Stéphanie Badiou

Centre Hospitalier Universitaire de Montpellier, Montpelhièr, Languedoc-Roussillon, France

Are you Stéphanie Badiou?

Claim your profile

Publications (76)201.11 Total impact

  • [Show abstract] [Hide abstract]
    ABSTRACT: Monoclonal gammopathies (MG) associated with HIV infection are frequent but their evolution and significance are uncertain in this population at high risk of lymphoproliferative disorder. Our aim was to describe the long term evolution of MG in HIV-infected subjects under antiretroviral therapy (ART). Retrospective study of HIV-1 infected adults, with a monoclonal (M) protein detected by serum protein electrophoresis (SPE) and confirmed by immunofixation. Logistic regression was used to analyze factors associated with peak disappearance. Between September 1997 and November 2012, 1219 EPS were performed on our HIV cohort, and 137 (11.3%) MG were detected. Seventy seven subjects met the inclusion criteria: 68% male, median age 41 years, 47% AIDS stage, median CD4 count 237/mm, 81% uncontrolled HIV infection with HIV viral load over 400 copies/mL, 32% chronic hepatitis C and 9% chronic hepatitis B. Eighteen subjects were not included because of previous or concomitant hemopathy. With a median follow-up of 6.8 years (IQR 3.9-9.1), 66.2% of subjects showed a peak disappearance. In multivariate analysis, MG disappearance was associated with HIV-virologic control (OR 5.98, CI95% [1.63;21.87], p 0.007) and absence of hepatitis C virus replication at end of follow-up (OR 10.16, CI95% [2.36;43.69], p 0.002). One subject developed a myeloma three years after the diagnosis of an IgA kappa MG. MG associated with HIV infection concerned a young population, and had favorable evolution on ART in most cases. M protein disappearance was associated with HIV-virologic control and absence of chronic viral C hepatitis.
    JAIDS Journal of Acquired Immune Deficiency Syndromes 06/2015; DOI:10.1097/QAI.0000000000000708 · 4.56 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: The use of reference change value (RCV) instead of reference interval emerged as an alternative approach for longitudinal interpretation of biological marker. Follow-up of creatinine variation in HIV-positive adults remains a challenge in order to prevent renal complications. To determine the long term RCV of creatinine in HIV-positive adults receiving anti-retroviral therapy (ART) according to the use of tenofovir or ritonavir. Longitudinal study of 24months that include 124 HIV-positive patients followed in HIV outpatient unit. Plasma creatinine was measured at 0, 6, 12 and 24months in order to calculate the RCV. In the whole group, a 24-month RCV of creatinine was 22.5%. Whatever the ART, the index of individuality was <0.6. Significantly higher RCV of creatinine was observed in patients receiving the association tenofovir and ritonavir (28%) compared to the patients receiving i) tenofovir without ritonavir (21.9%), ii) no tenofovir but ritonavir (22.2%), and iii) no tenofovir and no ritonavir (19.7%). The low value of index of individuality pinpointed that RCV should be used to identify critical change in serial creatinine results in HIV-positive adults. RCV of creatinine under ART was around 20% but reached 28% in case of association of tenofovir and ritonavir. Copyright © 2015. Published by Elsevier Inc.
    Clinical Biochemistry 01/2015; 48(9). DOI:10.1016/j.clinbiochem.2015.01.002 · 2.28 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: We sought to evaluate the added value of Heart Fatty Acid Protein assay (HFABP) for rapid diagnosis of acute myocardial infarction in a prospective cohort of emergency department (ED) patients with acute chest pain.Methods High-sensitivity cardiac troponin T (hs-cTnT, Roche Diagnostics, Meylan, France) and HFABP (Randox, Mauguio, France) were blindly assayed from venous blood samples obtained at admission. Diagnosis was made by 2 ED physicians using all available data and serial cardiac troponin I as the biochemical standard. Diagnostic performances of HFABP combined with hs-cTnT were assessed using logistic regression. Analysis was conducted in all patients and in patients without ST-elevation myocardial infarction.ResultsA total of 181 patients were included (age, 61 ± 17 years;male sex, 66%). Acute myocardial infarction occurred in 47 (25.9%) patients, including non–ST-elevation myocardial infarction (NSTEMI) in 31 (17.1%). The ROC AUC was 0.893 for hs-cTnT levels at presentation (95% CI, 0.812-0.974) and 0.908 (95% CI, 0.839-0.977) for the combination of hs-cTnT and HFABP, with no significant difference (p = 0.07). Adding HFABP to hs-cTnT increased both sensitivity and NPV for NSTEMI diagnosis, with about 13% and 3% increase, respectively, leading to a sensitivity of 97% and a NPV of 99%.Conclusion The assessment of HFABP at ED admission adds incremental value to initial hs-cTnT. The increase of sensitivity and NPV without sacrificing the specificity and PPV in patients with chest pain with non contributive ECG could potentially allow safe and early rule out of AMI without the need for further serial troponin testing.
