Fatal adverse drug reactions: A worldwide perspective in the World Health Organization pharmacovigilance database.

AIMS: Adverse drug reactions (ADRs) are important causes of death. However, the main involved drugs are relatively unknown. The present study was performed to characterise death-related drugs recorded in a large pharmacovigilance database during the last 10 years. METHODS: A retrospective analysis of VigiBase, the World Health Organization pharmacovigilance database, was performed investigating fatal ADRs registered between 1 January 2010 and 31 December 2019 in male and female patients aged ≥18 years and reported by physicians. Analyses were descriptive investigating age, sex and suspected drugs. Differences in reporting according to sex, age and continents were investigated using disproportionality analysis with calculation of reporting odds ratio and its 95% confidence interval. RESULTS: Among the 23 millions ADRs recorded in VigiBase, 3 250 967 were included with 43 685 fatal. They were reported mainly in patients older than 75 years. The 3 most frequently involved drug classes were antineoplastic/immunomodulating drugs followed by nervous system and cardiac drugs. The top 3 individual drugs were denosumab, lenalidomide and thalidomide with marked differences according to age, sex, continents and countries. The risk of reporting fatal ADRs was higher in males, in the Americas and in patients ≥65 years. CONCLUSION: Fatal ADRs registered in a large pharmacovigilance database during the last 10 years correspond to just over 1% of the total number of ADRs. They occurred more in males, after 65 years and with antineoplastic/immunomodulating drugs in general. Our study also highlighted, for the first time, important differences in fatal ADRs between continents and countries.

Risk comparison of beta-lactam-induced anaphylaxis: Therapeutic stratification analysis in a Vietnamese pharmacovigilance database.

WHAT IS KNOWN AND OBJECTIVE: There is limited data on the specific risks of anaphylaxis induced by beta-lactam drugs. The aim of this study was to compare the risks of reporting beta-lactam-induced anaphylaxis using the national pharmacovigilance database of Vietnam (NPDV). METHODS: The multivariate generalised linear regression model was applied for signal generation and comparison of beta-lactams. RESULTS: Between 2010 and 2016, there were 2,921 reports of anaphylaxis (19.93%) from 14,655 spontaneous reports of beta-lactam use in the NDPV. Anaphylaxis signal generation was also found for the subgroup J01D (cephalosporins and carbapenems) (ROR = 1.27 [1.16-1.39]) and beta-lactamase-sensitive penicillins (ROR = 1.74 [1.27-2.35]). In the third generation cephalosporin subgroup, different risks were identified for the following combinations of beta-lactams: 1) cefotaxime with cefoperazone+sulbactam; 2) cefixime/cefpodoxime/cefdinir with cefoperazone+sulbactam or ceftizoxime/cefoperazone/ceftazidime/ceftriaxone/cefotaxime. For the second generation cephalosporin subgroup, different risks were found for cefotiam compared to cefmetazole, cefaclor, cefamandole and cefuroxime. WHAT IS NEW AND CONCLUSION: These findings identified and highlighted the different anaphylactic risks caused by various beta-lactams in the main subgroups.

Abatacept in rheumatoid arthritis and the risk of cancer: a world observational post-marketing study.

OBJECTIVES: We aimed to investigate whether abatacept used in patients for RA was associated with an increased risk of reporting overall cancer and specific cancers, including breast, lung, lymphoma, melanoma and non-melanoma skin cancer when compared with other biologic DMARDs (bDMARDs). METHODS: We performed an observational study within VigiBase, the World Health Organization's global database of individual case safety reports, from 2007 to 2017 to compare the cases of cancer reported in RA patients exposed to abatacept with those reported in RA patients exposed to other bDMARDs. We conducted disproportionality analyses allowing the estimation of reporting odds ratios (RORs) with 95% CIs of the exposure odds among spontaneous reporting of cancers to the exposure odds among other reported adverse effects. RESULTS: We identified 15 846 adverse effects reported in RA patients who received abatacept and 290 568 adverse effects reported in RA patients treated with other bDMARDs. Compared with other bDMARDs, the use of abatacept was not associated with an increased risk of reporting cancer overall [ROR 0.98 (95% CI 0.91, 1.05)]. Analyses by specific cancer sites showed a significantly increased ROR for melanoma [1.58 (95% CI 1.17, 2.08)], but not for other specific cancer sites. CONCLUSION: Compared with other bDMARDs, exposure to abatacept in RA patients was only significantly associated with an increased risk of reporting melanoma. This increased risk is consistent with the properties of abatacept (CTLA-4 agonist) since it has an opposite action than ipilimumab, an antibody that blocks CTLA-4 and is approved for the treatment of malignant melanoma. TRIAL REGISTRATION: ClinicalTrials.gov (http://clinicaltrials.gov), NCT03980639.

