Friday, February 28, 2014

Cardiac Arrest: To the cath lab with ongoing chest compressions

A male in his 20's with no past medical history was found in cardiac arrest.  He had been complaining of chest pain.  Medics found him in ventricular fibrillation.  It is uncertain whether he had bystander CPR.  After one shock, he was in PEA, then VF again.  He had multiple doses of epinephrine and also amiodarone.

Of course he was undergoing chest compressions the entire time.  He had a King airway in place.

On arrival to the ED, approximately 50 minutes after arrest, he was still in full arrest.  A rhythm check showed asystole.  Chest compressions were continued with the LUCAS device.  He was intubated.  The ResQPod (inspiratory threshold device) was applied to the endotracheal tube to increase venous return and cardiac output.   He was given more epinephrine, bicarbonate, and calcium in a last ditch effort to resuscitate him.  End-tidal CO2 was 18-20, then improved to 30-40.

A rhythm check revealed a slow regular wide complex rhythm with ST elevation apparent on the monitor.  A cardiac ultrasound revealed no significant motion.   There was only a "flicker" of movement with each electrical complex. 

More Epi, bicarb, and atropine was given.  SpO2 was 80%. 

A 12-lead ECG was recorded at approximately 70 minutes after arrest:
There is a regular wide complex bradycardia at a rate of 37, with no visible P-waves.  The QRS has a right bundle branch block and left anterior fascicular block morphology.  This could be due to either: 1) an idioventricular escape from the posterior fascicle, or 2) a junctional escaped with RBBB and LAFB.  There is obvious ST elevtion in V2-V6 and uiin I and aVL, with reciprocal ST depression in II, III, and aVF.

[RBBB and LAFB, with STE, are associated with massive MI from left main or LAD occlusion.  See this incredible caseAnd also this amazing case of resuscitation after 68 minutes of CPR.]

Case continued:

Chest Compressions were continued.  The interventionalist was consulted.  The cath lab was activated and the interventionalist also recommended tPA until cath could be done (it was early AM, the team was not in-house, and angiogram/PCI would be delayed).  Bolus and infusion were started at about 80 minutes after arrest.

VBG at this time showed pH of 6.94, pCO2 72, bicarb 18 and 69% O2 saturation.  Lactate was 11.3 mEq/L.  Pulseox was mid 80s and etCO2 was 25 mmHg. 

Esmolol was given to help prevent further VF.  (See abstract below: esmolol may be beneficial for refractory VF.  I would be worried that it would be detrimental in PEA.)

The patient was transported to the cath lab at approximately 120 minutes after arrest, with LUCAS CPR ongoing.

Outcome:
Angiography showed occlusion of the LAD and it was opened and stented with a bare metal stent.  However, the patient never regained cardiac function.


Angiography and PCI during CPR


In this study from 2010, 11 of 36 patients who underwent PCI during LUCAS chest compressions survived neurologicially intact.

In this study from 2007, PCI was successful in all 11 patients, but only 3 of 11 survived the procedure and none of them were discharged alive.

Here is a nice summary of research on PCI during CPR, written by Ron Johannsen of Hennepin County Medical Center: They conclude:

"Currently available data do not support routine [my emphasis] activation of the catheterization laboratory for patients without ROSC, even if mechanical CPR is being performed. Whether use of the LUCAS™ or other devices will affect outcome is uncertain; we will have to await the completion of ongoing clinical trials. For patients with cardiac arrest in the catheterization laboratory, however, it seems reasonable to institute mechanical CPR when this is available. Alternatively, other means of cardiac support, such as early transition of patients to ECMO or assist devices, to allow intra-arrest PCI could be mobilized for selected patients."

Who should be considered for PCI during chest compressions, in my opinion? 

