Misplaced concreteness

Nephrology, more than other specialties is plagued by misplaced concreteness. We get false senses of precision because of the myriad of equations that spit out results to the milliliter. All of those equations from Kt/V, to water deficits, to IVF brain teasers depend on an estimate of total body water.

Everyone knows the rule of thumb that young males are 60% water, young females are 50% water and the percent body water falls as people age or get fatter.

Going beyond these rules of thumb, how is total body water measured empirically? The gold standard is heavy water dilution.

This works by giving a sample of heavy water and then waiting for it to equilibrate. Then the investigators measure the heavy water content of exhaled water vapor or a blood sample, the fraction of the water that is heavy water will be equivalent to the fraction of total water which is heavy water. Then since one knows the amount of heavy water given to the patient, one can calculate total body water.

When this is done, or when one reviews the primary literature, as was done in this study the numbers are a little different.

The drop in total body water (in red) in men never gets down to 50% as predicted for the elderly, and in women almost all of the numbers are below 50% and the trend to lower percentages through aging holds if you ignore the 9 women over the age of 70. Of note, these patients are not that obese, see BMIs in blue.
The above study lead to the development of the Watson equation to determine total body water. The Watson equation uses age, weight and height for men and height and weight for women:
  • Males: TBW (in liters) = 2.447 + (0.09156 × age) + (0.1074 × height) + (0.3362 × weight)
  • Females: TBW (in liters) = –2.097 + (0.1069 × height) + (0.3362 × weight)

This study in peritoneal dialysis patients (peritoneum empty) showed surprisingly close relationship (R=0.92) between deuterium dilution and the Watson equation:

Another study of PD patients demonstrated one of my pet peeves, the major effect of obesity has on total body water.

The chart is a bit difficult to understand. The Y axis shows the Watson calculation of total body water minus heavy water dilution. So negative numbers indicates cases where the Watson method underestimates TBW. When the Y-axis is positive the Watson calculation overestimates TBW. The X-axis expresses obesity as body fat over body water. Really? fat over water. You couldn’t just graph this versus BMI?

IV Fluid Brain Teaser: Salt versus Saline

Everyone knows that if you give a liter of saline, all of it remains in the extracellular compartment.

But what if you give a patient just the salt from the saline and none of the water? How much does the solute contribute to the increase in the extracellular volume? How does 154 mmols of NaCl affect the size of the extracellular and intracellular compartments?

Assume the patient is a 70 kg lean young male with a serum osmolality of 280 mOsm/kgH2O. Ignore any renal losses during the process.

For full credit fill out the following:

Total body water:
Size of the extracellular compartment:
Size of the intracellular compartment:


Step one calculate the total number of osmoles the patient has:

70 kg lean young male means 60% total body water or 42 liters
42 liters times 280 mOsm/Kg = 11,760 osmoles in the body

Giving 308 mosm of solute will increase that to 12,068. There is no additional water so dividing that by 42 liters gives us a new osmolality of 287 mOsm/Kg water.

Remember that even though the solute is trapped in the extracellular compartment, the osmolality is the same across all body compartments since water can flow from compartment to compartment.

Now we need to find out how much the extracellular compartment expands in osmoles.

Before the addition of solute the extracellular compartment should be one third of total body water, so 14 liters times osmoality of 280 is 3920 mOsmoles. Add 308 and then divide that by the new osmolality to give you the new volume:

That increased volume of course comes from the intracellular compartment, so it goes down by 0.7 liters. You can also get there by taking the original volume of 28 liters multiplying by 280 mOsm to get 7840 miliosmoles and divide that by the new osmolality of 287:
So the addition of 308 miliosmoles from the bag of saline will increase the extracellular compartment by 0.7 liters. Only 0.3 liters less than the increase you would get with a liter of 0.9 NS. 
It’s all about the salt.

Total body water: 42 liters
Size of the extracellular compartment: 14.7 liters
Size of the intracellular compartment: 27.3 liters

Please fund my #DreamRCT, it is just embarrassing how little evidence is found in hyponatremia

So I checked in at DreamRCT and noticed that my DreamRCT is no longer in the top five.

I’m a big boy and can take this (very minor) form of rejection but I do want to plead my case for a moment. You can read the entire description of my DreamRCT here. One of the dirty little secrets of nephrology is the almost total lack of prospective data on hyponatremia. There are a number of RCTs with regards to tolvaptan, conivaptan and other approved, and soon to be approved, vaptans. But after those there is an evidence desert populated by only a few mirages made up of case reports and retrospective analysis. From this scant data we have built a comprehensive and detailed model of how sodium acts in the body and the importance of osmoregulation. But thats like theoretical physics without a supercollider. We need to test the model with real data.

Think about the fact that hyponatremia is the most common electrolyte disorder. We order metabolic profiles on every patient, every day, but when it comes to interpreting those results we might as well be reading hieroglyphics.

