Introduction: Syncope is responsible for up to 5% of emergency department visits. Vasovagal syncope (VVS) is the most common subtype and can have significant quality of life implications as it is often recurrent. Clinicians treating VVS have limited treatment options available to them and often struggle with prognostication. The aim of our study was to identify patient-specific determinants of VVS improvement or cessation. Methods: Patients (pts) from the Prevention of Syncope Trials (POST) 1 and 2 were included in this study. All patients had VVS according to tilt table testing or a diagnostic point score. Patients had fainted ≥1 time in the previous year and all were followed for up to 1 year after enrollment. Data are presented as median (IQR). Complete responders (CR) did not faint in follow-up; partial responders (PR) fainted ≥1/year less than prior year but did not stop; and non-responders (NR) did not improve or stop. Results: There were 392 patients: 126 males, median age 34 (23,50) who had fainted for 10 (3,22) years and followed for a median of 363 (148,376) days. There were 225 CR (57%), 120 PR (31%) and 47 NR (12%). PR subjects were younger: 27 (24,33) years compared to CR (36 (32,42)) years and NR (36 (29,47)) years (p<0.05). Receiver operator characteristic analysis showed age predicted PR (AUC=0.62). Lifetime fainting frequency was 0.67 (0.14,2.00) faints per year, increasing to 4 (2,10) faints in the pre-year and decreasing to 0 (0,1.9) faints in the post-year (p<0.0001). Pts had similar syncope frequency in the distant past (PR, 1.14 faints/year; CR, 0.68 faints/year; NR, 0.58 faints/year) but PR pts worsened markedly prior to enrollment. PR subjects fainted much more in the prior year: 10 (6,18) faints compared CR (3 (2,3) faints, p<0.0001) and NR (2 (2,4) faints, p<0.05). Receiver operator characteristic analysis showed prior year faints predicted PR well (AUC=0.81). There was no significant interaction with treatment (metoprolol in POST 1, fludrocortisone in POST 2). Conclusion: After specialist consultation, 57% of VVS patients stop fainting and 31% improve incompletely without a significant treatment effect. Patients who will improve incompletely can be accurately selected based on younger age and more frequent syncope. Older patients with less frequent syncope are 83% likely to stop fainting. These findings will help counsel pts and select candidates for medical therapy.