Friday, July 24, 2020

A Complication of the COVID Era

Submitted and written by Gia Coleman MD and Roshan Givergis DO, edits by Meyers and Smith

A woman in her 30s was found crawling in the streets, altered on arrival to the ED. Here is her presenting ECG:

How would you interpret this EKG and what is on your differential?

At first glance, it appears to be a sinus rhythm with PR prolongation at a rate of about 75 bpm. The QRS may appear narrow but is in fact slightly wide (see figure below). The computer measured it to be 136 ms.

Perhaps the most striking finding in this EKG is the almost complete loss/flattening of the T waves. The computer calculated the QTC to be 427. Looking closer, the P waves, particularly in the precordial leads, are not of uniform morphology. Taking the flattened T waves into consideration, this made me question whether the P waves were only P waves, OR were P waves with superimposed U waves, OR only U waves. Most likely these are sinus P waves with superimposed U waves.

Meyers comment: I agree, I cannot tell where the T and/or U wave is or where it ends. I don't think a QT interval can be meaningfully calculated here.

Taking a global look at this EKG, everything just looks long, regardless of the computer calculations. The top two things on my differential when I see EVERYTHING stretched out like this is electrolyte abnormalities or toxins.

The patient then progressed into an irregular wide complex rhythm which unfortunately was not captured on paper. She was treated with an amp of bicarb and an amp of calcium with normalization of her rhythm. She did not lose a pulse.

This is her EKG after calcium and bicarb:

Likely back in sinus rhythm however the U waves are more discernible so it is difficult to see the P waves. The QRS is now narrow and it is regular. Compared to the first ECG, V2 and V3 now show more recognizable T and U wave morphology, showing that there are indeed U waves superimposed on the P waves. The QU interval, if calculated, would be extremely long, in the range of 600 msec.

The patient managed to tell the team through her stupor just prior to intubation that she had taken a bottle of hydroxychloroquine, estimated later to be approximately a 50-gram ingestion. She was given activated charcoal post intubation and started on high dose diazepam, as well as epinephrine for refractory hypotension.

Initial labs returned with a potassium of 3.2; calcium and magnesium were all normal. On repeat labs, potassium dropped to 1.6 with EKG as below.

Ignoring the digital conversion artifact in the inferior leads, there is likely atrial bigeminy. QRS is narrow and QTC by computer is 629 ms (not sure which algorithm) which results from the computer incorporating the U wave into the calculation. To measure the QTc, we must first differentiate the T wave from the U wave, which is easiest to do using V3. See the enlarged view of V3 below. The first beat in V3 is a sinus beat followed by a premature atrial beat resulting in a U wave and a P wave that are superimposed. This becomes more evident when looking at the second beat in V3 where you can clearly see a T wave and a separate U wave.
(Blue arrow = P wave, Red arrow = U wave, Green arrow = T wave)

Below is the most recent EKG, taken four days later.
You can see the P waves are much smaller now, and likely her normal P waves when compared to prior EKGs.
This confirms that what we saw previously were likely U waves superimposed on P waves. The patient was transferred from the Medical ICU down to the floor the following day.

In the COVID era, we should be familiar with signs of potential hydroxychloroquine toxicity if providers are still using and patients are still taking this medication as it has a very narrow therapeutic window and can be lethal.

Hydroxychloroquine/chloroquine are potassium and sodium channel blockers which can lead to QRS widening, ST changes, QTC prolongation and dysrhythmias.

Hypokalemia is a hallmark feature and degree of hypokalemia is associated with mortality risk. Exact mechanism is unclear but is thought to be due to an intracellular shift. While repleting the potassium may be necessary, Goldfrank’s advises caution with aggressive repletion as hyperkalemic complications have been noted later in the course when the potassium then shifts out of cells. Use of bicarb for widened QRS is controversial as it may exacerbate hypokalemia but is reasonable to use if the potassium level is normal taking into consideration the patient’s entire clinical picture.

Management is mainly supportive with early intubation, glucose therapy for refractory hypoglycemia, and epinephrine for hypotension. Several small human and animal studies have shown a benefit in using high dose diazepam (2mg/kg IV over 30 minutes followed by a high dose drip).

If you have a patient with a possible overdose, call your local poison control center at 1-800-222-1222.

For a more in-depth analysis of the QTc interval and U waves in another similar case, check out this article by Dr. Smith, in which cardiac arrest due to Hydroxychloroquine was treated with intravenous fat emulsion

Take home:
1. Know how to measure QTC, the computer is often wrong.

2. Pay close attention to QRS and QTC in potential tox patients.

3. Know the possible EKG findings of hydroxychloroquine as ubiquitous use in the COVID era can cause serious complications.


Barry, James David. "Antimalarials." Goldfrank's Toxicologic Emergencies, 11e Eds. Lewis S. Nelson, et al. McGraw-Hill, 2019,§ionid=21027 3082.

