Isolated ST Segment Depression: Not a Sign of Inferior Ischemia

I have long maintained that ST segment depression does not localize and the isolated ST depression in "inferior" leads is actually reciprocal to less obvious ST elevation in lead aVL or in "anterior" leads.

A reader alerted me to a 2010 paper that addressed this issue.  They looked at all ACS cases in a CCU over a 12 year period, chose those with isolated "inferior" ST depression without any ST elevation, and found that only 10% had inferior ischemia.  Most had LAD or first diagonal culprits, and most were due to occlusion or high grade thrombotic stenosis (i.e., cath lab indicated).

(The one example they showed was an LAD occlusion that would also have been identified by the LAD occlusion rules and formula).

Here is a link to the paper (unfortunately, no full text):

Isolated inferior wall ST segment depression as an early sign of acute anterior wall myocardial infarction

Here is a case of a first diagonal occlusion that demonstrates this fairly well:
This one does have minimal but significant ST elevation in aVL, so perhaps not the perfect example.

Here are some other  cases that demonstrate this.

Is there ST Segment Depression here?

A middle aged male with a history of Hypertension, Diabetes, Hyperlipidemia, and CHF due to severe concentric LVH presented with severe substernal chest pain and SOB.   The blood pressure was 170/120.

There were B-lines on chest ultrasound, diagnostic of pulmonary edema.

Here is his initial ED ECG:
Sinus rhythm.  There is high voltage diagnostic of LVH.  With a large S-wave in V2 and V3, there should be baseline ST elevation of LVH in these leads.  However, the J-point in V2 and V3 is isoelectric.  This is abnormal.  This is highly suggestive of relative ST segment depression.

So the previous ECG was sought:
Here we see the expected ST baseline ST elevation of LVH in V1-V3.  It is proportional and appropriate.

This confirms that the presentation ECG shows relative ST depression, consistent with ischemia.  There was worry for subendocardial ischemia vs. posterior STEMI.  The elevated blood pressure makes one lean heavily towards subendocardial demand ischemia.

A review of the chart showed there was an angiogram from one year prior with no significant obstructive lesions, mild plaque only.  This does not rule out ACS as the etiology of ischemia, but makes it far less likely.

The patient was placed on a high dose nitro drip, and his BP dropped to 150/90.  His symptoms resolved.

Another ECG was recorded 40 minutes after the first:
The relative ST depression is resolved.

The patient had a mild rise in troponin to 0.084 ng/mL which was deemed due to demand ischemia, not ACS.  Of course there could possibly be ACS, but I think it is unlikely.

Learning Point:

1. An isoelectric J-point may represent ST segment depression in a patient with baseline ST elevation, and in a patient who should, due to QRS abnormalities such as LVH, have baseline ST elevation.

ST elevation and depression: is it ischemia? The entire clinical context is critical.

This patient has a history of cardiomyopathy (EF 10%) and IVCD and heart failure and presented with altered mental status.   BP was 87/52, pulse 90, saturations 100%, and respirations 14.   An ECG was recorded:
There is sinus rhythm.  The P-wave in V1 has a huge negative component, diagnostic of left atrial enlargement.  The PR interval is 360 ms (the computer read it as 227 ms).  The QRS = 134 ms, so it is an IVCD (intraventricular conduction delay), but is not LBBB because of Q-waves in I and aVL.

There is a large amount of ST elevation in V3-V5, and ST depression in V6.  But there are also very deep S-waves in V3-V5 and a very tall R-wave in V6.  The highest ST/S ratio is about 10% (normal).  

There was an ECG from 4 months prior:
Atrial fib.  QRS = 118 ms.  Voltage was high then too, but ST segments were not.  So the ST elevation is new.

The physicians were worried about this and showed it to me.

I said that this is very unlikely to be ischemia because of:

1) the severe cardiomyopathy
2) the high voltage
3) relative clinical stability in a patient with a baseline EF of 10%
4) presentation without chest pain

I recommended a bedside echo and further eval. This was done and showed very poor LVF and no WMA.

Electrolytes and Venous Gas results returned:

Na = 117 mEq/L
K = 2.7 mEq/L
Chloride = 68 mEq/L  (this indicates metabolic alkalosis)
CO2 = 46 mEq/L
Anion Gap = 3 mEq/L (so there is no serious metabolic acidosis)
Venous pH = 7.58 (normal = 7.33), corresponds to an arterial pH of 7.65 (very alkalemic)
HCO3 = 45
Venous pCO2 = 47 (normal = 47), which corresponds to an arterial pCO2 of about 40 mmHg.

