A handheld twelve-lead electrocardiograph produced this tracing at 21:09 on 17 August 2026. The device reported a heart rate of 59 beats per minute and displayed the message Analysis successful. The question is whether the recording can carry any diagnostic weight — and the honest answer is that the machine's confidence and the tracing's usability are two different things.
No — this should not be accepted as a diagnostic twelve-lead. It is usable as a rhythm capture and nothing more. Rate, regularity and QRS width can be stated with reasonable confidence. Intervals, repolarisation, axis and every measurement specific to conduction system capture cannot.
That is not a criticism of the device. It is a statement about three separable, correctable defects that happen to be present simultaneously, two of which are fixed at the electrode and one of which is fixed in a settings menu.
What the tracing does support
The rate is 59 beats per minute and the rhythm is regular. QRS complexes are narrow and, importantly, consistent from beat to beat — the morphology does not alternate, wander or vary with the respiratory cycle. Consistency is itself an informative finding, because it argues that the underlying activation sequence is stable even though the recording of it is not.
The precordial leads, particularly V2 through V6, are the cleanest part of the tracing. R-wave progression across the anterior leads appears plausible. That word is deliberate: plausible is as far as this recording allows anyone to go, because assessing progression properly requires confidence in both the amplitude calibration and the electrode positions, and neither can be verified from what is on the screen.
Nothing in the tracing suggests an emergency. There is no wide-complex tachycardia, no chaotic ventricular activity, no complete absence of organised depolarisation. A tracing can be too noisy to interpret while still being clean enough to exclude the small number of findings that would demand immediate action, and that is the case here.
What is compromising it
Continuous fine baseline noise in every lead
A dense, high-frequency fuzz rides on all twelve channels. This is the classic combined signature of electromyographic activity — muscle tension, shivering or tremor — and dry or poorly adhering electrodes raising the skin-electrode impedance. The practical consequence is not cosmetic. The noise band is deep enough that in several leads the P wave, which is a low-amplitude deflection at the best of times, sits inside it rather than above it.
Fix: skin preparation and fresh electrodes; relax and fully support the limbs.
Rectangular saturation excursions in the limb leads
Several large, flat-topped, near-vertical-edged deflections lasting hundreds of milliseconds appear in the limb lead channels, most obviously in lead III and in the aVR/aVL rows, with one intruding into the V1 row. No physiological event produces that shape. These are front-end amplifier saturation and recovery, caused by an abrupt change in electrode-skin impedance driving the differential input beyond its dynamic range.
Because the excursions cluster in the limb leads while lead I retains recognisable complex morphology, the pattern points towards one limb electrode intermittently losing and regaining contact, rather than a global acquisition failure. That distinction is worth making, because a single reseated electrode can resolve the whole finding.
Fix: reseat the limb electrodes; strain-relieve the cables.
The AC notch filter is set to 50 Hz
The bottom-left corner of the display reads AC:50Hz. A mains notch filter is a narrow band-stop filter centred on the powerline frequency; its narrowness is the entire point, because it must remove interference without removing signal. If this recording was made on a 60 Hz supply — as it would be in North America — then the filter is centred on the wrong frequency, the interference passes through essentially unattenuated, and a slice of genuine signal energy near 50 Hz is discarded for nothing in return.
Powerline interference of this kind is easily mistaken for muscle tremor, and in the atrial portion of the tracing it can be mistaken for fine fibrillatory activity. Correcting the setting costs one menu selection and may account for a substantial share of the fuzz described above.
Fix: set the notch filter to the local mains frequency before re-recording.
Interface occlusion and missing calibration
The Analysis successful toast sits directly over the V3 channel, and the Stop button covers the V6 label. Two of twelve leads are therefore partly unavailable for review. A calibration step does appear at the start of each lead row, which is a point in the recording's favour. What is missing is the annotation of paper speed and gain, and in any case a photograph of an angled, curved display cannot be measured against the underlying grid with the accuracy that voltage criteria require. Any statement about hypertrophy or low-voltage complexes from this capture would therefore be unfounded.