Crouzette, J et al. “Experimental assessment of the protective activity of diazepam on the acute toxicity of chloroquine.” Journal of toxicology. Clinical toxicology vol. 20,3 (1983): 271-9. doi:10.3109/15563658308990070

Riou, B et al. “Treatment of severe chloroquine poisoning.” The New England journal of medicine vol. 318,1 (1988): 1-6. doi:10.1056/NEJM198801073180101

MY Comment by KEN GRAUER, MD (7/24/2020):
GREAT case, with superb explanation and dissection of serial ECGs in this case by Drs. Coleman, Givergis, Meyers & Smith. My comments are brief:
  • I’ll repeat the emphasis that was made on the importance of accurate assessment of QRS duration. As noted above — at 1st glance, the QRS complex appears to be “normal”. However, as was shown by the magnified view (2nd figure above) — vertical BLUE lines clearly document significant QRS prolongation (ie, to >0.13 sec). This is highly relevant — because QRS prolongation >100 msec is a most important predictor of serious sodium channel blocker toxicity (especially as a predictor of potential seizures).
  • PEARL #1: The reason the QRS complex “looks normal” in the 1st ECG shown above (which I’ll call ECG #1) — is that: i) QRS morphology does not resemble any known form of conduction block in ECG #1 (ie, not RBBB nor LBBB); andii) The initial portion of the QRS complex is completely normal (ie, it looks virtually identical to the initial portion of the QRS complex in serial tracings after QRS duration returned to normal). This terminal prolongation effect is commonly seen with sodium channel blockade and with selected other toxicities — in which QRS widening is primarily a result of widening of the latter portion of the QRS complex.
  • PEARL #2: Hydroxychloroquine/chloroquine is a sodium channel blocker. Prominent for its absence in ECG #1 are 2 of the characteristic signs of sodium channel toxicity = i) Right axis deviation; andii) Development of a terminal tall and wide R wave in lead aVR (felt to be due to delay that occurs predominantly in the right bundle branch). It should be appreciated that these 2 ECG signs should be specifically looked for when suspecting sodium channel toxicity — and, it’s insightful to know that despite ingestion of a very large amount of a potent sodium channel blocker (ie, hydroxychloroquine) — that these 2 characteristic ECG signs of sodium channel blockade were absent.

The 3rd 12-lead ECG shown above manifests a bigeminal rhythm. I found this to be the most fascinating tracing in this case.
  • As shown in the magnified view of lead V3 — the KEY that reveals “The Answer” to a problematic tracing will often be found within the pause in the rhythm (even when this pause is brief, as it is here). U waves are really only seen well in lead V3 of this 3rd 12-lead tracing. It is because of the pause that occurs after the early beat — that the P wave in front of the 3rd (and last) beat in this V3 lead becomes isolated, and clearly identified ( = the last BLUE arrow in this magnified view).
  • Knowing that this last BLUE arrow lies over the last P wave in this magnified V3 lead — tells us that the 2nd RED arrow lies over a U wave that is no longer superimposed on the next P wave. Therefore, it is only because of this short pause (that results from this bigeminal rhythm) — that we are finally able to establish with certainty where the U wave ends.
  • Rather than the T wave — I believe the 2nd GREEN arrow in this magnified V3 view lies over a coved (but not elevated) ST segment. I believe the T wave itself has a negative component, that falls in-between the 2nd GREEN and RED arrows, blending imperceptibly with the U wave.
  • Finally — Beyond-the-Core: This 3rd 12-lead ECG shows a bigeminal rhythm. Although difficult to tell because of the electrical interference artifact in the baseline and because each of the early-occurring P waves is superimposed upon a U wave — P wave morphology in all 12 leads for both the 1st and 2nd P waves in each group looks to be virtually the same. IF we could clearly identify a difference in P wave morphology between sinus P waves and the P waves preceding the early-occurring beats — then we could identify this rhythm as atrial bigeminy. But, when there is a bigeminal rhythm with each beat preceded by P waves with the same morphology — we need to include 3:2 SA block of the Wenckebach (Mobitz I) type in our differential diagnosis of the rhythm.


  1. amazing. thank you all.
    interesting that sodium bicarb, and calcium, which we (i) often might give, especially if considering a sodium channelopathy or toxin, might exacerbate the relative hypokalemia (drive the potassium into the cells) and possibly cause harm.

    1. Excellent point Tom! Fortunately the treating clinicians were carefully monitoring the patient's lab values — :)


DEAR READER: We welcome your Comments! Unfortunately — due to a recent marked increase in SPAM — we have had to restrict commenting to Users with a GOOGLE Account. If you do not yet have a Google account — it should not take long to register. Comments give US feedback on how well Dr. Smith’s ECG Blog is addressing your needs — and they help to clarify concepts of interest to all readers. THANK YOU for your continued support!

Recommended Resources