Which predilatation balloons provide the best lesion preparation prior to use of drug coated balloons in de-novo lesions? Results from the PREPARE study.

BACKGROUND: There is a lack of data on the clinical outcomes following the use of different strategies for lesion preparation prior to the use of drug-coated balloons (DCB). In this study, we have explored the clinical outcomes between different types of predilatation balloons: semi-compliant (SB), non-compliant (NB) and scoring balloons (ScB) used when preparing de-novo lesions prior to the use of DCB. METHODS: We retrospectively evaluated all patients who underwent treatment with DCB for de-novo lesions between 2011-2019 at 4 high-volume European centres. The measured study endpoints were cardiac-death, TV-MI, TLR and MACE. RESULTS: During the study period, 553 patients were treated with DCB for de-novo lesions, 327 with SB only, 172 with NB and 54 with ScB. There were some differences in the procedural characteristics between the 3 groups. Predilatation balloons were significantly larger in the ScB, and NB groups as compared to the SB (2.7 mm and 2.6 mm vs. 2.3 mm; P<0.001). The reference vessel diameter was significantly larger in the NB group as compared to the ScB and SB (2.6 mm vs. 2.2 mm and 2.3 mm; P<0.001). During the median follow-up duration of 547-days, there were no differences in the hard-clinical endpoints, however, TLR was significantly higher in the ScB as compared to SB and NB group (11% vs. 3.4% and 4.7%; P=0.02). CONCLUSIONS: The PREPARE study results do not suggest routine use of ScB prior to DCB in de-novo lesions.

Ischemia-guided vs routine non-culprit vessel angioplasty for patients with ST segment elevation myocardial infarction and multi-vessel disease: the IAEA SPECT STEMI trial.

BACKGROUND: In patients with multi-vessel disease presenting with ST elevation myocardial infarction (STEMI), the efficacy and safety of ischemia-guided, vs routine non-culprit vessel angioplasty has not been adequately studied. METHODS: We conducted an international, randomized, non-inferiority trial comparing ischemia-guided non-culprit vessel angioplasty to routine non-culprit vessel angioplasty, following primary PCI for STEMI. The primary outcome was the between-group difference in percent ischemic myocardium at follow-up stress MPI. All MPI images were processed and analyzed at a central core lab, blinded to treatment allocation. RESULTS: In all, 109 patients were enrolled from nine countries. In the ischemia-guided arm, 25/48 (47%) patients underwent non-culprit vessel PCI following stress MPI. In the routine non-culprit PCI arm, 43/56 (77%) patients underwent angioplasty (86% within 6 weeks of randomization). The median percentage of ischemic myocardium on follow-up imaging (mean 16.5 months) was low, and identical (2.9%) in both arms (difference 0.13%, 95%CI - 1.3%-1.6%, P < .0001; non-inferiority margin 5%). CONCLUSION: A strategy of ischemia-guided non-culprit PCI resulted in low ischemia burden, and was non-inferior to a strategy of routine non-culprit vessel PCI in reducing ischemia burden. Selective non-culprit PCI following STEMI offers the potential for cost-savings, and may be particularly relevant to low-resource settings. (CTRI/2018/08/015384).

Multivessel spontaneous coronary artery dissection: Clinical features, angiographic findings, management, and outcomes.