Paradoxical adverse drug reactions: descriptive analysis of French reports.

INTRODUCTION: Paradoxical adverse drug reactions (ADRs) are defined as being opposing reactions to the pharmacological effect of drugs in relation to its pharmacodynamic properties. Their diagnosis is difficult as they are relatively rare with atypical clinical presentation (with the possibility of being confused with drug ineffectiveness or the worsening of the underlying disease). This kind of ADR may be particularly subject to under-notification. The aim of the present study is to describe paradoxical ADRs using the French PharmacoVigilance DataBase (FPVDB). METHOD: We analysed all reports recorded in the FPVDB with drugs defined as "suspect" and which included the term "paradoxical reaction" (PR) (according to MedDRA classification) from 01/01/1984 to 12/31/2018. The drugs were classified according to the Chemical Therapeutic Anatomical Classification (ATC). RESULTS: We found 57 reports of PR, with half of them recorded between 2015 and 2018. The median age of patients was 46 years, mainly male (54%). The most frequently involved drugs were immunomodulating agents (n = 28, 49%) and psychotropics (n = 28, 49%). The leading paradoxical ADRs were psychiatric (anxiety, sleep and behavioural disorders) and skin-related. In 19 cases (33%), PR was related to benzodiazepines mainly occurring in patients in extreme ages (five cases in children and patients > 70, respectively, 53%). For psychotropic-induced PR (n = 28), known contributory factors (alcohol consumption, underlying psychiatric diseases) were found in 18 cases (64%). Paradoxical reactions with immunomodulating agents were mainly related to skin ADRs (n = 25). For psychotropics, paradoxical ADRs occurred rapidly after a mean delay of 1 day, predominantly following high doses. We also identified several "unexpected" paradoxical reactions, such as cognitive degradation with donepezil, or a return to impulsive smoking addiction with varenicline. CONCLUSION: This study highlights that pharmacovigilance databases like the French database make it possible to investigate the main characteristics of paradoxical reactions to drugs. This ADR was mainly found in the FPVDB with psychotropic drugs and immunomodulating agents. Moreover, pharmacovigilance databases enable the identification of some signs of "unexpected paradoxical reactions". In order to identify this type of ADR more effectively, work on awareness and harmonization is required to register these reports. The addition of the term "paradoxical reaction to the drug" to the list of other symptoms would facilitate their identification and analysis.

Muscle Damage Due to Fusidic Acid-Statin Interaction: Review of 75 Cases From the French Pharmacovigilance Database and Literature Reports.

BACKGROUND/AREA OF UNCERTAINTY: Statins, which reduce cardiovascular risk in both primary and secondary prevention, are one of the most widely prescribed therapeutic classes in the world. Usually well-tolerated, statin-associated muscle symptoms are a well-known adverse effect. Fusidic acid (FA) is a bacteriostatic antibiotic of interest in the treatment of methicillin-resistant Staphylococcus aureus infections. Cases of rhabdomyolysis, sometimes fatal, have been reported after coprescription of FA and a statin. DATA SOURCES/AREA OF UNCERTAINTY: We studied 75 cases of muscle damage related to interaction between FA and a statin reported in the French national pharmacovigilance database (43 cases) and from a literature review (32 cases). RESULTS: Cases were mostly men (72.5%), often overweight (mean body mass index: 29.4). The most commonly reported statins were atorvastatin (60%), simvastatin (22.7%), and rosuvastatin (8.0%). Muscle disorders appeared on average 30 days after initiation of FA. Symptoms were muscle weakness (82%), dark urine (71%), and myalgia (61%). Mean creatine kinase level at diagnosis was 43,890 UI/mL, and acute renal injury occurred more than half of the cases. Outcome was fatal in 22% of cases and 28% kept sequelae at the end of the follow-up (54 days). CONCLUSIONS: Muscle damage induced by interaction between FA and statin is a potentially life-threatening complication, leading to contraindication of this association in France. This is to be reminded especially because FA is about to get FDA approval and should soon be available in the United States.

Drug-induced osteoporosis/osteomalacia: analysis in the French and Spanish pharmacovigilance databases.

