Hyperkalemia has the follow effects on the electrophysiology of the heart:

    1. Increases Ikr current which produces rapid repolarization. Thus T waves have a narrowed base, usually < 250ms. This makes the T-waves appear peaked but is only evident in 22% of patients with hyperkalemia.
    2. Brings resting membrane potential closer to depolarization threshold. This makes sodium channels more excitable initially but this quickly results in increased refractoriness of sodium channels. This results in a slowed conduction system manifesting as prolonged PR interval and slowed phase 0 of depolarization manifesting as widened QRS.
    3. It is important to note that the resting membrane potential of the sinoatrial tract is dictated by the Ifunny channel which regulates sodium ions. Thus hyperkalemia can cause profound bradycardia, heart block, or junctional rhythm as its only EKG finding.
    4. Hyperkalemia affects myocytes, specifically atrial, more than conduction cells because phase 0/depolarization in myocytes is dictated by Na but phase 0/depolarzation in conduction cells is dictated by calcium. In severe hyperkalemia, sodium channel refractoriness can cause obliteration of p-waves because there is no atrial myocyte contraction. This is the principle by which mild hyperkalemia is sometimes used as a therapy for A-fib or dig induced atrial tachycardia.

Whether these changes happen and the order in which it happens is impossible to determine because of several factors: rate of potassium change, calcium concentration, pH, and sodium concentration. Ventricular fibrillation can be the first presenting rhythm of hyperkalemia as evidenced by this study in 1953 which gave 15g of oral KCl to healthy subjects to observe their EKG changes. Thus, hyperkalemia should be treated emergently for 1) K >6.5 mmol/L or 2) EKG manifestations of hyperkalemia regardless of the potassium.

Membrane Stabilizers

Calcium is beneficial even in patients who are normocalcemic. Calcium for injection is available as the chloride or gluconate salt, both 10% by weight. The preferred agent is the gluconate salt, since it is less likely than calcium chloride to cause tissue necrosis if it extravasates. The recommended dose is 1g intravenous over 1 mins; this is much faster and higher dose than people are used to which is 1g over 1hour. In unstable electrocardiographic rhythms (heart block or QRS widening), you should give 3g over 3min then reassess. The onset of action is <3 mins. The dose may be repeated in 5 mins if there is no improvement in the EKG. The duration of action is 30–60 mins. The biggest error I see is physicians forgetting that they can use more. I have not infrequently given 15g with good effect; there are case reports of the same. You can’t really hurt a patient with calcium. It is probably okay even in dig toxicity.

Hypertonic sodium chloride has been shown to reverse the EKG changes of hyperkalemia in patients with concurrent hyponatremia. Whether hypertonic saline is effective in the treatment of eunatremic patients has not been established.

Contrary to teachings in medschool, bicarbonate does not shift potassium intracellularly. Infusion of a hypertonic or an isotonic bicarbonate solution for 60 mins has no effect on potassium in dialysis patients, despite a substantial increase in serum bicarbonate concentration. Only after a 4-hr infusion was a small (0.6 mmol/L) decrease in PK is detectable. Bicarbonate grants the theoretical benefit of lowering resting membrane potential from action potential threshold and thereby act as a membrane stabilizer.

Potassium Shifters

Insulin activates Na, K-ATPase by recruitment of intracellular pump components into the plasma membrane. 10U of regular insulin given as a bolus IV along with 1 amp of D50 to anephric adult patients lowers potassium by about 0.6 mmol/L. The onset of action is <15 mins and the effect is maximal between 30 and 60 mins after a single bolus. After the initial bolus, a dextrose infusion should be started, since a single bolus of 25 g of dextrose has been shown to be inadequate to prevent hypoglycemia at 60 mins with 50-75% of patients becoming hypoglycemic (CBG 60s) at 30min-60min.

Normal volunteers receiving insulin infusion demonstrated a transient lowering of potassium that reached a nadir after 90 mins then rose. Thus, there is no benefit to continuous infusion or repeated bolus injections over a single bolus injection. Hypertonicity worsens hyperkalemic induced membrane instability so don’t give D50 in hyperglycemic patients.

Albuterol by intravenous infusion (0.5 mg over 15 mins) reduces potassium by about 1 mmol/L between 30 and 60 mins. Injectable albuterol is unavailable in the United States but you can use terbutaline SQ or epi drip with good effect. Nebulized albuterol at 10mg reduces potassium by about 0.6 mmol/L, at 20mg reduces potassium by 1.0 mmol/L. One MDI can decrease potassium by 0.4 mmol/L after 60 mins. The effect of high-dose therapy is apparent at 30 mins and persists for at least 2 hrs. Subcutaneous terbutaline (7 µg/kg body weight) reduces potassium in dialysis patients by an average of 1.3 mmol/L within 60 mins. Beta-blocked patients and 40% of patients not on b-blockers do not respond to beta adrenergy. There is currently no basis for predicting who will respond.