BACKGROUND: Spontaneous coronary artery dissection (SCAD) is a rare cause of acute coronary syndrome. Multivessel (MV) SCAD is a challenging clinical presentation that might be associated to a worse prognosis compared with patients with single-vessel (SV) involvement. METHODS: The Spanish multicentre nationwide prospective SCAD registry included 389 consecutive patients. Patients were classified, according to the number of affected vessels, in SV or MV SCAD. Major adverse events (MAE) were analyzed during hospital stay and major cardiac or cerebrovascular adverse events (MACCE) at long-term clinical follow-up. RESULTS: A total of 41 patients (10.5%) presented MV SCAD. These patients had more frequently a previous history of hypothyroidism (22% vs 11%, p = 0.04) and anxiety disorder (32% vs 16%, p = 0.01). MV SCAD patients presented more often as non-ST segment elevation myocardial infarction (73% vs 52%, p = 0.01) and showed less frequently type 1 angiographic lesions (12% vs 21%, p = 0.04). An impaired initial Thrombolysis In Myocardial Infarction (TIMI) flow 0-1 was less frequent (14% vs 29%, p < 0.01) in MV SCAD. In both groups, most patients were treated conservatively (71% vs 79%, p = NS). No differences were found regarding in-hospital MAE or MACCE at late follow-up (median 29 ± 11 months). However, the rate of stroke was higher in MV SCAD patients, both in-hospital (2.4% vs 0%, p < 0.01) and at follow-up (5.1% vs 0.6%, p = 0.01). CONCLUSIONS: Patients with MV SCAD have some distinctive clinical and angiographic features. Although composite clinical outcomes, in-hospital and at long-term follow-up, were similar to those seen in patients with SV SCAD, stroke rate was significantly higher in patients with MV SCAD.

Clinical Implications of the "Broken Line" Angiographic Pattern in Patients With Spontaneous Coronary Artery Dissection.

Spontaneous coronary artery dissection (SCAD) is a rare but increasingly recognized cause of acute myocardial infarction. Coronary angiography remains the best diagnostic tool; however, clinical suspicion and experience is required to interpret angiographic findings. This study sought to assess the clinical implications of the "broken line" (BKL) angiographic pattern in a large, nationwide, cohort of patients with SCAD. The Spanish SCAD registry (NCT03607981) prospectively enrolled consecutive patients with SCAD. All angiograms were centrally analyzed and the BKL pattern was systematically assessed. The BKL angiographic pattern was found in 64 of 389 patients (16%). Patients with the BKL appearance were more frequently female (97 vs 87%, p <0.05), presented more often as intramural hematoma (83 vs 58%, p <0.001), had longer lesions (47 ± 29 vs 36 ± 22 mm, p <0.01), and had severe tortuosity (25 vs 10%, p <0.01) but showed better initial coronary flow (thrombolysis in myocardial infarction flow 2.6 ± 0.8 vs 2.1 ± 1.2, p <0.01). Patients with BKL received more frequently conservative medical management (91 vs 76%, p <0.01). At late clinical follow-up (median 29 months, interquartile range 17 to 38) predefined adverse events (death, myocardial infarction, revascularization, recurrent SCAD, or stroke) occurred less frequently (3.5 vs 15%, p <0.05) in patients with the BKL appearance. The better clinical outcomes of patients in the BKL group persisted after adjusting for potential confounders (adjusted hazard ratio 0.2, 95% confidence interval 0.1 to 0.9, p <0.05). In conclusion, patients with SCAD presenting the BKL angiographic pattern are more frequently female and present more often as intramural hematoma with longer lesions and severe vessel tortuosity but have better coronary flow. Patients with the BKL morphology have a favorable prognosis (NCT03607981).

Subclavian angioplasty during coronary interventions using radial approach.

