Friday, April 8, 2016

STEMI with Life-Threatening Hypokalemia and Incessant Torsades de Pointes

A late middle-aged man presented with one hour of chest pain.  He had significant history of CAD with CABG x5, and repeat CABG x 2 as well as a subsequent PCI of the graft to the RCA (twice) and of the graft to the Diagonal.   Most recent echo showed EF of 60%.  He also had a history of chronic kidney disease, stage III.

He had recently had a NonSTEMI. Angio had shown some acute disease in the saphenous vein graft to the posterior descending artery off of the RCA.  He was managed medically with Clopidogrel.  Medics stated that he had not been taking his clopidogrel for 2 weeks.

He appeared to be in shock.  Bedside ultrasound showed no effusion and moderately decreased LV function, with B-lines of pulmonary edema.

Here is his ED ECG:
There is obvious infero-posterior STEMI.
What are you worried about in addition to his STEMI?
See below.

The corrected QT interval is extremely long, about 500 ms.   This suggests an electrolyte abnormality or a medication effect (acquired long QT).  There is also bradycardia.  Bradycardia puts patients at risk for "pause-dependent" Torsades de Pointes.  

Torsades in acquired long QT is much more likely in bradycardia because the QT interval following a long pause is longer still.  Thus, Torsades in acquired long QT is called "pause dependent": if there is a sinus beat after a long pause (which creates a longer QT interval), then an early PVC ("early afterdepolarization," EAD) is much more likely to occur during repolarization and to initiate Torsades.  The usual sequence is: sinus beat, then early PVC, then a long pause because the PVC was early, which then results in a particularly long QT, then another PVC with "R on T" that initiates torsades.  

Here are his medications, none of which prolong the QT:

--Imdur 60mg daily
--Furosemide 40mg BID
--Lisinopril 5mg daily
--Amlodipine 10mg daily
--Digoxin 125mcg q48
--Clopidogrel 75mg daily 
--Atorvastatin 20mg daily
--Metoprolol 100mg BID

Clinical Course

His potassium returned at 1.8 mEq/L.  This is dangerously low.  There is an abundance of literature linking K less than 3.5 to ventricular fibrillation in acute MI (see many references below).

The creatinine was 1.8 mg/dL.

The patient was intubated, given antiplatelet and antithrombotic therapy, 10 mEq of KCl IV was started, and sent to the cath lab.

At cath, he immediately had incessant Torsades de Pointes requiring defibrillation 7 times and requiring placement of a transvenous pacer for overdrive pacing at a rate of 80.  He was given amiodarone and lidocaine load and drip and K and Mg drips.  After pacing, there was no recurrence of Torsades.

After resuscitation, he was found to have a 90% thrombotic lesion in the same saphenous vein graft to the right posterior descending artery.  This was stented.

The patient stabilized.

This subsequent ECG was recorded after the K was up to 2.2 mEq/L:
The STE is resolved.
The QT is much shorter
There are now clear U-waves in V2 and V3

2 days later, this ECG was recorded with a K of 3.5:
There is atrial fibrillation.  The QT is much shorter still.

The patient stabilized and had a good outcome.


STEMI with hypokalemia, especially with a long QT, puts the patient at very high risk of Torsades or Ventricular fibrillation (see many references, with abstracts, below).  These two rhythms are often indistinguishable on the monitor or ECG.  If there is a pulse, you would call it Torsades.  If there is polymorphic VT with a long QT on the baseline ECG, then generally we call that Torsades, but Non-Torsades Polymorphic VT can result from ischemia alone.

However it is classified is not so important!  What is important is that the initial treatment is the same for both, especially if there are no pulses: defibrillation, as was done here (NOT synchronized cardioversion).

The fact that it was controlled with overdrive pacing and potassium and magnesium suggests that it was indeed Torsades, but, on the other hand, antidysrhythmics and potassium were also given.

See here for management of Polymorphic Ventricular Tachycardia, which includes Torsades.

Could the dysrhythmias have been prevented?

