re going to have nervous radiologists saying no to contrast. And despite single trials suggesting the risk is infinitesimally low, it will continue to be challenging to implement department wide change

[Unknown Member]

This is taken from the RANZCR guidelines:
 
Recommendations
R4. Intravascular iodinated contrast media should be given to any patient regardless of renal function status if the perceived diagnostic benefit to the patient, in the opinion of the radiologist and the referrer, justifies this administration.
 
R5. Emergency imaging procedures requiring contrast media administration e.g. acute stroke, acute bleeding, trauma etc. should not be delayed in order to obtain renal function testing results prior to the procedure.
 
R6. The risk of intravenous contrast media related acute kidney injury (CI-AKI) is likely to be non-existent for patients with eGFR greater than 45 mL/min/1.73m2 .No special precautions are recommended in this group prior to or following intravenous administration of iodinated contrast media.
 
R7. The risk is of intravenous CI-AKI is also very likely to be low or non-existent for patients with eGFR 30 – 45 mL/min/1.73m2 . Universal use of periprocedural hydration in this group to prevent the theoretical risk of CI-AKI cannot be recommended but patients with impaired function in this range that is acutely deteriorating rather than stable may benefit from this intervention.
 
R8. In patients with severe renal function impairment (eGFR less than 30 ml/min/1.73m2 ) or actively deteriorating renal function (acute kidney injury) careful weighing of the risk versus the benefit of iodinated contrast media administration needs to be undertaken. Consideration should be given to periprocedural renal protection using intravenous hydration with 0.9% saline (see relevant section). However, severe renal function impairment should not be regarded as an absolute contraindication to medically indicated iodinated contrast media administration.

[mario.mtz30_447]

Quote from RoleCall

But until the ACR, RSNA, ARRS make some definitive statement, we’re going to have nervous radiologists saying no to contrast. And despite single trials suggesting the risk is infinitesimally low, it will continue to be challenging to implement department wide change

Not sure I agree. There is a risk to doing anything. The nervous ones could review the current literature. Or a group member could come up with a protocol after reviewing the literature.

[khodadadi_babak89]

careful weighing of the risk versus the benefit of iodinated contrast media administration needs to be undertaken.

in the real world, this is a useless statement.

We also need to consider that we are functioning in our own little echo chamber here, and other physicians need to be told. They don’t get it, They just heard something on rounds once that contrast could kill kidneys, and so they order C- studies  

[toumeray]

Not to belabor this point but I mentioned previously the two Davenport articles, which are also cited by ACR contrast manual to justify the somewhat wishy washy recommendation for <30GFR
 
[link=https://pubmed.ncbi.nlm.nih.gov/23360737/]Contrast material-induced nephrotoxicity and intravenous low-osmolality iodinated contrast material – PubMed (nih.gov)[/link]
 
[link=https://pubmed.ncbi.nlm.nih.gov/23579046/]Contrast material-induced nephrotoxicity and intravenous low-osmolality iodinated contrast material: risk stratification by using estimated glomerular filtration rate – PubMed (nih.gov)[/link]
 
Since no one responded previously I will just summarize the results: 
“Intravenous low-osmolality iodinated contrast material had a significant effect on the development of post-CT AKI for patients with pre-CT SCr levels of 1.6 mg/dL (141.44 μmol/L) or greater ”
 
“IV LOCM is a nephrotoxic risk factor in patients with a stable eGFR less than 30 mL/min/1.73 m(2), with a trend toward significance at 30-44 mL/min/1.73 m(2). IV LOCM does not appear to be a nephrotoxic risk factor in patients with a pre-CT eGFR of 45 mL/min/1.73 m(2) or greater.”
 
Essentially, both trials proved a statistically significant and independent risk associated with contrast for the development of AKI.  Both were high powered, large scale trials (~17-20k patients).  Both used propensity 1:1 matching that was used by MacDonald to effectively eliminate bias (a necessary change).  In short, they followed more or less the same protocol as MacDonald and others did (studies which showed no risk at any GFR), but their results were the opposite at a <30 GFR (showing some risk for patients <30GFR).
 
Effectively ACR was presented with 4 very large well run and minimally biased trials, 2 of which saying no risk at any GFR, and 2 saying yes risk at <30 GFR.  So they have wishy washy recommendations as a result.
 
I believe the risk is real for <30 GFR and we do need to consider the benefits of the test.  It doesn’t have to be a hand wringing analysis of whether or not you need contrast for a CT CAP staging CT.  If the onc wants it, they probably need the contrast, but you can’t make a provably wrong decision either way.  I would withhold for GFR<30 diverticulitis or appy work up.  I would certainly withhold for objectively stupid orders like CT AP c+ for renal stone.  

