Letteratura scientifica selezionata sul tema "Pharmacoépidémiologie – France"

Background:Golimumab (GLM) is the latest anti-TNFα to be indicated for treating rheumatoid arthritis (RA), psoriatic arthritis (PsA) and axial spondyloarthritis (axSpA). The GO-PRACTICE study was performed in France at the request of the French Health Authorities, for the reevaluation of GLM in real-life.Objectives:The primary objective was to estimate GLM persistence at 2 years from initial prescription. This abstract focuses on a post-hoc analysis of the factors linked to GLM discontinuation in axSpA patients.Methods:Observational, prospective, multicenter study, that consecutively recruited adult patients with RA, PsA and axSpA who were newly prescribed GLM. Patients were followed-up for 2 years and outcomes data were collected at baseline (BL), 1 and 2 years. Patients’ sociodemographic characteristics, disease history, comorbidities and treatment history were also collected at BL. Persistence was estimated with the Kaplan-Meier method. Cox proportional hazard models were used to assess factors associated with persistence. Selected BL characteristics were studied in univariate models, where those associated withp-value <0.20 were included in multivariate analysis. Significance level was set atp<0.05.Results:478 patients with axSpA were included from Jan 2015 to Mar 2016. Mean age was 43 years and 55% were female; 61% of patients were biologic-naïve (BN, n=291) and 39% (n=187) were biologic-pretreated (BP). Median time-elapsed in years since axSpA diagnosis was 1.7 (range 0–45.1) and 6.9 (range 0.2–51.8) in BN and BP patients, respectively (P<0.001); 97% patients were prescribed 50 mg GLM monthly and co-treatments included DMARD (34%), corticosteroids (17%) and NSAIDs/analgesics (90%).Cumulative persistence probability of GLM at 2-years was 52.6% (Fig 1). Table 1 details the binary variables associated with GLM discontinuation atp<0.20. Among continuous variables, BL CRP level was associated withp<0.20. A multivariate analysis of these factors revealed that being female (HR 1.92, 95%CI 1.43–2.56,P<0.001) and being BP (HR 1.45, 95%CI (1.11–1.90),P=0.007) were risk factors for GLM discontinuation (Table 1).Table 1.Logistic model results for variables of interest and their link to GLM discontinuation in axSpAFactorModalitiesχ2(p)Hazard ratio (HR)95% CIHR following univariate analysis (p>0.20)AgeContinuous variable0.5201.000.99–1.02Disease duration0.4011.010.99–1.03Inflammatory bowel diseaseYes vs. No0.2770.740.43–1.28Gastrointestinal disease0.3441.270.78–2.06Uveitis0.2370.800.55–1.16Psoriasis0.2380.920.64–1.31 HR following multivariate analysis (variables with p<0.20 at univariate analysis)GenderFemale vs. Male< 0.0011.921.43–2.56Biologics historyPretreated vs. naïve0.0071.451.11–1.90Serum CRPContinuous variable0.1770.990.98–1.00DMARD historyYes vs. No0.0621.370.99–1.90Ongoing corticosteroids0.6931.080.73–1.61Anemia0.1701.820.78–4.24Kidney Disease0.5081.500.45–4.97Other physical illness0.4351.280.69–2.34Conclusion:2-year GLM persistence in axSpA patients was 52.6%. Females and those who were biologics-pretreated were at greater risk for discontinuing GLM before 2 years.Disclosure of Interests:Philippe Bertin Consultant of: MSD France, Philippe Goupille Grant/research support from: AbbVie, Amgen, Biogen, BMS, Celgene, Chugai, Lilly, Janssen, Medac, MSD France, Nordic Pharma, Novartis, Pfizer, Sanofi and UCB, Consultant of: AbbVie, Amgen, Biogen, BMS, Celgene, Chugai, Lilly, Janssen, Medac, MSD France, Nordic Pharma, Novartis, Pfizer, Sanofi and UCB, Speakers bureau: AbbVie, Amgen, Biogen, BMS, Celgene, Chugai, Lilly, Janssen, Medac, MSD France, Nordic Pharma, Novartis, Pfizer, Sanofi and UCB, Florence Tubach Grant/research support from: Florence TUBACH is head of the Centre de Pharmacoépidémiologie (Cephepi) of the Assistance Publique – Hôpitaux de Paris and of the Clinical Research Unit of Pitié-Salpêtrière hospital, both these structures have received research funding, grants and fees for consultant activities from a large number of pharmaceutical companies, that have contributed indiscriminately to the salaries of its employees. Florence Tubach didn’t receive any personal remuneration from these companies., Eric Lespessailles Consultant of: Amgen, Celgene, Lilly, MSD France, Novartis, UCB, Speakers bureau: Amgen, Celgene, Lilly, MSD France, Novartis, UCB, Naoual HARID Employee of: MSD France, Saannya Sequeira Consultant of: MSD France, Jean-Marie Fayette Consultant of: MSD France, Bruno Fautrel Grant/research support from: AbbVie, Lilly, MSD, Pfizer, Consultant of: AbbVie, Biogen, BMS, Boehringer Ingelheim, Celgene, Lilly, Janssen, Medac MSD France, Nordic Pharma, Novartis, Pfizer, Roche, Sanofi Aventis, SOBI and UCB, René-Marc Flipo Consultant of: Johnson and Johnson, MSD France, Novartis, Sanofi, Speakers bureau: Johnson and Johnson, MSD France, Novartis, Sanofi

Background:Objectives:To estimate the incidence rate of malignancies in biologic-treated RA patients, and to compared it to the general population and across different classes of biologicsMethods:We conducted an historical cohort study within the French the national claim database, named SNDS. This database prospectively records individual health resource of 86% of the entire French population (65 million inhabitants) since 2007. RA adult patients were identified based on ICD-10 code (M05 or M06). Patients with cancer history were excluded. Treatment exposures focused on incident first use of biologics including all anti-TNF, rituximab, abatacept, tocilizumab, ustekinumab, anakinra. To identify incident treatment periods, only patients who did not receive any biologics in the 1-year period before the index date were selected. In the base case analysis, exposure was defined with a 90-day latency after treatment initiation and a 180-day carry-over period after drug discontinuation.To compare the risk of malignancies between biologic treated patients and general population, Standardized incidence ratio (SIR [95%CI]) were calculated using FRANCIM (“France Cancer Incidence et Mortalité”) estimations as reference.To compare the risk of malignancies between biologics, a propensity score (including age, sex, year of first occurrence of RA code, date of treatment initiation, number of previous DMARDs, Charlson comorbidity index, diagnosis of tobacco and/or alcohol-associated