Reading an ECG: A Seven-Step Routine for Ward and Clinic
Forty tracings in the tray, a pre-operative list to clear, and no cardiology service to bounce the doubtful ones to. In that setting an ECG routine is not an academic ritual; it is the only thing that stops the abnormal tracing hiding among the normal ones at four in the afternoon. The routine below is built around what each part of the tracing lets you decide, and it is deliberately sequenced so that the findings which change management tonight surface before the ones that merely populate a report.
On this page
Why a fixed order beats pattern recognition
Experienced readers recognise patterns instantly, which is exactly the problem. Once a striking abnormality captures attention, the second abnormality on the same tracing is missed at a rate that has been demonstrated repeatedly. A fixed sequence protects against that, and it also produces reports that are comparable between clinicians and over time. The cost is about ninety seconds per tracing. The routine matters most on the tracings that look boring.
The decision points
1. Decide whether you trust the recording
Calibration at 10 mm per millivolt, paper at 25 mm per second, and a filter setting that has not amputated the ST segment. Half standardisation is used more often than clinicians realise and will convert genuine hypertrophy into a normal voltage report. Limb lead reversal produces an inverted P and QRS in lead I with a normal precordial progression, a combination that should trigger a repeat rather than a diagnosis of dextrocardia. Then find a previous tracing. Comparison changes interpretation more often than any single measurement on the current one.
2. Decide whether the rate explains the patient
Rate is trivial to calculate and is worth pausing over only when it is inconsistent with the clinical state. A rate of 105 in a comfortable patient at rest, or a rate of 48 in a breathless one, is a finding rather than a number. In irregular rhythms count over a full six seconds rather than dividing into a single interval.
3. Decide whether the rhythm needs anything doing tonight
Work through three questions in order: is there organised atrial activity, is the ventricular response regular, and is there a fixed relationship between the two. That sequence separates sinus rhythm, atrial fibrillation, flutter with variable conduction, atrial tachycardia and the junctional and ventricular rhythms without needing a differential list. In broad complex tachycardia, treat the rhythm as ventricular in origin until proven otherwise, particularly in any patient with structural heart disease; the alternative error is far more dangerous than the diagnostic one. Systematic rhythm interpretation at the bedside is covered in our arrhythmia bedside reference.
4. Decide whether conduction is safe to leave unmonitored
Measure the PR interval and classify any block precisely, because the management divide runs between block above and below the atrioventricular node. Progressive PR lengthening with a dropped beat in an asymptomatic patient is usually benign. Sudden non-conducted P waves with a constant PR, block with a broad escape, or alternating bundle branch patterns indicate infranodal disease and a pacing conversation. Then assess QRS duration: a broad complex demands an explanation, whether bundle branch block, pacing, pre-excitation, hyperkalaemia or a sodium channel blocking drug.
5. Decide how much weight to give the chamber signals
Axis, atrial abnormality and voltage criteria for hypertrophy are supportive rather than decisive. Voltage criteria have modest sensitivity and are distorted by body habitus, chest wall and lung disease, so hypertrophy suggested on a tracing should be confirmed by imaging rather than reported as fact. Left axis deviation with a small septal q pattern, right axis deviation in a breathless patient, and biatrial enlargement all direct the next investigation rather than concluding it. Poor R wave progression deserves a considered comment rather than a stock phrase, since prior infarction, lead placement, hypertrophy and chronic lung disease all produce it.
6. Decide whether repolarisation is ischaemic, structural or pharmacological
Judge ST deviation against the QRS that produced it. Discordant changes proportional to a broad or high-voltage complex are secondary and expected; changes disproportionate to the complex or concordant with it are not. Look at T wave morphology as well as polarity, since symmetry and base width carry the ischaemic signal. Then measure the QT and correct it, remembering that the common correction formula overcorrects at fast rates and undercorrects at slow ones. A corrected interval beyond 500 milliseconds in a patient on multiple prolonging agents is an action point, not an observation: review the drug chart, replace potassium and magnesium, and arrange monitoring. Broader rhythm risk assessment is set out in the heart failure, rhythm and risk hub.
