#hyperk

Just watched Amal Mattu (and Andy Neill's) great EKG review, this time on (spoiler alert) -- hyperkalemia! (If you're not watching these videos, you're missing out on free, amazing education from I think the best EKG teacher in the world.)

But I think there are two things worth mentioning:

You cannot -- and should not -- use an EKG to "rule out" hyperkalemia. I completely agree with Amal -- if you see a bizarre looking EKG, you should think tox, potassium or calicum derangement (I like to throw LBBB in there too), but a normal EKG won't rule out diddly squat in your patient. A few studies-in-point:

The Ability of Physicians to Predict Hyperkalemia From the ECG: Took patients with known hyperkalemia in the ED, had two physicians use the EKG to determine if the patient had hyperkalemia: sensitivity around 0.4; specificity around 0.85. Not great.

Electrocardiography is unreliable in detecting potentially lethal hyperkalaemia in haemodialysis patients: Took dialysis patients, got a pre-dialysis potassium level, and looked at T waves, R waves, ratios. Again, no correlation (but these patients had an average K of 4.9).

A few other cases:

K > 9, no EKG changes!!

K > 6, no quantitative EKG measurements to detect hyperkalemia on EKG

And secondofly -- I'd have to disagree on Amal's recommendation of bicarb.

Sodium bicarbonate does not work -- or at least, does not work well, or on its own -- for treating hyperkalemia. And in patients with fluid/volume issues (heart failure, renal failure -- you know, the typical people who get hyperkalemic), I always worry about giving a big intravenous hyperosmotic sodium bolus to these patients (as my nephrology professor used to remind us -- "water follows sodium"). (NB: This was brought to my attention by one of my co-chiefs, Kim Medlej, who finished a critical care fellowship last year at Harvard, and now practices in Lebanon, so all the credit is his.)

Quick summary: We're all taught bicarb works within 30 minutes, by intracellular shift/exchange of potassium ions for hydrogen ions, yada yada yada. That really doesn't appear to be the case. I think in the ED we're sometimes taught to just give them an amp or two of sodium bicarb, but that appears to have NEVER been studied. In the crashing/dying patient, yes, I give sodium bicarbonate, but I'm otherwise skeptical of the benefit and worried about the harm.

All the studies have really looked at bicarb infusions over hours, and if there's any change to be found, it's maaaybe at the 6 hour mark (after 6 hours of bicarb infusion, in patients who are already getting dialysis). Other studies with bicarb infusions show no statistically significant change, either. (One study that took patients and put them on a high or low dose bicarb infusion for an hour actually found a higher potassium levels after the infusion.) Probably the best study (Blumberg, 1992) found only a 0.5-0.7 drop, but they then attribute half the drop to the expansion of the ECF due to all the sodium the patients got.

Insulin definitely works. Albuterol works (but the studies are small and they usually give a good 10-20mg of it nebulized). There have been a few studies looking at combining bicarb + either of these other methods, and it looks like the bicarb probably DOES have some synergistic effect (it lowers the potassium more than just, say, albuterol alone). But by itself? Bicarb is probably pretty worthless.

Reviewing the literature, it seems like the insulin/D50, albuterol (? Lasix, not much literature on it) methods are the way to go. I know before I read this literature I felt better because I'd given the person kayexalate, or I'd given them bicarb, but really, the other methods are much more likely to keep the patient alive on the floor for 6 hours while they await their dialysis, without putting them into florid fluid overload.

To the stable, no dysrhythmias or severe symptoms patients I tend to give:

Regular Insulin 10 units IV with 1-2 amps D50

Albuterol 10mg nebulized

Calcium Gluconate 1-2g IV

Lasix if they make urine (pick your dose)

I've summarized the literature and we can send you the articles if you're curious:

Burnell, 1956 http://www.ncbi.nlm.nih.gov/pubmed/13367188 Looks like this is where a lot of it started. Many articles from the 70s/80s cite this one. There's very little on their methodology, but they have some pretty cool graphs that show an inverse relationship between pH and serum potassium concentration.

Schwarz, 1959 http://www.ncbi.nlm.nih.gov/pubmed/13629781 Case series of hyperK patients who had EKG changes who got better with bicarb. (Some of them got calcium as well, others required "5-10 grams of bicarb a day," others got bicarb + blood transfusion.)