I could find very little literature on the treatment of severe life-threatening hypokalemia.  There is particularly little on how to treat when the K is less than 2, and/or in the presence of acute MI.  Here are the American Heart Association Guidelines: 

Part 10.1: Life-Threatening Electrolyte Abnormalities

Treatment of Hypokalemia

"The treatment of hypokalemia consists of minimizing further potassium loss and providing potassium replacement.  IV administration of potassium is indicated when arrhythmias are present or hypokalemia is severe (potassium level of less than 2.5 mEq/L).  Gradual correction of hypokalemia is preferable to rapid correction unless the patient is clinically unstable.

"Administration of potassium may be empirical in emergent conditions.  When indicated, the maximum amount of IV potassium replacement should be 10 to 20 mEq/h with continuous ECG monitoring during infustion  A more concentrated solution of potassium may be infused if a central line is used, but the tip of the catheter used for the infusion should not extend into the right atrium.

"If cardiac arrest from hypokalemia is imminent (i.e., malignant ventricular arrhythmias are present), rapid replacement of potassium is required.  Give an initial infusion of 10 mEq IV over 5 minutes; repeat once if needed.  Document in the patient's chart that rapid infusion is intentional in response to life-threatening hypokalemia."

This last section is appropriate for this case.  Everyone is appropriately worried about giving K too fast.  How much does rapid infustion increase the K?  There is, again, little empirical data on this topic that I can find (see 2 studies below, which do not really answer the question).  Perhaps there are studies in animals that I have not found?  
Total Body Potassium: a 70 kg person has about 7500 mEq of total body K, but the extracellular fluid has only about 48 mEq!   Of course the difficulty with K replenishment is that the total body stores may be depleted by far more than can possibly be quickly repleted.  The estimated deficit associated with a serum decrease from 4.0 to 3.0 mEq/L is 100-200 mEq of total body K, and from 3.0 to 2.0, the associated loss is double, at 200-400 mEq.* [Sterns RH, et al. Internal potassium balance and the control of the plasma potassium concentration. Medicine (Baltimore) 1981;60:339-54].  

But 100 mEq given all at once would raise the serum K by 30 mEq/L (and be immediately fatal)!!

*The NEJM review referenced below (and ACLS, for what that is worth), states that, on average, in a "typical" 70 kg person, the serum K falls by 0.3 mEq/L for every 100 mEq total body deficit.  However, this review references the Sterns article above, which by my reading does not state this.

Here are some calculations for a safe rapid dose:
A 70 kg person has about 5 liters of blood, and 3 liters are serum (2 liters are RBCs).  If 10 mEq is given very rapidly, leaving no time for intracellular shift, then it will raise serum K by about 3.3 mEq/L.  If the patient is at 1.8, that will raise it to 5.1 mEq/L.  One need only get the K above 3.0 to greatly decrease risk (although in STEMI, the optimal level is about 4.0-4.5 mEq/L).  5 mEq rapid bolus would raise this patient's K from by 1.6, from 1.8 to 3.4 mEq/L.   The difficulty is in estimating the ongoing shift.  As you infuse K, it will start to shift into depleted cells and the serum K will fall again rapidly.  Thus, it is critical  in patients like this to repeatedly and rapidly, after each bolus, measure the K, and supplement as needed.

In the case presented, it is not clear to me that the 10 mEq of K was given rapidly.  I suspect it was set to go over 1 hours on a pump, which is the usual practice.  It would be difficult to get a nurse to give it faster!  However, in this case, it would be appropriate to give it over 5-10 minutes, with monitoring, then immediately measure the K again and be ready to give more.

Further complicating the issue is that severe hypokalemia can result in rhabdomyolysis and subsequent K release, with resulting hyperkalemia!

Here is review of hypokalemia from the NEJM, but it is mostly about etiology, and says little about rapid replacement in life-threatening hypokalemia EXCEPT to emphasize how dangerous rapid replacement is.

Supposedly, patients without ischemia are at low risk of complications from hypokalemia,  But it is not entirely without risk.  I saw this 30 year old woman with no cardiac disease who was resuscitated from ventricular fibrillation:
Classic Hypokalemia, with large U-waves.  K was 1.3 mEq/L.

Learning Points:

1. Severe hypokalemia in the setting of STEMI or dysrhythmias is life-threatening and needs very rapid treatment.  5-10 mEq over 5-10 minutes is appropriate for a K of 1.8 mEq/L.

2. Be certain that your laboratory value is accurate and that it corresponds with the ECG findings!  If the ECG shows no evidence of hypokalemia, it may be an artifactual value.  If the ECG shows no evidence, it is unlikely to be life-threatening!

