The cost of blogging

Michael Hyatt has an a good read about the costs of blogging and advises against getting bogged down in details about design, custom software and hosting. Find a way to cut through that BS inorder to start writing.

He recommends WordPress.com, premium version for $12-17 a year. He doesn’t say why he would choose that over Blogger (the host of PBFluids).

Today I received my bill for blogging for the next year here at PBFluids:

Blogging. Total bargain.

$10. That’s all.

Here is Dr. Ves on Medical Blogging. He is actually quoting Seth Godin and Tom Peters. This best sums up my feelings on blogging:

It doesn’t matter if anyone reads it. What matters is the humility that comes from writing it. What matters is the metacognition of thinking about what you’re going to say.

No single thing in the last 15 years professionally has been more important to my life than blogging. It has changed my life, it has changed my perspective, it has changed my intellectual outlook, it’s changed my emotional outlook.

And it’s free.

And the baby with the baboon heart. Or It’s an iPad world

I was giving my cardiorenal syndrome lecture on Friday and during the question and answer session one of the residents asked why furosemide drips were more effective than boluses. I explained about the results from this Cochrane review and this recent RCT. Unfortunately I had not read the table of contents from this weeks NEJM:
So of course one of the interns mentions the article and asked if I had read it. I copped to the truth but what happened next was incredible. Across the lecture room I could see dozens of iPads flick to life as nearly everyone started pulling up NEJM.org to check out the latest.

Medicine is magical and magical is art
The Boy in the Bubble
And the baby with the baboon heart

Cardiorenal syndrome

On the first Friday of every month I give a lecture to the residents at St. John Hospital and Medical Center. I like to do an electrolyte lecture but for March the chief resident asked me to talk about cardiorenal syndrome. In researching the lecture I came across this article by Claudio Ronco.

The article defines cardiorenal syndrome as any condition with simultaneous kidney and heart failure. He then goes on to subdivide cardiorenal syndrome into 5 types:

  1. Acute heart failure causing acute renal failure
  2. Chronic heart failure causing chronic kidney disease
  3. Acute kidney injury causing any type of acute cardiac dysfunction (including arrhythmia)
  4. Chronic kidney disease causing any chronic cardiac disease
  5. Any systemic condition that causes renal and cardiac dysfuction (e.g. sepsis)

This is terrible. Cardiorenal syndrome used to signify the unique cause of acute kidney injury where the decrease in function is due to apparent volume depletion in a patient that obviously overloaded. It named the only scenario where acute kidney injury responded to diuresis. It was unique and specific. Ronco comes along and says, yes I like your version of cardiorenal syndrome so I will make it type 1 in my new all purpose definition of cardiorenal syndrome. Now whenever there is cardiac dysfunction and simultaneous kidney dysfunction we can just call it cardiorenal syndrome.

It doesn’t have to be this way look at the example of hepatorenal syndrome. The syndrome does not refere to just any situation with simultaneous renal and liver dysfunction. It is a very specific diagnosis that only occurs with chronic liver disease and ascites. The patients must be oliguric, there is no non-oliguric HRS. Patients must be sodium avid and unresponsive to fluids and albumin. Additionally the patients cannot have laboratory or imaging evidence for an alternative cause of renal failure. Because of this definition hepatorenal syndrome identifies a very specific disorder, with a specific pathophysiology and unique prognosis and treatment options.

Ronco takes the beautiful and evocative name cardiorenal syndrome, strips it of all specificity and then tries to restore it by tacking on five different types. The fifth type 5 is the one that makes my brain explode. Sepsis, really? Acute kidney injury from sepsis that happens in the same patient who also suffers from sepsis induced cardiomyopathy should now be considered to have cardiorenal syndrome? Ronco is a man who has spent his life studying sepsis and acute renal failure, I can’t believe he is actually referring to that condition as CRS type 5.

I’m not buying what Ronco’s selling. Cardiorenal syndrome begins and ends with type 1 for me.

FYI: Here is the lecture (Keynote, PDF). It still needs some work. I’d like to add a section on ultrafiltration and I need to include the NEJM article on furosemide that was published yesterday.

Go green: recycle your organs

On of my patients, whose daughter had a kidney transplant, came into clinic wearing this T-shirt. Love it.
It says. “My child contains recycled parts. Be a hero, be a donor.”