BACKGROUND: In the past years, the percentage of percutaneous coronary angiography and coronary interventions using radial access had significantly increased due to its higher safety, lower risk of major bleeding, and hence lower cardiovascular mortality. Subclavian artery stenosis is one of the challenges that may be met during transradial coronary interventions, which may necessitate femoral access crossover or conversion. AIMS: To evaluate the feasibility and safety of performing subclavian angioplasty via radial access during complex coronary interventions using the forearm approach. METHODS: A series of patients with complex radial approach due to subclavian stenosis received subclavian angioplasty during the procedure. We included 48 patients out of 22 500 procedures performed from February 2009 to February 2020. All patients did not have alternative vascular access due to extensive peripheral arterial disease (previous history of iliac stenting or distal aortic occlusion, which makes femoral access crossover difficult; also the contralateral radial/ulnar artery was very faint or not detectable at all). RESULTS: Mean age was 72 (10) years and 67% of patients were males. Subclavian angioplasty was successfully done in all patients via ipsilateral radial access; 44 patients (91.7%) required subclavian stenting, and 4 patients were treated by subclavian angioplasty without stenting. Coronary angiography or intervention was perfectly achieved through the revascularized subclavian artery; coronary stenting was successfully done in 36 patients as indicated. CONCLUSIONS: It can be concluded that percutaneous subclavian artery angioplasty can be done safely and effectively to facilitate complex transradial coronary procedures with an acceptable immediate technical success, especially in patients without alternative vascular access. Also, we may conclude that subclavian angioplasty may be successfully performed in patients with symptomatic upper limb ischemia, via the radial approach.

Clinical outcomes in spontaneous coronary artery dissection.

OBJECTIVE: Spontaneous coronary artery dissection (SCAD) is an infrequent cause of acute coronary syndrome. Our aim was to assess adverse events at follow-up from a nationwide prospective cohort. METHODS: The Spanish Registry on SCAD (SR-SCAD) included patients from 34 hospitals. All coronary angiograms were analysed by two experts. Those cases with doubts regarding the diagnosis of SCAD were excluded. The angiographic SCAD classification by Saw et al was followed. Major adverse cardiovascular and cerebrovascular event (MACCE) was predefined as composite of death, myocardial infarction, unplanned revascularisation, SCAD recurrence or stroke. All events were assigned by a Clinical Events Committee. RESULTS: After corelab evaluation, 389 patients were included. Most patients were women (88%); median age 53 years (IQR 47-60). Most patients presented as non-ST-segment-elevation myocardial infarction (54%). A type 2 intramural haematoma (IMH) was the most frequent angiographic pattern (61%). A conservative initial management was selected in 78% of patients. At a median time of follow-up of 29 months (IQR 17-38), 46 patients (13%) presented MACCE, mainly driven by reinfarctions (7.6%) and unplanned revascularisations (6.2%). Previous history of hypothyroidism (HR 3.79; p<0.001), proximal vessel involvement (HR 2.69; p=0.009), type 2 IMH (HR 2.12; p=0.037) and dual antiplatelet therapy (DAPT) at discharge (HR 2.18; p=0.042) were independent predictors of MACCE. CONCLUSIONS: In this large prospective cohort of patients with SCAD, prognosis was overall favourable, with events mainly driven by reinfarctions or unplanned revascularisations. History of hypothyroidism, proximal vessel involvement, type 2 IMH and DAPT at discharge were associated with MACCE. TRIAL REGISTRATION NUMBER: NCT03607981.

Procedural Results and One-Year Clinical Outcomes of Treatment of Bioresorbable Vascular Scaffolds Restenosis (from the RIBS VII Prospective Study).

Currently, both drug-eluting stents (DES) and drug-eluting balloons are recommended in patients with in-stent restenosis (ISR) of metallic stents. However, the clinical results of repeated interventions in patients with restenosis of bioresorbable vascular scaffolds (BVS) remain unsettled. We sought to assess the results of interventions in patients with BVS-ISR as compared with those obtained in patients with ISR of DES and bare-metal stents (BMS). Restenosis Intrastent: Treatment of Bioresorbable Vascular Scaffolds Restenosis (RIBS VII) is a prospective multicenter study (23 Spanish sites) that included 117 consecutive patients treated for BVS-ISR. Inclusion/exclusion criteria were similar to those of previous RIBS studies. Patients in the RIBS IV (DES-ISR, n = 309) and RIBS V (BMS - ISR, n = 189) randomized trials, were used as controls. Most patients with BVS-ISR were treated with DES (76%). Patients with BVS-ISR were younger, had larger vessels, and after interventions had higher in-segment residual diameter stenosis (19 ± 13%, 15 ± 11%, 15 ± 12%, p <0.001) than those treated for DES-ISR and BMS-ISR, respectively. At 1-year clinical follow-up (obtained in 100% of patients) target lesion revascularization (6%) was similar to that seen in patients with DES-ISR and BMS-ISR (8.7% and 3.7%, p = 0.32). Freedom from death, myocardial infarction, and target vessel revascularization (primary clinical end point) was 8.5%, also similar to that found in patients with DES-ISR and BMS-ISR (14.2% and 7.4%, p = 0.09). Results were also similar when only patients treated with DES in each group were compared and remained unchanged after adjusting for potential confounders in baseline characteristics. Time to BVS-ISR did not influence angiographic or clinical results. This study demonstrates the safety and efficacy of coronary interventions for patients presenting with BVS-ISR. One-year clinical results in these patients are comparable to those seen in patients with ISR of metallic stents (ClinicalTrials.gov ID:NCT03167424).

