Electrocardiogram (ECG) interpretation

• The electrocardiogram (ECG) is a recording of the heart's electrical activity from electrodes on the skin. It shows rhythm, conduction, electrical axis, signs of ischaemia and infarction, and signs of chamber enlargement or electrolyte and drug effects.
• It does not show how well the heart pumps or how the valves work. A normal ECG never excludes disease.
• Read every ECG in the same order so nothing is skipped: rate, rhythm, axis, intervals (PR, QRS, QT), then the P waves, the Q waves, the ST segments and T waves, and the voltage.
• Always read the ECG against the patient. The same tracing means different things in a well athlete and in a patient with chest pain.
• In a patient with suspected acute coronary syndrome, the ECG is recorded at first medical contact and interpreted within 10 minutes.^[2]

By the end of this page you should be able to:

1. Explain what each part of the ECG trace represents and how the paper is calibrated.
2. Calculate heart rate from a regular and from an irregular rhythm.
3. Decide whether a rhythm is sinus and name the common non-sinus findings.
4. Work out the QRS axis from leads I, II and aVF and list the causes of deviation.
5. Measure the PR, QRS and QT intervals, correct the QT for heart rate, and say why published upper limits differ.
6. Recognise the evolution of ischaemic changes, pathological Q waves and the territory involved.
7. Recognise atrial enlargement and ventricular hypertrophy, with the voltage criteria and their limits.
8. Spot common artefacts and pitfalls, and say who needs an ECG.

Each heartbeat starts in the sinoatrial (SA) node, the heart's natural pacemaker, and spreads through the atria, the atrioventricular (AV) node, the His bundle and bundle branches, and the ventricular muscle. The ECG is the sum of these electrical events viewed from outside. A wave of depolarisation travelling toward a lead's positive electrode produces an upward deflection; one travelling away produces a downward deflection.

The parts of one beat are:

• P wave: depolarisation of the atria.
• PR interval: from the start of the P wave to the start of the QRS. It includes the pause while the impulse crosses the AV node.
• QRS complex: depolarisation of the ventricles. A downward first deflection is a Q wave; the first upward deflection is the R wave; a downward deflection after the R is the S wave.
• ST segment and T wave: ventricular repolarisation (recovery).
• QT interval: from the start of the QRS to the end of the T wave, which is the total time of ventricular depolarisation and repolarisation.

The twelve leads are six limb leads (I, II, III, aVR, aVL, aVF), which look at the heart in the frontal plane, and six chest (precordial) leads V1 to V6, which look at it in the horizontal plane. Each lead faces a different part of the heart, which is why changes in a group of neighbouring leads point to a region.

The standard paper speed is 25 mm per second. At this speed:

• one small square (1 mm) is 0.04 s (40 ms);
• one large square (5 mm) is 0.2 s (200 ms);
• five large squares are 1 second.

Vertical calibration is standardly 10 mm per millivolt, so 1 mm is 0.1 mV.^[1] Before you interpret voltages, check the calibration mark at the start of the trace. Machines can be set to half or double gain, and a half-height tracing will make normal hearts look low-voltage while a double-height one will make normal hearts look hypertrophied.

A fixed order protects you from "satisfaction of search", where you spot one abnormality and stop looking.

The heart rate is how many ventricular beats occur per minute. A resting adult rate under 60 beats per minute (bpm) is bradycardia and over 100 bpm is tachycardia; between those values the rate is normal. Both are descriptions, not diagnoses: athletes are slow, and fever, pain, anaemia, hypovolaemia and thyroid overactivity all speed the heart.

Regular rhythm: the 300 rule. At 25 mm/s one minute of tracing is 300 large squares (60 s at 5 large squares per second). So the rate is 300 divided by the number of large squares between two consecutive R waves. Two squares give 150 bpm, three give 100, four give 75, five give 60 and six give 50. If the R-R interval is not a whole number of squares the result is approximate.

Irregular rhythm: count and multiply. Most ECGs print a 10-second rhythm strip. Count the QRS complexes in the strip and multiply by 6, or count them in 6 seconds and multiply by 10. This averages out the variation, which the 300 rule cannot do.