Shifting potassium does not correct the problem except in insulin deficient states like uncontrolled diabetes. Albuterol and insulin should never be used without a plan for eliminating potassium. In fact, if the plan is to definitely remove potassium through lasix or dialysis, shifting the ion intracellular actually hides it from elimination. It has become a personal pet peeve to see insulin/D50 ordered reflexively for hyperkalemia, even mild cases. It only serves to hide a nasty surprise for the next provider after the insulin wears off in 2 hours. And hypoglycemia is a very real risk for euglycemic patients.

Potassium Elimination

Potassium is eliminated primarily through the nephrons. The major determinants are GFR, aldosterone, and alkalotic urine. A number of agents have been studied. Fludrocortisone is the treatment of choice for RTA type 4. Diamox can increased bicarb delivery into the tubule and enhance kaliuresis but also worsens metabolic acidosis.

However, neither regimens are in my mind as effective as a loop diuretic in the acute phase. If the patient makes urine, they should receive furosemide. It can be beneficial to preload them with several fluid boluses to increased GFR followed by a massive furosemide dose. Be sure to match UOP with fluid repletion, perhaps with bicarbonate gtt. The only side effect to furosemide is ototoxicity, which does not occur if lasix is given at <4mg/min (160mg dose should be given over 40min) or if the bolus dose is <60mg. Renal insufficiency is NOT a reason to abstain from lasix. It in fact demands more lasix because the drug has to make its way into the tubules to take effect.

I use LR for fluid loading even though it contains 4mEq/L. RCT comparing LR to NS in ESRD patients undergoing transplant showed higher hyperkalemia (5/26 patients developed K > 6) and acidosis with NS as well as worse renal function vs LR (0/25 pts developed K > 6). This is somewhat unsurprising since potassium excretion is inversely proportional to nephrons' intraluminal chloride concentration. Another RCT comparing LR to NS in ESRD patients undergoing renal transplantation showed similar results:


The effect of lasix induced kaliuresis can be greatly enhanced by sodium bicarbonate drip because this functions through the same mechanism as diamox without the danger of inducing preload azotemia or worsening metabolic acidosis (which is frequently concomitant in these patients). Chlorothiazide (1000mg) is also very useful with lasix especially in severe CKD. The two diuretics work synergistically for kaliuresis.

Futility of Kayexelate

The gut can be thought of as a large nephron and eliminates 10-20% of daily potassium. Sodium polystyrene sulfonate (SPS, kayexelate) has been traditionally used to enhance gut elimination. Our continued use of this drug is probably more based out of tradition rather than science.

The original study on the efficacy of SPS had no control groups and simultaneously treated patients with insulin, D50, potassium sparing diets, and bicarb. The subsequent study with a control group showed that SPS is no different than sorbitol, its solvent, in reducing serum potassium. A recent study replicated these findings and showed that in fact the best resin combination should probably be phenolphthalein, docusate, SPS. Another recent study also showed no effect on serum potassium at 12 hours postSPS but this was in normokalemic ESRD patients.

The safety concerns with SPS derives from many case reports of intestinal necrosis after oral agent or enema irrespective of sorbitol use. Given that the efficacy of kayexelate is minimal if at all and requires at several hours days to exert effect, I do not utilize it in the treatment of acute or emergent hyperkalemia. It is neither safe nor effective even according to nephrologists. Luckily there may be new resins that actually do work – Sodium Zirconium Cyclosilicate and Patiromer – but they have only been shown to work over the span of two days rather than hours.


In hyperkalemia, the first step is to obtain an EKG. Abnormal EKG should receive aggressive treatment with membrane stabilizers (several boluses of calcium gluconate) followed by repeat EKGs.

The next step is to rule out obstructive nephropathy with an ultrasound or foley. Ask the patient if they make urine. Patients who make urine need LR boluses followed by high-dose lasix with or without bicarb drip, with or without chlorothiazide. Patients who do not make urine need emergent/urgent dialysis with or without potassium shifters.

Potassium shifters should be used only in clinically or electrocardiographically unstable scenarios or if there will be significant delays to elimination strategies. All shifting strategies do not address the underlying etiology and, in fact, hide the hyperkalemia from elimination as well as generate a false sense of security.