Chest pain, normal ECG, and serial undetectable high sensitivity troponins. Should we do further testing?

Many authors state that if troponins are undetectable and ECG is normal, no further testing should be done.  For example, in JACC:

Bandstein N, Ljung R, Johansson M, Holzmann MJ. Undetectable high-sensitivity cardiac troponin T level in the emergency department and risk of myocardial infarction. J Am Coll Cardiol [Internet] 2014;63(23):2569–78. Available from: http://dx.doi.org/10.1016/j.jacc.2014.03.017

Conclusions

All patients with chest pain who have an initial hs-cTnT level of <5 ng/l and no signs of ischemia on an ECG have a minimal risk of MI or death within 30 days, and can be safely discharged directly from the ED.   

It is very dangerous to make such general definite statements: "All patients...."

Case

A 70-something with history of hyperlipidemia and chronic back pain presented to the emergency department with chest pain.  

Patient c/o left sided chest pain described as squeezing intermittently for 10 days.  One note stated that there is no association with exertion and that it comes without a known trigger, but another stated that there is association with exertion.  Last episode of chest pain was 2 days prior.  Stated able to go up stairs without difficulty.  No other associated symptoms.  Denied nausea, diaphoresis, shortness of breath, radiation of pain.   Pain started 1.5 weeks ago, patient notes more discomfort than pain; episodes last 20-30 seconds. Patient has known familial hx of cardiac disease (father with angina) 

ECG

What do you think?  What would you do?

Let's calculate a couple risk scores: HEART score and EDACS (ED Assessment of Chest Pain) Scores:

HEART Score for Major Cardiac Events: 4 points (Moderate risk)

Moderate Score (4-6 points)

Risk of Major Adverse Cardiac Events (MACE) at 30 days = 12-16.6%.

History —> 1 = Moderately suspicious (I put this because the pain lasts for only 20-30 seconds, and ischemic pain usually lasts a few minutes at least)

EKG —> 0 = Normal

Age —> 2 = ≥65

Risk factors —> 1 = 1-2 risk factors

Initial troponin —> 0 = ≤normal limit

EDACS:  22 points, Not low risk. 

Age —> 78 years

Sex —> 6 = Male

Diaphoresis —> 0 = No

Pain radiates to arm, shoulder, neck, or jaw —> 0 = No

Pain occurred or worsened with inspiration —> 0 = No

Pain is reproduced by palpation —> 0 = No

Here is interesting data from HIGH-STEACS, using Abbott Architect hs-cTnI assay.

In this study, using risk scores had very little utility

Chapman AR, Hesse K, Andrews J, et al. High-Sensitivity Cardiac Troponin I and Clinical Risk Scores in Patients With Suspected Acute Coronary Syndrome. Circulation [Internet] 2018;138(16):1654–65. Available from: http://dx.doi.org/10.1161/CIRCULATIONAHA.118.036426   

--Non-ischemic ECG

--Symptoms ＞3 hours to blood draw AND initial hs-cTnI ＜5 ng/L or

--Any duration of pain and inital value ＜URL (16 ng/L, women, 34 ng/L men) and 3 hour delta ＜3 ng/L.

--Identified 65% as low risk for death/MI at 30 days, with 99.7% NPV (Sensitivity 98.7%)

--Use of HEART or EDACS score:

   --NPV: HEART: 99.9 vs. 99.7 for EDACS

   --% Ruled out: Only 24% for HEART, 41% for EDACS.

--So using the scores resulted in minimal change in NPV, but many fewer ruled out.

Why use the scores then?

1. If pain is so brief that you could not possibly expect the troponin to be elevated (in this case, 20-30 seconds)

2. This study looked for death or MI, but 30 day MACE also includes PCI!

3. Many patients do not fit the protocol. Example: Initial trop 10 ng/L and delta of 5 ng/L.

Does a negative ECG and 2 undetectable troponins rule out ACS?  

Again, you can only expect troponins to be elevated if there was sufficient duration of pain.  No one has ever determined what is a sufficient duration.  