7. Decide what the tracing means for this patient, then say it
Finish by answering the referral question in one sentence. A report that lists findings without a conclusion transfers the decision back to the person least equipped to make it. If the tracing is normal for the clinical context, say so plainly, since an equivocal report generates unnecessary referrals and anxious patients.
When to escalate, transfer or call for help
- Same shift, with monitoring: complete heart block, Mobitz type two block, a broad escape rhythm, sustained broad complex tachycardia, or a corrected QT beyond 500 milliseconds with syncope.
- Urgent referral even when the patient feels well: pre-excitation with atrial fibrillation, a coved right precordial pattern, epsilon waves with right precordial T inversion, and voltage criteria with deep T inversion suggesting apical hypertrophic disease. These patients look well until the moment they do not, and the diagnostic detail sits in our electrophysiology essentials volume.
- Escalate the tracing, not the patient: photograph an equivocal ECG and send it to a colleague. Building that route before you need it is one of the more valuable things a doctor in an isolated post can do.
- Arrange further imaging rather than referral when voltage or repolarisation findings suggest structural disease in a stable patient; a targeted echocardiogram answers more than a cardiology appointment three months away, and the diagnosis and imaging hub outlines the sequence.
What to document
- Date, time and clinical context, including symptoms at the moment of recording.
- Rate, rhythm, axis, and the measured PR, QRS and corrected QT intervals as numbers.
- Whether a previous tracing was available, and what changed.
- Technical limitations, including any non-standard calibration or suspected lead placement error.
- An explicit conclusion answering the referral question, and the action taken or planned.
- For any drug-related interval finding, the drugs reviewed and the electrolytes checked.
Bench card
| Element | Usual adult range | Act when |
|---|---|---|
| Rate | 60 to 100 per minute | Rate is discordant with the clinical state |
| PR interval | 120 to 200 ms | Short with a delta wave, or long with dropped beats |
| QRS duration | Under 120 ms | Broad without a known cause, or newly broad |
| QRS axis | Minus 30 to plus 90 degrees | Marked deviation with new symptoms |
| Corrected QT | Under 450 ms in men, under 460 ms in women | Beyond 500 ms, or with syncope on prolonging drugs |
| ST segment | At the isoelectric line | Deviation disproportionate to or concordant with the QRS |
The seven-step order, condensed: trust the recording, check the rate against the patient, define the rhythm, classify conduction, weigh the chamber signals, judge repolarisation against the QRS, then state a conclusion. Practice sets that build this into a reflex are collected in the clinical ECG workshop and the rapid cardiology clinical reference. A stepwise tutorial version of the same routine is published as this systematic approach to reading an ECG.
Questions from the floor
Should I rely on the machine interpretation at the top of the tracing?
Use it as a prompt, never as a conclusion. Automated algorithms are reasonable at rate and interval measurement and unreliable at rhythm classification, particularly in atrial fibrillation with a regularised response, in paced rhythms and in artefact. Delete the machine statement from the record if you disagree with it, and write your own.
How much does a previous ECG really add?
More than any other single input. Established T wave inversion, an old bundle branch block or long-standing Q waves change the disposition of a chest pain patient entirely. Where old tracings are inaccessible, ask the patient to photograph any they hold; in international practice this succeeds more often than requesting records.
Which QT correction should I use in atrial fibrillation?
Average the measured interval over several beats spanning the range of RR intervals rather than trusting a single corrected value. When the decision concerns a prolonging drug, err towards caution, correct electrolytes and monitor rather than relying on the arithmetic.
How do I keep the routine when the department is overwhelmed?
Shorten the report, not the sequence. The order takes seconds; abandoning it is what produces the missed second finding. Reading in the same order every time also makes it far easier to teach juniors and to audit your own accuracy afterwards.
