ECG Filters, or My Least Educational Post Ever

The other day I was driving for the shift.  My partner and I responded to a high school for the report of syncope on the football field.  Practice was going on and one of the coaches had fainted.  In checking him out, my partner found out that coach was 55 years old, had felt weak and lightheaded, and thinks he just got overheated out in the sun.  His vital signs were normal and the coach was only complaining of residual weakness. 

We put him on the monitor.  My partner hit print (he apparently hadn’t read the Quick Six rant from a few weeks ago):

Dude was very worried about the inferior ST depression.  I  finished putting the precordial leads onto the patient.  The 12-lead that resulted, cut down for privacy of course:

Take a closer look at the two ECGs.  First, the third beat of Lead II in the print view:

Now, look at the second beat in Lead II from the 12-lead:

Do you see the difference?  The printed view has one and a half or two millimeters of ST depression.  In the 12-lead, though, the depression is gone.  Maybe there is half a millimeter of depression in the 12-lead, but there isn’t much.  What caused the difference?  (Hint: Look at the whole strips above, rather than the individual pictures.  The difference is spelled out.)

The signals collected by an ECG monitor are very small.  The heart’s electrical signal is around 1 millivolt, and many waveforms are much less than that.  So the ECG signal can be easily drowned out by other “noise.”  The extraneous signal can be from other muscular movements, nearby electrical equipment, nearby power lines, poor electrode-skin interface, and from equipment in the monitor itself.  The collected ECG signal must have the extra noise filtered out so that a clean picture of the heart’s electrical activity can be shown more clearly.

I need to take a minute here and explain that I am a little embarrassed.  You see, my plan for this week’s post was to explain how ECG filtering works.  So I spent all week researching the topic.  And failed.  Miserably.  I don’t understand the technical aspects of how filtering works on an ECG.  For those of you who are smarter than I am, or have a better technical background, there are several explanations on the internet.  Please take a few minutes to go check them out.  This one, for example.  Or this one.  Here is a powerpoint presentation in pdf form that tries to explain the topic.  All they did is confuse me and make me feel dumb.  So now I am sad.

Good friggin’ luck to you if you want to get into the specifics of signal filters.  If you can understand the topic (understand it, not just regurgitate it), then you are a better person than I.  If it is something that interests you, to read about log scales, low pass filters, Fourier Analysis, superposition, and phase distortion, well then knock your bad self out with the links above.

Here is what the average medic needs to know about ECG filtering: There is a difference between the print view and the 12-lead.  By the way, the screen on the monitor shows the print view.  The 12-lead is diagnostic.  You can make decisions based on a 12-lead.  It is the same view as on a 12-lead in the hospital.  The same thing cannot be said for the print view. 

Look at the bottom left of the printed ECG and the 12-lead above.  The print view is filtering at the 0.5-40 hertz level.  That is, there is more aggressive filtering going on to smooth the waveforms.  That filtering, though, can do weird things to the ST segments like, say, depress them by 2 millimeters.

The bottom left of the 12-lead shows the filtering is being performed at the 0.05-150 hertz level.  This is a diagnostic view.  It is going to result in more artifact, baseline wander, and general noise.  But it is also going to give you the accurate view of the patient’s heart’s electrical activity.  ST segments are accurate at this level of filtering.  In order to cut down on the noise in the ECG, we have to be very precise with electrode application, electrode location choice, vehicle movement, patient movement, and those kinds of things.  We have to make physical changes to improve the quality of the ECG, rather than relying on the monitor to filter out the noise for us. 

One way or the other, make sure that you are making decisions based on a diagnostic-level view of the patient’s ECG.  It would be pretty embarrassing to STEMI alert a patient based on falsely elevated print view ST segments. 

The print view can give you false positive and false negative findings.