INTRODUCTION: An effective cardiac resynchronization therapy (CRT) algorithm classifies each left ventricular (LV) pace as effective or ineffective based on electrogram morphology. Loss of capture, latency, pseudo-fusion during sinus rhythm, and rapid ventricular responses due to atrial fibrillation (AF) and premature ventricular conductions (PVCs) are potential causes of ineffective CRT pacing. However, little is known about the relationship between percentage of ineffective CRT pacing (%i-CRT) and CRT response. METHODS AND RESULTS: Among 136 consecutive CRT recipients, we evaluated 44 patients with an effective CRT algorithm. The primary outcome was the prevalence of CRT responders, defined as a reduction in the LV end-systolic volume by ≥ 10% 6 months after implantation. The median values of percent ventricular pacing and %i-CRT were 97.5% (interquartile range [IQR]: 96.4%-98.3%) and 0.3% (IQR: 0.1%-1.0%), respectively. Patients in the highest quartile of %i-CRT (%i-CRT ≥ 1%) showed a significantly lower prevalence of CRT responders (4 [36.4%] vs. 28 [84.8%], p = 0.004) than the remaining quartiles. The univariate analysis showed that %i-CRT ≥ 1% was negatively associated with the CRT response (odds ratio: 0.10, 95% confidence interval: 0.02-0.48). The most frequent cause of ineffective pacing was rapid ventricular responses due to AF and/or PVCs, whereas the cause of the highest %i-CRT in this cohort was latency. A total of three nonresponders with a high %i-CRT showed favorable responses after therapeutic interventions for their hidden causes. CONCLUSION: High %i-CRT could be associated with an unfavorable CRT response and provide insights into hidden causes and therapeutic targets.