disorders, number of hospitalizations for RA, cumulative corticosteroid dose) was calculated for each comparison. Hazard Ratios (HRs) for risk of cancer were estimated using Cox proportional hazard model using inverse probability of treatment weighting (IPTW) with propensity score. Exposure was considered as a time-dependent variable and propensity scores were estimated dynamically using pooled logistic regression reassessed for each new exposure.Results:Between 2007 and 2016, 31,792 patients (112,802 patient-years)- were exposed to biologics. The annual incidence rate of overall malignancies was 0.865 per 100 patients-years. Malignancies occurred in 730 patients exposed to anti-TNF, 235 patients exposed to another biologic and 11 exposed to both.As compared to the general population, biologic treated patients had an increased risk of lung cancer (SIR=1.35 [1.14;1.60]), a decreased risk of pancreatic cancer (SIR=0.52[0.31-0.85]) and no significant increased risk of invasive melanoma (SIR=1.15 [0.82;1.61]). Results were similar for anti-TNF-treated patients. Other biologics were not analyzed separately due to small sample sizes.The overall risk of malignancies and risk of lymphoma did not differ between anti-TNF and other biologics (analysed all together), or abatacept. Within the anti-TNF class, the overall risk of malignancies and risk of lymphoma did not differ between etanercept and monoclonal anti-TNF (table).Type of malignanciesHR [95% CI]p-valueHR [95% CI]p-valueHR [95% CI]p-valueAnti-TNF (ref) vs. other biologicsAnti-TNF (ref) vs. AbataceptMonoclonal anti-TNF (ref) vs. EtanerceptP-Y exposure83256 vs. 2564991770 vs. 468149620 vs. 36790All malignancies (excl. non-melanoma skin cancer)0.97 [0.81;1.17]p=0.71.27 [0.89;1.81]p=0.21.11 [0.94;1.32]p=0.2Solid cancer (excl. non-melanoma skin cancer)0.98 [0.80;1.20]p=0.81.23 [0.84;1.82]p=0.41.10 [0.92;1.32]p=0.3Lymphoma0.69 [0.32;1.46]p=0.31.73 [0.55;5.48]p=0.50.87 [0.49;1.57]p=0.7Conclusion:Using a large nationwide healthcare database, representative of the French population, the overall risk of malignancies did not seem to differ across the different classes of biologic. Among anti-TNF, the risk of malignancies of lymphoma did not differ between etanercept and monoclonal antibodies. The risk of organ specific cancers, except lung cancer, did not differ from that of general population.Disclosure of Interests:Raphaèle Seror Consultant of: BMS, Medimmune, Novartis, Pfizer, GSK, Lilly, Bruno Fautrel Grant/research support from: AbbVie, Lilly, MSD, Pfizer, Consultant of: AbbVie, Biogen, BMS, Boehringer Ingelheim, Celgene, Lilly, Janssen, Medac MSD France, Nordic Pharma, Novartis, Pfizer, Roche, Sanofi Aventis, SOBI and UCB, Alexandre Lafourcade: None declared, yann de-rycke: None declared, Xavier Mariette Consultant of: BMS, Gilead, Medimmune, Novartis, Pfizer, Servier, UCB, Florence Tubach Grant/research support from: Florence TUBACH is head of the Centre de Pharmacoépidémiologie (Cephepi) of the Assistance Publique – Hôpitaux de Paris and of the Clinical Research Unit of Pitié-Salpêtrière hospital, both these structures have received research funding, grants and fees for consultant activities from a large number of pharmaceutical companies, that have contributed indiscriminately to the salaries of its employees. Florence Tubach didn’t receive any personal remuneration from these companies.

Background:Objectives:To estimate the incidence rate of malignancies in csDMARD-treated RA patients and to compare it to that of general population and to biologic-treated RA patientsMethods:We conducted an historical cohort study within the national claim database that prospectively records individual health resource use of 86% of the French population (65 million inhabitants). RA adult patients were identified based on ICD-10 code (M05 or M06) between 2007-2016. Patients with previous cancer history were excluded. Treatment exposures were incident first use of any treatment: csDMARD (methotrexate, leflunomide, sulfasalazine, azathioprine, hydroxychloroquine) or biologics (anti-TNF, rituximab, abatacept, tocilizumab, ustekinumab, anakinra). To identify incident treatment periods, only patients who did not receive any treatment in the 1-year period before the index date were selected. Exposure was defined with a 90-day latency after treatment initiation and a 180-day carry-over period after drug discontinuation.To compare the risk of malignancies between csDMARD-treated patients and general population, standardized incidence ratio (SIR [95%CI]) were calculated using FRANCIM (“France Cancer Incidence et Mortalité”) estimations as reference.To compare the risk of malignancies between csDMARD and biologics treated patients, a dynamically propensity score (including age, sex, year of first occurrence of RA code, date of treatment initiation, number of previous DMARDs, Charlson’s comorbidity index, diagnosis of tobacco and/or alcohol-associated disorders, number of hospitalizations for RA, cumulative corticosteroid dose) was constructed using pooled logistic regression. Hazard Ratios (HRs) for risk of cancer were estimated using Cox proportional hazards model after dynamically propensity score matching. Exposure was considered as a time-dependent variable.Results:Between 2007 and 2016, 83,706 RA patients exposed to csDMARD (n=63,837) and/or biologics (n=19,727) were identified.As compared to the general population, csDMARDs treated patients had an increased risk of lung cancer (SIR=1.29 [1.14; 1.45]), invasive melanoma (SIR=1.52 [1.24; 1.86]) and a borderline increased risk of breast cancer (SIR=1.11 [1.01;1.22]). By contrast, they had a decreased risk of pancreatic cancer (SIR=0.68[0.51-0.9]) and liver cancer (SIR=0.43 [0.27; 0.67]). This later is due to a protopathic bias.After propensity score matching, analyses the risk of malignancies between csDMARD and biologics treated patients were conducted on 19727 patients in each group (mean age: 51 ±14 yrs; female: 74.6%). Malignancies occurred in 435 patients exposed to biologics and 332 patients exposed to csDMARD. The overall risk of malignancies (figure), risk of solid cancer (excluding non-melanoma skin cancer), lymphoma, and other hematologic malignancies did not differ significantly between csDMARD and all biologics (table). Regarding organ specific cancer, no difference was observed. Results were similar for biologic in monotherapy or