Calculating the fractional excretion of potassium and the urine potassium to creatinine ratio (using those crazy American units)

Since the demise of TTKG I have had to retrain my brain to determine if hypokalemia is due to renal wasting or extra-renal potassium losses (as well as intracellular shift). The is not so important in the evaluation of hyperkalemia as persistent hyperkalemia is always due to decreased renal clearance of potassium.

There are two methods of looking at renal potassium wasting, the first is Fractional Excretion of Potassium. Super easy calculation.

The values are from this Skeleton Key group article at the Renal Fellow Network. They pulled it from this study of 84 hypokalemic individuals: Fractional excretion of potassium in normal subjects and in patients with hypokalaemia. From the abstract:

The mean FEK+ in normal subjects was 8% (range 4-16%). FEK+ was positively correlated with serum potassium (r = 0.74, p < 0.0001) and inversely with serum creatinine (r = -0.51, p < 0.001). The mean FEK+ in patients with hypokalaemia of external origin was 2.8% (range 1.5-6.4%). On the contrary, the mean FEK+ in hypokalaemic patients in whom renal potassium loss was the main aetiologic factor for the pathogenesis of hypokalaemia was 15% (range 9.5-24%).

Even though the creatinine is measured in mg/dL and the potassium is measured in mEq/L the units don’t mess you up because the serum and urine creatinine units cancel each other out.

The other way to look at hypokalemia is urine potassium creatinine ratio. Just divide the urine potassium by the urine creatinine and if it is greater than 1.5 you have renal potassium wasting. But alas this only works if the urine creatinine is measured in mmol/L. I get urine creatinines in mg/dl, so to make this conversion you need to multiply urine creatinine by 88 to get micromol/L of creatinine and then divide it by 1000 to convert to mmol/L. In one step it looks like this:

The line in the sand of 2.5 comes from this study, Laboratory Tests to Determine the Cause of Hypokalemia and Paralysis published in JAMA Internal Medicine.

But for the grace of God…

I often get a call from a nurse saying something to the effect of, “Dr. Smith is ready to discharge this patient home as long as he is cleared by nephrology.”

I will tell you, that the pre-test probability that I will “clear” the patient for discharge is high. I want patients to go home and I feel that a lot of what we do in the hospital can be achieved as an outpatient. Send the patient home and I’ll see them in clinic. But one needs to be careful.

I didn’t discharge the patient, but received him when he was readmitted. The patient was admitted with bilateral pleural effusions and respiratory distress. Patient was diagnosed with COPD exacerbation and heart failure.

The patient was on furosemide 40 mg daily at home and this bumped to 80 mg IV twice a day on admission.

Nephrology was consulted to assist with diuretics and added some metolazone on day 2. That was a one time order and not repeated. Adding a thiazide, metolazone, to a loop diuretic part to increase diuresis is called sequential nephron blockade. One cause of diuretic resistance, is with chronic loop diuretic use, sodium resorption that occurs after the thick ascending limb can short circuit effective loop blockade. So by stacking diuretics that act distal to the loop increases the effectiveness of the loop diuretic.

The addition of a thiazide or K sparing diuretic to a loop diuretic is an example of sequential nephron blockade.

Robert Centor and I did a podcast about the concept a few years ago. We talked about this seminal article by Dave Ellison.

I would have been a bit nervous adding a thiazide in a patient who already had hyponatremia, but the following day the labs look okay.

The following day is the day of discharge. The morning potassium is replaced with a combination of oral and IV potassium as well as 2 grams of magnesium sulfate.

The nephrologist clears the patient for discharge as long as the potassium is normal. The potassium is almost normal and the patient is sent home. On a combination of furosemide and metolazone!

Three days later the patient collapsed at home and during resuscitation is found to have a cardiac arrhythmia and potassium of 1.6.

Know your diuretics. Respect the diuretics

Also keep in mind this propensity matched trial of metolazone in acute decompensated heart failure. They found a large fraction of the excess mortality found with metolazone could be explained by hypokalemia.