  1. Out of hospital arrest: Patients who:
    1. Have witnessed ventricular fibrillation arrest, especially if preceded by chest pain
    2. Receive bystander CP
    3. Are relatively young without comorbidities
    4. Are undergoing excellent CPR (e.g., LUCAS)
    5. Have short arrest to treatment times and can undergo immediate PCI.  
      1. This case had PEA
        1. But ventricular fibrillation was the initial rhythm
        2. Although there was PEA, it was unusual in that the electrical activity was very recognizable and organized and clearly due to STEMI.  
      2. The time from arrest to PCI was probably too long to expect a good outcome.
  2. ED arrest from STEMI/STEMI-equivalent: relatively young without comorbities, good compressions on LUCAS, or put on ECMO.
  3. Cath lab arrest


Fibrinolysis unlikely to be of benefit
tPA has not been shown effective for cardiac arrest, whether with PEA (suggesting pulmonary embolism) or otherwise.  It does not even decrease mortality due to STEMI-related cardiogenic shock.       


Utility of angiography and PCI in patients with ROSC:

Spaulding article on prediction of coronary etiology of arrest: http://www.nejm.org/doi/pdf/10.1056/NEJM199706053362302
84 patients resuscitated and had cardiac etiology, 60 patients with clinically significant CAD, 40 with occlusion, 9 of these had no ST elevation.  The only independent predictors of occlusion were ST-segment elevation (odds ratio, 4.3; 95 percent confidence interval, 1.6 to 2; P 0.004) and chest pain before the arrest (odds ratio, 4.0; 95 percent confidence interval, 1.3 to 10.1; P 0.016).  The presence of one of these two factors was associated with positive and negative predictive values of 0.63 and 0.74, respectively, and the presence of both with values of 0.87 and 0.61.  How expert were the readers of these ECGs at finding ischemia?

Successful angioplasty was the best predictor of survival:  the following factors were predictive of survival: absence of the need for inotropic drugs during transportation to the hospital (odds ratio, 3.6; 95 percent confidence interval, 1.1 to 11.8; P 0.03) and successful coronary angioplasty (odds ratio, 5.2; 95 percent confidence interval, 1.1 to 24.5; P 0.04).

In the study by Silfvast (1991), 78% of patients with ventricular fibrillation had "coronary disease" as the etiology.

Here is an outstanding powerpoint on the need for angiography and PCI after return of spontaneous circulation, written by Demetris Yannopoulos, who, along with Keith Lurie (here at Minneapolis Medical Research Foundation, associated with Hennepin County Medical Center and the University of Minnesota), is doing incredible bench research on cardiac arrest.  Here is an EMCrit podcast on their spectacular advances.

Which patients with ROSC after out of hospital arrest need immediate angiography/PCI? 

The case that Dr. Yannopoulos uses in his presentation to illustrate the "negative" ECG in a patient with arrest and LAD occlusion does, in fact, show ischemia: there is very suble STE in aVL and reciprocal ST depression in II, III, and aVF.  Many would miss this.  Very expert ECG readers would see it.

My guideline
  1. If another definite etiology of arrest is found, other than ischemia, angiography/PCI is not indicated.  
    1. If no other etiology is found, then patients with any of the following should go to PCI:
      1. Preceded by chest pain
      2. Ventricular fibrillation with no known prior cardiomyopathy or channelopathy
      3. Any ischemia on the ECG, as interpreted by a very expert reader.  
        1. Since most ECG interpreters are not very expert, then in most cases any patient without a known etiology should go to the cath lab 




Abstract on Esmolol, to be presented at Social Media and Critical Care 2014, Gold Coast, Australia:

Emergency Department Use of Esmolol in Refractory Ventricular Fibrillation

Brian Driver 1 David Plummer 1 Stephen W Smith 1
  1. Hennepin County Medical Center, Minneapolis, MN, United States
OBJECTIVE
We describe the outcomes for patients receiving esmolol during refractory ventricular fibrillation (RVF) in the emergency department (ED).
METHODS
A structured chart review in an urban academic ED of patients between January 2011 and March 2013 who received esmolol with an ED diagnosis of cardiac arrest (CA), ventricular fibrillation, or pulseless ventricular tachycardia, excluding patients who received esmolol before CA or after sustained return of spontaneous circulation (ROSC).  Cardiac rhythms, CA management, timing of ROSC, and patient outcomes were recorded.
RESULTS
Six male patients met inclusion criteria; one was excluded because esmolol was administered after sustained ROSC.  Four of five patients had out-of-hospital CA; all had automatic mechanical chest compressions delivered by a LUCAS™ device.  All patients received repeated doses of epinephrine, amiodarone, lidocaine, sodium bicarbonate, as well as other adjunctive medications. Defibrillation was attempted many times for each patient prior to esmolol administration (median = 6.5, range 4-10). Some had temporary ROSC, but no patient had sustained ROSC after administration of these medications and defibrillation. All patients had a rhythm of VF at the time of esmolol administration. An esmolol loading dose and infusion of 500 mcg/kg and 50-100 mcg/kg/min, respectively, was subsequently administered to all patients.  One patient with incessant VF achieved temporary ROSC and three others attained sustained ROSC after the administration of esmolol with repeat defibrillation; two survived to discharge with excellent neurologic outcomes.
CONCLUSION
Beta-blockade should be considered in all patients with RVF in the ED prior to cessation of resuscitative efforts.

Wednesday, February 26, 2014

Prolonged chest pain and Intraventricular Conduction Delay similar to Left Bundle Branch Block



A 67 year old male with no previous cardiac history or risk factors woke up at midnight with right-sided chest pain. It was the "worst pain of his life," and it radiated around his ribs, through to his back, and up into his neck, all on the right side. He was profusely diaphoretic, nauseated, and he vomited twice. He took some aspirin which gave him some minor improvement in symptoms. He "toughed out" the night, and when his pain was still not getting better about 8 hours later, he asked his wife to take him to the ED.   

He arrived at about 8.5 hours after onset of symptoms, still with severe chest pain, and had this ECG recorded:
What do you think?



There was a baseline ECG available from 3 years ago:

What is your interpretation? What would you do at this point? See below for interpretations.











Interpretation of presenting ECG:
Sinus tachycardia at about 105 bpm. The computerized QRS duration is 120ms, qualifying this for an Intraventricular Conduction Delay.  It has morphology very similar to LBBB: there is a wide complex with dominant S waves in the anterior leads, so we presumably can apply the same rules as full-blown LBBB.

The treating physicians documented that this ECG is "Sgarbossa negative."  There is no concordant ST depression in V1-V3, and no ST elevation equal or greater than 5mm.  Is there concordant ST elevation?  This depends on whether the QRS is mostly positive or mostly negative in lead V5; if positive, then the ST elevation is indeed concordant..

Lead V4 alone would make this ECG "modified Sgarbossa positive"(reference 1, reference 2): it has a 3mm S wave with 2.5mm of ST elevation, making a STE/S ratio of 0.83.

How about lead V5?  The S-wave and R-wave are of nearly equal voltage; on the other hand, the S-wave is slightly wider than the R-wave. What matters most?  Is it the voltage?  Or is it the integral (area under the curve) that matters most?   In any case, the difference in voltage and in area is not great, and therefore there should be almost no ST deviation in that lead. 

Stated in other words: The "area under the curve" or "integral" is all the area contained between the waveform and the isoelectric line.  In LBBB, ST-T is normally discordant to the majority of the QRS, but is that "majority" measure by voltage (in mm of amplitude), or is it best measured by area under the curve?  In this case, in V5 the R-wave amplitude is greater, but the S-wave area is greater.  In either case, the difference is small.

Thus, leads V4 and V5 are diagnostic of STEMI.

Furthermore, one expects a small R-wave in V1 and V2 in LBBB.  Instead, there are Q-waves.  There is also a notch on the ascending limb of the S-wave in V2 and also slightly in V3.  These are reminiscennt of  "Cabrera's sign" (a notch greater than 50 ms on the ascending limb of the S-wave in one of V3-V5).  These are signs of previous MI, or of well developed "subacute" acute MI.  