Here is an interesting study from Circulation:

Árnadóttir Á, Pedersen S, Bo Hasselbalch R, et al. Temporal Release of High-Sensitivity CardiacTroponin T and I and Copeptin After Brief InducedCoronary Artery Balloon Occlusion in Humans. Circulation [Internet] 2021;143(11):1095–104. Available from: http://dx.doi.org/10.1161/CIRCULATIONAHA.120.046574

Thirty-four patients (median age, 60 years [interquartile range, 51–64]; 15 men, 43%) with angiographically normal coronary arteries were randomly assigned into 4 groups with different durations of induced myocardial ischemia (0, 30, 60, 90 s). Ischemia was induced by inflating a balloon in the left anterior descending artery between the first and second diagonal branch.

Using the cTnT, hscTnI (Siemens), and hs-cTnI (Abbott) concentrations at 0 and 180 minutes, 1 (11%), 0, and 0 patients from the 60-s ischemia group and 5 (63%), 2 (25%), and 1 (11%) from the 90-s ischemia group, respectively, fulfilled criteria for a biochemical myocardial infarction.

So very brief TOTAL Occlusion of less than 90 seconds usually will not have elevated troponins. They did usually have significant deltas, however.

So what should we do for this patient who has elevated risk scores and a troponin that cannot be trusted because the duration of pain was too short?  

            --CT Coronary angiogram

The patient underwent CT Coronary Angiogram: Calcium score was 3600!

9 hours later:

Echo was normal

Angiogram:

Severe bilateral coronary calcium and severe two-vessel obstructive CAD involving the left main and ostial-proximal LAD

The patient underwent successful coronary artery bypass grafting

Here are 5 more cases of "Unstable Angina in the Era of High Sensitivity Troponin" 

Most of these could be discerned from the ECG

===================================

MY Comment, by KEN GRAUER, MD (5/17/2025):

===================================

When I had the opportunity to review today's case — the information I was provided with included the following:

• The patient is a 70-year old man.
• A "PMH" (Past Medical History) consisting of hyperlipidemia and chronic back pain — but without mention of known heart disease.
• A "CC" (Chief Complaint) of CP (Chest Pain).
• 2 historical accounts (apparently made by 2 different clinicians) of this patient's "HPI" (History of Present Illness). 
• In HPI Account #1 — CP was "squeezing" and left-sided, occurring intermittently over the previous 10 days and not associated with exertion.
• In HPI Account #2 — CP was more of a "discomfort" than pain — lasting only 20-30 seconds over the previous 1.5 weeks — with this CP made worse by exertion.

============================

Editorial Notes ( = My Perspective as "Devil's Advocate" on this Case ... ):

============================

Not knowing if more clinical information (or validation) was obtained since my original review of today’s case — I thought it worthwhile to present my thoughts based on the above information that was given me at the time of my initial interpretation of today's 2 ECGs.

• The difference for the HPI (as described in the notes from 2 different clinicians who each interviewed this same patient) — is the reason why I have never found numeric risk assessment calculations useful (The HEART and EDACS Scores applied in today's case revealing at least moderate risk — which practically speaking, should be immediately evident without formal scoring given the age of 70 for this man who comes to the ED because of new CP).
• Converting the HPI into a numeric assessment is a subjective exercise — which I find difficult to place trust in when 2 capable clinicians come up with different assessments of the KEY parameter as to whether or not CP is made worse with exercise. Missing from HPI Account #2 are the "fine points" in the history (ie, Even if exercise-induced CP is "short-lived" — if every time the patient walks just a little bit faster he has to stop because this slight extra effort regularly precipitates CP — this may indeed represent angina).
• Statistically — a history of angina regularly precipitated by extra effort in a 70-year old man suggests underlying coronary disease until you prove otherwise.
• Conceptually — there is a major misunderstanding inherent in the approach by many in the ED. Focus is typically placed primarily (if not solely) on ruling out an acute event, which is done by negative serial Troponins and non-acute serial ECGs. But rather than an acute OMI — many patients who seek acute medical care have non-infarction angina that does not necessarily require immediate cath, but which does require timely evaluation. In my experience — that evaluation often gets dropped once the patient is discharged with negative Troponins.

Were the 2 ECGs in Today's Case "Normal"?