3. In a 70 kg person, a 10 mEq bolus will raise serum K by 3.3 mEq/L in the absence of any intracellular shift

4. It is optimal to give such a bolus through a central line, but this may not always be possible.

5. It is difficult to correct K without also correcting low magnesium.  In this case, the Mg was 1.9 mEq/L (within normal limits)

6. Learn the management of Polymorphic VT, including Torsades.  


Two Articles on Rapid Replacement of Potassium

Efficacy and safety of potassium infusion therapy in hypokalemic critically ill patients.   1991 May;19(5):694-9

Objective: To evaluate the efficacy and safety of potassium replacement infusions in critically ill patients.
Design: Prospective cohort study.
Setting: Multidisciplinary critical care unit.
Intervention: Potassium chloride infusions (20,30, or 40 mmol in 100 mL normal saline over 1 hr) were administered to patients for serum potassium levels of <3 .5="" but="" style="font-family: times, 'times new roman', serif;">3.2 mmol/L (n = 26), 3.0 to 3.2 mmol/L (n = 11), and
Measurements and Results: All patients tolerated the infusions without evidence of hemodynamic compromise, ECG change, or new dysrhythmia requiring treatment. The mean maximum potassium increase was 0.5 +/- 0.3 mmol/L, 0.9 +/- 0.4 mmol/L, and 1.1 +/- 0.4 mmol/L in the 20-, 30-, and 40-mmol groups, respectively. The increase in serum potassium was maximal at the completion of the infusion and was significant (p < .05) compared with baseline in all groups. Peak potassium levels were the same in patients with normal renal function (n = 33) compared with those with renal insufficiency (n = 15).
Urinary excretion of potassium increased in all groups during the infusion and was significant (p < .05) in the 30- and 40-mmol groups, but was no greater in those patients who had received diuretics (n = 8) compared with those patients who had not (n = 40).
Conclusions: In the select group of hypokalemic patients studied, potassium infusions of 20 to 40 mmol delivered over 1 hr were safe to administer and effectively increased serum potassium levels in a dosedependent and predictable fashion. Furthermore, these results were independent of the patient's underlying renal function or associated diuretic administration. (Crit Care Med 1991; 19:694)

Concentrated Potassium Chloride Infusions in Critically Ill Patients with Hypokalemia

The Journal of Clinical Pharmacology.  Volume 34Issue 11pages 1077–1082, November 1994

Although concentrated infusions of potassium chloride commonly are used to treat hypokalemia in intensive care unit patients, few studies have examined their effects on plasma potassium levels. Forty patients with hypokalemia were given infusions of 20 mmol of potassium chloride in 100 mL of normal saline over 1 hour; 26 patients received the infusions through the central vein and 14 patients through the peripheral vein. Plasma potassium ([K]p) was measured at 15-minute intervals during and after the infusion in 31 patients. ΔK was defined as the difference between each potassium determination and baseline plasma potassium concentration. Continuous electrocardiographic recording was carried out during the infusion and during the 1-hour period immediately preceding the infusion. Mean baseline [K]p was 2.9 mmol/L and all subsequent plasma concentrations significantly increased from baseline. Mean peak [K]p was 3.5 mmol/L, [K]p (1 hour postinfusion) was 3.2 mmol/L, and mean postinfusion ΔK was 0.48 mmol/L (range −0.1–1.7 mmol/L). Arrhythmias, changes in cardiac conduction intervals, and other complications did not occur. The frequency of premature ventricular beats decreased significantly during the infusion compared with that of the control period. The high concentration (200 mmol/L) and rate of delivery (20 mmol/hr) of the potassium chloride infusions were well tolerated, decreased the frequency of ventricular arrhythmias, and did not cause transient hyperkalemia.

Literature on Hypokalemia as a risk for ventricular fibrillation in acute myocardial infarction.  All of it comes from the 1980's.  Use of diuretics is strongly associated with hypokalemia and ventricular fibrillation in myocardial infarction.