All of these ECG findings, along with the clinical scenario, are all but diagnostic of a subacute STEMI in the setting of LBBB.

Initial troponin I was 50 ng/ml. Cardiology was summoned. They took him immediately to the cath lab, where they found an acute thrombotic 99% stenosis of the proximal LAD with TIMI 1 flow. There were also thrombotic lesions of the mid LAD and D1. They aspirated the thrombi and placed 3 stents at these lesions. No further troponin data was available.

Here is his ECG later that day:
T wave inversions in the anterolateral leads, consistent with reperfusion. Some ST elevation persists, but no longer meets any criteria. The QRS appears a little bit shorter than previous.

And here is his ECG the next day:

Still shows T wave inversions with persistent ST elevation.  This persistence may portend the development of an LV aneurysm.

He recovered well and was discharged several days later.

Should thrombolytics be given?

Thrombolytics are still recommended up to 12 hours after the onset of pain. (Sorry, no full text here: this is an analysis of thrombolytic trials from 1983-1993, and found that if pain has been present for 6-12 hours, then a mean of 18 lives were saved per 1000 patients treated with lytics vs. placebo).

The ECG is, in fact, an even better measure of acuteness of a STEMI. 

When there is subacute STEMI, the thrombolytic decision must be made carefully with attention to both risk and benefit.  This is a large anterior STEMI with persistent pain and ST elevation.  There are Q-waves, and the highly elevated troponin I confirms prolonged infarct.  ST elevation is still present, however, and T-waves have not yet inverted, so there is still significant salvageable myocardium at risk.  Depending on the patient's risk factors for bleeding, and on the door to balloon time for transfer to a PCI institution, thrombolytics may be indicated.  


Take Home Points:
 - the modified Sgarbossa rule is more sensitive than the original
 - even when the ECG doesn't have a perfect LBBB, if there is significantly abnormal depolarization with a wide QRS it must still follow the same rules of appropriate discordance and proportionality
-Signs of old or subacute MI may also be seen in the setting of LBBB

Tuesday, February 25, 2014

Is This a Simple Paced Rhythm?

A very complex patient presented critically ill.  This was his initial ECG:

There are pacer spikes.  Is this a simple paced rhythm?

















There is a paced rhythm with a sine wave, diagnostic of hyperkalemia.  The potassium was 6.8mEq/L.

He was aggressively treated for hyperkalemia.  Another ECG was recorded 42 minutes later:
The QRS is still very wide (about 250 ms), but substantially narrower.  



Here is a recent baseline ECG:
This has a rather long baseline QRS at about 180 ms, but not nearly as long as with the high K.


A paced rhythm with a duration of greater than 180 ms is very long and, if it is chronically long, it is associated with poor prognosis.  

When the QRS in paced rhythm is acutely prolonged, as in this case, all of the etiologies of prolonged QRS should be on the differential: especially hyperkalemia and sodium channal blocking drugs.


Thursday, February 20, 2014

Complaint of Weakness. Alert. Hemodynamically Stable. What is it?

A male in his 60's, dialysis patient, complained of weakness.  He was alert and had no SOB or Chest pain.  BP was 85/57 with a pulse of 65.  Here is his ECG:

What is it?  See below.






















This is pathognomonic for hyperkalemia and is nearly end stage, almost a complete sine wave morphology.  It seems miraculous that a patient can have this ECG and not be in profound shock, but it is true. 

He may die at any moment though.

His potasssium was 8.8 mEq/L.

There were access problems, but he did get insulin, glucose, bicarb, calcium gluconate, and albuterol, all while preparing emergency dialysis. The treatment resulted in the following ECG 36 min later:

Although still scary, it is now much better.  Are there P-waves?  Perhaps; I find it hard to tell.

After another hour and more treatment including more calcium, and still awaiting dialysis, this ECG was recorded, with a corresponding K of 6.6 mEq/L:
It appears slightly narrower, but still scary


The patient's level of consciousness deteriorated and he was intubated, taken for dialysis, and survived.  Here is his post dialysis ECG which is nearly identical to his baseline:


My next post will show another example of severe hyperkalemia with sine wave morphology.

