Olivier Fourcade

Paul Sabatier University - Toulouse III, Tolosa de Llenguadoc, Midi-Pyrénées, France

Are you Olivier Fourcade?

Claim your profile

Publications (181)401.59 Total impact

  • [Show abstract] [Hide abstract]
    ABSTRACT: IntroductionBrain midline shift (MLS) is a life-threatening condition that requires urgent diagnosis and treatment. We aimed to validate bedside assessment of MLS with Transcranial Sonography (TCS) in neurosurgical ICU patients by comparing it to CT.Methods In this prospective single centre study, patients who underwent a head CT were included and a concomitant TCS performed. TCS MLS was determined by measuring the difference between the distance from skull to the third ventricle on both sides, using a 2 to 4 MHz probe through the temporal window. CT MLS was measured as the difference between the ideal midline and the septum pellucidum. A significant MLS was defined on head CT as >0.5 cm.ResultsA total of 52 neurosurgical ICU patients were included. The MLS (mean¿±¿SD) was 0.32¿±¿0.36 cm using TCS and 0.47¿±¿0.67 cm using CT. The Pearson¿s correlation coefficient (r2) between TCS and CT scan was 0.65 (P <0.001). The bias was 0.09 cm and the limits of agreements were 1.10 and -0.92 cm. The area under the ROC curve for detecting a significant MLS with TCS was 0.86 (95% CI =0.74 to 0.94), and, using 0.35 cm as a cut-off, the sensitivity was 84.2%, the specificity 84.8% and the positive likelihood ratio =5.56.Conclusions This study suggests that TCS could detect MLS with reasonable accuracy in neurosurgical ICU patients and that it could serve as a bedside tool to facilitate early diagnosis and treatment for patients with a significant intracranial mass effect.
    Critical care (London, England) 12/2014; 18(6):676. · 4.72 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: IntroductionEchocardiographic indices based on respiratory variations of superior and inferior vena cava diameters (¿SVC and ¿IVC, respectively) have been proposed as predictors of fluid responsiveness in mechanically ventilated patients but they have never been compared simultaneously in the same patient sample. The aim of this study was to compare the predictive value of these echocardiographic indices when concomitantly recorded in mechanically ventilated, septic patients.Methods Septic shock patients requiring hemodynamic monitoring were prospectively enrolled over a 1-year period in a mixed medical surgical ICU of a University Teaching Hospital (Toulouse, France). All patients were mechanically ventilated. Predictive indices were obtained by transesophageal and transthoracic echocardiography and were calculated as follows: (Dmax-Dmin) / Dmax for ¿SVC and (Dmax-Dmin) / Dmin for ¿IVC where Dmax and Dmin are the maximal or minimal diameter of SVC and IVC. Measurements were performed at baseline and after a 7 ml/kg volume expansion using a plasma expander. Patients were separated into responders (increase in cardiac index¿¿¿15%) and non-responders (increase in cardiac index¿<¿15%).ResultsAmong 44 included patients, 26 (59%) patients were responders (R). ¿SVC was significantly more accurate than ¿IVC to predict fluid responsiveness (the area under the ROC curve for ¿SVC and ¿IVC regarding assessment of fluid responsiveness were significantly different (0.74 (95% confidence interval (CI): 0.59 to 0.88) and 0.43 (95% CI: 0.25 to 0.61) respectively with P¿=¿0.012)). No significant correlation between ¿SVC and ¿IVC was found (r¿=¿0.005, P¿=¿0.98). The best threshold value to discriminate R from NR was 29% for ¿SVC with a 54% sensitivity and 89% specificity, and 21% for ¿IVC with a 38% sensitivity and 61% specificity.Conclusions¿SVC was better in predicting fluid responsiveness than ¿IVC in our cohort. It is worth noting that the sensitivity and specificity of ¿SVC and ¿IVC to predict fluid responsiveness were lower than those reported in the literature, highlighting the limits of using these indices in a heterogeneous sample of medical and surgical septic patients.