[farzadahmadimedrn710_43]

They showed that people who got contrast have sicker kidneys than people who didn’t get contrast. Not that the contrast damaged their kidneys. The very fact the patient was given contrast with GFR <30 means the treating team thought the patient had to be super sick and as worth the supposed “risk” of contrast. You can’t get rid of that confounding factor.
 
Ask yourself: why are there zero treatments for “contrast induced nephropathy” that have ever shown benefit in a randomized control trial?
 
Because CIN isn’t real and you’re treating a ghost disease.
 
ACR has never cited [link=https://www.mayoclinicproceedings.org/article/S0025-6196(15)00468-1/fulltext]https://www.mayoclinicpro…96(15)00468-1/fulltext[/link] in their contrast statements. Why did they ignore the strongest evidence to date that CIN is a **** diagnosis?
 

[toumeray]

CIN was massively overstated because of improper study design and significant statistical bias, which has been proven over and over.  But it likely is a real, albeit quite rare, phenomenon.  It probably only exists for GFR<30.  But there is a real, small, nephrotoxic effect of contrast.
 
The propensity 1:1 matching attempts to match patients with similar risk profiles (ie, CHF, GFR 26, pneumonia patient gets a contrast scan, matched to a CHF, GFR 26, with pneumonia patient who gets a noncontrast scan).  This is exactly what McDonald did in 2008 to prove the bias in older articles. 
 
You absolutely can get rid of the confounding factor, using the propensity matching.  That is exactly what McDonald did to prove no increased risk from contrast at any GFR.
 
Davenport also used the same method, except they got different results, showing some risk at <30GFR
 
ACR cited at least 3 articles from McDonalds team:
7. McDonald JS, McDonald RJ, Carter RE, Katzberg RW, Kallmes DF, Williamson EE. Risk of intravenous contrast material-mediated acute kidney injury: a propensity score-matched study stratified by baseline-estimated glomerular filtration rate. Radiology 2014;271:65-73.
8. McDonald JS, McDonald RJ, Comin J, et al. Frequency of acute kidney injury following intravenous contrast medium administration: a systematic review and meta-analysis. Radiology 2013;267:119-28.
9. McDonald RJ, McDonald JS, Bida JP, et al. Intravenous contrast material-induced nephropathy: causal or coincident phenomenon? Radiology 2013;267:106-18. 10. Newhouse JH, Kho D, Rao QA, Starren J. Freq
 
McDonald has published numerous articles all with similar methods and conclusions, just in slightly different settings ie different patient populations.  They all say pretty much the same thing, no increased risk from contrast.  You’re article also says the same thing, and honestly doesn’t add much information. 
 
They could add a fourth reference and reach the same conclusion.  Two large research teams with similar methods and conflicting results.  One team showing no risk at any GFR, one team showing risk at <30 GFR
 

[farzadahmadimedrn710_43]

The risk profiles aren’t the same despite the propensity matching. A patient with GFR <30 that they decide to use contrast on and a patient with GFR < 30 that they decide to not use contrast on are different levels of sick. It takes a lot to get nervous IM/ED docs to override their training that contrast is bad to order a contrast study for a CKD patient. The less sick ones got the non-contrast study. It’s a variable you can’t just easily harvest from the raw data of EPIC charts to make a giant retrospective study like a ICD code such as pneumonia, age, gender, etc. 
 
What Davenport did is segregate sick from less sick patients and then showed that sicker patients have more AKI. Not that contrast causes AKI.

[toumeray]

I guess we just disagree on whether or not the propensity scoring method works, or what papers used a method that works.  I think if the risk profiles were as different as you suggest this would come through in many variables such as length of hospital stay, length of ICU stay, additional CECT’s within 5 days of index CT.  However the propensity score system by definition ensures all of these variables are similar between the two groups, which suggests they are of similar “sickness” levels.  
 
To be clear the article you cited [link=https://www.mayoclinicproceedings.org/article/S0025-6196(15)00468-1/fulltext]Risk of Acute Kidney Injury, Dialysis, and Mortality in Patients With Chronic Kidney Disease After Intravenous Contrast Material Exposure – Mayo Clinic Proceedings[/link] also uses a propensity score system which mostly contains the same covariates but with some additional ones.  I take it you do not believe this propensity score system to be biased, only the Davenport article?  Why is the McDonald propensity score system better?

[toumeray]

Bumping this. I did not mean the questions I asked in my last post to be antagonistic in any way, I am genuinely interested in the metanalysis/ detailed review of the methods of these 2 articles re Davenport vs McDonald. Based on my review of their publications, they seem to be in a tit for tat rebuttal to each others work, so is one better than the other at isolating the effects of IV contrast and nephrotoxicity in GFR<30? If so, why?