INTRODUCTION: Osteomalacia and osteoporosis are two metabolic bone disorders that increase the risk of fracture due to several causes. In terms of drugs, apart from corticosteroids, which are known to induce bone disorders, several other drugs used in chronic disease management have also been linked with an increased risk of osteoporosis and osteomalacia. PURPOSE: The aim of this study was to describe spontaneous reports of drug-induced osteoporosis and osteomalacia in the French (FPVDB) and Spanish (SPVDB) pharmacovigilance databases. METHODS: Data were provided by the FPVDB and SPVDB. All reports of osteoporosis and osteomalacia recorded from 1985 up to 31 December 2015 inclusive were selected. Taking the time to onset of bone loss into account, all cases occurring in less than 1 month were excluded. RESULTS: A total of 369 reports (44 cases of osteomalacia, 325 cases of osteoporosis) were registered in the FPVDB and 64 (22 cases of osteomalacia, 42 cases of osteoporosis) in the SPVDB. In France, the top 5 drugs involved in the onset of osteoporosis were corticosteroids accounting for approximately half of the reports (n = 170) followed by systemic antiviral (n = 87), antacid (n = 29), antiepileptic (n = 27) and antithrombotic (n = 24) drugs. The 2 main classes of drugs implicated in osteomalacia were systemic antiretroviral drugs for half of the reports (n = 21) and antiepileptic drugs (n = 15). In Spain, corticosteroids were involved in 35.7% of reported cases of osteoporosis (n = 15) followed by systemic antiviral drugs (n = 12). There was no spontaneous report for antacid drugs. For osteomalacia, the 2 main drug classes were systemic antiretroviral drugs (n = 18, 81.8%) followed by antiepileptics (n = 2, 9.0%). In both countries, concomitant administration of systemic corticosteroids with other suspected drugs did not significantly modify the time to onset of drug-induced osteoporosis. CONCLUSION: Despite some differences between the French and Spanish PVDBs, our data consistently show that bone loss is not only restricted to glucocorticoids but also involves antivirals, antiepileptic drugs, antacid drugs or antidepressants. Further analysis might prove useful in exploring the characteristics of drug-induced bone loss on a larger scale.

Drug-induced Stevens-Johnson syndrome and toxic epidermal necrolysis in vietnamese spontaneous adverse drug reaction database: A subgroup approach to disproportionality analysis.

WHAT IS KNOWN AND OBJECTIVE: Despite the numerous studies investigating drug-induced Stevens-Johnson syndrome/toxic epidermal necrolysis (SJS/TEN), the understanding and quantitative data in developing countries remain limited. The study aimed to describe and quantify the drug-related risk of SJS/TEN in a resource-limited context using the Vietnamese spontaneous reporting database (VSRD) of adverse drug reactions. METHODS: Spontaneous reports relating to medium- and late-onset severe cutaneous adverse reactions (MLOSCAR) and SJS/TEN recorded in the VSRD from 2010 to 2015 were retrospectively analysed. The demographic characteristics and drug information were described and compared between SJS/TEN and other MLOSCAR reports. The drug-induced SJS/TEN signals were estimated using subgrouped disproportionality analysis with calculation of the reporting odds ratio (ROR) and the respective 95% confidence interval (CI). RESULTS: The VSRD received 2,849 MLOSCAR reports, 136 of which focus on SJS/TEN over a 6-year period. About 60% of SJS/TEN patients were male, and the majority of them were adults (mean age 42.5 ± 22.9). Up to 91.8% of drugs induced SJS/TEN within 1-28 days, and 45% SJS/TEN cases were evaluated as life-threatening. Positive signals were generated with carbamazepine (n = 25, ROR [95% CI] = 11.99 [7.07-19.92]), allopurinol (n = 15, ROR [95% CI] = 4.2 [2.20-7.59]), traditional/herbal medicines (n = 7, ROR [95% CI] = 2.76 [1.12-5.86]), colchicine (n = 4, ROR [95% CI] = 6.22 [1.69-18.72]), valproic acid (n = 3, ROR [95% CI] = 8.71 [1.89-30.19]) and meloxicam (n = 3, ROR [95% CI] = 7.09 [1.55-24.29]), which are well known for SJS/TEN. Cefixime (n = 5, ROR [95% CI] = 3.34 [1.13-8.00]) and paracetamol (n = 22, ROR [95% CI] = 5.23 [3.10-8.49]) also generated positive signals despite their popularity in Vietnam. WHAT IS NEW AND CONCLUSION: This first Vietnamese population-based study has highlighted original characteristics and signals of drug-induced SJS/TEN, which are relatively consistent with other worldwide data and typical for a developing country.