associated with csDMARD.Type of malignanciesHR [95%CI] csDMARD (ref) vs. all biologicsp-valueAll malignancies (excl. non-melanoma skin cancer)0.99 [0.86;1.14]p=0.9Solid cancer (excl. non-melanoma skin cancer)0.95 [0.82;1.11]p=0.5Lymphoma1.35 [0.72;2.53]p=0.3Other hematologic malignancies1.18 [0.56;2.49]p=0.7Conclusion:Using a large nationwide representative healthcare database, the overall risk of malignancies and the risk of organ-specific cancers and hematologic malignancies in biologic treated RA patients did not differ from that of patients treated with csDMARD. Compared to general population, patients treated with csDMARD had an increased risk of lung cancer and melanoma, but a decreased risk of pancreatic cancer.Disclosure of Interests:Raphaèle Seror Consultant of: BMS, Medimmune, Novartis, Pfizer, GSK, Lilly, Alexandre Lafourcade: None declared, yann de-rycke: None declared, Bruno Fautrel Grant/research support from: AbbVie, Lilly, MSD, Pfizer, Consultant of: AbbVie, Biogen, BMS, Boehringer Ingelheim, Celgene, Lilly, Janssen, Medac MSD France, Nordic Pharma, Novartis, Pfizer, Roche, Sanofi Aventis, SOBI and UCB, Xavier Mariette Consultant of: BMS, Gilead, Medimmune, Novartis, Pfizer, Servier, UCB, Florence Tubach Grant/research support from: Florence TUBACH is head of the Centre de Pharmacoépidémiologie (Cephepi) of the Assistance Publique – Hôpitaux de Paris and of the Clinical Research Unit of Pitié-Salpêtrière hospital, both these structures have received research funding, grants and fees for consultant activities from a large number of pharmaceutical companies, that have contributed indiscriminately to the salaries of its employees. Florence Tubach didn’t receive any personal remuneration from these companies.

Background The increasing use of psychotropic drugs to treat anxiety and depressive disorders (ADDs) is concerning. According to the study, ‘Etude Pharmacoépidémiologique de l'Impact de Santé Publique des modes de prise en charge pour 3 groupes de pathologies’ (EPI3)-LASER, adult ADD patients who consult a general practitioner prescribing homeopathic medicines (GP-Ho) report less psychotropic drug use and are marginally more likely to experience clinical improvement than those receiving conventional care. We determined whether these observations also apply to patients ≥ 65 years old in the EPI3 cohort. Methods The EPI3-LASER study, conducted in France between March 2007 and July 2008, was a nationwide, observational survey of the three most common reasons for primary care consultation, including ADD, and the impact of the GPs' prescribing preferences: homeopathy (GP-Ho), conventional medicines (GP-CM) or mixed prescriptions (GP-Mx). This sub-analysis included 110 patients ≥ 65 years old with ADD from the EPI3 cohort who consulted either a GP-CM or GP-Ho. Socio-demographic and medical data and details of any medications prescribed were collected at inclusion. Information regarding the patients' functional status (Hospital Anxiety and Depression Scale [HADS)]) was obtained via a telephone interview 72 hours after inclusion, and at 1, 3 and 12 months post-inclusion. Medication use and outcome were determined over the 12-month period. Differences between the GP-CM and GP-Ho groups were assessed by multivariate logistic regression analysis. Results One hundred and ten patients were recruited and 87 (79.1%) with ADD (HADS ≥ 9) at the 72-hour interview were evaluated (age range: 65–93 years, 82.8% female). Patients who consulted a GP-Ho were more likely (odds ratio [OR] = 10.38, 95% confidence interval [CI]: 1.33–81.07) to have clinical improvement (HADS < 9) after 12 months than those in the GP-CM group. Patients who consulted a GP-Ho reported less psychotropic drug use (OR = 22.31 [95% CI: 2.20–226.31]) and benzodiazepine use (OR = 60.63 [95% CI: 5.75–639.5]) than GP-CM patients. Conclusions Management of ADD patients aged ≥ 65 years by GP-Ho appears to have a real public health interest in terms of effectiveness and lower psychotropic drug use.

Ces travaux de thèse présentent une approche de pharmaco-épidémiologie explorant l'exposition médicamenteuse chez des patients atteints de lymphome tout au long de leur parcours de soin. L'utilisation des données du Système National d'Information Inter-régimes de l'Assurance Maladie (SNIIRAM) nous a permis de quantifier cette exposition médicamenteuse et de déterminer les facteurs associés dans le contexte de la vie réelle. Les travaux de recherche réalisés au cours de cette thèse répondaient à 3 objectifs. Dans un premier objectif, nous avons souhaité valider l'utilisation du SNIIRAM à des fins de recherche pour améliorer la robustesse des futures études menées sur le lymphome. Après avoir défini plusieurs algorithmes d'identification des cas incidents de lymphome, nous avons testé leur validité par confrontation aux données cliniques exhaustives du Registre des cancers du Tarn. Les paramètres de performance obtenus permettent de considérer les données disponibles dans le SNIIRAM comme un outil puissant pour mener des études pharmaco-épidémiologiques ou médico-économiques sur le Lymphome. Le second objectif était d'explorer l'existence d'une surconsommation de médicaments psychotropes au cours de la phase active de traitement par rapport à une population témoin, sous l'hypothèse d'une initiation accrue de ces médicaments pour pallier les complications psychologiques associées à la prise en charge du Lymphome. Par ailleurs, la chronicisation fréquemment observée de ce type de consommation peut conduire à terme à des complications potentiellement évitables. Pour répondre à cet objectif, nous avons réalisé une étude selon une approche " new-user design ", à partir d'une cohorte de patients incidents de lymphome identifiés dans les données du SNIIRAM à l'échelon régional. Nous avons observé un taux d'initiation supérieur au taux observé en population générale ou chez des patients atteints d'autres pathologies mettant en jeu le pronostic vital (infarctus du myocarde). En fonction de la classe thérapeutique étudiée, 20 à plus de 50% des patients restaient exposés de façon inappropriée (au-delà des recommandations) à ces médicaments. Le troisième objectif a donc été d'explorer les déterminants associés à une utilisation prolongée de médicaments psychotropes dans la phase de l'après-cancer à partir des données de l'Echantillon généraliste des bénéficiaires (EGB). [...]