The TTKG is dead, now what?

Halperin has declared the TTKG dead.

And therefore never send to know for whom the bell tolls; It tolls for the TTKG

However we still need to assess patients for hypokalemia and differentiate between renal and extra-renal losses.

Measuring a fractional excretion of potassium (FEK) doesn’t physiologically make sense. The idea behind the fractional excretion calculation is calculating what percentage of the filtered potassium (in this case, but can be anything) ends up in the urine. But potassium doesn’t work that way. Essentially all of the filtered potassium is reabsorbed in the proximal tubule and thick ascending limb of the loop of Henle so that the fractional excretion of potassium is zero at that point. Then in the late distal convoluted tubule and the medullary collecting duct all of the potassium that is destined for the toilet is secreted. So all of the potassium that is cleared by the kidney is secreted but the distal nephron/tubules not filtered by the glomerulus. That said the FEK is just a calculation and you can do it. I reviewed the the best data on it here:

FERE: Fractional excretion of random electrolytes

So what calculation do I use? I use the TTKG, but that’s because I’m a dinosaur. What I should be doing is the urine potassium to creatinine ratio.

The answer is 13 mEq/g creatinine.

In this study of hypokalemic periodic paralysis versus patients with increased renal potassium excretion, the K:Cr ratio neatly divided the two groups.

the dividend line here was 2.5 mmol K/mmol Cr or 22 mEq/g Cr

If you know of a better reference for the potassium to creatinine ratio, tweet me up.

You knew that proteinuria is protective against amphotericin induced hypokalemia. Right?

All of you #NephMadness players crying into your coffee about Proteinuria getting beat out by Patient Reported Outcomes need to understand that proteinuria isn’t always bad*.

*I am being sarcastic here, proteinuria is always bad, and the only reason I am writing this post is because of this interesting quirk where it appears to be protective.

Proteinuria protects against amphotericin b induced hypokalemia. In patients on amphotericin, heavy proteinuria, a protein concentration of 3 g/L (3+ on dipstick), is protective against amphotericin b induced hypokalemia.

The study was done on normal formulations (as opposed to liposomal preparations) of amphotericin B.

Amphotericin B is highly protein bound. With standard doses, the normal amphotericin concentration in the urine will be 1-2 micromol/L. With 3+ urine protein, the albuminuria concentration is over 40 micromol/L, and this is apparently enough to bind and neutralize amphotericin’s collecting duct toxicity. Amphotericin’s anti-fungal property comes from its ability to tear open fungi cell membranes. Unfortunately it does a doozy on the membranes of the collecting tubules as well, allowing potassium to flow down its contraction gradient from the cells to the tubular fluid (and out in the urine). Similarly hydrogen flow from the tubular fluid back into the cells causing metabolic acidosis. It is an unusual cause of renal potassium loss without increased aldosterone levels.

For you #NephMadness geeks, toad bladder was instrumental to working out the mechanism for amphotericin induced hypokalemia.

FERE: Fractional excretion of random electrolytes

Magnesium:

  • 142 controls: 1.8% (range 0.5-4%)
  • 74 hypomagnesemic
    • Extra-Renal origin 1.4% (range 0.5-2.7%)
    • Renal origin 15% (range 4-48%)
  • Authors conclusion: >4% per cent is indicative of inappropriate renal magnesium loss

Potassium

  • 312 normal subjects: 8% (range 4-16%)
  • 84 hypokalaemic patients
    • Extra-renal origin: 2.8% (range 1.5-6.4%)
    • Renal origin: 15% (range 9.5-24%)
  • Authors conclusion: >6.5% per cent is indicative of inappropriate renal potassium loss

We’ve got one! Finding a functional adrenal adenoma

A year ago, a slender, 40 year old, white female presented to my clinic with new onset elevated blood pressure. The hypertension was discovered during a routine visit for a minor injury. The family practitioner refused to believe the vitals and kept having the patient return for follow-up visits before resigning himself to the diagnosis. Surprisingly, this otherwise healthy woman, was resistant to multiple medications. He began to suspect a more sinister diagnosis and initiated a work-up for secondary hypertension and referred her to me.