Patterns of direct oral anticoagulant drug prescription in France in 2010-2013: a study in the Midi-Pyrénées area.

AIM: The aim of our study was to study the pattern of prescription of direct-acting oral anticoagulants (DOACs) according to the French recommendations. METHODS: We performed a cross-sectional study using anonymous data of patients covered by the French National Health Insurance information system (SNIIRAM) from 1 January 2010 to 31 December 2013 in the area of Midi-Pyrénées (southwest of France). RESULTS: Of the 355,608 patients identified, 325,216 (91.5%) were included, of whom 22,142 received at least one DOAC. About 39.1% (8,652 patients) had DOAC in an orthopedic indication, 46.5% (10,303 patients) in a cardiac indication, and 16.1% (3568 patients) in an indeterminate indication. Overall, guidelines were largely followed as for renal function monitoring, prescribing in orthopedic indications, in cardiac indications in patients aged 80 years and older, and in the case of concomitant use of verapamil. However, inappropriate prescriptions were observed for cardiac indications, and for dosage adjustments in orthopedic indications, with respect to both the age of patients (75 years and older) and those taking verapamil or amiodarone concomitantly. Guidelines were more followed in women and patients aged 80 or more. CONCLUSIONS: Among patients receiving DOACs, 58% were exposed to a prescription falling outside the guidelines. This study on DOAC prescription patterns revealed insufficiencies in the compliance with the French guidelines in certain indications.

Does substitution of brand name medications by generics differ between pharmacotherapeutic classes? A population-based cohort study in France.

PURPOSE: The purpose of this study was to measure the rate of substitution failure to generic antiepileptic drugs (AEDs) compared to two other pharmacotherapeutic classes (neuroleptics, beta-blockers). METHODS: We conducted a cohort study involving beneficiaries of the French health insurance system from January 2009 to November 2012. Substitution failure to generic drugs was estimated by the rate of switchback (i.e. from generic drug back to its branded drug). We selected the patients who had a dispensation of a branded AED for 60 days or more during the 90 days preceding the generic substitution. Cox proportional hazard regression was used to model time to switchback for antiepileptics vs. other therapeutic classes in the 90 days after generic substitution, adjusting for age, gender and polytherapy. RESULTS: The cohort included 6727 patients of whom 1947 were exposed to AEDs, 2398 to neuroleptics and 2382 to beta-blockers. The switchback rate was 62% for AEDs. AED users were more likely to switch back as compared to beta-blocker (crude hazard ratio 1.87; 95% CI 1.68-2.07 for patients under 75) or neuroleptic users. The same observation was made in patients above 75 years (crude hazard ratio 1.36; 95% CI 1.16-1.60). CONCLUSIONS: Compared to beta-blocker users, AED users were more likely to switch back to the branded drug, whereas this difference was not observed with neuroleptics. These results could reflect a poor acceptance of switching AEDs to generic compounds in France.

[Improving adverse drug reaction reporting by general practitioners through clinical research assistants visits]. / Apport de la visite d'assistants de recherche clinique aux cabinets de médecins généralistes sur la notification des effets indésirables médicamenteux.

The main limitation of adverse drug reactions (ADRs) reporting, particularly by general practitioners (GP), to the regional pharmacovigilance centers is under-reporting. The Midi-Pyrénées Regional Pharmacovigilance Center (South western, France) sets up regular visits by a clinical research assistant (CRA) to GP in order to increase the number of ADR reports. The aim of this pilot study was to assess the effect of regular visits of a CRA in GPs offices on the rate of ADR reporting. After one year, CRA visits permit a two-fold increase in ADR reporting.

Atropinic (Anticholinergic) Burden in Parkinson's Disease.

Use of atropinic drugs remains controversial in Parkinson's disease (PD) because there is insufficient evidence about their efficacy and they can induce serious adverse drug reactions. Atropinic risk scales were developed to help to identify atropinic drugs in prescription forms and to evaluate their burden in clinical practice. In the present review, we discuss the few studies investigating atropinic burden in PD and present the results of our study indicating that atropinic drugs are still widely prescribed in PD (almost 3 of 5 prescriptions) with a clinically significant atropinic burden in around 1 of 6 PD patients. Drugs mainly responsible for high values of atropinic burden were those used for nonmotor symptoms. Clinically significant atropinic burdens were mainly induced by associations of several "low-risk" drugs. Physicians must be aware that in addition to classical atropinic antiparkinsonian drugs, many others (psychotropics) can contribute to increased atropinic burden in PD patients. © 2016 International Parkinson and Movement Disorder Society.