Monday, February 17, 2014

ST Elevation and a Positive Troponin: Is it Myopericarditis or STEMI?

A 51 yo male smoker with no other CV risk factors and no previous heart disease presented to a hospital without PCI capability complaining of 1.5-2 hours of severe epigastric pain radiating to the shoulders.  He was diaphoretic and agitated.  He had been treated for "flu" manifesting as fever and sore throat for 5 days.  He was afebrile.  He had no pericardial friction rub.

Here is his initial ECG:

There is sinus rhythm with a QTc of 434 ms.  There is a possibly significant Q-wave in lead III. There is ST Elevation in leads I, II, aVF (+minimal in lead III) and V4-V6.  The T-waves are not definitely hyperacute, but also possibly slightly enlarged.  The ST vector is approximately 30 degrees, between leads I and II, opposite aVR, which has the most ST deviation (in this case, ST depression).  And it is towards V5.  There is no reciprocal ST depression in lead aVL (if there were, the ST vector would have to be greater than 60 degrees).  There is no definitely excessive PR depression.  There is ST depression in V1 also. Is there ST depression in lead V2?  There is some artifact and wandering baseline.  If there is ST depression in V2, it is very important.  There is no appreciable PR depression.
The obvious differential diagnosis here is inferolateral +/- posterior STEMI vs. Myo- or pericarditis (or baseline ST elevation -- early repolarization).

Favoring STEMI:  

ST depression in V1 and possibly V2.  One should never be confident in an ECG diagnosis of (myo-) peri-carditis if there is ANY ST depression.  There may be ST depression in localized (as opposed to the much more common diffuse) pericarditis, but this is relatively rare.

Favoring Myopericarditis: 

Absence of reciprocal ST depression in lead aVL. (In our study of 150 patients with inferior STEMI, 33 had ST elevation in leads V5 and V6 (inferolateral STEMI).  All of them had some ST depression in lead aVL). 

This is a very difficult ECG.  I always say that "you diagnose pericarditis at your peril" because, if there is no rub or effusion, and the ECG is not absolutely typical of pericarditis, and the clinical scenario is consistent with STEMI, then you must act as if it is a STEMI.

The best way to rapidly differentiate is to do a high quality echocardiogram or, of course, just go to  the cath lab.

Clinical Course

A stat echo showed no pericardial effusion, but technique was inadequate to find or exclude a wall motion abnormality.

The initial troponin I returned at 8 ng/mL.  (myopericarditis?  or STEMI?)


With suspicion of STEMI, the patient was promptly transferred to a PCI-capable hospital
(door-to-door time: 30 minutes) and had this ECG on arrival:
Now there is much less ST deviation.  This would be rapid evolution for myopericarditis, and more typical of a coronary thrombus (or spasm) that has lysed (resolved).


He underwent coronary angiography by a very experience angiographer which ruled out coronary occlusion and had normal coronary arteries.  However, there was no IVUS (intravascular ultrasound) to absolutely prove absence of extralumenal plaque with culprit).

After angiography the patient was taken back to the hospital where he first presented
and was treated as for myo-pericarditis with anti-inflammatories.  A repeat high quality echocardiogram showed minimal pericardial effusion in the posterior and lateral sides. There was no evidence of wall motion abnormality.

After angiography, troponin I peaked at 18.120 ng/ml with a rapid fall thereafter to 0.143 ng/ml on day 4.

Here is the ECG on day 2:


All ST deviation is resolved.  There is T-wave inversion in I and aVL.  This evolution is consistent with either myopericarditis or STEMI

And here is the ECG on day 3 (48 hours):
Unchanged

No MRI or biopsy was done to prove or disprove myocarditis.  There was a mildly elevated CRP and ESR.


