Things have been coming in pairs: electrolyte free water and hypernatremia

First we had the highest creatinine followed by the lowest creatinine.

Then we had a case of hyponatremia/SIADH that we evaluated with the concept electrolyte free water and now we have a case of hypernatremia that we also evaluated with electrolyte free water.

I have a special fondness for dysnatremia formulas that work with either hyper- or hyponatremia because there is an elegance in using a model that works at both extremes.
Both the change of sodium formula and electrolyte free water calculation work as well with hypernatremia as they do with hyponatremia.
The case: 56 year old African American nursing home resident with a history of bipolar disease. She presents with altered mental status and initial labs reveal acute kidney injury and hypernatremia.
Body weight 70 kg
Na 177
Cr 4.18
On exam the patient had obvious hypovolemia and the elevated sodium reveals dehydration. The primary team appropriately uses half normal saline to correct both of these deficiencies.
The sodium came down to 167 the following day and after that they have been unable to further correct it. The patient remained in the ICU for 4 days prior to us being consulted for persistant hypernatremia. The creatinine rapidly corrected over four days to 1.6.
Additional data:
Admission urine sp grav 1.011
Admission urine Na 10
Admission urine osmolality 330
On the fourth day of the admission the urine output was 1,500 over the prior 8 hours
The primary team had been using the water deficit formula to estimate the amount of fluid to give the patient to correct the sodium over 2 days:
The water deficit equation asks how much water will be needed to dilute all the solutes to some ideal. In many books and programs the ideal sodium is fixed at 140 mmol/L. I usually use 145 mmol/L, which is the least amount of change in sodium to get a normal sodium. I do this because the downside of correcting sodium is all on the side over correction and inducing cerebral edema.
The equations often wont let you determine the % body water and fill in 60%. This is a large source of error because in the United States we’re all fat and fat people are relatively anhydrous. Also, since typical internal medicine patient is about 100, and old people are likewise anhydrous 60% is a over estimation of total body water. My fellow estimated the total body water to be 35 liters.
The calculated free water water deficit is 7.7 liters. I usually administer half of that in the first 24 hours unless the sodium is very high, as in this case, and then I would have given about a third of the volume in the first day and correct the sodium over three days. I shoot for a change of about 12 mmol/l per day.
The equation worked well initially with the sodium going from 177 to the mid 160’s but after that they stalled. The reason the sodium stopped improving was that they cured the patient of her renal failure and the urine output increased. Neither the team nor the water deficit formula accounted for this.
We can account for the urine output by replacing the electrolyte free water clearance:
So one would have to add 2.8 liters of electrolyte free water to the free water deficit calculation to account for urinary loss of electrolyte free water. If one were to recalculate the water formula using the partially corrected sodium of 167 you get:
If you plan to give half that in the first day that is 2.65 liters. Compare that to the electrolyte free water loss of 2.8 liters and it immediately becomes obvious why the sodium remained stable for days.
Teaser: An elevated electrolyte free water clearance in the presence of hypernatremia is presumptive evidence of diabetes insipidis. I will save that discussion for my next post.

Lecture at Providence Hospital on Electrolytes

I am trying to do a monthly lecture for the Providence internal medicine residents on electrolytes. I gave my second one last Friday. It was an interesting case we had of hypernatremia on the consult service last summer.

I did this lecture in Keynote and I am blown away by how good it presents through SlideShare. Really impressive.

iPhone Medical Applications

I have four medical applications on my iPhone, of which I use two. Here is a quick review.

To show how the iPhone equipped physician approaches clinical problems I will use the DB’s Medical Rants most recent acid-base problem. He presents a case with the following information:

49-year-old man, previously in good health, presents after a few weeks of progressive weakness and dizziness. He admits to polyuria. Your job is to extensively discuss his lab tests.

The first step in my mind is to fully interpret the ABG. To do this we will use the application ABG.

