Leadless Pacemaker
End-of-Service (EOS)
6-Minute Walk Test
Question & Answer
Question
During a 6MWT performed at low battery status, what is the incidence of exertional non-capture or intermittent capture (verified by ambulatory ECG/EGM) and does the burden of capture failure correlate with shorter 6MWD, greater HR variability, higher NT-proBNP the next day, or post-test symptoms?
Answer (Prospective Monitoring Protocol & Analysis Plan)
TL;DR
Implement simultaneous 6MWT + ambulatory ECG/EGM logging near EOS to quantify exertional capture failure (ECF).
Express ECF as % paced beats without capture and/or episodes per minute.
Expect that higher ECF burden associates with shorter 6MWD, more erratic minute-to-minute HR,
elevated next-day NT-proBNP, and more symptoms. Use adjusted models and event-level analyses to test these relationships.
1) Verification & Definitions
Non-capture (beat-level)
- Pacing artifact present with no ensuing paced QRS (or morphology inconsistent with capture) within vendor-specific capture window.
- EGM markers: pace delivered (Vp), no capture confirmation if available.
Intermittent capture (episode-level)
- ≥3 consecutive non-captured paced beats or ≥5 non-captured beats within 10 s during walking.
- Flag episodes with start/stop timestamps to align with HR/SpO2 traces.
ECF Burden
Beat-level: non-captured paced beats / total paced beats (%)
Episode-level: episodes per minute; total duration (s)
2) Data Collection (same day unless noted)
- Ambulatory ECG: 3‑lead Holter or patch (250–500 Hz). Sync clock with LP programmer.
- Device EGM/logs: capture threshold, output, safety/backup events, sensed R-wave, impedance, RR settings.
- 6MWT signals: continuous HR, SpO2, laps; Borg at finish; interruptions noted.
- Symptoms: presyncope, dizziness, dyspnea, chest pain during/after test (binary + severity scale).
- NT-proBNP: drawn at baseline (within 7 days pre-test) and next day (24±6 h).
3) HR Variability (operationalized for 6MWT)
- Minute-to-minute variability: SD of 60‑s average HR across the 6 minutes (
SD6×60s).
- Short-term variability: RMSSD from 1‑s HR series (artifact-corrected), reported as
RMSSDwalk.
- Instability index: coefficient of variation (CV) of 5‑s HR bins during the walk.
4) Primary Endpoints
Incidence
- Proportion of patients with any ECF during 6MWT at EOS.
- Median (IQR) ECF burden (% non-capture) among those with ECF.
Correlates
- Association of ECF burden with 6MWD (m).
- Association with HR variability indices (SD6×60s, RMSSDwalk, CV).
- Association with ΔNT-proBNP (next day − baseline).
- Association with post-test symptoms.
5) Statistical Plan
- Incidence: Exact binomial CI for any ECF; report episode rate (Poisson with exposure = walk time).
- Correlations: Linear/quantile regression for 6MWD and HRV; logistic regression for symptoms; linear model for ΔNT-proBNP.
- Adjustment set (pre-specified): age, sex, BMI, hemoglobin, eGFR, LVEF, pulmonary function, β‑blocker dose, chronic capture threshold, programmed output, RR slope/sensitivity.
- Nonlinearity: Restricted cubic splines for ECF burden and chronic threshold.
- Event-level analysis: Align ECF episodes with 10‑s windows of HR/SpO2; mixed models with random subject intercepts.
- Sensitivity: Exclude tests with corridor interruptions; repeat using %-predicted 6MWD.
6) Expected Patterns (Hypothesis)
- Higher ECF burden → shorter 6MWD (dose–response).
- Higher HR variability during walk when ECF is present (irregular capture → irregular RR pacing).
- ΔNT-proBNP > 0 the next day in those with substantial ECF, reflecting transient hemodynamic stress.
- Symptoms more frequent with ECF episodes >10 s or with cluster frequency >2/min.
7) Operational Safety
- Pre-walk interrogation: verify capture at programmed output; document safety/backup behaviors.
- Stop rules: presyncope, chest pain, malignant arrhythmia, SpO2 < 85%.
- Have external pacing/defib resources immediately available during EOS testing.
8) Sample Size (planning)
- For detecting a per 5% ECF decrease of 8–12 m in 6MWD (SD ≈ 60 m), ~70–110 participants provide 80% power (two-sided α=0.05) depending on covariate R².
- Precision goal for incidence: to estimate 15% incidence with ±6% margin, n≈140.
9) Reporting Template
Incidence any ECF: ____% (95% CI ____ to ____)
ECF burden (non-captured/total paced): median ____% (IQR ____–____)
6MWD vs ECF: β per 5% ECF = ____ m (95% CI ____ to ____), p = ____
HR variability vs ECF: ΔRMSSDwalk = ____ ms; ΔSD6×60s = ____ bpm
ΔNT-proBNP (ng/L): ____ (adjusted β per 5% ECF = ____)
Symptoms (OR per 5% ECF): ____ (95% CI ____ to ____)
Event-level: ECF windows associated with HR drop of ____ bpm and SpO₂ drop of ____ % (mixed-model p = ____)
10) Caveats
- Accurate beat-level annotation requires high-fidelity ECG and synchronized clocks; reconcile programmer time drift.
- Wearable-derived HR can mask capture failure timing; prefer ECG/EGM for primary ECF definition.
- NT-proBNP affected by renal function and circadian variation—ensure consistent timing and adjust for eGFR.