    Critical care (London, England) 09/2014; 18(5):473. · 4.72 Impact Factor
  • Annales Françaises d Anesthésie et de Réanimation 09/2014; 33:A229-A230. · 0.84 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Introduction Le patient polytraumatisé présente une dépense énergétique majeure, un hypercatabolisme associé à une protéolyse, une lipolyse et un déficit en micronutriments. Son statut en carnitine a été peu étudié. Cette amine quaternaire est pourtant l’élément clé de la β-oxydation des acides gras à chaînes longues et la lipolyse est une source majeure d’énergie durant cette phase d’hypercatabolisme liée à l’agression tissulaire aiguë. Chez l’Homme, 75 % de la carnitine provient de l’alimentation mais les produits de nutrition artificielle en sont dépourvus. Matériel et méthodes Étude observationnelle sur 38 patients polytraumatisés (dont 18 avec traumatisme crânien) hospitalisés en réanimation polyvalente. La carnitine plasmatique libre a été dosée au 3e jour post-traumatique, sur des reliquats de prélèvements de sang restant issus des bilans réalisés pour la prise en charge habituelle, chez ces patients recevant progressivement une nutrition entérale à partir de j2. Résultats La carnitine plasmatique est abaissée chez 95 % des patients avec une médiane est de 18 μmol/L [11–47] pour une norme entre 36 et 46 μmol/L. Cette diminution est plus importante chez les polytraumatisés cérébro-lésés : 7,72 [11–36] vs 21,5 [11–47] μmol/L (p = 0,031). En analyse univariée, on observe une relation entre la carnitinémie et l’IMC (p = 0,049), le DFG estimé par le CKD-EPI (p = 0,0171) et l’urée sanguine (p = 0,0022). Pour les patients présentant les carnitinémies les plus diminuées (< 18 μmol/L), en analyse multivariée, les facteurs en cause sont l’association à un traumatisme crânien (p = 0,0151) et l’urée plasmatique (0,0371) ( Fig. 1). Discussion La concentration plasmatique de carnitine libre est abaissée précocement chez le patient polytraumatisé et de façon significativement plus importante chez les cérébro-lésés. Différents mécanismes peuvent être évoqués : diminution de la synthèse de novo et des apports exogènes, augmentation de l’élimination, majoration de l’utilisation et/ou association des divers mécanismes. La relation entre la carnitinémie et le DFG évoque augmentation de l’excrétion urinaire déjà montrée par Cerderblad puis Davis. La majoration de l’utilisation de la carnitine peut être incriminée chez les traumatisés crânien présentant la dépense énergétique la plus élevée. Une adaptation métabolique ne peut être exclue. Les conséquences de ce déficit peuvent être évoquées en prenant en compte l’effet neuro-protecteur potentiel de la carnitine.
    Annales francaises d'anesthesie et de reanimation 09/2014; 33:A24–A25. · 0.77 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Introduction Certaine études suggèrent que lors d’une encéphalopathie hépatique, une hypertension intracrânienne (HTIC) peut dans certains cas être observée [1]. Toutefois, l’incidence de l’HTIC lors d’une encéphalopathie hépatique reste inconnue. Évaluation de l’incidence de l’hypertension intracrânienne (HTIC) par méthodes non invasives chez les patients présentant une encéphalopathie hépatique et son évolution avec le traitement de la maladie hépatique. Matériel et méthodes Étude prospective, observationnelle au sein d’un CHU. Des patients présentant une encéphalopathie hépatique (critères de West Haven) ont été inclus. Les patients ont bénéficié d’une estimation non invasive de la pression intracrânienne par mesure échographique du diamètre des enveloppes du nerf optique (DENO) (HTIC si > 0,58 cm) et mesures des vélocités cérébrales dans les artères cérébrales moyennes (ACM) au Doppler transcrânien (DTC) de l’inclusion et jusqu’à 4 jours après le début de l’étude (j4). Les tests de Mann-Whithney ou Kruskal-Wallis ont été appliqués quand ils étaient appropriés (p significatif < 0,05). Résultats Vingt-neuf patients ont été inclus. Le DENO médian à j0 était de 0,54 cm (extrêmes de 0,42 à 0,60 cm). Il n’y avait pas de différences ou pour les vélocités au DTC entre les différents stades d’encéphalopathie pour les DENO à j0 (p = 0,13) et à j1 (p = 0,41), ni pour les index de pulsatilité (p = 0,38 pour les IP droits et p = 0,39 pour les IP gauches), entre les différents groupes d’encéphalopathie clinique à j0 et à j1 ( Fig. 1). De même, l’analyse de l’évolution lors du traitement de l’encéphalopathie ne retrouvait pas de différences significatives entre les valeurs des DENO (p = 0,7) et les index de pulsatilité (IP) des ACM entre j0 et j4 (IP droit p = 0,4 et IP gauche p = 0,5). Discussion La mesure par des techniques non invasives ne confirment pas la présence d’HTIC chez les patients en encéphalopathie hépatique.