ABG

This simply named program is an ABG calculator that runs through the standard algorithms for detecting multiple primary acid-base abnormalities. Can’t remember Winter’s Formula. As long as you don’t have boards coming up you can just plug’n chug and turn DB’s ABG into the following:

This does two of the calculations that DB describes at length:

  1. Winter’s formula (16 * 1.5 + 8 ±2) shows that the predicted pCO2 is 30-34. The patient’s CO2 is 33 so the patient has isolated and appropriately compensated pCO2 of 33. ABG displays this information in the second line when it describes the acid-base disorder as “Compensated metabolic acidosis.” It does not describe a second primary condition such as respiratory acidosis or alkalosis.
  2. Gap-Gap or delat-delta. The patient has a dramatically elevated anion gap at 27 (15 over the upper limit of normal of 12) but his bicarb of 16 is only 8 below normal. The difference between the delta gap and the delta anion gap is 7 (15-8) when this is added to the normal bicarbonate you get 31; so the patient had a pre-existing metabolic alkalosis with a bicarbonate of 31. ABG displays this information as the corrected bicarbonate.

The next step is adjusting his sodium for the hyperglycemia. To do this we will use Mediquations though Medical  Calc works just as well.

Mediquations
DB, in his discussion, states that he has unpublished data proving that no formula is effective at adjusting the serum sodium for the hyperglycemia. For those of us without his unpublished data should adjust the sodium using Katz’s traditional conversion (pdf of a letter to JAMA discussing adjusting sodium for hyperglycemia in DKA. Katz’s original conversion was discussed in a letter to the NEJM) of a drop in Na of 1.6 for every 100 the glucose is over 100 mg/dL. Nephrology fellows should additionally be aware of Hillier’s data showing the sodium falling 2.4 for every 100 of glucose. Both Mediquations and Medical calculator adjust the sodium using Katz’s conversion.

Of coarse you wouldn’t know it was Katz’s conversion because even if you tap on “More Info,” Mediquation does not provide the reference. Likewise you will not get the reference with Medical Calc.

Though DB did not explore free water defecits in his discussion of the case this is a clinically relevent point. You can use Mediquation to calculate the water deficit.


I feel that using ABG and Mediquations will make you a more effective physician without forcing you to memorize equations used only periodically.

Polyuria, polydipsia

The set up

Patient is the ICU, you receive a consult for hypernatremia resistant to therapy for three days.

Na 174
Cl 128
BUN 38

K 3.2
HCO3 27
Cr 1.6
glucose 128

urine lytes:
Na 56
K 12
Body weight 45 kg
urine output 2,800

Step one


Calculate the fluid deficit:
We assume the increase in sodium is entirely due to loss of water. If the patient is hypotensive, tachycardic orthostatic or has other clues that he maybe volume depleted, the situation is more complex. We calculate the percentage that the sodium is elevated above an imagined “ideal sodium.” Many equations use 140 as the idealized sodium, I tend to use 145. The percentage the sodium is above the ideal is identical to the percentage of the total body water that is needed to lower the sodium to the target level.

174-145 / 145 = 0.2 (the sodium is 20% above 145)

Multiple that by the estimated total body water (weight x % body water) = 5.4 liters

I used 60% for estimated total body water. The patient looks like a young boy. Rose suggests lowering the estimated % body water by 10% so 50% would be okay also. In the elderly and obese this number can go below 50%. 

Step two


Going from 174 to 154 is too much change in 24 hours. The speed limit is 12 mmol/L/day so we will target a change from 174 to 162.

I use a modified water deficit formula. I divide 12 (the change in sodium we are looking for) by the target sodium (current sodium minus 12):

12/(174-12) = 0.08 x total body water = 2.0 liters or just over 80 mL/hour


Step three

Calculate the electrolyte free water clearance. The reason the ICU team repeatedly failed to correct the sodium is that they corrected the deficit without accounting for the ongoing water losses. In this case it is renal losses. If you just add the urine volume to the water deficit you will over estimate the amount of water needed because the urine contains sodium and potassium. To account for the electrolyte content of the urine we calculate the free water clearance.

Urine Na + K divided by the serum Na will give the ratio of urinary electrolytes to serum solute. 
56 + 12 / 174 = 39%
The urine has 39% electrolyte content of plasma, another way of thinking about this is that 39% of the urine volume is isotonic and the remainder (61%) is pure water. The 61% is what we are interested in; multiply the urne output by 61% this is the volume of water we need to give the patient to account for his ongoing renal losses. 
Then multiply this by the urine output (2,800) to get the electrolyte free water clearance, 1705 ml. This is another 71 mL/hour

Wrap up

Final fluid orders: 154 mL/hour of water, this can be given as D5W in the IV or preferably oral flushes. This calculation does not insensible losses.