Abstract Aims A dual-chamber pacemaker with closed-loop stimulation (CLS) mode is effective in reducing syncopal recurrences in patients with asystolic vasovagal syncope (VVS). In this study, we explored the haemodynamic and temporal relationship of CLS during a tilt-induced vasovagal reflex. Methods and results Twenty patients underwent a tilt test under video recording 3.9 years after CLS pacemaker implantation. Three patients were excluded from the analysis because of no VVS induced by the tilt test (n = 1) and protocol violation (n = 2). In 14 of the remaining 17 patients, CLS pacing emerged during the pre-syncopal phase of circulatory instability when the mean intrinsic heart rate (HR) was 88 ± 12 b.p.m. and systolic blood pressure (SBP) was 108 ± 19 mmHg. The CLS pacing rate thereafter rapidly increased to 105 ± 14 b.p.m. within a median of 0.1 min [inter-quartile range (IQR), 0.1–0.7 min] when the SBP was 99 ± 21 mmHg. At the time of maximum vasovagal effect (syncope or pre-syncope), SBP was 63 ± 17 mmHg and the CLS rate was 95 ± 13 b.p.m. The onset of CLS pacing was 1.7 min (IQR, 1.5–3.4) before syncope or lowest SBP. The total duration of CLS pacing was 5.0 min (IQR, 3.3–8.3). Closed-loop stimulation pacing was not observed in three patients who had a similar SBP decrease from 142 ± 22 mmHg at baseline to 69 ± 4 mmHg at the time of maximum vasovagal effect, but there was no significant increase in HR (59 ± 1 b.p.m.). Conclusion The reproducibility of a vasovagal reflex was high. High-rate CLS pacing was observed early during the pre-syncopal phase in most patients and persisted, although attenuated, at the time of maximum vasovagal effect. Registration ClinicalTrials.gov identifier: NCT06038708