    American Journal of Emergency Medicine 11/2014; 33(3). DOI:10.1016/j.ajem.2014.11.012 · 1.27 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: This study aimed to evaluate a new ion-exchange HPLC system to measure hemoglobin A1c (HbA1c) in comparison to two other widely used HPLC systems. Analytical performance and hemoglobin variants detection was assessed on the new Arkray/ELITech ADAMS HA-8180V®, in parallel to Bio-Rad Variant(TM) II Turbo2.0® and Tosoh Bioscience HLC®-723G8. A method comparison and concordance for the classification of patients according to the American Diabetes Association (ADA) categories was performed using kappa test. ADAMS HA-8180V® demonstrated excellent within-run imprecision (0%) and between-run imprecision: ≤1.21% using blood sample and ≤0.94% using quality controls. Method comparison of ADAMS HA-8180V® with the two other systems yielded very high coefficient correlation (r > 0.995). A very good agreement (kappa value ≥0.81) was observed between methods for the classification of patients according to the ADA categories. In addition, all systems were able to detect the presence of most classical variants such as HbC, HbS, HbD, and HbE. The Arkray/ELITech ADAMS HA-8180V® demonstrated high analytical performance similar to previous systems such as Biorad(TM) Variant II Turbo 2.0® and Tosoh Bioscience HLC®-723G8. The three systems allow a high-quality HbA1c measurement and appeared interchangeable for diabetes diagnosis or for the therapeutic monitoring of patients without hemoglobin variants.
    Journal of Clinical Laboratory Analysis 11/2014; 28(6). DOI:10.1002/jcla.21705 · 1.04 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Objectives The study was designed to evaluate the performance of five automated immunoturbidimetric assays to quantify urinary albumin, each corresponding to a combination of a reagent and an analyzer (Olympus on AU640®, Roche on Cobas Integra, Abbott on Architect, Ortho-Clinical Diagnostics Vitros on Fusion and Siemens on Dimension). Design and methods To assess imprecision, albumin was measured in three urinary pools with a mean value of 25, 66 and 131 mg/L. One hundred and eight patient urine samples were then used to compare each turbidimetric method using Passing-Bablok regression and Bland-Altman analyses. Concordance of the albumin / creatinine ratio (ACR), according to the albuminuria classifications proposed by the KDIGO, was calculated to test the agreement between the different assays. Results All immunoturbidimetric methods evaluated in this study exhibited acceptable imprecision (CV < 6%). Mean values for 108 urine samples varied from 0.5 to 762.2 mg/L. Significant differences were found (p < 0.05) between all methods except between Olympus and Ortho (p = 1.0) and between Abbott and Roche (p = 0.12). Regarding the albuminuria categories based on the ACR proposed by the KDIGO, only the classification obtained with the Roche method was significantly different from the four other methods (p < 0.001). Conclusions We demonstrated that all assays were not strictly equivalent which could affect ACR categories in clinical practice, suggesting the need for harmonization of commercial methods.
