Introduction: The recurrence of Paroxysmal Atrial Fi-brillation (PAF) is likely to be associated with more se-vere cardiovascular (CV) diseases and symptoms than de novo episode.

Objective: The identification of clinical and paraclini-cal differences between recurrent and de novo episode of PAF for the patients that need hospitalization for this event.

Methods: A retrospective study of 104 patients hospi-talized for an episode of PAF in the Cardiology Clinic of SCU “Sf. Pantelimon” Bucharest, between 23.04.2015 and 22.06.2016. Demographical, clinical, anamnestic and paraclinical data (EKG, echocardioghrapy) regar-ding the acute episode were collected. Particularities have been identified.

Results: 58.65% of patients have been found with recur-rent PAF. The average age was 70 years. Female patients constitute the majority (72.13%); same data for de novo episode. The presence of CV risks factors: hypertension (85.24%), dyslipidemia (50.81%), obesity (34.42%) and type 2 diabetes mellitus (31.14%) was also similar. Smo-king (30.23% vs. 18.03%) was higher in de novo type. HF, CAD and stroke was the most CV diseases associa-ted; just for HF we identify statistically near significant differences (p=0.07) between the two types, HF being most associated with recurrent PAF. The most frequent symptoms were palpitations, dyspnea and angina in hi-gher percent for the recurrent type, with statiscally near significant differences (p=0.08) for the palpitations. History of HF was statistically significant (SS) associ-ated with dyspnea in both types (p=0.02 vs. p=0.05, de novo vs. recurrent); history of CAD was SS associated with angina and dyspnea for de novo (p=0.02, p=0.02) and dyspnea for the recurrent (p=0.04); history of stro-ke was SS associated with symptoms (angina, fatigabi-lity and vertigo) only for the recurrent type (p=0.06, p=0.04, p=0.06). SS differences were noticed between the two types regarding the CHA2DS2-VASc score ≥4 (p=0.004); recurrent PAF being most associated with higher CHA2DS2-VASc scores. No SS diferrences were noticed for HAS-BLED score ≥3 between the two types. Conclusions: Recurrent PAF was present among the majority of the patients. SS differences between the two types were noticed regarding the presence of HF (higher in the recurrent episode) and the associated symptoms with CV disease history (CAD and stroke). Recurrent PAF has associated a CHA2DS2-VASc score ≥4 and a HAS-BLED score ≥3 in important percent, higher than in de novo, with SS difference for the CHA-2DS2-VASc score ≥4.