RCTs are difficult and expensive but there are particular areas where we should require them prior to treatment. One of these corners is when we treat people with no symptoms and we are effectively treating a number. High cholesterol, high blood pressure, and low sodiums are three such areas, however hypertension and hypercholesterolemia have both gone through the right of passage called a randomized controlled trial. I am confident that my patients with asymptomatic hypertension benefit from treatment. My patients with coronary disease and hypercholesterolemia will live longer and better with treatment with a high potency statin. On the other hand, patients with sodiums of 129 and no apparent symptoms are supposed to be at higher risk of falls, have a higher mortality from heart and liver failure. Does treating them reduce these risks?


We can do better. Hyponatremia is the most common electrolyte disorder found in patients, we owe it to them to have real, prospective, data to answer these questions.

Go to UKidney to vote for my trial: No hyponatremia modification in asymptomatic hyponatremia. Thanks.

DreamRCT begins

DreamRCT is a creative writing project for nephrologists. The assignment is to scour the landscape of nephrology knowledge for a corner that is dominated by dogma and retrospective evidence. Once the target is identified, the writer needs to summarize the gaps, and think up a creative way to shine science’s greatest flashlight on the subject, a randomized controlled trial.

We have recruited 16 people to submit DreamRCTs which were published today on MedPage Today. Thanks Ivan, Kristina and Elbert. It is a amazing collection of creativity; there are trials on kidney stones, electrolytes, dialysis, proteinuria and lupus. Please go check them out; read them and see which are great and which should be relegated to The Journal of Craptology.

After reading the DreamRCTs move on over to Jordan Weinstein’s excellent UKidney where it is time to channel your inner Mark Cuban and play Shark Tank with the DreamRCTs.

Which trial should be funded which shouldn’t. How much should each trial be awarded. You will get $100,000 to distribute among the trials. Think KickStarter meets NIH. You will not be alone in this endeavor. We have recruited an expert panel of clinical researchers to score the trials. At the conclusion of the contest we will look at how the experts spent their cash and how the crowd did. We will also award a small prize to the funder whose distribution best matches the expert panel.

DreamRCT only works if we get a critical mass of people to participate. Please check out the trials at MedPage Today and then go to UKidney to vote with your (completely virtual, don’t ask me for a refund) dollars. Announce the project at morning report, assign your fellows to vote and then submit discuss and submit your own DreamRCT, because in the end DreamRCT is not just a game but a shorthand expression for what we need to do to fix nephrology and heal our patients.

#OUWB Renal question: The bad quiz question

Apparently there was a weekly quiz and one of the questions was as follows:

And I received an e-mail asking e to answer this question. Lets go through it item by item. The stem sets up a patient with diarrhea induced metabolic acidosis. This is a cause of non-anion gap metabolic acidosis due to GI loss of bicarbonate.
Choice A. This is wrong. The filtered load of bicarbonate is dependent solely on the plasma bicarbonate concentration. The lower bicarbonate concentration seen in all metabolic acidosis would cause decreased not increased filtered load.
Choice B. This is right. Ammonia is produced in the proximal tubule in response to metabolic acidosis and hypokalemia. This is why ammonium excretion is able to be up-regulated due to an acid load. Titratable acid is fixed and can’t accommodate an increased acid load. The ammonia production varies depending on metabolic need and chronic diarrhea would up-regualte ammonia production so it could be converted to ammonium in the medullary collecting duct to help clear the excess acid load.
Choice C. This is right. Hydrogen secretion is increased in the distal nephron in response to the metabolic acidosis. This is needed to replace the bicarbonate lost in the stool. Every hydrogen in secreted int eh distal nephron synthesizes a new bicarbonate molecule for the body.
Choice D. This is wrong. Diarrhea causes a non-anion gap or normal anion gap metabolic acidosis.
Choice E. This is right. Hydrogen secretion in the distal nephron is stoichiometrically equivalent to producing new bicarbonate, one cannot happen without the other.
Some of the social media action around this question:
What am I missing here? It looks like B, C and E should all be correct. #NephPearls pic.twitter.com/ztg94B4mR5

— Joel Topf (@kidney_boy) August 24, 2015

@kidney_boy @Nephro_Sparks @hswapnil Nothing. Questions fail. Perhaps author trying to write an “all are true except” question.

— Graham Abra, MD (@GrahamAbra) August 24, 2015

@S_brimble @kidney_boy C leads to E Both are D/T B

— د علي السهو (@alialsahow) August 24, 2015

@kidney_boy If the distal nephron still commences with the TALoH and ends with the papillary collecting duct, B-C-E are all right

— Fra Ian (@caioqualunque) August 24, 2015

@alialsahow @NephJC I agree with you, B, C and E are all correct.

— Joel Topf (@kidney_boy) August 24, 2015

@GrahamAbra @Nephro_Sparks @hswapnil and we wonder why students find nephrology so hard.

— Joel Topf (@kidney_boy) August 24, 2015