Hospital re-admission associated with adverse drug reactions in patients over the age of 65 years.

CONTEXT: Adverse drug reactions (ADRs) are responsible for 5 % of hospital admissions, but hospital re-admission induced by ADRs remains poorly documented. OBJECTIVE: The aim of this study was to estimate the rate of hospital re-admission and the factors associated with re-admission in the patients over the age of 65 years. Secondary, we described the characteristics of cases of ADRs leading to re-admission for drugs other than chemotherapy agents. METHODS: Data were extracted from hospital discharge summaries provided by the Department of Medical Information of Toulouse University Hospital. All patients over the age of 65 years admitted to the hospital in 2010 for an ADR, identified from ICD-10 codes, were selected. All subsequent admissions of members of this cohort within 1 year of discharge following the index admission were reviewed retrospectively. The risk factors associated with hospital re-admission for ADRs were analyzed. Medical records were used for descriptive analysis of re-admission due to drugs other than chemotherapy agents. RESULTS: We found that 553 of the 1000 patients admitted for ADRs in 2010 were re-admitted to hospital within 1 year. Among them, 87 cases were re-admitted for ADRs (estimated rate of 87/1000 re-admission for an ADR within 1 year). A comparison of the patients re-admitted for ADRs (n = 87) with those of patients re-admitted for other causes (n = 410) suggested that only cancer increased the risk of re-admission for ADRs (OR = 7.69 [4.59-12.88] 95 % CI). ADRs due to the same drug combination were the suspected cause of repeat admission in half the cases (other than chemotherapy). Hospital re-admission was considered avoidable in four cases (22 %). CONCLUSION: This study shows an estimated rate of re-admission for an ADR around 87/1000 within 1 year, and the same drug combination were the suspected cause of repeat admission in half the cases. At least, 11 % of cases were avoidable.

Cyberpharmacovigilance: What is the usefulness of the social networks in pharmacovigilance?

Internet is changing the way people learn about health and illness. Health websites are among the most popular resources on the web. The sharing of patient experiences on a website could be an interesting source of information about adverse drug reactions (ADRs) or could generate "signals". Three examples of the use of patient forums in pharmacovigilance are discussed: (1) analysis of perceptions of risk by patients before and after withdrawal of benfluorex, and then following media coverage; (2) comparison of oral antineoplasic-induced ADRs collected on patient websites with those registered in the French pharmacovigilance database; (3) analysis of data found in the social media about drugs and pregnancy.

[Proton pump inhibitors: Real indication or trivialized prescription?] / Les inhibiteurs de pompe à protons : vraie indication ou prescription banalisée ?

PURPOSE: Protons Pump Inhibitors (PPIs) are widely used in France. The aim of our study was to determine the rate of exposure to PPIs in old patients and to assess the appropriateness of their prescription according to French guidelines. METHODS: We performed a descriptive study from 1st June to 30th August 2016, including all patients admitted in the department of geriatric post emergency of the university hospital of Toulouse with a prescription of PPI. Data concerning age, name of PPI, dose, indication, duration of prescription and the modification of PPI prescription were collected. RESULTS: Among 375 patients admitted during this period, 134 (35.7) were exposed to PPIs with a mean age of 85.9±6.6years. About one third of them were exposed to PPI more than one year (N=49). Prescription was inappropriate for the criteria "dose and indication" and "duration" for respectively 59 (44.0%) and 15 (11.2%) cases. In 50 cases (37.3%), the PPI was renewed, stopped in 69 cases (51.5%) and the dose was reduced for 15 patients (11.2%). CONCLUSION: According to our data, the prescription of PPI was no appropriate in about 40% of included patients. A reassessment of their prescription with awareness of patients should be necessary to improve the good utilization of these "popular" drugs and to prevent some serious adverse reactions after long exposure.

Which pharmacovigilance for vaccines?