This thesis presents a pharmacoepidemiology approach to describe drug utilization in lymphomas during their whole care pathway. The use of the French claims database (Système National d'Informations Inter-Régimes de l'Assurance Maladie (SNIIRAM)) allows to exhaustively quantify this drug utilization in real life conditions. This thesis consists of three mains objectives. First, we aimed to develop validated algorithms for the identification of incident cases of lymphoma. For the validation, we used data from a regional Cancer Registry as the gold standard. The purpose of this validation study was to enhance validity of future studies conducted on lymphomas in the SNIIRAM database. The results of this study associated to strengths of this database demonstrate that this approach is of great interest to conduct pharmacoepidemiological or medico-economic studies in lymphomas. Second, we aimed to estimate the incidence of use of psychotropic drugs during the active treatment phase of lymphoma in comparison with control groups. Indeed, the increased probability of developing anxio-depressive disorders after diagnosis could lead to an increased initiation of psychotropic drugs and a potential inappropriate chronic use of these drugs after initiation. Such inappropriate chronic use can unnecessarily expose patients to adverse event. For this aim, we conducted a new-user cohort study using data from the SNIIRAM database. The results of this study revealed that the initiation rate of these drugs is higher than in the general population or for life-threatening diseases such as myocardial infarction. Moreover, we observed an inappropriate prolonged use for a significant fraction of patients (20% to more than 50% according to therapeutic class). On the basis of these findings, the third objective was to identify factors associated with prolonged use of these drugs during survivorship. This study was conducted using data from the General Sample of Beneficiaries (EGB). [...]

La thrombopénie immunologique (TI) est une maladie auto-immune rare. Son épidémiologie est mal connue. Le traitement de la TI aiguë repose sur les corticoïdes et les immunoglobulines. Passée la phase aiguë, plusieurs traitements sont possibles, dont la splénectomie, le rituximab et les agonistes du récepteur à la thrombopoiétine. L'exposition aux traitements de la TI en vie réelle n'a jamais été évaluée, et leur risque infectieux jamais comparé. Des vaccinations sont recommandées avant le rituximab et la splénectomie, mais la couverture vaccinale n'est pas connue. Nous avons constitué deux matériels complémentaires : 1) la cohorte French Adult Immune Thrombocytopenia : a pHarmacoepidemiological study (FAITH) est la cohorte des patients adultes incidents ayant une TI primaire traités de façon persistante (plus de 3 mois), bâtie dans le Système National d'Information Inter-régimes de l'Assurance Maladie (SNIIRAM) au niveau national ; 2) le registre Cytopénie Auto-immune : Registre Midi-PyrénéEN (CARMEN) est une étude observationnelle en vie réelle suivant les patients incidents de TI dans la région Midi-Pyrénées. Grâce à ces deux cohortes, nous avons décrit l'épidémiologie de la TI incidente en France, l'exposition aux traitements et la couverture vaccinale chez l'adulte, et évalué le risque infectieux des traitements et l'effet protecteur des vaccins
Immune thrombocytopenia (ITP) is a rare autoimmune bleeding disorder. ITP epidemiology is not well known. ITP treatment is based on glucocorticoids, and intravenous polyvalent immunoglobulin in case of severe bleeding. ITP becomes persistent (lasting more than 3 months) or chronic (more than 12 months) in about 70% of adults. In that case, non-corticosteroid treatments are suggested, mostly splenectomy, rituximab and thrombopoietin receptor agonists. The use of these treatments have never been assessed in the real life practice as well as their effectiveness and safety, particularly as regards the risk of infection. Vaccinations are recommended before splenectomy or rituximab. However, the vaccination rates and their effectiveness has not been assessed. We build two complementary materials: 1) the French Adult Immune Thrombocytopenia: a pHarmacoepidemiological study (FAITH) is the cohort of all incident primary ITP adults persistently treated (more than 3 months), built in the French health Insurance system database (Système national d'information inter-régimes de l'Assurance Maladie, SNIIRAM) at the national level; 2) the Cytopénie Auto-immune : Registre Midi-PyrénéEN (CARMEN) registry that includes and follows all incident ITP adults in the French Midi-Pyrénées region. Thanks to these two cohorts, we could assess the epidemiology of incident ITP in France; describe the exposure ITP treatments; assess the vaccination coverage in rituximab treated and splenectomized patients in France; and assess the risk of infection according to ITP treatments and the protective effect of the vaccines in this population

En France, environ un million de personnes seraient touchées par l’insuffisance cardiaque (IC) ; on recense près de 70 000 décès liés à l’IC, et plus de 150 000 hospitalisations et cela, malgré une prise en charge thérapeutique bien codifiée. Ces chiffres devraient s’accroitre dans les années futures du fait notamment du vieillissement de la population.L’objectif de ce travail était d’étudier l’utilisation des traitements pharmacologiques indiqués dans le traitement de l’IC (beta bloquant, inhibiteur de l’enzyme de conversion, anti-aldostérone, antagoniste des récepteurs à l’angiotensine II, diurétiques, digoxine, ivabradine) en situation réelle de soin, et d’identifier les facteurs cliniques ou pharmacologiques associés à la survenue d’un épisode de décompensation cardiaque.Un premier travail a permis de mesurer la fiabilité des bases de données médico-administratives françaises pour identifier des patients IC.Une deuxième étude a permis d’estimer que 17 à 37% de patients IC n’étaient exposés à aucun traitement de l’IC dans l’année suivant une première hospitalisation pour IC.Les troisième et quatrième parties de cette thèse ont mis en évidence qu’environ un quart des patients IC étaient réhospitalisés dans les 2 ans suivant une première hospitalisation. Les principaux facteurs cliniques prédictifs de cette réhospitalisation étaient l’âge, l’hypertension artérielle, la fibrillation auriculaire et le diabète. L’association retrouvée entre l’utilisation de fer bivalent et la réhospitalisation pour IC, souligne l’importance du risque lié à la présence d’une anémie ou d’une déficience en fer dans la survenue d’un épisode de décompensation cardiaque.Ces résultats permettent de reconsidérer la prise en charge thérapeutique chez les patients IC et mettent en avant la nécessité de renforcer la surveillance des patients les plus à risque de décompenser leur IC