Thiazide-lnduced Hypokalemia: Association With Acute Myocardial Infarction and Ventricular Fibrillation

Ten of 59 patients (17%) were receiving a thiazide preparation at the time of an acute myocardial infarction and ventricular fibrillation. Hypokalemia was present in seven of eight patients (87%) receiving thiazides, whereas it was observed in only one of 38 patients (2.6%) not receiving these medications. If hypokalemia is present in patients receiving thiazides who have had an acute myocardial infarction, it should be corrected so as to remove this predisposing cause of ventricular fibrillation.

(JAMA 239:43-45; 1978)

Malignant arrhythmia in relation to serum potassium in acute myocardial infarction.

Serum Magnesium and Potassium in Acute Myocardial InfarctionInfluence on Ventricular Arrhythmias

Henryk Kafka, MD; Lorrie Langevin, RN; Paul W. Armstrong, MD
Arch Intern Med. 1987;147(3):465-469. doi:10.1001/archinte.1987.00370030069014.

Over a 13-month period, serum potassium and magnesium levels were measured in 590 patients admitted to a coronary care unit. Hypokalemia, often in the absence of diuretic use, occurred In 17% of the 211 patients with acute myocardial infarction. Patients with acute myocardial infarction and a potassium level of less than 4.0 mEq/L (4.0 mmol/L) had an increased risk of ventricular arrhythmias (59% vs 42%). Because hypokalemia is common in acute myocardial infarction and is associated with ventricular arrhythmias, routine measurement of serum potassium levels and prompt correction are recommended. Hypomagnesemia occurred in only 4% of the patients, but It was more common in the group with acute myocardial infarction than in the group without myocardial infarction (6% vs 3%). Ventricular arrhythmias occurred in ten of the 13 patients with both acute myocardial infarction and hypomagnesemla, but eight of these patients also had low serum potassium levels. This low incidence of hypomagnesemia does not justify routine measurement of serum magnesium levels. However, the mean level (2.5±0.4 mg/dL [1.03 ± 0.16 mmol/L]) in a reference population of healthy volunteers was unexpectedly high and suggests that the low incidence of hypomagnesemia In our population may not be applicable to other centers and may reflect a higher magnesium content in our geographic area of southeastern Ontario.

(Arch Intern Med 1987;147:465-469)

Frequency of hypokalemia after successfully resuscitated out-of-hospital cardiac arrest compared with that in transmural acute myocardial infarction 

To evaluate the prevalence of hypokalemia in out-of-hospital cardiac arrest, the initial serum potassium and arterial pH values were reviewed from 138 consecutive patients resuscitated from cardiac arrest. For comparison, the same variables were reviewed for 62 consecutive patients who had transmural acute myocardial infarction (AMI) without cardiac arrest. The mean serum potassium level was lower after resuscitation from cardiac arrest (3.6 ± 0.6 mEq/liter) than during AMI (3.9 ± 0.5 mEq/liter) (p < 0.005). The incidence of hypokalemia (potassium less than 3.5 mEq/liter) was greater in patients sustaining cardiac arrest (41%) than in patients who had AMI without cardiac arrest (11%) (p < 0.001). Hypokalemia was common after cardiac arrest regardless of the occurrence of AMI at the time of arrest. Hypokalemia after cardiac arrest was independent of arterial pH, epinephrine or bicarbonate therapy during resuscitation, or prior therapy with diuretic drugs, digoxin or propranolol. In 10 patients with marked hypokalemia, the serum potassium level returned to normal rapidly (16 hours) during the hospitalization even though only 29% of the predicted potassium requirement was infused before its normalization. Thus, hypokalemia is prevalent immediately after out-of-hospital cardiac arrest, whereas it is uncommon in AMI in the absence of cardiac arrest. The cause and electrophysiologic consequences of this hypokalemia are unknown; in most cases, it is apparently caused by a shift of potassium from the intravascular compartment rather than a total body depletion of potassium.

Sixty patients with a first acute myocardial infarction and no current treatment with cardioactive drugs were included in a prospective study of the relationship between serum potassium concentration and the early occurrence of ventricular tachycardia and premature ventricular contractions (PVCs). Serum potassium level (range 2.5 to 5 mmol/liter) was estimated 3.8 +/- 2.5 hr (mean +/- SD) after the onset of the infarction, and Holter monitoring was performed during the subsequent 12 hr. In multivariate analysis, serum potassium level was negatively and age positively related to ventricular tachycardia. Among the subclasses of PVCs (frequent unifocal, multifocal, couplets, bigeminy), serum potassium concentration was negatively related to the frequent unifocal subclass; hypertension was related to couplets and to the presence of any of the subclasses, and serum aspartate aminotransferase concentration was related to multifocal PVCs. Heart failure leading to death was related to all subclasses of PVC. Serum potassium concentration is an independent inverse predictor of the occurrence of ventricular tachycardia and frequent unifocal PVCs early in acute myocardial infarction.