Safety of digoxin in nonagenarian patients with atrial fibrillation: lessons from the Spanish Multicenter Registry.

BACKGROUND: The association between digoxin and mortality is an unclear issue. In older patients with atrial fibrillation (AF), where use of digoxin is frequent, the evidence of its safety is scarce. Our aim is to assess the safety of digoxin in nonagenarian patients with AF. METHODS: We evaluated data from 795 nonagenarian patients with non-valvular AF from the Spanish Multicenter Registry. We analyzed the relationship between digoxin and all-cause mortality with the Cox proportional-hazards model. RESULTS: Follow-up was 27.7 ± 18.3 months. Mean age was 92.5 ± 3.8 years, and 71% of nonagenarian patients were female. Digoxin was not associated with increased risk of mortality [adjusted hazard ratio (aHR) = 1.16, 95% CI: 0.96-1.41,P = 0.130]. However, we found a significant increase in mortality in the subgroup with estimated glomerular filtration rate (eGFR) < 30 mL/min per 1.73 m 2 (aHR = 2.01, 95% CI: 1.13-3.57,P = 0.018), but not in the other subgroups of eGFR (30-59 mL/min per 1.73 m2 and ≥ 60 mL/min per 1.73 m2). When exploring the risk of mortality according to sex, male subgroup was associated with an increase in mortality (aHR = 1.48, 95% CI: 1.02-2.14,P = 0.041). This was not observed in females subgroup (aHR = 1.03, 95% CI: 0.81-1.29,P = 0.829). Based on the presence or absence of heart failure, we did not find significant differences (aHR = 1.20, 95% CI: 0.87-1.65,P = 0.268 vs. aHR = 1.15, 95% CI: 0.90-1.47,P = 0.273, respectively). CONCLUSIONS: In our large registry of nonagenarian patients with AF, we did not find an association between digoxin and mortality in the total sample. However, in the subgroup analyses, we found an increase in mortality with the use of digoxin in men and in patients with an eGFR < 30 mL/min per 1.73 m 2.

Long-term clinical, angiographic, and optical coherence tomography findings of Mg-based bioresorbable scaffold in patients with acute coronary syndrome.