    Annales francaises d'anesthesie et de reanimation 09/2014; 33:A15. · 0.77 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Introduction Le monitorage de la pression de perfusion cérébrale (PPC) est recommandé chez le cérébro-lésé. Ce monitorage invasif, via la pose d’un capteur de PIC, peut parfois nécessiter un délai pour être disponible. Le Doppler transcrânien (DTC) est un outil non invasif et rapide d’utilisation. Trois auteurs ont proposé une estimation de la PPC (ePPC) à partir des vélocités au DTC et des paramètres hémodynamiques 0005, 0010 and 0015. Nous avons réalisé une étude observationnelle prospective afin de comparer les performances de ces 3 formules pour prédire la PPC réelle du patient. Matériel et méthodes Parmi les patients, 120 cérebro-lésés ayant bénéficié du monitorage invasif de la PPC ont été inclus. Ils ont bénéficié d’une mesure bilatérale des vélocités dans l’artère cérébrale moyenne par DTC, permettant l’analyse de 109 mesures (mesures impossibles chez 11 patients). La précision de la prédiction de la PPC suivant les 3 formules : 1) Czosnyka : PPCe = (PAM × VD) / VM + 14 ; 2) Bellner : PPCe = 89,646–8,258 × IP et 3) Edouard : PPCe = (VM / (VM-VD)) × PAM–PAD) a été évaluée en les comparant à la méthode de référence : la PPC mesurée de manière invasive (PPC = PAM − PIC) à l’aide d’une régression linéaire avec calcul des intervalles de prédiction à 95 %. Un sous-groupe de patients de > 60 ans a été isolé afin d’étudier l’effet de l’âge sur les performances du DTC. La précision était définie par la moyenne des valeurs absolue des différence avec la méthode de référence. Résultats L’équation de Czosnyka et Bellner sont très proches dans la population générale (respectivement précision à 8,5 et 8,6 mmHg et r2 à 0,43 et 0,56) ( Fig. 1). Chez les patients de plus 60 ans, la formule de Bellner, cette fois ci, est plus concordante (précision à 7,2 mmHg et r2 à 0,73). Discussion La méthode de Coznycka et de Bellner se rapprochent le plus de la méthode de référence dans la population générale. Cependant, chez le patient de plus de 60 ans, c’est la méthode de Bellner qui est la plus précise et ce, possiblement, parce que cette formule de calcul s’affranchit de la PAM.
    Annales francaises d'anesthesie et de reanimation 09/2014; 33:A30–A31. · 0.77 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Background.(It has been suggested that the complementary use of echocardiography could improve the diagnostic accuracy of lung ultasonography (LUS) in acute respiratory failure (ARF) patients. Nevertheless, the additional diagnostic value of echocardiographic data when coupled to LUS is still debated in this setting. The aim of the current study was to compare the diagnostic accuracy of LUS and an integrative cardiopulmonary ultrasound approach (TUS) in patients with ARF. Methods.We prospectively recruited patients consecutively admitted for ARF in ICU (University Teaching Hospital) over a 12-month period. Inclusion criteria were age at least 18 years and the presence of criteria of severe ARF justifying ICU admission. We compared both LUS and TUS approaches and the final diagnosis determined by the panel of experts using machine-learning methods to improve the accuracy of the final diagnostic classifiers. Results.One hundred thirty-six patients were included (age 68 ± 15 yr; sex ratio 1). A 3 dimensional PLS-mod (partial least square and multinomial logistic regressions) was developed and subsequently tested using an independent sample of patients. Overall, the diagnostic accuracy of TUS was significantly greater compared to LUS (p < 0.05; learning and test sample). Comparisons between ROC curves shown that TUS significantly improve the diagnosis of cardiogenic edema (p < 0.001; learning and test samples), pneumonia (p < 0.001; learning and test samples) and pulmonary embolism (p <0.001; learning sample). Conclusion.This study demonstrated for the first time a significantly better performance of TUS compared to LUS in the diagnosis of ARF. The value of TUS approach was particularly important to disambiguate cases of hemodynamic pulmonary edema and pneumonia. We suggest that the patient's bedside use of artificial intelligence methods in this setting, could pave the way for the development of new clinically relevant integrative diagnostic models.