In France, around one million persons would be affected by heart failure (HF); there are nearly 70 000 deaths related to HF and more than 150 000 hospitalizations despite a well defined treatment management. These numbers should increase in the next years due in particular to the ageing of the population.The objective of this work was to study the use of the pharmacological treatments indicated in HF (beta-blocker, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, aldosterone antagonist, diuretics, digoxin, ivabradine) in real-world setting and to identify the clinical or pharmacological predictors associated with a new episode of cardiac decompensation.A first work has enabled to estimate the accuracy of French claims databases in identifying HF patients.A second study estimated that 17 to 37% HF patients were not exposed to any HF treatment in the year following an incident HF hospitalization.The third and fourth parts of this thesis showed that almost one forth of HF patients was rehospitalized within the 2 years following a first hospitalization. The main clinical predictors of rehospitalization were age, high blood pressure, atrial fibrillation and diabetes. The association found between bivalent iron use and HF rehospitalization underlines the importance of the risk related to anemia or iron deficiency in the occurrence of a cardiac exacerbation episode.These results allow to reconsider the treatment management of HF patients and highlight the need to reinforce the surveillance of patients with a highest risk of cardiac exacerbation

Les patients présentant un trouble psychiatrique ont une diminution de leur espérance de vie en lien avec un taux élevé de suicide et aux comorbidités non psychiatriques. De plus, les patients ayant fait une tentative de suicide (TS) sont particulièrement à risque de récidive et de décès prématuré.Premièrement, nous avons tenté d'identifier des facteurs de risque de récidive avec des approches statistiques innovantes. Ainsi, nous avons pu montrer que patients ayant un trouble de l'usage d'alcool et ayant consommé de l'alcool lors de leur TS, les patients souffrant d'un trouble anxieux, ceux ayant fait plus de 2 TS et enfin ceux qui consomment des benzodiazépines et/ou des hypnotiques ont un risque élevé de récidive.Ensuite, nous avons estimé le taux de décès par suicide après une TS à l'aide d'une méta-analyse. Nos résultats ont montré un taux de 2,8% à 1 an. Puis, nous avons recherché les causes de décès dans la cohorte Vigilans. Un an après la TS, les causes les plus fréquentes de décès étaient le suicide et les causes cardiovasculaires.Enfin, nous avons montré un exemple d'évaluation d'un dispositif de prévention du suicide. Nous avons évalué l'efficacité de la formation d'agents sentinelles dans les maisons de retraite. Nos résultats montrent une amélioration des connaissances sur la crise suicidaire et une diminution du nombre de TS après la formation.En conclusion, la diminution de la morbi-mortalité des suicidants passe par une prise en charge globale : par la prévention du suicide, mais également par la prise en charge des pathologies non psychiatriques
Patients with a psychiatric disorder have a decreased life expectancy associated with a high rate of suicide and non-psychiatric diseases. In addition, patients who have attempted suicide (SA) are at particular risk for re-attempt and premature death. First, we tried to identify risk factors for re-attempt with novel statistical approaches. Thus, patients with an alcohol use disorder and with an acute alcohol use during their SA, patients with an anxiety disorder, those who had more than 2 SAs and those who consumed benzodiazepines and/or hypnotics had a high risk of re-attempt.Then, we estimated the rate of death by suicide after SA using a meta-analysis. Our results showed a rate of 2.8% at 1 year. Then, we searched for causes of death in the Vigilans cohort. One year after SA, the most common causes of death were suicide and cardiovascular diseases.Finally, we showed an example of an assessment of a suicide prevention program. We evaluated the effectiveness of training gatekeepers in nursing homes. Our results showed an improvement in knowledge about the suicidal crisis and a decrease in the number of SAs after the training.In conclusion, the reduction of the morbidity and mortality of those who have attempted suicide requires a global management: by suicide prevention, but also by the management of non-psychiatric diseases

La douleur chronique constitue une problématique majeure et complexe, qui impacte profondément les individus en altérant leurs capacités fonctionnelles et émotionnelles ainsi que leur qualité de vie. Selon une nouvelle terminologie du IASP adoptée en 2017, trois types de douleur chronique sont identifiés : nociceptive, nociplastique et neuropathique. En particulier, la douleur neuropathique, définie comme une douleur liée à une lésion ou une maladie affectant le système somato-sensoriel, est l’objet de trois enjeux majeurs : 1/ la nécessité d’une meilleure connaissance de son épidémiologie, qui doit reposer sur une uniformisation des définitions et outils utilisés, les estimations actuelles restant très hétérogènes, 2/ le besoin d’évaluer dans la vie réelle l’utilisation, l’efficacité et l’adéquation des thérapeutiques médicamenteuses avec les dernières recommandations, afin d’optimiser la qualité de la prise en charge de la douleur neuropathique dans une perspective de meilleure personnalisation et 3/ l’impériosité de développer de nouvelles stratégies thérapeutiques plus efficaces et mieux tolérées, la douleur neuropathique étant largement reconnue comme l’un des syndromes douloureux les plus difficiles à prendre en charge de manière optimale. La première partie de ce travail s’est attachée à actualiser les données épidémiologiques françaises concernant la prévalence de la douleur chronique avec ou sans caractéristiques neuropathiques en utilisant pour la première fois dans cette thématique la méthode de capture-recapture, à partir de l’exploitation des données de l’EGB : en 2015, la prévalence de la douleur chronique avec ou sans caractéristiques neuropathiques dans la population générale adulte française était estimée à 27.2% [IC 95% : 26.1-28.4] et celle de la douleur chronique neuropathique à 7.20% [IC 95% : 6.38-8.22]. Cette méthode a démontré son intérêt majeur, compte tenu non seulement de la rapidité et de la fiabilité des estimations obtenues, mais également de la facilité de sa mise en œuvre et de son coût moindre comparativement aux études ou enquêtes de terrain, tout en s’affranchissant des limites classiques de ces dernières en termes de représentativité et de généralisabilité.La deuxième partie de notre travail a proposé de réaliser une évaluation pharmaco-épidémiologique