Which beats count? In atrial fibrillation or heart block the atrial and ventricular rates differ. The pulse and the QRS rate matter for output, so quote the ventricular rate and describe the atrial activity separately.

Ask whether the rhythm is sinus, and if not, what it is.

Sinus rhythm means the impulse starts in the SA node. The signature: each QRS is preceded by a P wave, each P wave is followed by a QRS, and the P wave is upright in lead II and inverted in aVR (because the atrial wave spreads down and to the left, towards lead II and away from aVR). The normal P-wave axis is about 0 to +75 degrees in the MSD Manual;^[4] the Brazilian guideline allows up to +90 degrees.^[6]

Sinus rhythm is not always perfectly regular. In young, healthy people the rate normally speeds on inspiration and slows on expiration (sinus arrhythmia), and the P waves keep their normal shape. This variation is physiological, autonomic and phasic with breathing.^[6]

Then ask five questions:

1. Is it regular? Mark the R-R intervals on a piece of paper and slide it along the strip.
2. Are there P waves, and how do they relate to the QRS? No visible P waves with an irregularly irregular rhythm suggests atrial fibrillation; a saw-tooth baseline suggests atrial flutter; more P waves than QRS complexes suggests AV block (see bradyarrhythmias and AV block).
3. Is the QRS narrow or wide? A narrow QRS (under 120 ms) means the ventricles were activated through the normal conduction system, so the rhythm is supraventricular (see supraventricular tachycardias). A wide QRS in tachycardia is ventricular tachycardia until proven otherwise (see ventricular arrhythmias).
4. Are there extra beats? Premature beats, early and followed by a pause, are common and often harmless.
5. Is the patient stable? An unstable patient with an abnormal rhythm needs treatment before further analysis.

The QRS axis is the average direction of ventricular depolarisation in the frontal plane, measured in degrees on the hexaxial diagram: 0 degrees points to the patient's left (lead I), +90 points down (aVF), and angles above the horizontal are negative.

A lead gives an upward QRS when the average wave of depolarisation travels toward it. So if the QRS is mainly upward in both lead I and aVF, the axis lies between 0 and +90 degrees and is normal. The three-lead check below gives the quadrant.

Lead II needs care: an axis between 0 and −30 degrees has a small upward QRS in lead II and is still normal. The axis is only truly "left" when lead II is predominantly downward.

Causes of left axis deviation. The MSD Manual names block of the left anterior fascicle (left anterior hemiblock, axis about −60°) and inferior myocardial infarction.^[4] The Brazilian guideline defines left anterior fascicular block as an axis of −45° or beyond, an rS pattern in II, III and aVF and a QRS under 120 ms.^[6] Left ventricular hypertrophy can also shift the axis leftward (the Romhilt-Estes score counts left axis deviation as a feature).^[4]

Causes of right axis deviation. The MSD Manual states that it occurs in any condition that raises pulmonary pressure and causes right ventricular hypertrophy.^[4] Block of the left posterior fascicle (axis beyond +90°) is another cause, but it is diagnosed only when right ventricular hypertrophy has been excluded, and it is more likely in a slender person or after an old lateral infarction.^[6] Always consider lead reversal first (see Pitfalls).

Pre-excitation (see Wolff-Parkinson-White syndrome) can change the axis, depending on where the accessory pathway lies. Axis deviation also occurs with bundle branch blocks (see bundle branch blocks) and may be a normal variant in tall or short individuals.

The PR interval measures the time from the start of atrial activation to the start of ventricular activation. It is long when the AV node conducts slowly (first-degree AV block: every P wave still conducts, only more slowly), and short when the impulse bypasses the node through an accessory pathway or arises from the AV junction. Pre-excitation often also shows a slurred upstroke to the QRS called a delta wave.

A QRS of 0.12 s or more means the ventricles were not activated quickly through both bundle branches.^[4] Look at the shape in V1 and V6 to tell left from right bundle branch block; the details are in bundle branch blocks. A new left bundle branch block in a patient with symptoms of acute coronary syndrome needs urgent assessment, because it prevents reliable reading of the ST segments. In patients with high clinical suspicion of ongoing ischaemia, the 2023 European Society of Cardiology (ESC) guideline says to manage them as for ST-elevation myocardial infarction (STEMI) whether the block is known to be old or not.^[2]

The QT interval shortens as the heart rate rises, so a QT measured at a fast rate cannot be compared directly with one at a slow rate. The QT is therefore corrected for rate (QTc). Two formulas dominate.