Fix: export or print the recording rather than photographing the live screen.
Why this matters more in a conduction system pacing patient
In a person without a device, a noisy twelve-lead is an inconvenience: the rhythm is usually still readable, and the questions that a clean tracing would answer can often wait. In a person with a left bundle branch area pacing system, the calculus changes, because the measurements that characterise the device's behaviour are precisely the ones that live closest to the noise floor.
Conduction system capture is not assessed from the general shape of the complex. It is assessed from fiducial points measured to within a few milliseconds: the onset of the QRS, the peak of the R wave in V6, the timing of the pacing stimulus itself, the presence or absence of a low-amplitude terminal deflection in V1. Baseline noise does not merely make these harder to read — it makes the boundaries genuinely ambiguous, and an ambiguous fiducial point produces a confidently wrong number rather than an obviously missing one.
| Assessment | Supported here? | Why |
|---|---|---|
| Rate and regularity | Yes | R-peaks are unambiguous despite the noise |
| Narrow versus wide QRS | Yes | Gross width is well above the noise band |
| Beat-to-beat morphology stability | Yes | Comparison is relative, not absolute |
| P-wave presence and morphology | Partly | Low amplitude; buried in noise in several leads |
| PR and QT intervals | No | Onsets and offsets are not resolvable |
| Precise QRS duration | No | Earliest onset and latest offset are blurred |
| Frontal plane axis | No | Limb leads carry the saturation artefact |
| ST segment and T wave | No | Baseline is not stable enough to define the isoelectric line |
| Voltage criteria | No | Speed and gain unannotated; a photographed screen is not measurable |
| Pacing stimulus identification | No | A stimulus of a few milliseconds is indistinguishable from noise spikes |
| V6 R-wave peak time | No | Requires QRS onset and R peak to within a few milliseconds |
| Stimulus to LV activation time | No | Requires an unambiguously timed pacing artefact |
| Terminal r′ in V1 | No | Low-amplitude late deflection sitting at the noise floor |
The unanswered question in this tracing
There is one observation worth flagging without over-reading it. The displayed rate is 59 beats per minute. In a paced patient, a rate a fraction below a round number invites a specific question: is this an intrinsic rhythm running just under the programmed lower rate limit, is the lower rate limit programmed below 60, or is a rate hysteresis or sleep-rate function active at 21:09 in the evening?
This tracing cannot answer that, because it cannot show whether the beats are paced, fused or intrinsic. Device interrogation can answer it in seconds. That asymmetry — a question raised by the surface ECG and settled only by the device — is the same pattern discussed at greater length in the ABC Farma page on the incremental value of device interrogation beyond symptoms and the twelve-lead ECG.
The device displayed Analysis successful. That message reports that the algorithm completed without an internal error — it is a statement about the software, not a certification of the recording. Automated interpretation algorithms are validated on signals meeting quality assumptions this tracing does not meet, and their behaviour in paced patients is a known limitation area. A successful analysis on an unusable signal should be read as a reminder of that gap, not as reassurance.
How to re-record it properly
- Set the notch filter to the local mains frequency — 60 Hz in North America and much of South America, 50 Hz across Europe, Africa and most of Asia. Do this first; it costs nothing.
- Prepare the skin. Remove surface oils with alcohol, abrade lightly with a preparation pad or dry gauze, and shave hair at any electrode site. This single step removes more artefact than any filter.
- Use moist electrodes from a sealed package. Dried gel raises impedance by orders of magnitude and is the most common cause of both fine noise and intermittent saturation.
- Reseat the limb electrodes specifically, given where the saturation excursions cluster, and make sure each cable is strain-relieved so movement does not pull at the connector.
- Position the person supine, warm and relaxed, arms fully supported, hands open rather than gripping. Isometric forearm contraction is a potent electromyographic source, and cold provokes shivering that is indistinguishable from tremor on the trace.
- Wait several seconds after application for the electrode-skin impedance to settle before starting acquisition.
- Export or print rather than photographing the screen, so the calibration pulse, gain and paper speed are preserved and no interface element covers a lead.
