AIMS: To assess treatment decision and outcome in patients referred for transcatheter aortic valve implantation (TAVI) in addition to predictive factors of mortality after TAVI. METHODS: Three-centre prospective observational study including 358 patients. Endpoints were defined according to the Valve Academic Research Consortium. RESULTS: Of the 358 patients referred for TAVI, TAVI was performed in 235 patients (65%), surgical aortic valve replacement (AVR) in 24 (7%) and medical therapy (MT) in 99 (28%). Reasons to decline TAVI in favour of AVR/MT were patient preference (29%), peripheral vascular disease (15%) and non-severe aortic stenosis (11%). The logistic EuroSCORE was significantly higher in patients who underwent TAVI and MT in comparison with those undergoing AVR (19 vs. 10%, p = 0.007). At 30 days, all-cause mortality and the combined safety endpoint were 9 and 24% after TAVI and 8 and 25% after AVR, respectively. All-cause mortality was significantly lower in the TAVI group compared with the MT group at 6 months, 1 year and 2 years (12% vs. 22%, 21% vs. 33% and 31% vs. 55%, respectively, p < 0.001). Multivariable analysis revealed that blood transfusion (HR: 1.19; 95% CI: 1.05-1.33), pre-existing renal failure (HR: 1.18; 95% CI: 1.06-1.33) and STS score (HR: 1.06; 95% CI: 1.02-1.10) were independent predictors of mortality at a median of 10 (IQR: 3-23) months after TAVI. CONCLUSIONS: Approximately two-thirds of the patients referred for TAVI receive this treatment with gratifying short- and long-term survival. Another 7% underwent AVR. Prognosis is poor in patients who do not receive valve replacement therapy.