• Bazett: QTc = QT divided by the square root of the R-R interval (in seconds).
• Fridericia: QTc = QT divided by the cube root of the R-R interval.

Bazett over-corrects at fast heart rates (making the QTc look too long) and under-corrects at slow rates.^[3] In a study of patients, at a heart rate of 80 bpm or more the Bazett QTc averaged 438 ms against 414 ms by Fridericia, and in that study Fridericia corrected best across the range, so its authors favour it for hospital QT monitoring.^[3] The Brazilian guideline limits Bazett to heart rates of 60–90 bpm and prefers linear formulas (Framingham, Hodges) outside that range.^[6] Whichever formula is used, it should be named in the report. Bazett remains the formula printed by most machines.

Measure the QT in a lead with a clear T-wave end, usually lead II or V5, to the end of the T wave; if the lead selection or U wave is in doubt, repeat in another lead and take the longer value only after careful inspection. Average several beats in atrial fibrillation.

Upper limits differ. There is no single universal cut-off.

• The MSD Manual gives normal QTc as 350–450 ms in men and 360–460 ms in women.^[4]
• The study above used 450 ms (men) and 470 ms (women) as its clinical standards, and measured upper limits of normal in a healthy subset of 472 ms (men) and 482 ms (women) with Bazett, but 448 ms (men) and 468 ms (women) with Fridericia.^[3]
• The Brazilian guideline accepts QTc up to 450 ms in men and 470 ms in women, and calls a QT under 340 ms short.^[6]
• Many teaching texts quote a single value of about 440 ms.

The ischaemic ECG changes reflect altered electrical behaviour of damaged heart muscle. Read them region by region, because ischaemia affects the muscle supplied by a particular coronary artery.

The leads look at the heart from different directions. The usual groupings are: inferior wall, leads II, III and aVF; lateral wall, leads I, aVL, V5 and V6; anterior wall (with the septum), leads V1 to V4.^[6] Changes must be present in at least two contiguous leads to count.^[1] A mirror-image change in the opposite wall (reciprocal change) supports a true infarct.

Two special sets of extra leads exist for walls the standard 12 leads see poorly.

• Posterior wall: ST depression in V1 to V3 with a positive terminal T wave suggests posterior (inferobasal) ischaemia, often from the circumflex artery. Extra leads V7 to V9 can be recorded to look for it; ST elevation of 0.5 mm or more there counts (1 mm or more in men under 40).^[1][2] The Brazilian guideline advises that the term "posterior wall" is no longer used, because V7 to V9 reflect the lateral wall.^[6]
• Right ventricle: in suspected inferior infarction, record right-sided leads V3R and V4R, where ST elevation of 0.5 mm or more (1 mm or more in men under 30) points to right ventricular involvement. A lack of elevation does not exclude right ventricular infarction.^[1][2]

After a coronary artery is blocked, the ECG changes in a sequence. Treat the timings as approximate because they vary with reperfusion and the size of the infarct.

1. Hyperacute phase. Prominent, symmetrical T waves in at least two contiguous leads appear first and may come before ST elevation.^[1]
2. ST elevation. The J point (where the QRS ends and the ST segment begins) rises. Standard teaching links ST elevation to full-thickness injury and ST depression to injury confined to the inner layer of muscle (subendocardial). Prompt reperfusion makes the elevation fall quickly.^[1]
3. T-wave inversion. As the ST elevation settles, the T waves become inverted. This inversion is now attributed to reperfusion or oedema rather than to ongoing ischaemia.^[6]
4. Q waves. New Q waves begin to form from hours after the insult, and are often established within a few days.^[1] They often persist, because they reflect lost viable muscle, but they can regress.^[1]

If symptoms persist or the first ECG is non-diagnostic, the Fourth Universal Definition advises serial ECGs, for example every 15 to 30 minutes for 1 to 2 hours.^[1]