The initial work-up showed a aldosterone of 16 but the renin was not done. She also had modestly elevated metanephrines, but not high enough to suggest a pheochromacytoma. Her blood pressure typically ran 140-160/100 with labetalol 100 mg bid, but she admitted to being forgetful regarding her medications.

One of the findings that stood out for me was the hypokalemia on the initial labs

We repeated the renin-aldo ratio and did a EKG. Unfortunately she had LVH. For me, this ruled out white coat syndrome. The demonstration of end-organ damage also helped the patient see that this condition was “real” and after that she was compliant with the medical therapy.

The repeat Aldo was only 3 with a fully suppressed renin at 0.15. This is an aldosterone-renin ratio (ARR) of 20, however, I was taught a low total aldosterone ruled this diagnosis out. In other words, one needs an elevated aldosterone, not just a suppressed renin to make the diagnosis of primary hyperaldosteonism. This always made sense to me but the Endocrine Society states that this is not always true and questions the requirement for a high aldosterone:

Against a formal cutoff level for aldosterone are the findings of several studies. In one study, seated plasma aldosterone levels were less than 15 ng/dl in 36% of 74 patients diagnosed with PA after screening positive by ARR defined as more than 30 and showing failure of aldosterone to suppress during fludrocortisone suppression testing (FST), and in four of 21 patients found by AVS to have unilateral, surgically correctable PA.

Her potassium remained low at 3.1 despite potassium supplementation. She was breast feeding at the time so we did not use an ACEi or ARB and were successfully treating her blood pressure with a combination of nifedipine XL and labetalol.

The low aldosterone appeared to rule-out primary hyperaldo but with the unexplained hypokalemia I ordered a third ARR and hit pay-dirt

An ARR of close to 300 with a sky-high aldosterone of 29. Remember, when you calculate the aldosterone-renin ratio make sure the units are correct:
  • aldosterone in nanograms per deciliter
  • renin measured as plasma renin activity (PRA) in nanograms per milliliter per hour
With a positive ARR, the endocrine society recommends a confirmatory test. There are four recommended tests, all of which are variations on attempts to suppress endogenous aldosterone via sodium loading or fludrocortisone suppression. I did not do this. I feel that the critical diagnosis to make is the functional adenoma that is surgically curative. Whether the patient has bilateral hyperplasia or simply aldosterone driven hypertension that doesn’t meet the criteria for primary aldosterone is not important to me because I’m going to treat both of those conditions identically, with spironolactone or eplerenone.

So we proceeded with the work-up for a functional adenoma and sent her for a CT scan. We found a 1 x 2 cm left adrenal mass.

Here is where it gets tricky. This sounds like a functional adenoma, however functional adrenal adenomas are rare diagnosis, and even in the presence of documented hyperaldosteronism, non-functional incidentalomas are too common (0.35-5%) to assure that a CT finding of an adrenal mass represents a functional adenoma. Following a CT scan, you can neither rule-out nor rule-in the diagnosis of a surgically correctible functional adenoma. Patients still need to get adrenal vein sampling. Here is the experience from University of Texas Southwestern:

Twenty patients had unilateral CT abnormalities, and 14 (70%) of them lateralized to the same side (concordant). Of the remaining 6 patients with unilateral CT abnormalities (3 left and 3 right), 1 patient each lateralized to the opposite side and 2 patients each had bilateral hypersecretion. Only 5 of 15 patients (33%) with bilateral CT abnormalities showed concordant bilateral aldosterone hypersecretion. The other 10 patients (67%) demonstrated unilateral hypersecretion. Of the 5 patients with normal-appearing adrenal glands on CT, 1 patient each lateralized to 1 side, and the other 3 patients had bilateral hypersecretion.

