Last week I wrote this tweetorial
I had been thinking about this for a long time.
When you hear #TWDFNR talks about AKI the two things that are usually hammered are:
- The uselessness of the FENa and other urinary excretion indices.
- The low yield of the kidney ultrasound.
I actually like both of those tests, despite the pundit class dumping on them.
The FENa is vilified because in trials that look at its ability to separate out purely hemodynamic decreases in GFR from acute tubular necrosis it does poorly. This is especially disappointing because all of the board exams and med student level lectures on FENa says this is just what the test is all about. I concede that FENa in isolation is a bad test to determine the etiology of AKI. But as part of the global assessment of the patient, getting a FENa can be incredibly powerful.
Saying the FENa is useless is like testing to see if a blinded physician given only an aucultory exam could diagnose community acquired pneumonia. I think there is a good chance auscultation would fail that hurdle. But that doesn’t mean we should abandon our stethoscopes. The auscultory exam is one of a number of studies we do when assessing a patient who may have pneumonia. It is a valuable part of the global patient assessment.
Similarly, the FENa is part of my global assessment of a patient.
For example, I have a patient with cirrhosis and acute kidney injury. I look at the urine and see a mixture of hyaline casts and granular casts. Now, you need to be careful about granular casts in cirrhosis. As the bilirubin climbs, it can stain innocent hyaline casts to look brown, and they can start to look somewhat like muddy-brown casts. I don’t think that is what is happening on this slide, because I can see simultaneous dirty brown and hyaline casts. Then we checked the urine sodium and it was 50. This is off diuretics. This is remarkably high for a patient with cirrhosis and ascites. But there it is. The urine sodium is high and in agreement with U/A findings. I make a diagnosis of acute tubular necrosis.
It’s an N of 1, but that’s all we have in clinical medicine, one N of 1 study after another.
The kidney ultrasound story is a much simpler than the FENa story. The argument that kidney ultrasound is of low yield and low cost effectiveness is due to the exrtemely low yield of U/S to find obstruction. In almost every case that you find obstruction, you suspected obstruction, and using it in cases without a clinical suspicion is like Acestes aiming at a non-existent target.
But the thing is, the treatment of obstruction is not dialysis, and if you fail to identify and correct the obstruction, no mixture of IV fluids, avoiding nephrotoxins, and regulating blood pressure will fix the obstruction. That obstruction will lead to renal failure and dialysis. In a world of $60,000 a month maintenance therapies, I refuse to miss even one easily correctable (but otherwise irreversible) cause of kidney failure.
The real #TWDFNR in AKI are:
- Urinary protein to creatinine ratio
- Intact PTH
The fallacy of the Protein Creatinine ratio was the subject of the afore mentioned tweetorial.
Some great comments and discussion from the tweetorial:
The problem with PTH in AKI is that it is useless. I see fellows and attendings ordering this and I have no idea what to do with it. Some people try to use the KDIGO PTH guidelines for ESKD. This seems to be totally evidence-less. PTH is an acute phase reactant in AKI. Part of the AKI syndrome is a drop in calcium, and after that PTH rises, just like it is supposed to. I don’t see any reason to suppose this secondary hyperparathyroidism is pathologic. I have seen some people order it to try to distinguish acute from chronic CKD. The reasoning being, that chronic kidney disease would have a high PTH and acute would not. This is not the case both acute and chronic kidney disease can have secondary hyperparathyroidism.
Stop ordering PTH in AKI. Stop checking urinary protein to creatinine ratio and start doing urinalysis and microscopy.
The Y-axis is “Impressions.”This is not impressions like Symplur does (used to do?). This is not the tweets multiplied by the number of followers. Twitter is in the unique position to know how many times any particular tweet is delivered to a device. So your cousin who lost her twitter password in 2014 and the sock puppet account that Eugene Gu abandoned in medical school don’t get counted as impressions. The first Tweet had 47,000 impressions. The second had 5,700. That first step is a doozy.From there things were surprisingly stable. 
More than 3,000 people read pretty much the entire stream. I am quite satisfied. 3000 people is a lot of grand rounds.