BACKGROUND: This study sought to evaluate the clinical outcomes of patients treated with magnesium-based bioresorbable scaffolds (MgBRS) in the context of acute coronary syndromes (ACS) at long-term follow-up (24 months). The study also aims to investigate the MgBRS performance by angiography and the healing and bioresorption pattern by optical coherence tomography (OCT) at 18 months. METHODS: Between December 2016 and December 2018, a total of 90 patients admitted for ACS and treated with MgBRS (Magmaris, Biotronik AG, Bülach, Switzerland) were enrolled in a multicenter prospective study. Clinical follow-up was performed in all patients at 24 months and angiographic and OCT follow-up in 51.5% of patients at 18 months. Serial OCT was available in 33 patients (36.7%). RESULTS: At a 2-year follow-up, 88.8% were free of symptoms, no cardiac death was reported, and the device-oriented composite event (DOCE): consisting of cardiac death, target vessel myocardial infarction, and target lesion revascularization (TLR) was 13.3%. Stent thrombosis and TLR were observed in 2.2 and 11.1%, respectively. Binary restenosis was observed in 21.7% of cases and in-stent late lumen loss was 0.61 ± 0.75 mm. By serial OCT imaging, the minimal lumen area was significantly reduced greater than 40% (from 6.12 ± 1.59 to 3.5 ± 1.55 mm2, p < .001). At follow-up, area stenosis was 44.33 ± 23.07% and half of the patients presented indiscernible struts. The principal observed mechanism of restenosis was scaffold collapse. CONCLUSIONS: At long-term follow-up, MgBRS implantation in ACS patients showed a high rate of DOCE, mainly caused by clinically driven TLR. MgBRS restenosis was caused by scaffold collapse in most of the cases.

Incidence, predictors of bleeding and prognosis of bleeding in anticoagulated nonagenarian patients with atrial fibrillation.

BACKGROUND: The prevalence of atrial fibrillation (AF) increases with age. The prescription of anticoagulation in very elderly patients is controversial and sometimes underused. Our objective is to report the incidence and predictors of major bleeding in anticoagulated nonagenarian patients with non valvular atrial fibrillation (NVAF). METHODS: We analyzed a large multicentre registry of anticoagulated nonagenarian patients diagnosed with NVAF from three health areas of Spain, between 2013 and 2017. Predictors of major bleeding were studied with a competing risk analysis and the impact of major bleeding with a time-dependent mortality analysis. RESULTS: The incidence rate of major bleeding was 5 per100 person-year (95% Confidence Interval [CI]: 4.59-6.35), similar in the group of patients with vitamin K antagonists (VKAs) and direct oral anticoagulants (DOAC). In the VKAs group we found as predictors of major bleeding: previous admission for bleeding (sub-distribution hazard ratio [sHR] 3.25, 95% CI: 1.48-7.13), creatinine (sHR 1.38, 95% CI: 1.16-1.64,) and control out-of-range INR (sHR 1.90, 95% CI: 1.02-3.55). In DOAC group, male sex (sHR 1.92, 95% CI: 1.18-3.13) and the history of previous admission for bleeding (sHR 2.60, 95% CI 1.33-5.06) were found as a predictor. The HAS-BLED was not associated with major bleeding. Major bleeding was associated with increased mortality in both VKAs and DOAC groups without significant differences. CONCLUSIONS: We found an incidence rate of major bleeding with relative low values, similar in those treated with VKAs and those treated with DOAC, with different predictors of major bleeding in each group. Major bleeding was associated with increased mortality, with no significant difference by oral anticoagulation therapy (OAT).

Spontaneous coronary artery dissection in Spain: clinical and angiographic characteristics, management, and in-hospital events.

INTRODUCTION AND OBJECTIVES: Spontaneous coronary artery dissection (SCAD) is an uncommon cause of acute coronary syndrome. The characteristics and in-hospital clinical course of patients with SCAD in Spain remain unknown. METHODS: We present data from consecutive patients included in the national prospective SCAD registry. Angiographic analysis was performed in a centralized core laboratory. RESULTS: Between June 2015 and April 2019, we included 318 patients with SCAD (358 lesions) from 31 centers. Median age was 53 years, and 88% were women. The most frequent presentation was non-ST-segment elevation acute myocardial infarction (53%). The most frequently involved artery was the left anterior descending coronary artery (44%), predominantly affecting the distal segments (39%) and secondary branches (54%). Most lesions (62%) appeared on angiography as intramural hematoma, without double lumen. Conservative management was selected as the initial approach in most patients (78%). During the index admission, 6% of patients had a major adverse event and 4 patients (1.3%) died. Independent predictors of adverse events were initial management with percutaneous coronary intervention (OR, 5.97; P=.004) and angiographic presentation as intramural hematoma (OR, 4.96; P=.028). CONCLUSIONS: In Spain, SCAD affects mainly middle-aged women. In most patients, the initial management strategy was conservative with excellent in-hospital survival. Initial management with percutaneous coronary intervention and angiographic presentation as intramural hematoma were related to the presence of in-hospital adverse events. Registered at ClnicalTrials.gov (Identifier: NCT03607981).