New favorite author at Renal Fellow Network

I’m a big fan of the Renal Fellow Network but one of the consequences of the Post-Nate structure, with a large cohort of authors is variable quality. One of the new horses is a first year fellow at Stanford, Graham Abra. He is doing a great job. Take a look at the work he has done this year. Great stuff. His post on alimentary azotemia is about as good a post as I have ever read on RFN. I can’t wait for him to finish his work on Kt/V.

Keep on writing Graham, you’re hitting it hard.

Creatinine, BUN and GFR: part two

Part one focussed on the fact that with a stable creatinine the amount of creatinine produced is equivalent to the amount of creatinine excreted in the urine. Then it showed how the general clearance formula can be rearranged to solve for the serum creatinine rather than the GFR.

The interesting concept, and the original question, is why does the creatinine rise when the GFR falls. Looking at the clearance formula if we decrease the GFR to 45 mL/min and keep the creatinine excretion fixed at 1,400 mg per day, the only way to balance the equation is to increase the serum creatinine.

In summary we have an equation with three variables:
  1. Clearance is the independent variable, and we are setting it at 45 ml/min
  2. Creatinine excreted is fixed at 70 mg/kg or 1,400 mg
  3. Serum creatinine
So if the GFR falls the only variable that can respond is the serum creatinine and in the above example it rises to 2.1 (remember to multiply the calculation by 100 to convert from mg/ml to mg/dL).

The only way for the kidney to excrete the daily creatinine load is to allow the creatinine to rise. The increase in serum creatinine allows the kidney to clear the daily creatinine load.

But this doesn’t really answer why the creatinine rises with a falling GFR. The secret comes from the efficiency of ultrafiltration as the source of clearance. Excluding secretion in other parts of the nephron clearance is provided by filtration at the glomerulus.

Substances filtered at the glomerulus are found in the ultrafiltrate at the same concentrations they are found in the plasma. So a liter of ultrafiltrate will have 140 mEq of sodium and 4 mEq of potassium. These examples should make it clear that ultrafiltration is much more efficient for excreting substances found at high concentration. Americans consume about 180 mmol of sodium a day (4140 mg), this can be cleared with less than 1.5 liters of ultrafiltration. Potassium intake is around 50 mmol per day, clearing this much potassium requires 12 liters of ultrafiltrate. Note: sodium and potassium handling do not depend on ultrafiltration because of extensive reabsorption and secretion that largely overwhelm the effect of ultrafiltration.

Let’s look at the patient at steady state with, 1,400 mg of creatinine production, a GFR of 100 and a creatinine of 0.97. He suddenly loses half his renal function and now has a GFR of only 50 mL/min. Looking at the clearance formula, the only things that changes at first is the GFR. For the first moments after the loss of GFR the serum Cr will still be 0.97. We can solve for amount of creatinine excreted by the kidney at the GFR:

So with a GFR of 50 and a serum creatinine of 0.97, only 698mg, or just under half, of the creatinine created is excreted by the kidneys. It is impossible for the kidneys to clear the daily creatinine load with a GFR of 50 and a serum Cr of 0.97. The 702 mg of creatinine that are not excreted, remain behind and serves to increase the serum the creatinine. If the patient has 60% body water, his total body creatinine initially was 407 mg of creatinine (0.97 mg/dl x 420 dL body water) and the additional retained creatinine will raise his serum creatinine to 2.6 (407mg + 702mg divided by the same 420 dL).

The next day, armed with the higher serum creatinine of 2.6, the same GFR allows the body excrete 1,929 mg of creatinine, more than the daily creatinine load. The resulting creatinine is then 1.4 mg/dl. Ultimately if you carry this calculation forward the creatinine will stabilize at 2.16 mg/dl.

Understanding the equations and calculations is not as important as understanding that higher serum creatinines allow more creatinine to be cleared by ultrafiltration, in fact the only way for the kidney to excrete the same daily creatinine load at lower GFRs is by allowing the serum creatinine to rise.

Think of a rising creatinine as not so much a complication of renal failure but as an adaptation to renal failure.

