Case study
ECG Lead Reversal Waveform Diagnosis Case Study
ECG lead reversal case study for electrode mapping, Einthoven consistency, simulator evidence, signal-quality checks, risk controls, corrective action, and release validation.
An ECG acquisition chain can pass noise, sampling and leakage-current verification while still displaying the wrong physiological view if the leads are mapped incorrectly. Lead reversal is a system problem: electrodes, cable labels, connector pinout, firmware channel mapping, derived-lead equations, display labels, user workflow and release tests must all agree.
This case study follows a ward ECG monitor after an accessory cable change. The device reports stable waveforms with acceptable signal-quality flags, but clinicians notice that lead I is inverted and the automated interpretation becomes inconsistent with the patient’s prior record. The engineering question is not to diagnose a patient. It is to determine whether the device is acquiring and labelling the ECG leads correctly for its stated monitoring claim.
The case is educational, simplified and not clinical advice. Real medical-device decisions require the applicable standards, risk management file, intended-use statement, clinical workflow, regulatory evidence, usability validation, service procedures and qualified clinical review.
Case Context
A biomedical engineering team is reviewing a new reusable ECG trunk cable and lead-wire set for a bedside monitor. The monitor supports waveform display, heart-rate trending and stored 12-lead review. It is not a standalone diagnostic cardiology workstation, but the displayed lead labels are still safety-critical because clinicians use the waveforms to make decisions and to decide when a formal diagnostic ECG is needed.
After the cable update, three recordings show the same suspicious pattern:
- lead I appears inverted compared with previous records;
- leads II and III appear exchanged;
- aVL and aVR have unexpected polarity;
- precordial leads keep plausible R-wave progression;
- signal quality remains green and no lead-off alarm is active.
The central decision is:
Is this a patient-specific waveform difference, a low-quality acquisition artefact, or a lead mapping fault that requires a release hold?
The evidence points to a lead mapping fault.
Simplified Evidence
Use the following simplified evidence from the event review.
| Evidence | Observation | Engineering interpretation |
|---|---|---|
| affected devices | only units using the new cable lot | accessory or cable mapping is plausible |
| signal quality flag | normal | the device sees stable signals, not a lead-off condition |
| skin-electrode impedance | 7 to 14\ \text{k}\Omega | contact quality is not the main fault |
| lead I morphology | inverted relative to simulator and prior record | right-arm and left-arm electrode roles may be swapped |
| leads II and III | appear exchanged | consistent with right-arm and left-arm reversal |
| precordial leads V1 to V6 | progression remains plausible | chest electrode positions are probably not the primary error |
| alternate cable | waveform returns to expected polarity | device front end is not generally failed |
The important feature is repeatability. A random noisy recording would not produce the same lead relationship across devices, patients and a calibrated simulator.
Lead Definitions
For the limb electrodes:
- RA is the right-arm electrode potential;
- LA is the left-arm electrode potential;
- LL is the left-leg electrode potential.
The standard limb leads are:
Therefore:
The augmented leads are:
These equations are a useful engineering check, but they must be interpreted carefully. Einthoven consistency can still hold after a physical lead swap because the displayed channels may remain mathematically self-consistent while being assigned to the wrong electrode names.
Step 1: Check Whether the Waveform Is Internally Consistent
At a representative QRS peak from the device recording, the limb lead amplitudes are:
| Displayed lead | Measured amplitude |
|---|---|
| I | -0.79\ \text{mV} |
| II | +0.39\ \text{mV} |
| III | +1.18\ \text{mV} |
Use Einthoven’s relation:
Substitute the displayed values:
This matches displayed lead II:
Engineering Comment
The recording is internally consistent. That does not clear the device. It only shows that the displayed limb leads are related by the expected vector equation. A right-arm and left-arm reversal can pass this check because the wrong channels can still form a consistent triangle.
Step 2: Compare Against a Known Simulator Mapping
A calibrated ECG simulator is configured for a known normal limb-lead morphology. The expected QRS peak amplitudes are:
| Reference lead | Expected amplitude |
|---|---|
| I | +0.80\ \text{mV} |
| II | +1.20\ \text{mV} |
| III | +0.40\ \text{mV} |
| aVR | -1.00\ \text{mV} |
| aVL | +0.20\ \text{mV} |
| aVF | +0.80\ \text{mV} |
The device records:
| Displayed lead | Recorded amplitude |
|---|---|
| I | -0.79\ \text{mV} |
| II | +0.39\ \text{mV} |
| III | +1.18\ \text{mV} |
| aVR | +0.21\ \text{mV} |
| aVL | -0.98\ \text{mV} |
| aVF | +0.81\ \text{mV} |
If there were no reversal, the residuals would be:
| Lead | Recorded minus expected |
|---|---|
| I | -1.59\ \text{mV} |
| II | -0.81\ \text{mV} |
| III | +0.78\ \text{mV} |
| aVR | +1.21\ \text{mV} |
| aVL | -1.18\ \text{mV} |
| aVF | +0.01\ \text{mV} |
The root-mean-square residual is:
That residual is too large for a calibrated simulator test.