ECG Changes in Acute Myocardial Infarction: Bedside Reference
It is three in the morning in a district hospital. The nurse hands you a tracing on a 45-year-old man with an hour of chest pain, and there is 1.5 mm of ST elevation in the inferior leads that could be ischaemia or could be his baseline. There is no cardiologist rostered, the fibrinolytic is in the fridge, and the nearest interventional centre is two hours away in good weather. The clinical question is not what the ECG shows in a textbook sense; it is whether this tracing commits you to reperfusion in the next ten minutes. This reference is organised around that decision and the ones that follow it.
On this page
What the tracing is actually telling you
Occlusion produces a predictable sequence, but the sequence is fast and you may catch it at any point. Tall, broad, symmetrical T waves with a widened base appear within minutes and are frequently dismissed as normal variant. Convex ST elevation follows. Loss of R wave voltage and pathological Q formation begin within hours and continue over days, T waves invert as the injury current resolves, and after weeks a tracing may show nothing but Q waves and poor progression. The clinically useful consequence is that the absence of ST elevation on a tracing taken twenty minutes into symptoms tells you very little, and a single ECG is never a negative test in a convincing history.
The decision points
1. Verify the recording before you interpret it
Check calibration, paper speed, filter settings and lead placement. Reversed limb leads and high V1 and V2 placement generate spurious patterns that have sent many patients to unnecessary angiography and, more dangerously, have masked real disease. If the appearance is unexpected, repeat the tracing with the electrodes checked before acting on it.
2. Ask whether the pattern meets reperfusion criteria, and whether that is the right question
Conventional thresholds require new ST elevation at the J point in two contiguous leads, with sex and age specific cutoffs in V2 and V3 and a lower threshold elsewhere. But a meaningful minority of acute occlusions never meet those thresholds. The de Winter pattern of upsloping ST depression with tall symmetrical T waves in the precordial leads, hyperacute T waves without elevation, and isolated posterior infarction all represent occluded arteries and all are commonly delayed. Treat the clinical picture combined with a convincing pattern as an indication to act rather than waiting for millimetres to accumulate.
3. Extend the tracing when the standard twelve leads are equivocal
Two additional acquisitions cost ninety seconds and change management more often than most investigations you will order all night. Right sided leads, particularly V4R, identify right ventricular involvement in inferior infarction and change fluid and nitrate decisions immediately. Posterior leads V7 to V9 confirm the infarction suggested by tall R waves and ST depression in V1 to V3. Make both routine in inferior and in isolated anterior ST depression presentations.
4. Localise, then predict what is going to go wrong
Localisation is not an academic exercise; it tells you which complication to prepare for. Inferior infarction predicts bradyarrhythmia, atrioventricular block and, when the right ventricle is involved, profound hypotension with nitrates. Proximal anterior infarction predicts pump failure, bundle branch and high-grade conduction block, and mechanical complications. Elevation in aVR with widespread depression in a shocked patient suggests left main or severe multivessel disease and is a transfer decision rather than a fibrinolysis decision.
5. Use reciprocal change as your confirmation tool
Reciprocal ST depression in a territory opposite the elevation is the most reliable bedside discriminator between infarction and its mimics. Pericarditis produces widespread saddle-shaped elevation with PR depression and no reciprocal pattern except in aVR. Early repolarisation gives concave elevation with notched J points in a young patient with stable serial tracings. Left ventricular hypertrophy and left bundle branch block produce discordant elevation proportional to the QRS, and in the latter the concordance rules attributed to Sgarbossa, with the proportionality modification, remain the practical tool. Takotsubo and myocarditis will fool you and can only be excluded downstream.
6. Repeat the tracing rather than waiting for the biomarker
In an ongoing presentation, serial ECGs at ten to fifteen minute intervals with continuous ST monitoring will resolve the question faster and more safely than a second troponin. Evolution across serial tracings is diagnostic in itself; a static equivocal pattern in a resolving patient is reassuring in a way that a single tracing can never be. Systematic technique for this is set out in our clinical ECG interpretation reference.