This actually represents a superimposed respiratory alkalosis: when the HCO3 = 45, the patient should compensate with a decrease in minute ventilation, such that the arterial pCO2 should be about 57 (0.9 x HCO3 + 15), and venous pCO2 should be slightly higher, around 62-64 mmHg, not 47 mmHg.

For a comprehensive presentation on Acid Base disturbances, see my 55 minute lecture on the topic.

So the patient has profound metabolic disturbances [hypokalemic metabolic alkalosis without any respiratory compensation (and thus a superimposed resp alkalosis)], and this explains the ECG findings.  After some treatment, this ECG was recorded 23 hours after the first:
It is normalizing

At 36 hours, after more electrolyte correction:
Looking more and more like the old ECG

At 3.5 days, there was some tachycardia:
This was diagnosed as SVT, but it is, in reality, sinus tachycardia.  See image below with arrows.

Arrows point to P-waves. Note the P-wave morphology, and PR interval, are the same as in the slower sinus rhythms.

Learning Points:

Cardiomyopathy, especially with metabolic derangement, can result in profound PseudoSTEMI patterns.

Paced rhythm. Are there hyperacute T-waves? See the T-wave evolution, even in paced rhythm.

This was sent to me by John Larkin, from Australia, who has a great ECG Blog called "ECG of the Week," where he posted it, and graciously allowed me to post it here.

The ECG is from an elderly male with multiple co-morbidities including significant cognitive impairment, cardiac failure and diabetes. He presented to the Emergency Department with several hours of chest pain and has a pacemaker in-situ for an unknown indication.

Here is his ECG:
John's thoughts on the ECG are as follows:

·         Regular V-paced rhythm at 60 bpm with LAD.
·         Lead V2 has concordant ST elevation but less than 1mm
·         Lead V3 has excessive discordant ST elevation (ratio -0.33)
·         Lead V4 has excessive discordant ST elevation (ratio -0.32)
·         Lead V5 I suspect has excessive discordance but the baseline wander and native P wave superimposition makes it difficult to measure
·         The T waves in leads V2-6 are huge compared with the QRS voltages
·         Evidence of non-conducted native atrial activity

My thoughts

I was less certain of LAD occlusion because of the profound upward concavity, and I didn't think any lead met a consistent ratio of 0.25, though I think V3 did meet a ratio of 0.20.  But was very suspicious of the very large T-waves.

One should keep in mind that in our LBBB studies (we have two now, with data forthcoming on the second, and larger, validation study) an ST/S ratio of 0.20 is more sensitive though slightly less specific, than 0.25. 

But a ratio of 0.20 is still nearly 90% specific.

All this data is for LBBB and I do believe that it applies to paced rhythm.  


The patient was not taken for angio given their extensive co-morbidities (including GI bleed) following discussion with both cardiology and patient’s family. He did have a troponin rise and was treated with maximal medical therapy.

Although he did not have an angiogram, followup ECGs, echos, and troponin prove that he did, indeed, have an LAD occlusion.

His initial troponin I was 6.7 ng/mL.

Echo showed EF of 30% with large amount of severe segmental systolic dysfunction.

This was 130 minutes later:
Now lead V4 has a ratio of 0.25

This was recorded at 6 hours:
The hyperacute T-waves are gone.  The ST segments have nearly resolved.

This one is at 20 hours
Still more resolution of ST elevation and T-waves

T-wave starting to invert

In some data we have recently collected and will publish, we have shown that reperfusion of an occluded artery in LBBB, just like in normal conduction, results in reperfusion T-waves in many, if not most, cases.

Is this also true in paced rhythm?

This one is at 43 hours
More T-wave Inversion (reperfusion T-waves vs. evolution of completed MI)
345 hours
Much deeper T-wave inversion
454 hours
T-waves are deeper still

Learning Points

1.  Hyperacute T-waves may be seen in paced rhythm
2.  The Modified Sgarbossa criteria are likely applicable in paced rhythm as well as in LBBB
3.  The ECG often evolves T-wave inversion in reperfusion of an occluded artery in both LBBB and in Paced rhythm.

Paced rhythm, Modified Sgarbossa negative, but with a Wall Motion abnormality.

This case was posted by my colleagues at the Hennepin Ultrasound Blog, as Apical Wall Motion Abnormality or Electrical Asynchrony?

But there was no comment on the ECG, which is very interesting.