de l’utilisation en vie réelle des traitements recommandés dans la douleur neuropathique et de leur persistance, dans une cohorte de patients douloureux neuropathiques incidents, à partir de l’exploitation des données du SNIIRAM. Entre janvier 2013 et décembre 2016, 29 756 patients incidents ont été identifiés et moins de la moitié d’entre eux (44.1%) avaient bénéficié de la primo-prescription d’un traitement recommandé en première ligne en monothérapie, environ un tiers des patients (32.8%) ont reçu une première prescription correspondant à une association d’au moins deux molécules recommandées. Enfin, moins de la moitié des patients (46.3%) étaient encore traités avec au moins un des médicaments recommandés à 1 an et environ un tiers (33.5%) l’étaient encore à 2 ans ; la persistance aux traitements recommandés était plus faible chez les patients traités en monothérapie et chez ceux présentant une douleur neuropathique associée à un cancer, mais meilleure chez les patients souffrant de dépression.Enfin, dans une dernière partie, préclinique, nous avons évalué le potentiel antalgique de l’agomélatine, un antidépresseur commercialisé il y a quelques années (VALDOXAN®) dont le mécanisme pharmacodynamique original repose sur une activité agoniste au niveau des récepteurs mélatoninergiques MT1/MT2 et antagoniste au niveau des récepteurs 5-HT2C. L’agomélatine a réduit de manière dose-dépendante l’hyperalgésie induite dans trois modèles animaux de douleur neuropathique, à la fois lors d’une administration unique et répétée, sans phénomène de tachyphylaxie ou de tolérance, avec une efficacité au moins égale à celle de la gabapentine. (...)
Chronic pain is a major and complex problem, which has a profound impact on individuals by altering their functional and emotional capacities as well as their quality of life. According to new IASP terminology adopted in 2017, three types of chronic pain are identified: nociceptive, nociplastic, and neuropathic. In particular, neuropathic pain, defined as pain related to an injury or disease affecting the somatosensory system, is the subject of three major issues: 1/ the need for a better knowledge of its epidemiology, which must be based on a standardisation of the definitions and tools used, as current estimates remain very heterogeneous, 2/ the need to evaluate in real-life the use, efficacy and adequacy of drug therapies with the latest recommendations, in order to optimize the quality of neuropathic pain management with a view to better treatment personalization and 3/ the urgency to develop new, more effective and better tolerated therapeutic strategies, neuropathic pain being widely recognized as one of the most difficult painful syndromes to manage optimally. The first part of this work focused on updating French epidemiological data on the prevalence of chronic pain with or without neuropathic characteristics by using for the first time in this theme the capture-recapture method, based on the use of EGB database: in 2015, the prevalence of chronic pain with or without neuropathic characteristics in the general French adult population was estimated at 27.2% [CI 95%: 26.1-28.4] and chronic neuropathic pain at 7.20% [CI 95%: 6.38-8.22]. This method has demonstrated its major interest, given not only the speed and reliability of the estimates obtained, but also the ease of its implementation and its lower cost compared to conventional studies or field surveys, while getting rid from the traditional limits of the latter in terms of representativeness and generalizability.The second part of our work proposed to carry out a pharmaco-epidemiological evaluation of the real-life use of recommended treatments in neuropathic pain and their persistence in a cohort of incident painful neuropathic patients, based on the use of the SNIIRAM database. Between January 2013 and December 2016, 29 756 incident patients were identified and less than half of them (44.1%) had received a recommended first-line treatment as monotherapy, and approximately one third of patients (32.8%) received a first prescription corresponding to a combination of at least two recommended molecules. Finally, less than half of patients (46.3%) were still treated with at least one of the recommended drugs at 1 year and about one-third (33.5%) were still treated at 2 years; persistence to the recommended treatments was lower in patients treated with monotherapy and in those with neuropathic cancer pain, but better in patients with depression.Finally, in a third preclinical section, we evaluated the analgesic potential of agomelatine, an antidepressant marketed a few years ago (VALDOXAN®) whose original pharmacodynamic mechanism is based on agonist activity at the MT1/MT2 melatoninergic receptors and antagonist at the 5-HT2C receptors. Agomelatine reduced dose-dependently the hyperalgesia induced in three animal models of neuropathic pain, both after single and repeated administration, without tachyphylaxis or tolerance, with an efficacy at least equal to that of gabapentin. This effect was exerted through the spinal involvement of melatoninergic receptors, α-2 adrenergic and serotoninergic 5-HT2C, without involvement of ß-adrenergic receptors. Finally, an isobolographic analysis concluded that the combination of agomelatine and gabapentin was additive. In conclusion, this work demonstrated the major interest of using Health Insurance database to address pharmaco-epidemiological issues in neuropathic pain and offered a new and interesting therapeutic perspective through agomelatin in optimizing the management of neuropathic pain

Malgré la fréquence d’utilisation des psychotropes chez les personnes plus âgées et leur propension aux événements indésirables liés aux médicaments, peu d’études se sont intéressées au respect des règles de bon usage des psychotropes dans cette population. L’objectif de ce travail était d’évaluer les modalités d’utilisation des psychotropes et le respect des recommandations de bon usage publiées en France. Deux types de population ont été inclus : un échantillon clinique constitué de patients âgés d’au moins 65 ans hospitalisés dans deux Pôles de psychiatrie à Bordeaux, et un échantillon de la population générale âgée d’au moins 65 ans, à partir d’une base de données de l’assurance maladie française. En population clinique, les règles de bon usage étaient plutôt bien respectées, avec, pour les benzodiazépines, la prescription d’une produit à demi-vie courte dans près de trois quarts des prescriptions et une posologie adéquate dans près de deux tiers des cas. Toutefois, un arrêt de traitement n’était jamais prévu dès l’initiation du traitement et le rythme de prescription était discontinu dans seulement un tiers des cas. Dans la population générale âgée, la durée et la surveillance biologique des traitements antidépresseurs étaient conformes chez moins de 20% des sujets, alors que près des trois quarts des sujets initiant un traitement par benzodiazépine recevaient une durée adéquate. En revanche moins de la moitié des délivrances concernaient une benzodiazépine anxiolytique à demi-vie courte. L’étude de l’impact de la publication des recommandations n’a pas montré qu’elle permettait une amélioration de l’utilisation des psychotropes. Il est nécessaire d’accompagner la publication des recommandations d’autres mesures, afin d’insister sur le bon usage des psychotropes, notamment la durée nécessaire de traitement antidépresseur et le choix d’un anxiolytique à demi-vie courte
