Comparing Serial Kardia 6L ECGs in a Leadless Pacemaker Patient: Unreadable vs. Unclassified, Non-Capture, and Circadian Threshold Variation

The clinical question

The clinical question, unpacked: what is meaningfully different between these two tracings, and does the difference warrant action?

Why Kardia labels paced ECGs "Unreadable" or "Unclassified"

Before interpreting either tracing, the on-device labels themselves need to be demoted as clinical signal. The AliveCor Kardia classifier runs on Lead I only and is trained predominantly on native narrow-QRS rhythms. Paced rhythms violate two core assumptions: the QRS is wide and non-sinus, and P-waves are often obscured or dissociated.

The practical implication in a pacemaker patient is:

• "Unreadable" typically means the rate-detection stage failed entirely — usually because pacing spikes, wide QRS, or baseline artifact prevented stable R-peak extraction.
• "Unclassified" means a rate was extracted but the rhythm did not fit any trained class (sinus, AF, bradycardia, tachycardia).

Neither label is diagnostic. Both require the tracing to be read manually. For patients with pacemakers, the device-reported rate and the visible morphology across all six leads carry the clinical weight.

Reading Tracing A (April 9): the baseline paced picture

The dominant finding on the April 9 6L is uniformity. Every QRS is wide, the morphology is LBBB-like with a left-superior axis — deeply negative in leads II, III, and aVF, positive in I and aVL — which is the canonical signature of apical or septal RV pacing. R-R intervals are regular across the 30-second recording, and the visible rate on the strip is in the mid-60s, consistent with the patient's programmed base rate with modest sensor-driven augmentation.

The "Unreadable" label here is a classifier artifact, not a clinical finding. The tracing represents the patient's chronic paced baseline.

Reading Tracing B (April 19): three findings that travel together

The April 19 tracing differs from the baseline in three linked ways. None in isolation is pathognomonic; together they define a distinct clinical pattern.

1. Rate below the programmed lower rate limit

The device-reported rate is 50 bpm, and the strip visually confirms slower cycle lengths than the April 9 recording. For an Aveir VR typically programmed with a lower rate limit of 60 bpm (VVIR), a sustained rate of 50 bpm across a 30-second window cannot be explained by rate-responsive down-adjustment at rest — the lower rate limit should act as a floor.

2. Irregular R-R intervals

Lead II on pages 4 and 5 shows clear beat-to-beat variability: longer cycles followed by shorter cycles, without the metronomic regularity that characterizes consistent VVIR pacing at base rate. Pacing at a fixed sensor-indicated rate should not produce irregularly irregular R-R intervals in the absence of sensing, capture, or conduction abnormalities.

3. Beat-to-beat QRS morphology variability

Most complexes retain the expected paced morphology (wide, LBBB-like, superior axis). A subset, however, show a different initial QRS vector and narrower width in leads I and aVL, with a less deeply negative deflection in the inferior leads. These are morphologically compatible with either intrinsic conducted beats (implying transient recovery of AV conduction or a junctional escape), fusion beats (intrinsic activation colliding with paced activation), or pseudofusion (pacing spike falling on an already-initiated intrinsic beat).

Side-by-side summary

Differential diagnosis for the April 19 findings

In a chronically RV-paced patient with complete heart block as the original indication, the leading explanations for the April 19 pattern, in descending order of likelihood, are:

1. Intermittent loss of capture with junctional or ventricular escape. Irregular wide complexes at a rate below the programmed floor, with interspersed narrower beats, fit this pattern best. Capture failure in the early-morning window is classical.
2. Functional non-capture due to circadian capture threshold rise. Capture thresholds vary diurnally with vagal tone, cortisol, and sleep-state autonomic shifts, typically peaking between 03:00 and 07:00. A device programmed with a narrow safety margin above threshold can transiently fail to capture during this window and return to normal capture by mid-morning.
3. Undersensing producing competitive pacing and pseudofusion. Accounts for morphology variability without requiring true capture failure, though less likely to produce a sustained reported rate of 50 bpm.
4. Atrial arrhythmia with variable AV conduction. Less likely given complete heart block as the underlying substrate, but not excluded from a 6L alone if AV conduction has partially recovered.
5. Sensor or programming issue. A reprogrammed lower rate limit or sensor malfunction can explain the rate but not the morphology variability.