Step 3: Test the Right-Arm and Left-Arm Reversal Hypothesis
For an RA/LA reversal:
Apply that mapping to the expected simulator amplitudes:
| Displayed lead under RA/LA reversal | Predicted amplitude | Recorded amplitude | Residual |
|---|---|---|---|
| I’ | -0.80\ \text{mV} | -0.79\ \text{mV} | +0.01\ \text{mV} |
| II’ | +0.40\ \text{mV} | +0.39\ \text{mV} | -0.01\ \text{mV} |
| III’ | +1.20\ \text{mV} | +1.18\ \text{mV} | -0.02\ \text{mV} |
| aVR’ | +0.20\ \text{mV} | +0.21\ \text{mV} | +0.01\ \text{mV} |
| aVL’ | -1.00\ \text{mV} | -0.98\ \text{mV} | +0.02\ \text{mV} |
| aVF’ | +0.80\ \text{mV} | +0.81\ \text{mV} | +0.01\ \text{mV} |
The RMS residual for the reversal hypothesis is:
Engineering Comment
The right-arm and left-arm reversal hypothesis fits the simulator data far better than the no-reversal hypothesis:
This is strong engineering evidence of a mapping fault. It also explains why the signal-quality indicator stayed normal: the signals were real, stable and within range, but assigned to the wrong lead labels.
Step 4: Separate Signal Quality From Signal Meaning
The acquisition quality checks are not useless; they simply answer a different question.
| Check | Result | What it proves | What it does not prove |
|---|---|---|---|
| electrode impedance | pass | contact is plausible | electrode labels are correct |
| lead-off detection | pass | no open lead is detected | conductor positions are correct |
| SNR | pass | waveform is not dominated by noise | displayed lead identity is correct |
| sampling rate | pass | waveform timing is preserved | cable mapping is correct |
| latency | pass | display delay is controlled | morphology labels are correct |
| Einthoven consistency | pass | derived limb leads are mathematically coherent | physical electrode assignment is correct |
This distinction matters for risk management. A device can produce a clean, low-noise, well-sampled, mathematically consistent but wrongly labelled waveform.
Step 5: Root Cause Review
The investigation finds two contributing causes.
| Cause | Evidence | Control weakness |
|---|---|---|
| accessory cable lot has crossed RA and LA conductors at the trunk connector | alternate cable removes the pattern; continuity test confirms crossed pins | incoming inspection checked continuity, but not electrode-name-to-pin mapping |
| release test used only heart-rate and lead-off checks | test record shows simulator heart rate passed | verification did not include a known morphology mapping check |
The device electronics did not fail in the usual sense. The failure was at the interface between hardware labelling, accessory procurement, verification scope and release evidence.
Step 6: Risk and Release Decision
The risk is not limited to a cosmetic display error. Wrongly labelled leads can affect waveform interpretation, alarm review, clinician trust, escalation decisions and comparison with prior records.
| Failure mode | Effect | Severity | Occurrence | Detection | RPN |
|---|---|---|---|---|---|
| RA/LA lead reversal not detected by release test | displayed limb leads are wrong | 8 | 3 | 6 | 144 |
| signal-quality flag remains green during mapped fault | user receives false confidence | 7 | 3 | 7 | 147 |
| automated interpretation consumes wrongly labelled leads | misleading review statement | 8 | 2 | 6 | 96 |
| service cable replacement lacks mapping verification | error can recur after maintenance | 7 | 4 | 5 | 140 |
The engineering decision is:
Hold release of the cable lot and any software configuration that accepts it, quarantine affected accessories, add lead mapping verification with a known simulator morphology, update incoming inspection and service tests, and release only after the mapping fault is corrected and regression evidence is complete.
Corrective Action
The corrective action must address both the physical cable and the verification process.
- Quarantine the affected cable lot and identify every installed unit that used it.
- Perform continuity tests from each labelled electrode snap to the trunk connector pin.
- Run a simulator morphology test that checks lead polarity and expected limb-lead relationships.
- Confirm precordial lead mapping separately; a correct chest progression does not clear limb leads.
- Update incoming inspection to verify electrode-name-to-pin mapping, not only electrical continuity.
- Update service procedure after cable replacement to include a documented simulator snapshot.
- Add software or user-interface logic that flags improbable lead patterns where technically justified.
- Review the risk file and complaint/service records for possible field exposure.
Release Criteria
Release requires evidence that the physical, digital and user-facing states agree.
| Criterion | Required evidence |
|---|---|
| cable mapping | labelled electrode snaps trace to the approved connector pins |
| simulator morphology | known reference waveform displays correct polarity and lead relationships |
| limb-lead equations | displayed and derived leads satisfy equations after correct mapping |
| chest leads | V1 to V6 mapping is verified independently |
| signal-quality state | lead-off, impedance and signal-quality flags do not mask mapping faults in the release test |
| software configuration | channel map, firmware version and display labels match the approved design record |
| risk controls | lead reversal is linked to a detection or mitigation control in the risk file |
| service workflow | cable replacement procedure includes mapping verification and acceptance evidence |
| field action | affected units are identified, corrected and documented |
The release gate should not be satisfied by a heart-rate reading alone. Heart rate can be correct while lead identity is wrong.
Transferable Lessons
Lead reversal is a good example of a biomedical engineering failure that survives generic signal-quality checks. The waveform can be stable and the equations can still look coherent.
For ECG acquisition reviews, engineers should separate:
- signal integrity: noise, bandwidth, saturation, sampling and latency;
- electrode contact: impedance, lead-off detection and motion sensitivity;
- channel identity: electrode-to-pin mapping, firmware channel map and displayed labels;
- clinical workflow risk: how users interpret, compare and act on the labelled waveforms;
- release evidence: simulator morphology, accessory traceability and service verification.
This case is distinct from a front-end verification project. The front end may be good. The failure is that the system assigns real signals to the wrong physiological labels, which makes the acquisition evidence unsafe for the intended use.