Despite a high frequency of use and a tendency to present with adverse events of drugs, few studies assessed compliance with guidelines related to proper use of psychotropic drugs in the older population. The aim was to assess the patterns of psychotropic drugs use in this population, as well as the compliance with French guidelines. Two population samples were included, a clinical sample of older psychiatric inpatients and a sample of the older general population using claims database of the national health insurance. In clinical population, compliance with guidelines was rather good, with prescription of a short half-life benzodiazepine in nearly three quarters of prescriptions and adequate dosage in nearly two third of cases. However, treatment discontinuation was never specified at the time of treatment initiation and rhythm of prescription was discontinuous in only one third of cases. In the older general population, duration of antidepressant treatment and biological monitoring was appropriate in only 20 % of patients, whereas nearly three quarters of subjects initiating a benzodiazepine treatment were treated over an appropriate duration. However, less than half of them had received a benzodiazepine anxiolytic of short half-life. The assessment of the impact of practice guidelines publication found no improvement of psychotropic drugs use. Other interventions should accompany guidelines publication in order to underline the importance of proper use of drugs, particularly antidepressant treatment duration and use of short half-life benzodiazepine drugs

Contexte – Les cancers de la prostate, du sein, et colorectaux sont parmi les cancers les plus fréquents dans les pays développés, et, même si plusieurs facteurs de risque sont aujourd’hui bien établis pour ces cancers, leur étiologie reste encore largement à expliquer. L’inflammation chronique est fortement suspectée de jouer un rôle dans la survenue de ces cancers et la présence, dans les tissus tumoraux, d’infiltrats inflammatoires localisés pouvant être considérés comme des lésions précancéreuses, contribue à renforcer l’hypothèse d’un lien possible entre inflammation chronique et cancers. Dans ce contexte, de nombreuses études épidémiologiques se sont intéressées au rôle des Anti-Inflammatoires Non Stéroïdiens (AINS) dans les cancers. En effet, les médicaments ayant des propriétés anti-inflammatoires comme les AINS, dont l’aspirine, et les anti-inflammatoires inhibiteurs sélectifs de la cyclo-oxygénase 2 (COX-2), pourraient diminuer le risque de survenue de ces cancers.Objectifs – L’objectif général de cette thèse a été d’étudier le rôle de la consommation d’AINS, dont l’aspirine, les AINS usuels et les inhibiteurs sélectifs de la COX-2 dans la survenue des cancers de la prostate, du sein et colorectaux.Population et méthodes – Ce travail s’est appuyé sur les données de l’Echantillon Généraliste des Bénéficiaires (EGB) de l’Assurance Maladie pour les trois cancers d’intérêt et sur les données d’une étude cas-témoins réalisée en population générale dans le département de l’Hérault (EPICAP) pour le cancer de la prostate. Pour les données de l’EGB, une cohorte fixe de 426 410 personnes présentes au 1er janvier 2007 a permis d’identifier les cas incidents entre 2008 et 2012 à partir de différents algorithmes. L’exposition aux AINS a été identifiée à partir du 1er janvier 2005 jusqu’à la date de fin d’observation : date de survenue du cancer, date de décès ou date de censure fixée au 31 décembre 2012. Un temps de latence d’au moins un an a été défini entre l’exposition aux AINS et la survenue du cancer d’intérêt. Pour les données d’EPICAP, 819 cas incidents de cancer de la prostate et 879 témoins de population générale, de même âge en moyenne que les cas, ont été interrogés en face-à-face, à l’aide d’un questionnaire standardisé, notamment sur leur consommation d’AINS.Résultats – A partir de la cohorte issue de l’EGB, des résultats préliminaires montraient une augmentation du risque de cancer de la prostate (RR=1,30 [1,17-1,46]) et du sein (RR=1,29 [1,14-1,46]) chez les patients exposés aux AINS et une absence d’association pour les cancers colorectaux (RR=0,92 [0,82-1,05]). En revanche, une association négative était observée pour les cancers de la prostate (RR=0,85 [0,74-0,96]) et colorectaux (RR=0,77 [0,66-0,90]) lorsque le temps de latence considéré était de six ans. L’étude EPICAP a montré que la consommation d’AINS était associée négativement au cancer de la prostate (OR=0,77 [0,61-0,98]). Cette association était plus prononcée pour une fréquence de consommation quotidienne (OR=0,75 [0,33-0,92]) ou d’une consommation pluriquotidienne (OR=0,38 [0,18-0,79]), et pour une durée entre 5 à 10 ans (OR=0,55 [0,33-0,92]). L’association était renforcée pour une molécule ayant une activité anti-COX-2 préférentielle (OR=0,48 [0,28-0,79]). Enfin, une association négative était également observée pour les cancers de la prostate de haut grade (Gleason score =7 (4+3) ou GS>7) avec un OR de 0,62 [0,41-0,95].Conclusion – L’ensemble de ce travail de thèse a montré que la consommation d’AINS semblait être associée négativement à la survenue du cancer de la prostate et aux cancers colorectaux. Pour le cancer de la prostate cette thèse s’est appuyée sur deux bases de données et deux méthodologies différentes, permettant d’appréhender les limites et les forces de chacune