    Clinical Biochemistry 11/2014; 47(16-17). DOI:10.1016/j.clinbiochem.2014.07.014 · 2.28 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Objectives To report the analytical performances of the Radiometer AQT90 FLEX® cTnT assay (Neuilly-Plaisance, France) and to evaluate the concordance with hs-cTnT results from central laboratory for the diagnosis of acute myocardial infarction (AMI) at baseline and during a short follow-up among unselected patients admitted in emergency room or cardiology department. Design and Methods Analytical performances of AQT90 FLEX® cTnT immunoassay included imprecision study with determination of a coefficient of variation at 10% and 20%, linearity, and limit of detection. The concordance study was based on samples obtained from 170 consecutive patients with chest pain suggestive of acute coronary syndrome (ACS) admitted in the emergency room or cardiology department. The kinetic study (within 62 additional samples 3 hours later) was based on absolute delta criterion and the combination of relative change of 30% with absolute change of 7 ng/L. Results The cTnT assay from Radiometer was evaluated as clinically usable, although less sensitive than the Roche hs-cTnT assay as demonstrated by the concordance and the kinetic studies. Conclusions In non selected population, the cTnT AQT Flex© assay on AQT90© with kinetic change at 3 hours, provides similar clinical classification of patients, particularly for AMI group as compared to central laboratory hs-cTnT assay and could be suitable for clinical use.
    Clinical Biochemistry 10/2014; 48(4-5). DOI:10.1016/j.clinbiochem.2014.09.027 · 2.28 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Background: A new particle-enhanced turbidimetric immunoassay, Quantitative Microsphere System (QMS) everolimus, was evaluated using an Indiko analyzer (Thermo Fisher Scientific). Methods: Analytical performances (imprecision, linearity, limit of detection, and limit of quantification) of this new immunoassay were evaluated. The method was compared with an ultra-performance liquid chromatography-tandem mass spectrometry (UPLC/MS/MS) method and with the previously available Innofluor Certican immunoassay on 120 whole-blood samples from 74 transplant recipients. Passing-Bablok regression and Bland-Altman plot were used for method comparisons. Results: Within- and between-day coefficients of variation were <10% at mean levels of 4.3, 8.2, and 15.9 ng/mL, respectively. This assay could be used at everolimus concentrations between 1.7 and 20 ng/mL. We obtained a Passing-Bablok regression of y = 0.99 × -0.15 (r = 0.95) when comparing with the UPLC/MS/MS method and of y = 0.81 × -0.09 (r = 0.94) when comparing with Innofluor assay. The mean and limits of agreement (mean ± 1.96 SD) of the difference between QMS everolimus and UPLC/MS/MS were -0.13 (from -2.0 to 1.74) ng/mL and were -1.19 (from -3.39 to 1.01) ng/mL between QMS everolimus and Innofluor. Conclusions: The QMS everolimus method on Indiko analyzer demonstrated reliable and reproducible performances allowing routine monitoring in transplant patients.
    Therapeutic drug monitoring 09/2014; 37(2). DOI:10.1097/FTD.0000000000000133 · 2.38 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Cardiac troponin level measured by high-sensitivity assays (hs-cTn) in the elderly is frequently found higher than the 99th percentile upper reference limit, making the diagnosis of acute coronary syndromes (ACS) difficult. This study aimed at: 1) identifying determinants of hs-cTnT levels in an unselected population of elderly subjects; and 2) assessing the prognosis value of increased hs-cTnT in elderly people free of ACS. Hs-cTnT was measured in 591 individuals aged over 65 years without suspicion of ACS. Comorbidities were assessed using the Cumulative Illness Rating Scale for Geriatrics (CIRS-G). C-reactive protein, α Age, gender, cardiac CIRS-G, estimated glomerular filtration rate (p
    Clinical Chemistry and Laboratory Medicine 08/2014; 53(5). DOI:10.1515/cclm-2014-0121 · 2.71 Impact Factor
  • Médecine et Maladies Infectieuses 06/2014; 44(6):27-28. DOI:10.1016/S0399-077X(14)70118-0 · 1.24 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Rheumatoid factor (RF) consists of autoantibodies and because of its heterogeneity its determination is not easy. Currently, nephelometry and Elisa method are considered as reference methods. Due to consolidation, many laboratories have fully automated turbidimetric apparatus, and specific nephelemetric systems are not always available. In addition, nephelemetry is more accurate, but time consuming, expensive, and requires a specific device, resulting in a lower efficiency. Turbidimetry could be an attractive alternative. The turbidimetric RF test from Diagam meets the requirements of accuracy and precision for optimal clinical use, with an acceptable measuring range, and could be an alternative in the determination of RF, without the associated cost of a dedicated instrument, making consolidation and saving blood possible.
