Patients, that doctors see as high risk for AKI, get shifted away from IV contrast. So the “no contrast group” is enriched with high risk patients. The ideal way to deal with this is a randomized controlled trial. But there isn’t one. The next best thing is propensity scoring.pic.twitter.com/wFOFTgst0F
-
Show this thread
-
Individual patient have variable propensity for developing AKI following contrast. Patients with low GFR have higher risk. Patients with diabetes have higher risk. We can give every patient a propensity score based on that risk.pic.twitter.com/0eUGmfr2XK
1 reply 1 retweet 4 likesShow this thread -
The beauty of randomization is the risk gets balanced between groups.pic.twitter.com/22TVCsIySx
2 replies 1 retweet 7 likesShow this thread -
The problem with observational controlled trials is clever doctors keep trying to avoid contrast in their high risk patients, so that means the “No contrast” group is fortified with an additional helping of high-risk patients.pic.twitter.com/tYlISBgj9N
1 reply 1 retweet 7 likesShow this thread -
To fight this, statisticians do propensity matching where both sides get the same number of patients from every risk category.pic.twitter.com/XyPuIcEIRs
1 reply 1 retweet 5 likesShow this thread -
When this was done for contrast, the protective effect of contrast went away but there was no detectable excess risk of AKI from contrast. If contrast causes AKI it does so, so rarely as to be undetectable by the most sophisticated analysis available.pic.twitter.com/KiIK76HXsh
1 reply 7 retweets 21 likesShow this thread -
In the conclusion of Hinson’s propensity scored analysis, the authors wrote (https://www.annemergmed.com/article/S0196-0644(16)31388-9/abstract …)pic.twitter.com/wwMPvoBJWF
1 reply 2 retweets 7 likesShow this thread -
GE saw this and started a protocol to randomize patients. The indication was monitoring endovascular repair of abdominal aortic aneurysms. Patients with aortic stents go for yearly imaging, either CT w/o contrast + U/S or contrast CT. There is equipoise in imaging modality choice
1 reply 1 retweet 5 likesShow this thread -
The protocol can be seen at http://clinicaltrials.gov NCT03119662 but it was abandoned due to insufficient recruitment. It turns out that even though contrast researchers have equipoise, the wider physician community does not.pic.twitter.com/tn6bvabFCS
2 replies 1 retweet 6 likesShow this thread -
So without a randomized clinical trial this is what we are left with. The highest quality epidemiology studies are unable to detect any increase in AKI, so those bumps in creatinine following contrast are almost certainly run of the mill AKI and not contrast induced AKI.pic.twitter.com/RZztPZTj7Q
7 replies 14 retweets 56 likesShow this thread
any insight from these studies into the effect to which Non-con patients might have done worse because of delay or missed diagnosis?
-
-
Replying to @grepmeded
Joel Topf, MD FACP Retweeted Joel Topf, MD FACP
Here is the best guess on this: https://twitter.com/kidney_boy/status/1199719247250952192?s=21 …https://twitter.com/kidney_boy/status/1199719247250952192 …
Joel Topf, MD FACP added,
0 replies 2 retweets 0 likesThanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Replying to @grepmeded @kidney_boy
I'm sure everyone can think of some incidents from previous experience. I certainly can. But the plural of anecdote is not data.
0 replies 0 retweets 0 likesThanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.