    Chest 08/2014; · 7.13 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Postoperative acute kidney injury (AKI) is a cause of morbidity and mortality. Its diagnosis requires better markers than variations in diuresis or postoperative serum creatinine.
    European journal of anaesthesiology. 07/2014;
  • [Show abstract] [Hide abstract]
    ABSTRACT: Preoperative flushing of an anesthesia workstation is an alternative for preparation of the anesthesia workstation before use in malignant hyperthermia-susceptible patients (MHS). We studied in vitro, using a test lung, the washout profile of sevoflurane in 7 recent workstations during adult and, for the first time, pediatric ventilation patterns. Anesthesia workstations were first primed with 3% sevoflurane for 2 hours and then prepared according to the recommendations of the Malignant Hyperthermia Association of the United States. The flush was done with maximal fresh gas flow (FGF) with a minute ventilation equal to 600 mL × 15, to reach a sevoflurane concentration of <5 parts per million. After flush, 2 clinical situations were simulated in vitro to test the efficiency of preparation: decrease of FGF from max to 10 L/min, or decrease of minute ventilation to 50 mL × 30, to simulate the ventilation of an MHS infant. We report washout delays for MHS patients for previously studied workstations (Primus®, Avance®, and Zeus®) and more interestingly, for machines not previously tested (Felix®, Flow-I®, Perseus®, and Leon®). An increase of sevoflurane concentration was observed when decreasing FGF (except for flow-I® and Leon®) and during simulation of MHS infant ventilation (except for Felix®). This descriptive study strongly suggests that washout profiles may differ for each anesthesia workstation. We advise the use of maximal FGF during preparation and anesthesia. Required flushing times are longer when preparing an anesthesia workstation before providing anesthesia for MHS infants.
    Anesthesia and analgesia 05/2014; · 3.08 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Transcutaneous oxygen pressure (PtcO2) value in response to an increase of FiO2 or oxygen challenge test (OCT) in ventilated patients has been reported to be related to peripheral perfusion and outcome during septic shock. However, patients with sepsis related acute respiratory distress syndrome (ARDS) could demonstrate compromised arterial oxygenation with OCT impairment decoupled to circulatory failure. The aims of this study were to confirm the prognosis value of OCT, and to explore the influence of respiratory status on OCT results. Prospective study. Intensive care unit in tertiary teaching-hospital. Fifty six mechanically ventilated patients with septic shock criteria were studied. PtcO2 was measured at baseline and after OCT, at ICU admittance (T0) and 24h later (T24). Survival at day 28, hemodynamic and respiratory parameters were analyzed and compared according to outcome and respiratory status. Central hemodynamic parameters or static transcutaneous data did not differ between survivors and non survivors at enrolment. OCT was statistically different at T24 according to outcome (p<0.001) but sensitivity was low (53%). Moreover, patients with low OCT results at T24 exhibited more severe respiratory failure (p<0.01). OCT at T24 is related to outcome but is influenced by the severity of respiratory failure. Our results suggest considering with caution hemodynamic management based on OCT in septic shock patients with altered pulmonary function.