7. Decide reperfusion strategy on geography and clock
If timely primary angioplasty is achievable, activate transfer and start antithrombotic therapy. If it is not, fibrinolysis followed by transfer for angiography is the pathway, and the decision should be made in minutes rather than after a departmental discussion. Assess reperfusion at sixty to ninety minutes on ST resolution rather than on symptom relief, since pain is an unreliable marker. Failure of resolution mandates rescue angiography, not a second dose. First-hour sequencing for hospitals without on-site intervention is laid out in our cardiac emergency first hour protocols.
When to escalate, transfer or call for help
- Activate transfer immediately, before further investigation: ST elevation meeting criteria, new left bundle branch block with a convincing history, de Winter or hyperacute patterns, or elevation in aVR with shock.
- Prepare pacing: inferior infarction with second degree block or a ventricular rate under 40 per minute. Transcutaneous pads on before transport, not during it.
- Escalate for haemodynamic support: right ventricular infarction with hypotension unresponsive to volume, or anterior infarction with pulmonary oedema. Stabilisation approaches are covered in the acute cardiac care worldwide reference and across the interventional and acute care hub.
- Seek a second opinion on the tracing rather than waiting: a photograph sent to a colleague resolves most equivocal patterns in under five minutes, and that habit is worth building deliberately before the night you need it.
What to document
- Time of symptom onset, time of first medical contact, and time of every ECG, written on the tracing itself.
- Lead placement confirmation and whether right sided or posterior leads were recorded.
- The specific pattern, with the leads named, and the reciprocal changes present or absent.
- Comparison with any previous tracing, and where that tracing came from.
- The reperfusion decision, the reasoning, and the time therapy was given or transfer activated.
- ST resolution at sixty to ninety minutes, expressed as a proportion, and the action taken if resolution failed.
Bench card
| Leads involved | Territory | Usual culprit | Prepare for |
|---|---|---|---|
| V1 to V4 | Anteroseptal | Left anterior descending | Pump failure, conduction block |
| V1 to V6, I, aVL | Extensive anterior | Proximal left anterior descending | Shock, mechanical complication |
| I, aVL, V5 to V6 | Lateral | Circumflex or diagonal | Often subtle; check posterior leads |
| II, III, aVF | Inferior | Right coronary, sometimes circumflex | Bradycardia, block, nitrate sensitivity |
| ST depression V1 to V3 with tall R | Posterior | Circumflex or right coronary | Confirm with V7 to V9 before dismissing |
| V4R elevation | Right ventricle | Proximal right coronary | Volume, avoid nitrates and diuretics |
| aVR elevation with widespread depression | Global subendocardial | Left main or severe multivessel | Transfer for angiography, not lysis |
Patterns that should never be filed as non-diagnostic: hyperacute T waves in an ongoing presentation, de Winter upsloping depression, biphasic or deeply inverted precordial T waves in a pain-free patient with recent symptoms, and any new bundle branch block with a convincing history. Structured pattern libraries for each sit in the ECG reference for physicians and the wider cardiology reference library. A more didactic walk through the same electrocardiographic sequence is published as this guide to ECG changes in acute myocardial infarction.
Questions from the floor
Can an infarction occur with a genuinely normal ECG?
Yes, and it happens often enough to matter. Circumflex territory infarction is electrocardiographically silent in a substantial proportion of cases, and very early presentations may show nothing. Serial tracings, posterior leads and biomarkers together are what close the gap; a single normal tracing does not.
How do I approach ST elevation in a patient with a paced rhythm?
Apply the same concordance principles used in left bundle branch block, recognising that they are less well validated in ventricular pacing. In practice, a convincing history with concordant elevation or excessively discordant elevation should be treated as an occlusion and transferred.
Is ST depression ever an indication for immediate reperfusion?
Isolated posterior infarction is, once confirmed on posterior leads. Widespread depression with aVR elevation is an indication for urgent angiography rather than fibrinolysis. Ordinary anterior ST depression from demand ischaemia is neither.
What if the elevation persists weeks later?
Persistent elevation with Q waves in a pain-free patient several weeks after a large anterior infarction suggests ventricular aneurysm and warrants echocardiography, assessment for mural thrombus and a review of anticoagulation, rather than repeat reperfusion.