Background – Prostate, breast, and colorectal cancers are among the most common cancers in developed countries. Many risk factors have been identified over the years but could explain only a part of the new cases. Chronic inflammation is highly suspected to play a role in the carcinogenesis of those cancers and the presence of inflammatory infiltrate in tumoral tissue, considered as precancerous lesions, reinforced this hypothesis. In this context, several epidemiological studies have investigated the potential role of Non-steroidal anti-inflammatory drugs (NSAIDs) in cancer occurrence. Indeed, NSAIDs such as aspirin and non-aspirin NSAIDS including selective inhibitors of cyclo-oxygenase 2 (COX-2) may decrease the incidence of those cancers.Objectives – The main objective of the thesis was to investigate the role of NSAIDs use including aspirin, non-aspirin NSAIDs and selective inhibitors of COX-2 in the occurrence of prostate, breast and colorectal cancers.Population and methods – This work was based on the General Sample of health insurance Beneficiaries (EGB) for the three localizations of cancer and on the data of a population-based case-control study carried out in the département of Herault (EPICAP) for prostate cancer. In the EGB study, a cohort of 426 410 persons present in the database in January 1st, 2007 allowed to identify incident cases between 2008 and 2012 based on different algorithms. Exposure to NSAIDs was determined from January 1st, 2005 until the end of the follow up defined as either cancer incident date, date of death, or censure date fixed as December 31st, 2012. A latency of at least one year between the beginning of exposure to NSAIDs and the cancer occurrence was taken into account. For the EPICAP study, 819 incident prostate cancer cases and 879 population-based controls, frequently matched by age to the cases, were face-to-face interviewed using a standardized questionnaire, specifically on their NSAIDs use.Results – From the EGB cohort, preliminary results showed a positive association between all NSAIDs use and prostate or breast cancer occurrence (RR=1,30 [1,17-1,46], RR=1,29 [1,14-1,46], respectively), while no association was found with colorectal cancer occurrence (RR=0,92 [0,82-1,05]). These associations became negative associations when a latency of six years was taken into account in prostate and colorectal cancer (RR=0,85[0,74-0,96], RR=0,77 [0,66-0,90], respectively). In the EPICAP study, NSAIDs use was negatively associated with prostate cancer (OR=0,77 [0,61-0,98]). This association was more pronounced with daily intake (OR=0,75 [0,33-0,92]) or more than once a day (OR=0,38 [0,18-0,79]), and for a duration of five to ten years (OR=0,55 [0,33-0,92]). The negative association was reinforced for preferential anti-COX-2 NSAIDs (OR=0,48 [0,28-0,79]), and for patient with high grade prostate cancer (Gleason score, GS=7 (4+3) or GS>7 : OR=0,62 [0,41-0,95]).Conclusion – This work showed that NSAIDs use was negatively to prostate and colorectal cancer occurrence. For prostate cancer, this thesis was based on two different databases (a medical and administrative database and a case-control study) and used two different methodologies, allowing comparison about strengths and limits of both

Dans le cadre du programme commun d’études pharmaco-épidémiologiques de la caisse nationale de l'assurance maladie et de l’agence nationale de sécurité du médicament, visant à évaluer l'impact sanitaire en France de l'exposition in utero à l’acide valproïque à partir des bases de données médico-administratives (BDMA) françaises, l’objectif de cette thèse était d’étudier l’utilisation des antiépileptiques pendant la grossesse et les risques de malformations congénitales et de troubles neuro-développementaux associés chez l’enfant. Le premier volet de cette thèse a consisté à formaliser et publier un algorithme d’identification des grossesses spécifiquement adapté aux BDMA françaises. L’application de cet algorithme à la description de l’utilisation des antiépileptiques pendant la grossesse a permis d’estimer à 6,7‰ la prévalence de l’utilisation des antiépileptiques pendant la grossesse et de montrer une baisse de l’utilisation des antiépileptiques de première génération, en particulier de l’acide valproïque, au bénéfice des antiépileptiques de deuxième génération entre 2007 et 2014. Dans le deuxième volet de cette thèse, l’exposition in utero à l’acide valproïque a été retrouvée associée à une augmentation du risque d’un grand nombre des malformations congénitales majeures (MCM) étudiées, avec une relation dose-effet pour les MCM les plus fréquentes, et l’exposition in utero au topiramate à une augmentation du risque de fentes oro-faciales. Des signaux relatifs à la prégabaline, au clonazépam et au phénobarbital ont aussi été identifiés. Dans le troisième volet de cette thèse, l’exposition in utero à l’acide valproïque a été retrouvée associée à une augmentation du risque de chacun des événements neuro-développementaux précoces étudiés versus lamotrigine, avec une relation dose-effet, à l’inverse des autres antiépileptiques. La réalisation d’études pharmaco-épidémiologiques à partir des BDMA françaises a permis aux autorités sanitaires de fournir rapidement des données sur l’utilisation des antiépileptiques pendant la grossesse en France. La réalisation de ces études a aussi permis de participer à l’enrichissement de la littérature observationnelle internationale sur les conséquences de l’exposition in utero aux antiépileptiques pour l’enfant à naitre
The works of this thesis have been carried out within a programme of pharmacoepidemiological studies initiated by the National Agency of Medicine and Health Product Safety (ANSM) and the National Health Insurance fund (Cnam) in order to evaluate the public health situation in relation to prenatal exposure to valproic acid in France on the basis of the French health care databases. The objective of this thesis was to study antiepileptic drug (AED) use during pregnancy and the risks of congenital malformations and neurodevelopmental disorders associated with prenatal exposure to these drugs. In a first study, we developed an algorithm to identify pregnancy episodes and related outcomes using the French health care claims databases and applied it to study AED use during pregnancy between 2007 and 2014. Over the study period, 6.7 per 1000 pregnancies were exposed to an AED. The use of newer AEDs increased concomitantly with the decreased use of valproic acid and the other older AEDs. In a second study, prenatal exposure to valproic acid was found to be associated with a wide range of malformations among those investigated, with a dose-response relationship for half of them, and prenatal exposure to topiramate with an increased risk of cleft lip with or without cleft palate. Signals concerning pregabalin, clonazepam and phenobarbital have also been identified. In a third study, prenatal exposure to valproic acid was found to be associated with increased risks of all early neurodevelopmental outcomes investigated compared with lamotrigine, with a dose-response relationship. Prenatal exposure to the other AEDs was not associated with an increased risk of any of these neurodevelopmental outcomes versus lamotrigine. Conducting pharmacoepidemiological studies based on the French health care databases enabled the health authorities to rapidly provide data on the use of AED during pregnancy in France. It also brought additional evidence to the international observational literature on the consequences of prenatal exposure to AEDs for the unborn child