The authors did not provide a 2×2 table to determine sensitivity or specificity (insert rant regarding surgical literature here) so I put one together. This is how I interpreted the data above:

  • Positive test: 20 with unilateral findings, 14 true positives and 6 false positives (I considered the CT scan identifying the wrong affected adrenal as being a fail)
  • Negative test: 15 patients with bilateral findings, 5 were true negatives and 10 were false negatives
  • Negative test: 5 patients with normal adrenals, 2 lateralized, false negatives and 3 true negatives
The two-way table looks like this:
What? You’re still using Epocrates’ medical
calculator? Don’t be a tool, get a tool, MedCalc
It should be apparent that a CT scan looks truly terrible at diagnosing a functional adenoma. A negative predictive value of only 40%. Ughh! Note: these numbers assume the adrenal vein sampling is a valid gold-standard.

We sent her for adrenal vein sampling to see if the aldosterone secretion lateralizes. It did with a 20-fold increase in aldosterone on the left side. Because aldosterone levels can be unreliable due to dilution and technique, it is recommended that an adjusted aldosterone (aldo/cotisol) exceed the contralateral adrenal by three fold. In our case, it was 10-fold.

She went for an laparoscopic left adrenalectomy and is now normotensive off all medications.

The endocrine society had published consensus recommendations on screening, diagnosis and treatment of primary hyperaldosteronism. I love it when important articles are available in PDF for free.

Highest TTKG with hypokalemia

Patient with a lifelong history of hypokalemia. He came to me for a second opinion, his previous nephrologist had been nudging up his potassium dose on every visit and the patient was now on 70 mEq of KCl daily and was getting uncomfortable with endlessly increasing doses of potassium.

At the time I saw him these were his labs (he had decreased his potassium supplementation to 20 mEq/day):

  • Blood
    • sodium: 128
    • glucose: 90
    • potassium: 2.8
    • Creatinine: 0.9
    • BUN: 11
    • Magnesium: 1.8
    • Calculated osmolality: 265
  • Urine 
    • sodium: 135
    • potassium: >100
    • Osmolality: 637
Trans-tubular potassium gradient: 14.9. That’s crazy high for a patient with hypokalemia, one should expect it to be less than 2 for hypokalemia of extra-renal origin, and only 7 or 8 for hypokalemia from hyperaldosteronism. Halperin et al. were not able to get the TTKG that high even when they took normokalemic patients and doped them with fludrocortisone and 50 mEq of oral potassium. 
And that 14.9 is assuming the urine potassium is 100, our lab doesn’t do serial dilutions so who knows what the actual potassium is? 120? 140?
I’m still waiting for the renin and aldo but I smell some Bartter’s

Hypokalemia and rhabdomyolysis

Hypokalemia can induce rhabdomyolysis. The purported mechanism is that hypokalemia antagonizes one of the natural causes of exercise induced vasodilation. Normally, during exercise muscles release intracellular potassium causing local pockets of hyperkalemia which triggers vasodilation and increases perfusion to the active myocytes. Total body potassium depletion and hypokalemia decrease local hyperkalemia preventing the vasodilation which results in tissue hypoxia and rhabdomyolysis.

 In The Fluid and Electrolyte Companion we illustrated it thusly:

I remember thinking how funny it was that we used bowling to represent exercise (though we did slide in those runners behind the bowler).  Oh clipart, how much nerdtainment you have given me.

This past week-end one of my patients experienced this. He is a high school student who loves to fish. He has congenital type 2 RTA, so is predisposed to hypokalemia. The exercise that triggered his rhabdo:  fishing. He presented to the ER following a day of fishing with complaints of muscle cramps and weakness. His potassium was reported as less than 2, with CPKs in the 800 range.

One of the unexpected consequences of the rhabdomyolysis was that his cramping continued after the potassium was corrected. In fact, it was actually much worse muscle spasm and tetany of the hands. Turns out, this second round of neuromuscular symptoms weren’t due to hypokalemia but rather rhabdomyolysis induced hypocalcemia. He had a normal total calcium and a low ionized calcium. The muscular symptoms responded to a gram of IV calcium gluconate.