Contrast FFR plus intracoronary injection of nitro-glycerine accurately predicts FFR for coronary stenosis functional assessment.

BACKGROUND: Fractional flow reserve (FFR) is the "gold standard" for assessing the physiological significance of coronary disease. In the last decade, several alternative adenosine-free indexes have been proposed in order to facilitate the dissemination of the functional evaluation of coronary stenosis. Our aim was to investigate whether radiographic contrast plus intracoronary nitroglycerin (cFFR-NTG) can predict functional assessment of coronary stenosis offering superior diagnostic agreement with FFR compared to non-hyperemic indexes and contrast mediated FFR (cFFR). METHODS: Three hundred twenty-nine lesions evaluated with pressure wire in 266 patients were prospectively included in this multicenter study. RESULTS: The ROC curves for cFFR-NTG using an FFR≤0.80 showed a higher accuracy in predicting FFR (AUC=0.97) than resting Pd/Pa (AUC=0.90, P<0.01) and cFFR (AUC=0.93.5, P<0.01). A significant (P<0.01) strong correlation was found between FFR and the four analyzed indexes: Pd/Pa (r=0.78); iFR/RFR (r=0.73); cFFR(r=0.89) and cFFR-NTG (r=0.93). cFFR-NTG showed the closest agreement at Bland-Altman analysis. The cFFR-NTG cut off value >0.84 showed the highest negative predictive value (88%), specificity (91%), sensitivity (94%) and accuracy (92%) of the studied indexes. CONCLUSIONS: Submaximal hyperemic adenosine-free indexes are an efficient alternative to adenosine for the physiological assessment of epicardial coronary disease. The most accurate index in predicting the functional significance of coronary stenosis using FFR as reference was cFFR-NTG.

Spontaneous coronary artery dissection and Takotsubo syndrome: comparison of baseline clinical and angiographic characteristics and in-hospital outcomes.

BACKGROUND: Spontaneous coronary artery dissection (SCAD) and Takotsubo syndrome (TTS) constitute two relatively common nonatherosclerotic causes of acute coronary syndrome particularly frequent in women. METHODS: This study sought to compare the baseline clinical and angiographic characteristics and in-hospital outcomes of patients from two large prospective registries on SCAD and TTS (the prospective nation-wide Spanish SCAD Registry and a prospective single-center TTS registry). RESULTS: A total of 318 SCAD and 106 TTS consecutive patients were included. Most patients in both groups (88%) were women. Patients in the TTS group were older [74 (interquartile range, IQR 67-81) vs. 53 years-old (IQR 47-60), P < 0.001] and presented a higher prevalence of cardiovascular risk factors. Precipitating triggers were more frequent in TTS (56% vs. 42%, P = 0.009) but emotional stress was more common in the SCAD group (25% vs. 15%, P = 0.037). TTS patients showed a reduced release of cardiac biomarkers but had more severe left ventricular dysfunction (ejection fraction <50%: 73% vs. 12%, P < 0.001). In-hospital major adverse cardiovascular events occurred more frequently in TTS patients (12% vs. 4.7%, P < 0.001). Notably, TTS patients showed more frequently congestive heart failure (10% vs. 0.6%, P < 0.001), atrial fibrillation (11% vs. 1%, P < 0.001) and had a higher all-cause in-hospital mortality (5.7% vs. 1.3%, P = 0.032). CONCLUSION: TTS patients are older and present a higher prevalence of some cardiovascular risk factors than patients with SCAD. TTS is linked to a worse in-hospital prognosis with higher mortality.