A male in late middle age with a history atrial fibrillation, significant renal insufficiency, and implanted single chamber right ventricular pacemaker, but no known coronary disease, presented with 2 hours of sudden onset chest pain.  It felt like heavy pressure.  The pacer was placed 2 months prior, and the patient had no ECG recorded after placement to establish a baseline.

Here is his ED ECG:
There is a paced rhythm with proportionally discordant ST segments.   There are no concordant ST segments.  Thus, there is no evidence of STEMI.

Unlike Left Bundle Branch Block, there is comparatively little data on the accuracy of the Sgarbossa criteria in diagnosing acute MI (better, acute coronary occlusion) in the setting of a paced rhythm, and no data on the Modified Sgarbossa criteria.  There are two small studies showing good specificity of the traditional Sgarbossa criteria, but poor sensitivity.  There are 2 case reports (both by me and co-authors, references below) of STEMI diagnosed in paced rhythm by proportionally excessive discordant ST elevation (modified Sgarbossa criteria).   I have posted other cases in which STEMI was easily diagnosed in the setting of paced rhythm.  Here is one.  Here is another.  But we really don't know the sensitivity of concordant STE or proportionally excessively discordant STE in paced rhythm.  I suspect it is as good in paced rhythm as in LBBB.

So the ECG in this case does not meet any Sgarbossa criteria

However, this patient has very suspicious symptoms and you do not want to miss a coronary occlusion.

So a bedside echo was done; here is the parasternal long axis view:

It looks as if the apex is not moving well.

Here is the apical view:

There is clearly an apical wall motion abnormality.

And this apical view shows the WMA clearly as well.

A review of the patient's chart showed that the last formal echo had been normal, but had also been done prior to placement of the pacemaker.    Right ventricular pacing results in an abnormal sequence of activation and so there may appear to be a wall motion abnormality.

So the ECG shows no indication of occlusion, but the echo shows a new wall motion abnormality that can be completely due to the pacemaker.

The patient has a relative contraindication to coronary angiogram (renal insufficiency).

It was decided not to take the patient to the cath lab.  Troponins were negative and the pain resolved.

Formal ultrasound done later had the same findings.

This ECG was recorded 4 hours after the first:
The T-waves are slightly taller, but there are no abnormal ST segments.

Learning points:

1. Use the modified Sgarbossa criteria in paced rhythms.  The specificity is good. The sensitivity is unknown but probably similar to sensitivity in LBBB, which I believe to be as good as ST elevation in normal conduction, probably about 70-75% sensitive for coronary occlusion (though the sensitivity was much higher in our case control studies, which probably do not accurately reflect clinical practice).

2. The ECG in this instance was more reliable than formal echocardiography

Here is the ECG from the first case I link to:
This shows both concordant ST elevation in V2, and proportionally excessively discordant STE in V3.  There is also concordant ST depression in V6.

One difference between LBBB and Paced rhythm is that, in paced rhythm, the QRS in V5 and V6 is almost always negative (but positive in LBBB).  Therefore, any STEMI that manifests in V5 and V6 in LBBB will usually manifest by concordant ST elevation in these leads, whereas in paced rhythm, it must be excessively discordant ST elevation.


1. Sgarbossa EB, Pinski SL, Gates KB, et al. Early electrocardiographic diagnosis of acute myocardial infarction in the presence of ventricular-paced rhythm.(full text pdf) GUSTO-I investigators. Am J Cardiol. 1996;77:423–4.

2.  Maloy KR, Bhat R, Davis J, Reed K, Morrissey R.   Sgarbossa criteria are highly specific for acute myocardial infarction with pacemakers. (full text link)  West J Emerg Med. 2010;11(4):354-357.

modified Sgarbossa rule. (full text link) Ann Emerg Med. 2012;60(6):766-776.

4. Schaaf SG, Tabas JA, Smith SW. A patient with a paced rhythm presenting with chest pain and
hypotension. JAMA Intern Med. 2013;173(22):2082-2085.

5. Ukena C, Mahfoud F, Buob A, Böhm M, Neuberger H-R. ST-elevation during biventricular pacing. Europace 2012;14(4):609-611.

6. Karumbaiah K, Omar B. ST-elevation myocardial infarction in the presence of biventricular paced rhythm. J Emerg Med. 2013;45(2):e35-e40.

7. Walsh B.  Smith SW.  A Patient with a Biventricular Paced Rhythm Presenting With Chest Pain.  Challenges in Clinical Electrocardiography. JAMA Internal Medicine.  April 6, 2015.