Hypokalaemia and ventricular fibrillation in acute myocardial infarction (full text pdf; Br Heart J 50:525-529)

Serum potassium concentrations obtained on admission to hospital were inversely related to the incidence of ventricular fibrillation in 289 women and 785 men with acute myocardial infarction, 92 of whom developed ventricular fibrillation. Hypokalaemia (serum potassium concentration less than or equal to 3.5 mmol/l) was found in 122 patients (11.4%). The incidence of ventricular fibrillation was significantly greater in patients with hypokalaemia compared with those classified as normokalaemic (serum potassium concentration greater than or equal to 3.6 mmol/l) (17.2% v 7.4%). The increased risk of ventricular fibrillation in the hypokalaemic group was about the same for women and men. While they were in hospital patients with hypokalaemia developed ventricular fibrillation significantly earlier than did normokalaemic patients (median 0.3 hours v 7 hours). Hypokalaemia was more common in women (17.3%) than in men (9.2%), and 55% of the hypokalaemic patients had been treated with diuretics before admission compared with 22% of the normokalaemic group. Hypokalaemia on admission to hospital predicts an increased likelihood and early occurrence of ventricular fibrillation in patients with acute myocardial infarction.


  1. Sorry for repeating the same words, but this is another istructive case, with countless interesting points and literature.

    With respect to calculation of the Qtc in the presence of atrial fibrillation, I always wonder what is the best method.
    On the one hand AHA/ACCF/HRS Recommendations for the Standardization and interpretation of the Electrocardiogram Part IV: The ST segment, T and U Waves, and the QT interval -A Scientific Statement from the American Heart Association Electrocardiography and Arrhythmias Commitee, Council on Clinical Cardiology; the American College of Cardiology Foundation; and the Heart Rhythm Society--Circulation 2009; 119:e241-250 at Recommendation pag.e246 state: “In addition, rate correction of the QT interval should not be attempted when RR interval variability is large, as often occurs with atrial fibrillation…” while other authors suggest to calculate Qtc by averaging the QT intervals over 10 beats or by taking the average of QT intervals (Al-Khatib SM, LaPointe NM, Kramer JM, Califf RM. What Clinicians Should Know About the QT Interval. JAMA 2003;289:2120).
    What is your opinion on this topic? Many thanks.

    1. Mario,
      Hard to say because I don't know of any data to support an opinion on this. That is to say, one would have to have a comparison of techniques based on outcome (incidence of TdP in cohorts with long QT based on one or the other correction methods.) That is difficult data to get.

  2. Thanks Steve for this fascinating case with so many learning points.

    It has puzzled me for some time the need to correct QT for HR since it is at slow HR the risk of TdP increases. My understanding was that it is used as a way to risk stratify patients with congenital long QT syndrome once reversible causes were excluded.

    The initial ECG seems to be tricky measuring the actual QT. In the first few complexes it's not obviously as prolonged as it appears in the last and 3rd last.

    The other thing I've pondered about regarding K is the effects of changes in K on the resting membrane potential. As serum K falls you would expect intracellular K to have fallen as well.

    Add to that the effect of digoxin and you have to wonder what contribution digoxin had contributing to the dysrhythmias even if not at a toxic level.

    Interesting they chose to give amiodarone which would have further increased QT.

    I would have been pretty nervous inserting a pacing wire with that irritable myocardium. I think I would have gone with a Beta agonist.


    1. Andre,
      The correction is more important to calculate when there is NOT bradycardia: if it is long, then bradycardia will exacerbate risk.
      It is the longest QT that matters, as any prolonged beat will be susceptible to an EAD.
      One always gets confused trying to understand underlying ion flux!
      I agree about amio, would not give. Lidocaine.
      Isoproterenol is now rare and costs $20,000 per dose!


Recommended Resources