Coronary artery aneurysms, insights from the international coronary artery aneurysm registry (CAAR).

BACKGROUND: Coronary Aneurysms are a focal dilatation of an artery segment >1.5-fold the normal size of adjacent segments. Although some series have suggested a prevalence of 0.3-12%, data are lacking. In addition, they are not mentioned in practice guidelines. Our aim was investigate its prevalence, management and long-term outcomes. METHODS AND RESULTS: The coronary artery aneurysm registry (CAAR) involved 32 hospitals across 9 countries in America and Europe. We reviewed 436,467 consecutive angiograms performed over the period 2004-2016. Finally, 1565 patients were recruited. Aneurysm global prevalence was 0.35%. Most patients were male (78.5%) with a mean age of 65 years and frequent cardiovascular risk factors. The main indication for angiogram was an acute coronary syndrome, 966 cases. The number of aneurisms was ≤2 per patient in 95.8% of the cases, mostly saccular, most frequently found in the left anterior descending and with numbers proportional with coronary stenosis. Aortopathies were related with more aneurysms too. Most patients received any revascularization procedure (69%), commonly percutaneous (53%). After a median follow-up of 37.2 months, 485 suffered a combined event (MACE) and 240 died. Without major differences comparing CABG vs PCI, MACE and death were more frequent in patients who received bare metal stents. CONCLUSIONS: Coronary artery aneurysms are not uncommon. Usually, they are associated with coronary stenosis and high cardiovascular risk. Antiplatelet therapy seems reasonable and a percutaneous approach is safe and effective.

Vitamin K Antagonists and Direct Oral Anticoagulants in Nonagenarian Patients With Atrial Fibrillation.

OBJECTIVES: Nonagenarian patients are underrepresented in clinical trials that have evaluated oral anticoagulation in patients with atrial fibrillation (AF). The aim of this study was to assess the pronostic impact of oral anticoagulation in patients with AF age ≥90 years. DESIGN: Retrospective multicenter study of nonagenarian patients with AF. SETTING AND PARTICIPANTS: A total of 1750 nonagenarian inpatients and outpatients with nonvalvular AF between January 2013 and December 2018 in 3 Spanish health areas were studied. METHODS: Patients were divided into 3 groups based on antithrombotic therapy: nonoral anticoagulants (30.5%), vitamin-K antagonists (VKAs; 28.6%), and direct oral anticoagulants (DOACs; 40.9%). During a mean follow-up of 23.6 ± 6.6 months, efficacy outcomes (death and embolic events) were evaluated using a Cox regression analysis and safety outcomes (bleeding requiring hospitalization) by competing-risk regression. Results were complemented with a propensity score matching analysis. RESULTS: During follow-up, 988 patients died (56.5%), 180 had embolic events (10.3%), and 186 had major bleeding (10.6%). After multivariable adjustment, DOACs were associated with a lower risk of death and embolic events than nonanticoagulation [hazard ratio (HR) 0.75, 95% confidence interval (CI)] 0.61‒0.92), but VKAs were not (HR 0.87, 95% CI 0.72‒1.05). These results were confirmed after propensity score matching analysis. For bleeding, both DOACs and VKAs proved to be associated with a higher risk (HR for DOAC 1.43; 95% CI 0.97‒2.13; HR for VKA 1.94; 95% CI 1.31‒2.88), although findings for DOACs were not statistically significant (P = .074). For intracranial hemorrhage (ICH), only VKAs-not DOACs-presented a higher risk of ICH (HR 4.43; 95% CI 1.48‒13.31). CONCLUSIONS AND IMPLICATIONS: In nonagenarian patients with AF, DOACs led to a reduction in mortality and embolic events in comparison with nonanticoagulation. This reduction was not observed with VKAs. Although both DOACs and VKAs increased the risk of bleeding, only VKAs were associated with higher ICH rates.