Frequently asked questions
Can a 12-lead ECG with baseline noise, saturation artefact and a 50 Hz notch filter be interpreted in a patient with left bundle branch area pacing?
Not as a diagnostic twelve-lead. It supports a rhythm statement — 59 beats per minute, regular, narrow QRS, stable beat-to-beat morphology — and nothing further. Continuous baseline noise buries the P wave in several leads, rectangular saturation excursions dominate the limb leads, and the mains notch filter is set to 50 Hz, which lets 60 Hz interference through untouched if the recording was made on a 60 Hz supply. Two leads are covered by interface elements, and although a calibration step is present, paper speed and gain are not annotated and a photographed screen cannot be measured. In a conduction system pacing patient this matters more than usual, because paced QRS duration, V6 R-wave peak time, stimulus to left ventricular activation time and stimulus identification all depend on millisecond-scale fiducial points at the noise floor. Re-record rather than interpret.
What causes the rectangular, flat-topped deflections seen in the limb leads?
Front-end amplifier saturation and recovery. Flat tops, near-vertical edges and durations of several hundred milliseconds are not produced by cardiac tissue; the shape belongs to an amplifier whose input has been driven beyond its dynamic range by an abrupt change in electrode-skin impedance. An electrode losing and regaining contact, a dried gel pad, a tugged cable or a moving connector will all do it. Their concentration in the limb leads, with lead I largely spared in morphology, points to one particular limb electrode rather than a whole-system failure.
Does a notch filter set to 50 Hz matter if the recording was made on 60 Hz mains?
Yes. The notch is a narrow band-stop filter centred on the powerline frequency, and its selectivity is what makes it safe to use. Centred at 50 Hz on a 60 Hz supply, it rejects a slice of genuine signal while the actual interference passes through unattenuated. The visible result is a persistent fine oscillation on every lead that is easily mistaken for tremor, or in the atrial portion of the trace for fine fibrillatory waves.
Which LBBAP measurements are lost first when a tracing is noisy?
The ones defined by a precise instant rather than a general shape. V6 R-wave peak time needs QRS onset and the R peak located to within a few milliseconds. Stimulus to left ventricular activation time needs the pacing artefact itself to be visible and timed. Paced QRS duration needs the earliest onset and latest offset across leads, exactly where noise blurs the boundary. The terminal r′ in V1 is a low-amplitude late deflection near the noise floor. Distinguishing paced from fused from intrinsic beats needs a stimulus a few milliseconds wide. All of these fail long before the rhythm becomes unreadable.
How should a handheld 12-lead be re-recorded to remove this artefact?
Set the mains notch filter to the local supply frequency first. Then work on the electrode-skin interface, which causes both the fine noise and the saturation: degrease with alcohol, abrade lightly, shave where needed, and use moist electrodes from a sealed package. Reseat the limb electrodes and strain-relieve the cables. Position the person supine and warm, arms fully supported, hands relaxed rather than gripping. Wait several seconds for impedance to settle before acquiring, and export or print the result so the calibration pulse, gain and paper speed are preserved.
Where this leaves the recording
The correct disposition for a tracing like this is not interpretation and not dismissal. It is repetition. Nothing about the underlying rhythm looks alarming, the artefacts have identifiable and inexpensive causes, and a properly acquired tracing five minutes later would answer every question this one raises. What should not happen is that a noisy recording is read as evidence of pathology — muscle tremor read as atrial fibrillation, a saturation excursion read as an arrhythmic event, or a mains-contaminated baseline read as ischaemic ST change. Those errors have consequences, and they all begin with accepting a signal that should have been rejected.
For anyone with a conduction system pacing device, the second half of the answer is that the surface ECG was never the whole picture. The questions this tracing raises — whether beats are paced or intrinsic, what the lower rate limit is doing at this hour, whether capture has changed — belong to the device, and a related discussion of that division of labour is set out in the companion pages on persistent frontal axis findings after LBBAP and on what interrogation adds beyond the twelve-lead.
See the annotated walkthrough of the original screen capture →