    Shock (Augusta, Ga.) 03/2014; · 2.87 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: The purpose of this study was to determine the best estimate of glomerular filtration rate (GFR) to adjust vancomycin (VAN) dosage in critically ill patients. Seventy-eight adult intensive care unit patients received a 15 mg/kg loading dose of VAN plus a 30 mg/kg/day continuous infusion. Steady-state concentration was measured 48 hours later and the dose was adjusted to obtain a target concentration ranging from 20 to 25 mg/l. GFR was estimated by measured creatinine clearance (CLCR), Cockcroft, Modification of Diet in Renal Disease and Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equations. The required dose providing the target concentration was 36±17 mg/kg/day. The first dosage had to be increased in 51% of all patients and in 84% of trauma patients (highest GFR), but had to be decreased in 17% of patients. The closest relationship between clearances of vancomycin was observed with CKD-EPI to GFR. The correlation between clearances of vancomycin and measured CLCR was significant but was rather poor with Cockcroft and Modification of Diet in Renal Disease equation. On the Bland and Altman plots, measured CLCR provided a lower bias but a larger confidence interval and a weaker precision than CKD-EPI. For VAN dose adjustments in intensive care unit patients, Cockcroft formula and Modification of Diet in Renal Disease should be used with caution. In clinical practice, the physician does not have at their disposal the patient's measured CLCR when prescribing. The CKD-EPI appears to be the best predictor of clearances of vancomycin for calculation of a therapeutic VAN regimen.
    Anaesthesia and intensive care 03/2014; 42(2):178-84. · 1.40 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Data from previous studies indicate that optimal conditions for intubation are met 120 seconds after administration of 0.15 mg.kg-1 cisatracurium (ED95 x 3) following the induction of anesthesia. The aim of this study was to compare the doses required for complete paralysis after induction of anesthesia in ICU patients with the dose used in patients undergoing elective surgery. Seventeen ICU patients undergoing percutaneous tracheostomy and 17 patients undergoing an elective surgical procedure under muscle relaxation were included. In both groups, an initial intravenous bolus of cisatracurium besylate was given at a dose of 0.15 mg.kg-1 followed by repeated boluses of 0.03 mg.kg-1 every four minutes. The objective was to obtain no response to the train-of-four (TOF). The contractile response of the corrugator supercilii muscle was monitored every minute by observing the TOF in response to a peripheral nerve stimulator with a constant current set to 60 mA. After the initial dose of cisatracurium, none of ICU patients (0/17) versus 15/17 of the elective surgery patients were completely paralyzed (P < 0.0001). There was a delay in the onset of neuromuscular blockade among the ICU patients. The cumulative doses of cisatracurium were significantly higher in the ICU group with 38 +/- 14 mg (that is, 10 +/- 4.7 ED95) versus 11 +/- 2 mg (that is, 3 +/- 0.3 ED95) in the elective surgery group (P < 0.0001). The dosing of cisatracrurium for ICU patients, which is based on the dose recommended for elective anesthesia, is unsuitable because the onset is too slow. This phenomenon is probably caused by changes in the pharmacodynamics and pharmacokinetics. These data suggest that neuromuscular monitoring should be used in the ICU.
    Annals of intensive care. 02/2014; 4(1):3.
  • Minerva anestesiologica 02/2014; · 2.82 Impact Factor
  • Annales francaises d'anesthesie et de reanimation 01/2014; · 0.77 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Dabigatran is a direct thrombin inhibitor indicated for stroke and systemic embolism prevention in patients with non-valvular atrial fibrillation. No reversal agent exists, but hemodialysis has been proposed as dabigatran removal method. We report a case of an 80-year-old man presenting hemorrhage with dabigatran overdose caused by obstructive acute renal failure. Before nephrostomy, several hemodialysis sessions were necessary to remove dabigatran probably because of its large volume of distribution.
    Annales francaises d'anesthesie et de reanimation 12/2013; · 0.77 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: The technology of anesthesia ventilators has substantially progressed during last years. The choice of a pediatric anesthesia ventilator needs to be led by multiple parameters: requirement, technical (pneumatic performance, velocity of halogenated or oxygen delivery), cost (purchase, in operation, preventive and curative maintenance), reliability, ergonomy, upgradability, and compatibility. The demonstration of the interest of pressure support mode during maintenance of spontaneous ventilation anesthesia makes this mode essential in pediatrics. In contrast, the financial impact of target controlled inhalation of halogenated has not be studied in pediatrics. Paradoxically, complex and various available technologies had not been much prospectively studied. Anesthesia ventilators performances in pediatrics need to be clarified in further clinical and bench test studies.