The Fourth Universal Definition of Myocardial Infarction (MI) and the 2023 ESC guideline use these thresholds for new J-point elevation, in two or more contiguous leads (absent left ventricular hypertrophy or left bundle branch block).^[1][2]

Healthy young men can have J-point elevation up to 2.5 mm in V2 or V3, and it falls with age, which is why the V2–V3 limits are higher.^[1]

Other patterns:

• ST depression of 1 mm or more in six or more leads, with ST elevation in aVR and/or V1, suggests left main or multivessel disease, particularly with haemodynamic compromise.^[2]
• ST depression of 0.5 mm or more in V1 to V3 may indicate circumflex occlusion.^[1]
• Symmetrical deep T inversions (over 2 mm) in anterior leads are associated with significant occlusion of the left anterior descending artery.^[1]
• Bundle branch block or paced rhythm hides ST changes, so with a typical story treat it as STEMI.^[2]

A small Q wave is normal in many leads (it is the first, left-to-right septal activation). A Q wave becomes suspicious when it is wide, deep, or in an unexpected lead. Sources differ on the exact line.

• The Fourth Universal Definition counts as pathological: any Q wave of 0.02 s or more, or a QS complex, in V2–V3; and a Q wave of 0.03 s or more and at least 1 mm deep, or a QS complex, in I, II, aVL, aVF or V4–V6, in two contiguous leads of a group (I and aVL; V4 to V6; II, III and aVF).^[1] Minor Q waves of 0.02 to under 0.03 s and at least 1 mm deep suggest old infarction if the T waves in the same lead group are inverted, and Q waves wider than 0.04 s are the most specific. A QS complex in V1 is normal, and septal Q waves under 0.03 s and under a quarter of the R-wave height are normal in leads I, aVL, aVF and V4 to V6. In lead III a Q wave under 0.03 s and under a quarter of the R wave is normal when the axis is between −30° and 0°.^[1]
• The MSD Manual states that normal Q waves are shorter than 0.05 s in all leads except V1 to V3, where any Q wave is abnormal.^[4] This differs from the Fourth Universal Definition on V1 and V3.
• The Brazilian guideline counts a Q wave of 40 ms or more in two contiguous leads, with or without an amplitude over 25% of the QRS complex.^[6]

For exams, use the lines taught by your own school. The principle is the same: a wide, deep Q wave in a group of contiguous leads means dead muscle.

In the chest leads the R wave normally grows from V1 to about V4 or V5 while the S wave shrinks. Poor R-wave progression means the R waves stay small across the anterior leads. It can follow anterior infarction, but it is also seen with a misplaced lead, left ventricular hypertrophy, chronic lung disease, left bundle branch block and in healthy people, so on its own it is nonspecific.

When a chamber is overloaded its muscle thickens or stretches and it makes more electrical signal. The ECG shows this as taller waves, longer waves, or altered ST-T segments. These signs are specific rather than sensitive, so a normal ECG does not rule out enlargement. Echocardiography is the test for size and thickness (see heart failure).

• Right atrial enlargement (P pulmonale): tall, peaked P waves; the Brazilian guideline gives a P wave over 2.5 mm (0.25 mV) in lead II, and an initial positive deflection over 1.5 mm in V1.^[6] The MSD Manual quotes over 2 mm in leads II, III and aVF.^[4] The right atrium activates first, so a taller right-atrial component dominates the early P wave.
• Left atrial enlargement (P mitrale): a P wave of 120 ms or more in lead II, often notched (two peaks 40 ms or more apart), and a terminal negative component of the P in V1 at least 1 mm deep and 40 ms wide.^[6] The Romhilt-Estes score counts the same V1 deflection as a sign of left atrial change.^[4] A normal P wave is at most 2.5 mm tall and 110 ms wide.^[6] The left atrium depolarises last, so its enlargement lengthens the P wave.

Thickened left ventricular muscle produces larger voltages in leads that face the left ventricle (V5, V6, I, aVL) and deeper S waves in leads that face away (V1, V2).