STEMI with normal angiogram: Even in cases where there is significant intraluminal plaque, most of the plaque is extraluminal.  When there is no visible plaque on angiography (a "lumenogram"), there still may be much extralumenal plaque which can ulcerate, producing thrombus, which occludes the artery, resulting in ischemia and ST elevation, but then spontaneously lyses.  The ischemia resolves, the wall motion recovers, the angiogram shows no stensosis and possibly even no ulcerated plaque.  Only use of intravascular ultrasound (IVUS) would be able to demonstrate plaque.  Coronary spasm may also result in ischemia, ST elevation, and a normal angiogram.

Troponin: Does a rapid rise and fall of troponin indicate STEMI with reperfusion rather than myocarditis?  One would think so, but I have not been able to find any hard data on this.  I have seen many cases of purported, but not proven, myocarditis that had a rapid rise and fall.  As inflammation is a more constant process than ischemia and reperfusion, I would have guessed that it would have a more steady elevation of biomarkers.  Does anyone know?  The "state of knowledge" paper referenced above makes no comment on this.
In this study of 625 patients with presumed STEMI who had angiograms within 2 hours of the ECG, but not simultaneous with the ECG, 26 had completely normal angiograms. The final clinical diagnosis by 3 cardiologists (not by MRI or biopsy) was STEMI in 7 (1.2%) and myopericarditis in 5 (0.8%).  Normal angiograms (no stenosis at all) seem to occur in about 4.5% of patients with Non-STE-ACS, but they still have substantial subsequent morbidity (2% death or MI at 1 year; TIMI score stratifies them well)

Finally, there is this paper by Sarda et al. from JACC in which they studied 45 patients with apparent STEMI but with normal coronary arteries.  They used Indium scintigraphy to assess whether it was MI or myocarditis.  In 35 of 37 cases with adequate followup, it was myocarditis.  In 2, it was MI.  However, this article states that scintigraphy has poor sensitivity and specificity for myocarditis: "Current state of knowledge on aetiology, diagnosis, management, and therapy of myocarditis: a position statement of the European Society of Cardiology Working Group on Myocardial and Pericardial Diseases."

Was the initial management correct?  Yes.  This could have been STEMI and should be managed as such.  It still might have been STEMI.  The ST depression in V1 and V2 is particularly worrisome.  Had it been possible to get a high quality echocardiogram at the time of the ST elevation, and it had been completely normal, then it would have been reasonable to diagnose myopericarditis without an angiogram. 

What is the diagnosis?  Is it clearly and unequivocally myocarditis? The diagnosis can't be made for certain without MRI, biopsy or at least a pericardial friction rub.  There was an effusion later on echo, strongly supporting the diagnosis. Additionally, profound PR depression (or elevation in aVR) would be very supportive.



Sunday, February 9, 2014

Dyspnea on Exertion and a Positive Troponin

A woman in her 50's complained of dyspnea on exertion for two days.  Her lungs were clear on exam.  Her first troponin I returned at 0.1 ng/mL (elevated, upper limit of normal = 0.30 ng/mL).   Here is her ECG:
Sinus tachycardia.  The QRS duration is 103 ms and there is some left axis deviation.  What is the likely diagnosis (before I give more clinical information)?  -- See below

More data on this patient:

She only had a h/o well controlled AIDS and hypertension.  She had no fever or cough. Her oxygen saturations were 92% on room air.  On exam, she had clear lungs, normal heart sounds, and no peripheral edema.   A chest x-ray was normal.  A bedside echo showed "grossly normal function, good global contractility, no obvious wall motion abnormalities, no pericardial effusion and no B lines present in her lungs."  They did not comment on the right ventricle.

The BNP was significantly elevated.  Hemoglobin and electrolytes normal.

Clinical course

After admission, a D Dimer was ordered and returned at 6x the upper limit of normal.  A CT pulmonary angiogram showed multiple bilateral superior and inferior segmental pulmonary arteries with central filling defects (extensive pulmonary emboli.)  There was leftward bowing of the interventricular septum consistent with right heart strain.

The clinicians were cognitively steered down the ACS pathway because of the elevated troponin, and also down the heart failure pathway due to the elevated BNP.  

