BACKGROUND: Cardioinhibitory vasovagal response is uncommon during the tilt test (TT). Heart rate variability (HRV) by use of spectral analysis can distinguish patients with that response. OBJECTIVE: To compare the HRV in patients with cardioinhibitory vasovagal syncope (case group - G1) with that in patients without syncope and with negative response to TT (control group - G2). METHODS: 64 patients were evaluated (mean age, 36.2 years; 35 men) and submitted to TT at 70 degrees, under digital Holter monitoring. The groups were paired for age and sex (G1, 40 patients; G2, 24). RESULTS: In G1, 21 patients had a type 2A response and 19 had type 2B, with mean TT duration of 20.4 minutes. There was a greater low frequency (LF) component (11,6 versus 4,5 ms2, p=0.001) and a lower low/high frequency ratio in the supine position (3,9 versus 4,5 ms2, p=0.008) in G1, with no difference during TT between the groups. Applying the receiver operating characteristic curve for cardioinhibitory response, the area under the curve was 0.74 for the LF component in the supine position (p = 0.001). The following were observed for the cutoff point of 0.35 ms(2) for the LF component: sensitivity, 97.4%; specificity, 83.3%; positive predictive value, 85.3%; negative predictive value, 96.9%; and positive likelihood ratio, 5.8. CONCLUSION: HRV in the supine position allowed identifying patients with syncope and cardioinhibitory response with a high negative predictive value and likelihood ratio of 5.8.