Fraley, 1977 http://www.ncbi.nlm.nih.gov/pubmed/24132 Methods: Took 14 hyperK patients, gave them bicarb infusions over 4-6 hours. Checked K every hour. Results: Divided groups retrospectively into "constant pH" and "changed pH" groups. Both groups showed decreases in their potassium, ~1.6-1.8mmol/L (never seen this significant of a drop reproduced).

Blumberg, 1998 http://www.ncbi.nlm.nih.gov/pubmed/3052050 Methods: Took 10 HD patients, checked their K (along with other labs), gave them a bunch of different agents for changing K (bicarb, insulin, epi drip, regular dialysate), and then checked their labs after an hour. For bicarb, it was 8.4% in water, 4mmol/min, for 1 hour only. They also tried a isotonic bicarb infusion of 1.4%. Results: The K actually went UP after both bicarb infusions.

They conclude that bicarb didn't work, but in the past it's worked over longer periods of time. So then they do ...

Blumberg, 1992 http://www.ncbi.nlm.nih.gov/pubmed/1552710 Methods: Took 12 hyperK (>5.8) patients on dialysis, gave a bicarb (8.4% in free water) infusion 4mmol/min x1 hour, then 1.4% bicarb in water infusion 0.5mmol/min hours 2-6 and checked potassium levels throughout the time on dialysis. Also checked an EKG. Results: Average K was 6.0. K dropped at 4-6 hours, by 0.5-0.7, and they believe that half of the drop is probably due to the huge sodium load and increase in the extracellular fluid compartment.

Allon, 1996 http://www.ncbi.nlm.nih.gov/pubmed/8840939 Methods: Took 8 HD non-HyperK patients, put them through different combinations to lower their K (bicarb infusion, saline infusion, bicarb+insulin, saline+insulin, bicarb+albuterol, saline+albuterol). Results: Bicarb or saline infusions didn't work. Anything with insulin or albuterol the combination worked, lowered them from 0.5-0.8, depending on the group. Of note, bicacrb + albuterol worked better than saline + albuterol (see Kim, 1997).

Kim, 1997 http://www.ncbi.nlm.nih.gov/pubmed/8852501 Methods: Took 9 HD hyperK patients, gave them separate or combined bicarb infusions (1/2 hour long) along with nebulized albuterol, checked K before and after. Thought maybe there would be combined/synergistic effects of the two meds. Results: Bicarb alone didn't change the potassium. Salbutamol alone dropped the K by 0.6, and salbutamol + bicarb dropped the K by 0.9.

Kaplan, 1997 http://www.ncbi.nlm.nih.gov/pubmed/9043534 Methods: Took 8 dogs, gave potassium infusion until they got conduction disturbances, then backed down on the K, and gave either bicarb infusion (1.05% over 1 hour), bicarb bolus (8.4% over 5 minutes, then saline), or "saline" therapy (hypertonic saline 8.4% bolus + normal saline). Measured K before and after. Results: Saline worked just as well as bolus. Infusion worked better than both (but not statistically significant). Change was 1-2mmol/L.

Review Articles:

Kim, 2002: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3054237/ Don't recommend bicarb, especially as a single agent, especially in dialysis patients. "Should not be used."

Weisberg, 2008: http://www.ncbi.nlm.nih.gov/pubmed/18936701 Definitely doesn't work short-term, but might still be useful for temporizing hyperK. "It has now been clearly demonstrated that short-term bicarbonate infusion does not reduce PK in patients with dialysis-dependent kidney failure, implying that it does not cause K shift into cells. Infusion of a hypertonic or an isotonic bicarbonate solution for 60 mins has been shown to have no effect on PK in dialysis patients, despite a substantial increase in serum bicarbonate concentration."

Rachoin, 2010: http://www.ncbi.nlm.nih.gov/pubmed/21661096 "When treating hyperkalemic patients, hospitalists should use sodium bicarbonate to potentiate urinary elimination of potassium and should consider administering it either with acetazolamide or a loop diuretic, anticipating a lowering effect after a few hours.26 It should be avoided in patients with volume overload and anuria. Immediate translocation of potassium into cells is best achieved by insulin and b-2 agonists."