Sources differ on the details of the Romhilt-Estes score: the Brazilian guideline gives the QRS duration point as over 90 ms and the intrinsicoid deflection time as over 40 ms.^[6] The criteria do not agree with one another, and they find only a fraction of the true cases. A pooled analysis cited in a 2024 review gave sensitivities of about 22% for Sokolow-Lyon, 26% for Cornell and 43% for Peguero-Lo Presti, with studies varying widely.^[5] Reviews describe the Peguero-Lo Presti criterion as less specific than Sokolow-Lyon and Cornell.^[5] Voltage criteria are also altered by body size, age, chest wall thickness and sex. Echocardiography or cardiac magnetic resonance imaging (MRI) is the definitive test.

Strain pattern. In LVH, ST depression and T-wave inversion in the left-sided leads (V5, V6, I, aVL) indicate that the thickened muscle is under pressure overload and is repolarising abnormally. The Romhilt-Estes score awards 3 points for typical ST-T change (1 point if the patient takes digoxin, which causes its own ST depression).^[4] This pattern can mimic ischaemia.

Normally the left ventricle dominates the QRS. When the right ventricle thickens, right-sided forces become prominent: right axis deviation and a tall R wave (or an R-dominant complex) in V1, with a deep S wave in the left chest leads. The Brazilian guideline lists these signs: axis to the right of +110°; R/S ratio of 1 or more in V1–V2 with a deep S in V5–V6; a qR or qRs pattern in V1, one of the most specific signs; an rsR' pattern in V1–V2; and right precordial strain. One set of criteria (Seattle) adds R in V1 plus S in V5–V6 over 10.5 mm with an axis over 120°.^[6] The ECG is insensitive for RVH. Right atrial P pulmonale often accompanies it. RVH occurs in conditions that raise pulmonary pressure.^[4]

A wrong ECG diagnosis is often a wrong recording.

• Limb lead reversal. Swapping the left and right arm electrodes makes lead I look upside down (negative P wave, negative QRS and negative T), which is impossible in normal sinus rhythm with a normal heart. Other swaps produce changes that can mimic an inferior infarct or a false axis. If lead I is inverted with a normal chest-lead progression, repeat the recording.
• Misplaced chest leads. Placing the electrodes too high or too low alters the R-wave progression and the ST segments.
• Muscle tremor and movement. Shivering, Parkinsonian tremor and a patient who is cold or tense produce a fast, irregular baseline that can look like atrial fibrillation or flutter; tremor can also mimic ventricular fibrillation.^[6] Warming and relaxing the patient, and supporting the arms, helps.
• Mains interference. Interference from nearby mains-powered equipment and cables, or from poor contact.^[6] Re-prepare the skin and check the leads.
• Baseline wander. A slow drift from breathing, limb movement or a loose electrode distorts the ST segments.
• Poor skin contact. Clean, dry, hair-free skin gives a better signal.
• Machine reading. Automated interpretation is a prompt, not a diagnosis. Machines often miscalculate the QT and mislabel rhythm. Every ECG should be checked by a person.
• Normal variants and mimics. Young men can have ST elevation in V2–V3 within the normal limits above.^[1] Left ventricular hypertrophy and bundle branch blocks both alter the ST segments and the criteria for infarction.^[2] Pericarditis is one of the non-infarction causes of ST elevation.^[1] Its elevation tends to be widespread rather than confined to a single coronary territory.
• A normal first ECG. A single early ECG can be normal in acute coronary syndrome. The ESC guideline says to repeat it as necessary, especially if symptoms eased before the first recording.^[2]

The ECG is cheap, quick and non-invasive. It is most clearly indicated in:

• Suspected acute coronary syndrome. Record a 12-lead ECG at first medical contact and have it interpreted within 10 minutes; repeat as needed.^[2]

The MSD Manual lists what the 12-lead ECG can establish, and these findings define the other situations in which it is useful: arrhythmias, atrial and ventricular enlargement, hypertrophy, inherited arrhythmia syndromes such as Brugada and long QT syndrome, pre-excitation, and myocardial ischaemia or prior infarction.^[4] A baseline tracing is also useful for later comparison, for example in people with known heart disease or those starting QT-prolonging drugs; local protocols set the other indications, such as before surgery.