However: This is a classic case of PE until proven otherwise:


1. When there is hypoxia with a normal chest x-ray, put PE near the top of the differential.  Also consider: hypoventilation (including low respiratory drive), vasodilators and sepsis (both of which cause shunting, with V-Q mismatch), and, most commonly, asthma.

2. ACS infrequently is associated with sinus tachycardia (usually when it is severe ischemia with cardiogenic shock and pulmonary edema)

3.  Positive troponins in unselected ED patients, especially at a low level, are more commonly due to demand ischemia than to ACS.  Think CHF, hypertension, sepsis, PE, etc.

4. BNP is elevated in PE as well as in heart failure, from right ventricular strain.

5. This ECG is classic for PE: 1) sinus tach 2) right ventricular conduction delay (R' in V1) 3) T-wave inversions in BOTH precordial leads and in lead III.  4) S1Q3T3 (a tiny R-wave in III is equivalent to a Q-wave).  (see the primer on the ECG in PE below)


I was sent this ECG without clinical information and asked what it was.  My first thought was "pulmonary embolism."  All because of the above ECG findings.

Here are other cases of acute right heart strain:

Case 1
Case 2
Case 3 (is it acute right heart strain or not? -- with echocardiogram)
Case 4
Case 5 (with echocardiogram)


Primer on the ECG in Pulmonary Embolism:

These are findings of acute right heart strain, and could be seen in any condition which results in a rapid rise in pulmonary artery pressure. This includes hypoxia because of "pulmonary hypoxic vasoconstriction" 

The ECG is not sensitive for PE, but when there are findings such as S1Q3T3 or anterior T-wave inversions, or new RBBB, then they have a (+) likelihood ratio and the S1Q3T3, or even just the T3, may help to differentiate Wellens' from PE. 

Stein et al. found normal ECGs in only 3 of 50 patients with massive PE, and 9 of 40 with submassive PE.  Today, however, that number would be higher because we diagnose more of the submassive PEs that have minimal symptoms.

This is a paper worth reading: Marchik et al. studied ECG findings of PE in 6049 patients, 354 of whom had PE.  They found that S1Q3T3 had a Positive Likelihood Ratio of 3.7, inverted T-waves in V1 and V2, 1.8; inverted T-waves in V1-V3, 2.6; inverted T-waves in V1-V4, 3.7; incomplete RBBB 1.7 and tachycardia, 1.8. Finally, they found that S1Q3T3, precordial T-wave inversions V1-V4, and tachycardia were independent predictors of PE. 

What is an S1Q3T3?  Very few studies define S1Q3T3.  It was described way back in 1935 and both S1 and Q3 were defined as 1.5 mm (0.15 mV).  In the Marchik article, (assuming they defined it the same way, and the methods do not specify this), S1Q3T3 was found in 8.5% of patients with PE and 3.3% of patients without PE.

Kosuge et al. showed that, when T-waves are inverted in precordial leads, if they are also inverted in lead III and V1, then pulmonary embolism is far more likely than ACS.  In this study, (quote) "negative T waves in leads III and V1 were observed in only 1% of patients with ACS compared with 88% of patients with Acute PE (p less than 0.001). The sensitivity, specificity, positive predictive value, and negative predictive value of this finding for the diagnosis of PE were 88%, 99%, 97%, and 95%, respectively. In conclusion, the presence of negative T waves in both leads III and V1 allows PE to be differentiated simply but accurately from ACS in patients with negative T waves in the precordial leads."

Witting et al. looked at consecutive patients with PE, ACS, or neither. They found that only 11% of PE had 1 mm T-wave inversions in both lead III and lead V1, vs. 4.6% of controls.  This does not contradict the conclusions of Kosuge et al. that when T-wave inversions in the right precordial leads and in lead III are indeed present, then PE may indeed by more common.  In my experience, this is true, but needs validation in a study of similar methodology. Supporting Kosuge, Ferrari found that anterior T-wave inversions were the most common ECG finding in massive PE.