Remember Nate Hellman

Nate started the most important innovation in nephrology education since NephSAP, the Renal Fellow Network. Nate died, tragically, a year ago this past Sunday. We all stand on the shoulders of giants and Nate passed long before his work was done. In addition to thinking of Nate, we should also thank Matt Sparks and Conall O’ Seaghdha for picking up the pieces and transforming RFN from what was largely a one man show into the institution it has become.

Creatinine, BUN and GFR: part one

Question: What is the most basic concept in clinical nephrology?

Answer: As renal function falls, the creatinine and BUN rise.

For the purpose of this post the renal function is synonymous with glomerular filtration rate.

Think about every lab measurment in clinical medicine and think about how the normal range changes as the GFR falls from 100 mL/min to 10 mL/min, a 90% reduction of renal function.

  • How much does sodium change? 
    • Not at all.
  • How much does potassium change? 
    • In the absence of ACEi or other drugs which alter normal renal handling or an extreme change in diet, it doesn’t change at all.
  • Phosphorous? 
    • Maybe a 25% bump from the low 4s to the mid 5s.
  • White blood cell count? 
    • Not at all.
  • Albumin? 
    • Not at all.
  • Lipase? 
    • Not at all.
  • SGPT? 
    • Not at all.
In the broad world of clinically relevant biochemical tests, essentially none are readily affected by changes in glomerular filtration rate. BUN and creatinine (and cystatin C) stand alone in their exquisite sensitivity to changes in GFR. Of course this is not a weird coincidence, those labs are clincially relevent precisely due to their sensitivity to changes in GFR. But why is it, that as GFR falls, the creatinine rises?
Imagine a 70 kg male. Men, on average, generate 20 mg of creatinine per Kg body weight, so our patient will generate 1400 mg of creatinine and, if the renal function is stable, all of that creatinine is excreted by the kidneys every day. It makes no difference if the GFR is 10 or the GFR is 100, all of the creatinine generated is excreted.

None of it lingers.

None of it accumulates in some creatinine depot in the subcutaneous fat or lateral horn of the cerebral ventricles.

This has to be true because if some of creatinine hung around and accumulated, the serum creatinine would rise. By definition, stable renal function means the creatinine doesn’t rise. So our 70 kg man generates 1,400 mg of creatinine and excretes 1,400 mg of creatinine.

Once you know the amount of a substance excreted we cane solve for the plasma concentration using the standard clearance formula.
Using the 1,400 mg of creatinine, assuming a modest urine output of 1 liter and assuming a GFR of 100 the equation looks like this:
We can rearrange the equation to solve for the plasma creatinine:
and if you do the calculation you get a creatinine of 0.97 mg/dL. The neat part of the equation is that it is totally independent of the urine volume. If the patient excretes the 1400 mg of creatinine in 2 liters rather than 1 as we claculated above, the urine creatinine concentration falls by half (same amount of total creatine dissolved in twice as much urine), the urine volume doubles and the serum creatinine remains the same.
On the other end of the renal function spectrum, the poor patient with a GFR of 10, looks like this:
This gives him a serum creatinine of 9.7 mg/dL. The creatinine went from 0.97 to 9.7 with a change in GFR from 100 to 10. Imagine if any other electrolyte had a ten-fold change associated with a drop in the GFR? Raise your hand if you have seen a potassium of 40.

President Bush is 5’11”

You can use this spreadsheet below to predict the serum creatinine based on different GFRs, urine volumes and creatinine production. Try different urine volumes and see how that doesn’t affect the serum creatinine (the reason is that the numerator in the clearance formula is simply solving for the mass of creatinine excreted. Concentration of X multiplied by the volume gives amount of X dissolved in the solution.) . Use the spreadsheet to discover what Shaquile O’Neal’s serum creatinine is. Assume 20 mg/kg body weight, a weight of 147 kg, and a GFR of 120.

If you want to edit and use the equation image files download this Word file. Double click the equations to launch the equation editor.

Great site with lots of resources on hypertension

The Michigan Department of Community Health has put together a great resource, High Blood Pressure University, which links to resources all about hypertension. The links are divided into three campuses:

  1. Professionals for medical providers
  2. Community for church’s and community organizations
  3. Patient for people with high blood pressure
Take a look at the Vital Signs fact sheet from the professional campus. This is from the CDC and summarizes some of the broad epidemiology of lipids and hypertension.