The other problem I have always had with numeric risk assessment scores on patients with CP — is that ECG assessment is reduced to a numeric rating. In today's case — the HEART Score rated the initial ECG as "normal" ( = a 0 numeric rating).

For clarity in Figure-1 — I've reproduced the 2 ECGs in today's case.

• Unfortunately — no information is provided as to the presence or relative severity of CP at the time that these 2 ECGs were recorded. As we often emphasize — this is critical information, since our interpretation of ECG #1 would dramtically differ if the patient had ongoing 10/10 CP vs if the CP that prompted his ED visit had greatly decreased (or completely resolved) at the time ECG #1 was recorded.

While I completely agree that ECG #1 is non-diagnostic of an acute event — I maintain that it is not "normal".

• While true that T wave inversion is not necessarily an abnormal finding in leads III and aVF when the QRS complex is predominantly negative in these leads — IF this T wave inversion was a new finding that correlates to reduced CP at the time ECG #1 is recorded — then this ST-T wave finding may represent reperfusion changes.
• And while there is no ST segment deviation in lead II of ECG #1 — the 10 looks we have of the ST-T wave in the long lead II rhythm strip of ECG #1 suggest that there is nonspecific (ie, nondiagnostic) ST segment straightening in this lead.
• As we often emphasize — there is normally slight, gently upsloping ST elevation in leads V2 and V3. Instead — there is suggestion of nonspecific ST segment flattening with a non-elevated ST segment in lead V3 (isoelectric J-point seen within the 2 BLUE arrows).
• In addition — I thought the ST-T wave in lead V2 (and possibly to a lesser extent in neighboring leads V3,V4) looked a little bit more “bulky” that I would normally expect given QRS size and appearance in these leads (ie, Could this represent a hyperacute T wave?).
• To Emphasize: The above ECG findings are nondiagnostic of an acute event. They are extremely subtle. But in my opinion, in a 70-year old man who presents to the ED for new CP — ECG #1 is not a "normal" tracing — and depending on the timing and relative severity of symptoms when this tracing is recorded, may represent reperfusion ST-T wave changes in a patient with an infero-postero OMI (or even hyperacute changes in the anterior wall).

In my opinion — ECG #2 supports my suspicion that this patient may have had brief coronary occlusion and/or may have significant underlying coronary disease.

• KEY Point: We are told that ECG #2 was obtained ~9 hours after ECG #1. Unfortnately, we don't know if additional ECGs were obtained in the interim, nor are we told what the patient's symptoms were like during these 9 hours.
• Imagine this patient's CP is less at the time ECG #2 is recorded. Lead-by-lead comparison between the 2 tracings in Figure-1 shows subtle-but-real increased size of the negative T waves in leads III and aVF — and increased size of the positive T waves in leads I and aVL (BLUE arrows in ECG #2). If anything — the ST segment in leads II and V3 is flatter in ECG #2. I think it unlikely that all of these changes are the result of random variation. But unfortunately, we have lost our ability to try to correlate these ECG changes clinically — because there is no mention in the chart of the presence or relative severity of CP at the time ECG #2 is recorded.
• P.S.: The V1 and V2 electrodes have been placed too high on the chest in ECG #2 (as suggested by the negative P wave and rSr' in these leads, with strong resemblance of QRST morphology of V1,V2 with aVR — as I review in My Comment at the bottom of the page in My Comment in the December 18, 2022 post). Recognizing this electrode lead misplacement is relevant to interpretation of ECG #2 — because it invalidates our assessment and comparison of leads V1,V2 in this repeat tracing.

BOTTOM Line: The need for cardiac cath followed by CABG was prompted in today's case by the finding of a very high CAC (Coronary Artery Calcium) Score on CT angiogram.

• While fully acknowledging that my comments are made in the comfort of my home office with large computer screen and relaxing armchair — I suspect the need for cardiac cath could probably have been made much sooner (and without need for CAC Scoring on CT Angiogram) — simply by more focused history and more timely serial ECGs correlated to relative severity of the patient's symptoms. I believe it entirely possible that diagnostic ECG changes may have occurred during the time period between ECGs #1 and #2.

Figure-1: Comparison between the 2 ECGs in today's case.