Sunday, April 22, 2018

The blood sample is hemolyzed. Twice. Is there hyperkalemia on the ECG?

This patient presented from dialysis because his shunt was malfunctioning.  He had no definite indication for dialysis unless his K was high.

But every time we drew the K level, it was "hemolyzed."

So we looked closely at his ECG for evidence of hyperkalemia.
Is there evidence of hyperkalemia?
Here is his previous ECG when K is normal and all his medications were the same at that time.
Are they different?

The objective evidence of hyperkalemia on that top ECG is the prolonged PR interval.  It is over 300 ms!  (The PR in the previous is 218 ms) This prolonged AV conduction puts the patient at risk if his K were to rise further. 

Further, one might be tempted to blame a prolonged PR interval on vagal tone, especially at a slow heart rate of 60 beats per minute.  But since the sinus rate in the old ECG is the same as the new ECG, there is no difference in vagal tone.

So we ordered an admission bed and he underwent dialysis.  A bit later, we obtained non-hemolyzed blood and the K was 5.5 mEq/L, which is elevated but not extremely so.  Nevertheless, the patient is on carvedilol (as he was at the time of the ECG with a PR of 218 ms), and the combination could lead to significant AV conduction delay.

I am a believer that if the ECG truly shows no evidence of hyperkalemia, that there will be no short term adverse events from that hyperkalemia.

The caveat is that one must be able to see all those possible abnormalities.  Some are exceedingly subtle, but real.

Case 3 on this post illustrates this problem very well:

HyperKalemia with Cardiac Arrest. Peaked T waves: Hyperacute (STEMI) vs. Early Repolarizaton vs. Hyperkalemia

There was an article on this published a year ago in Western J Emerg Med:

Severe Hyperkalemia: can the ECG risk stratify for short-term adverse events?

Pendell Meyers wrote a fine FOAM article on this topic on EmCrit:


  1. NICE pickup by Dr. Smith that the PR interval in the 1st ECG is longer (than it was when serum K+ was normal) — as an objective sign of hyperkalemia. I found myself going back-and-forth multiple times comparing lead-to-lead between these 2 tracings. I DO think there is a difference — namely that the peak of many T waves in the 1st ECG are slightly more pointed (specifically in leads I, II, and in each of the chest leads). This is subtle … but together with the increase in PR interval, I believe it DOES support prediction of a probable slight increase in serum K+ level from assessment of this initial ECG. The helpful clinical point — is that these patients often WILL be back (!), as many of them are frequent visitors to the hospital near where they reside. Many of these patients tend to have a “repetitive pattern” — in that their ECG manifests certain characteristic features they they develop when having exacerbation of a chronic condition (be that recurrent ischemia and/or hyperkalemia). Clinicians who regularly work at institutions where these patients frequent can “get to know their ECG” — and I found that when I was in practice, such familiarity often allowed me to immediately recognize subtle ECG changes in comparison to previous tracings that otherwise would have gone unnoticed. So the NEXT time this particular patient returns to the ED with slight increase in serum K+ — it should be that much EASIER to quickly recognize this from repeat ECG. THANKS to Dr. Smith for presenting this interesting case!

  2. Replies
    1. Just the opposite!! Hypokalemia does not lengthen the PR interval. If, significant, it leads to sagging ST segments, U-wave (what often appears to be an impossibly long QT interval, but is really a QU), long QT.

      See these posts:

  3. Do you think that the differences in V4 and V5 are relevant?...

    The decrease in R wave amplitud or that the T waves are more peaked?


    1. I do think so, but that is difficult to see. The PR interval is objective.

  4. Steve...

    A very good topic and well-presented. The article by Dr. Meyers was also very informative. In my years of practice, I saw many cases of hyperkalemia that were very subtle because we had a huge dialysis population (many of whom were non-compliant) who often tolerated high levels of K+ on a daily basis. Their ECGs could be very misleading. It was always my philosophy to treat if there was ECG evidence of hyperK+ on the basis that by the time changes appeared on the ECG catastrophe was not far behind. Unfortunately I have seen a few articles in the last couple of years that suggest there is no need to treat hyperkalemia until the value of the serum K+ has reached a particular level. Although I have retired from clinical medicine, I still believe that is not a good approach. Also, the T waves in the first ECG - though not very impressive - were becoming more symmetrical and beginning to peak.


  5. Hey Dr. Smith, as a nephrology resident I get this scenario quite a lot. malfunctioning shunt i.e. insufficient hemodialysis, tendency towards hyperkalemia, usually not keeping recommended diet and especially "chronic, unspecific symptoms", which cannot be described by most patients. one thing to keep in mind is, that some labs get the message "hemolized" if the k is absurdly high, which lets me draw the blood sample myself in these patients, so I know if there is a mistake in the preanalysis.
    without knowing the ECG under normal k the dysmorphic QRS-complexes in I, II, aVL and V6 made me think, that something's wrong, even though they are below 100ms as far as I can see.

  6. Hey Dr. Smith, is there maybe an ECG recorded after dialysis to compare the PR intervalls? Thanks for the great blog!

    1. Adam,
      I looked! None was recorded. Alas.....


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