Vaccines are drugs. Like all medicines, they are submitted to pre-clinical tests and then clinical trials. These tests and trials are essential but unfortunately insufficient because, for example, they are necessarily too short, they include too few subjects not representative of future treated patients... These points underline, once again, the importance for both patients and their doctors of the pharmacovigilance after marketing authorization. Phar- macovigilance of vaccines involves first, like all other medicines, the reporting of adverse drug reactions to the regional pharmacovigilance centers. For vaccines, it can be first spontaneous (i.e. not solicited or passive) and sought (encouraged active, as did for the HIN vaccine) involving not only health professionals but also the whole population (patients, relatives... ). Spontaneous reporting remains the only methodfor an early warning detection. The pharmacoepidemiological methods (case-control, cohort studies, expected- observed method, Self Controlled Cases-Series...) are secondarily used to confirm or deny a signal suggested by spontaneous notifications. More than elsewhere, these studies are also needed to quantify the populational risk. Pharmacovigilance is essential for a modern, clinical and medical evaluation of the benefit of vaccine. Knowledge of vaccine pharmaco- vigilance data also allows to recall the excellent benefits harms balance of these drugs, much higher than that of many other drug classes.

Atropinic burden of prescriptions forms in patients with Alzheimer disease: a cross-sectional study in a French PharmacoVigilance Database.

AIM: Atropinic drugs in patients with Alzheimer disease (AD) can decrease the effects of anticholinesterase drugs and/or induce adverse drug reactions (ADRs). Several atropinic risk scales defining an atropinic burden of drugs were proposed but were little used in AD patients. METHODS: All ADRs' notifications of AD patients registered in the Midi-Pyrénées PharmacoVigilance Database between 1999 and 2013 were analyzed using Anticholinergic Drug Scale (ADS) and Anticholinergic Duran's list. The primary objective was to quantify atropinic burden in AD patients and the secondary one to investigate associated factors. RESULTS: Among the 475 notifications, at least one atropinic drug was found in 282 notifications (59.4%) according to ADS and 214 (45.1%) according to Duran. Mean number of atropinics per notifications was 0.9 ± 0.9 (ADS) and 0.7 ± 0.9 (Duran). Mean atropinic burden per notifications was 1.2 ± 1.5 (ADS) and 0.9 ± 1.3 (Duran). Atropinic burden ≥ 3 was found in 87 notifications (18.2%) according to ADS and 50 (10.5%) according to Duran. There was no association between atropinic burden and age of patients. The number of drugs is associated to a high atropinic burden. CONCLUSION: The present work found an association between an atropinic drug and an anticholinesterase agent in around 1 out of 2 AD patients and a clinically significant atropinic burden (≥ 3) in around 1 to 2 AD patients out of 10. The benefit harm balance of atropinic drugs must be discussed before each prescription in AD patients.

Drugs and Sleep Apneas? A review of the French Pharmacovigilance database.

The present work reviews the case reports of drug-induced sleep apnea recorded in the French pharmacovigilance database. Notifications are very rare (around 1/100 000 notifications). This paper shows that sleep apnea can be aggravated or revealed by some drugs. Main drugs involved were psychotropics (benzodiazepines, neuroleptics) and opioids.

[Factors Associated with Direct Oral Anticoagulants versus Vitamin K Antagonists in Patients with Non-valvular Atrial Fibrillation]. / Facteurs associés à la prescription d'anticoagulants oraux directs versus anti-vitamines K dans la fibrillation atriale non valvulaire.

OBJECTIVE: Describing the factors associated with direct oral anticoagulants (DOA) prescription in patients with atrial fibrillation (AF). METHOD: This study was performed in Toulouse on a cohort of patients received in rhythmology consultation, treated with vitamin K antagonists (VKA) or DOA for AF. A multivariate model was performed using logistic regression to describe the factors associated with DOA prescription and secondly, those associated with discontinuation of the anticoagulant. RESULTS: Among the 140 patients included, 96 (66%) were treated with VKA and 48 (34%) with DOA. Recent AF diagnosis (OR 7.52, 95% CI [2.41;23.29], p = 0.001), previous exposure to VKA (OR 17.11, 95% CI [4.48;60.91], p<0.001), and no current exposure to anti-platelet agents (APA) (OR 7.69, 95% CI [1.22; 50.00], p = 0.030) were associated to DOA prescription. Discontinuation of the anticoagulant (n=24) was associated to DOA intake (OR 2.71, 95% CI [1.21; 6.08], p = 0.016). DISCUSSION: DOA are less prescribed than VKA in patients treated with APA. DOA switch to VKA was not systematic in patients diagnosed for a long time. However, international normalized ratio (INR) values were stable in most of patients treated with VKA at the switching to DOA. A more powerful study would confirm the factors associated with DOA prescription.