What the surface tracing cannot tell you — and what device interrogation will

A Kardia 6L is a screening tool, not a confirmatory one. For this patient, the next step is not another surface recording but formal device interrogation focused on the 05:41 timestamp on April 19. Specifically:

• Marker channel at the event timestamp will label each event as paced capture, paced non-capture, sensed, or noise, definitively distinguishing the differential items above.
• Automatic capture threshold trend over the preceding 2–4 weeks will show whether thresholds are climbing globally or cycling circadianly.
• Programmed output versus measured threshold will reveal whether the safety margin has narrowed to the point where morning threshold rises cross programmed output.
• Stored electrograms will confirm or exclude true loss of capture versus fusion/pseudofusion.

Clinical relevance for the LBBAP upgrade discussion

In patients with high RV-pacing burden who are already being evaluated for upgrade to left bundle branch area pacing (LBBAP), evidence of intermittent non-capture represents a second, independent indication for intervention — distinct from the chronic remodeling and biomarker trajectories that typically drive the upgrade discussion. A patient who is both developing pacing-induced cardiomyopathy physiology and experiencing episodic capture reliability problems has more reason to move, and to move sooner, than a patient with remodeling alone.

Teaching points

• Kardia "Unreadable" and "Unclassified" labels in a pacemaker patient are classifier limitations, not clinical findings. Always read the tracing manually.
• A device-reported rate below the programmed lower rate limit in a pacemaker patient is, by itself, an indication for device interrogation.
• Circadian variation in capture threshold is real and is most pronounced between approximately 03:00 and 07:00. Serial early-morning tracings are disproportionately informative.
• Beat-to-beat morphology variability in a previously uniformly-paced rhythm suggests fusion, pseudofusion, escape, or non-capture. The surface ECG narrows the differential; the device resolves it.
• In patients being considered for LBBAP upgrade, evidence of capture unreliability on a chronic RV lead adds urgency independent of remodeling findings.

Frequently asked questions

Why does Kardia label a paced ECG as "Unreadable" or "Unclassified"?

The Kardia on-device algorithm runs on Lead I and is trained on native narrow-QRS rhythms. Wide paced QRS complexes and obscured P-waves break its detection assumptions, so paced tracings default to "Unreadable" (no rate extracted) or "Unclassified" (rate extracted but rhythm does not match a trained class). In a pacemaker patient, neither label carries clinical weight.

What does an RV-paced QRS morphology look like on a Kardia 6L?

Apical or septal RV pacing produces a wide LBBB-like QRS with a left-superior axis: negative in II, III, and aVF, positive in I and aVL. When capture is consistent, the morphology is uniform beat-to-beat.

How do you recognize intermittent loss of capture on a 6L ECG?

Look for a reported rate below the programmed lower rate limit, pacing spikes not followed by the expected paced QRS, R-R irregularity in a previously regular paced rhythm, and intermixed narrower complexes representing escape or conducted beats.

Why would capture thresholds rise in the early morning?

Capture thresholds follow a circadian pattern driven by vagal tone, cortisol nadir, and sleep-state autonomic changes. Thresholds typically peak between 03:00 and 07:00. A device programmed with a narrow safety margin can transiently fail to capture in this window and recover spontaneously later.

When should a Kardia tracing trigger formal device interrogation?

When it shows a rate below the programmed lower rate limit, new rhythm irregularity in a previously paced-regular patient, visible non-captured pacing spikes, beat-to-beat morphology variability, or correlated symptoms such as presyncope.

References and further reading