    Annales francaises d'anesthesie et de reanimation 11/2013; · 0.77 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: The management of cerebral perfusion pressure (CPP) is the one of the main preoccupation for the care of paediatric traumatic brain injury (TBI). The physiology of cerebral autoregulation, CO2 vasoreactivity, cerebral metabolism changes with age as well as the brain compliance. Low CPP leads to high morbidity and mortality in pediatric TBI. The recent guidelines for the management of CPP for the paediatric TBI indicate a CPP threshold 40-50mmHg (infants for the lower and adolescent for the upper). But we must consider the importance of age-related differences in the arterial pressure and CPP. The best CPP is the one that allows to avoid cerebral ischaemia and oedema. In this way, the adaptation of optimal CPP must be individual. To assess this objective, interesting tools are available. Transcranial Doppler can be used to determine the best level of CPP. Other indicators can predict the impairment of autoregulation like pressure reactivity index (PRx) taking into consideration the respective changes in ICP and CPP. Measurement of brain tissue oxygen partial pressure is an other tool that can be used to determine the optimal CPP.
    Annales francaises d'anesthesie et de reanimation 11/2013; · 0.77 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: We describe the case of a 19-year-old male diagnosed with Reye syndrome within the context of viral pericarditis and salicylate ingestion. He presented a fatal brain oedema without liver failure. Brain biopsies obtained during a decompressive craniectomy led to the diagnosis.
    Annales francaises d'anesthesie et de reanimation 10/2013; · 0.77 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: To evaluate the feasibility of guidewire detection in right cardiac cavities by transthoracic echocardiography (TTE) in order to detect catheter misplacement and to optimize central venous catheter (CVC) positioning. Ultrasonic control for catheter tip positioning was compared to that by chest X-ray (CXR). We conducted a monocentric prospective observational study (January-November 2010). All consecutive patients undergoing CVC insertion were included. The puncture was performed using the landmark method or ultrasound guidance. TTE was performed during the procedure to follow the arrival of the guidewire in the right cardiac cavities. Catheter misplacement was defined as an aberrant position on the postprocedural CXR (catheter positioning in ipsilateral or contralateral veins). The primary endpoint was the prediction of catheter misplacement by guidewire detection in the cardiac cavities. The secondary endpoint was the optimization of the catheter tip placement in the superior vena cava. A total of 98 patients received 101 CVC. The guidewire was visualized in 92 cases. In five cases, the guidewire was not seen in the right cardiac cavities and CXR showed catheter misplacement. In four cases, poor echogenicity led to the ultrasound examination being abandoned. Catheter misplacement was detected by TTE with a sensitivity of 96 % (CI 90-98 %), a specificity of 83 % (CI 44-97 %), a positive predictive value of 98 %, and a negative predictive value of 55 %. Likelihood ratios were LR+ 5.7 (CI 0.96-34.4) and LR- 0.05 (CI 0.02-0.14). Guidewire removal under TTE avoided an excessively distal position of the catheter tip in all cases. TTE is a reliable tool to detect catheter misplacement and to optimize catheter tip positioning during the procedure of CVC insertion.
    European Journal of Intensive Care Medicine 09/2013; · 5.17 Impact Factor
  • Annales Françaises d Anesthésie et de Réanimation 09/2013; 32:A330. · 0.84 Impact Factor

Publication Stats

1k Citations
401.59 Total Impact Points

Institutions

  • 1997–2014
    • Paul Sabatier University - Toulouse III
      Tolosa de Llenguadoc, Midi-Pyrénées, France
  • 2004–2013
    • University of Toulouse
      Tolosa de Llenguadoc, Midi-Pyrénées, France
    • Centre Hospitalier Universitaire de Toulouse
      • Service de Pharmacologie Clinique
      Tolosa de Llenguadoc, Midi-Pyrénées, France
  • 2009
    • Université Paris-Sud 11
      Orsay, Île-de-France, France
  • 1998
    • Unité Inserm U1077
      Caen, Lower Normandy, France
  • 1995
    • French Institute of Health and Medical Research
      Lutetia Parisorum, Île-de-France, France