    Clinical laboratory 04/2014; 60(4):685-8. DOI:10.7754/Clin.Lab.2013.130515 · 1.13 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: As the implementation of POC analyzers in clinical department is growing and HbA1c has been included in diagnostics, it appears mandatory to evaluate the accuracy of the POC method to be used. The aim of our study is to evaluate the analytical performance and usefulness of the Afinion analyzer, and the impact of HbF, hemoglobin variants, and changes of reagent lots. The Afinion analyzer (Axis-Shield, Oslo, Norway) is based on affinity separation methodology. Analytical performances including imprecision studies and correlation with currently used HPLC were performed on Biochemistry laboratory. We evaluated the effects of presence of variants on the results of HbA1c. Two reagent lots were evaluated on samples without variant as well as on samples with in the presence of hemoglobin variants HbA/S. The practicability of the system was also tested installing the instrument on pediatric consultation. Imprecision studies of Afinion were acceptable. Linear regression analysis indicated good correlation between HbA1c values with Afinion (r > 0.97) in comparison with HPLC method. Overall, the Afinion overestimated the HbA1 compared to the MenariniHA8140. These differences were confirmed when the estimated average glucose was calculated. In addition, the variance of cartridge lots requires its use with caution and the problem of lot number instability must be resolved, requiring an adjustment of the calibration by the manufacturer. The Afinion analyzer met the acceptance criteria of analytical performance, requires minimal operator interaction and presents good correlation with the laboratory method. Beyond the analytical performance, the different regression lines for different lots is a matter of concern. The reproductibility of the production of the different reagent lots for POC devices appears inadequate for its use in a clinical department.
    Clinical laboratory 04/2014; 60(3):369-76. DOI:10.7754/Clin.Lab.2013.130209 · 1.13 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Until now, specific proteins have traditionally been analyzed by immunonephelometry requiring spe- cific equipment. Due to laboratory consolidation, turbidimetry is commonly used for the determination of specific proteins. Biochemistry analyzers have a wide range of turbidimetric assays available to a single workstation, which makes them particularly efficient. For our laboratory consolidation, we installed two distinct Cobas8000® modular analyzer series. However, the turbidimetric albumin from Roche did not give satisfactory results in re- gard to the traceability to the reference measurement material. Analytical assays of turbidimetric albumin from Diagam (Liège, Belgique) on c502/Cobas8000®, includ- ing imprecision studies, accuracy using the ERM®-DA470k/IFCC, and correlation with nephelometry and colo- rimetry, were performed. Total precision for the immunoturbidimetric assays was consistently better than 2.3% CV and linear throughout the dynamic range of the assays. Correlation of serum albumin by turbidimetry is in good agreement with nephelometry. The mean ± SD of the reference material ERM-DA470k was 37.6 ± 0.25 g/L, with a bias of 1.07%. Turbidimetric albumin from Diagam meets the requirements of accuracy and precision for optimal clinical use and could be an alternative without the associated cost of a dedicated instrument in the context of lab- oratory consolidation.
    Clinical laboratory 01/2014; 60(10):1769-73. DOI:10.7754/Clin.Lab.2014.131120 · 1.13 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Transplant patients have an increase incidence of lymphoproliferative disorders. The development of sensitive immunoassay for serum free light chains (FLC) determination has improved the diagnosis of monoclonal gammapathy. However, metabolism of FLC largely depends on renal function which could lead to misinterpretation of results. The aim of this study was to assess polyclonal FLC in kidney transplant recipients in relation to markers of renal function. 97 renal transplant patients (median transplantation follow-up of 2.2years) were included in this study. Kappa FLC, Lambda FLC and cystatin C were measured by particle-enhanced turbidimetric immunoassay (PETIA) using kits from Binding Site (Birmingham, UK) on SPAPLUS® analyzer. Glomerular filtration rate was measured using urinary clearance of 99mTc-DTPA and estimated using CKD-EPI creatinine 2009, CKD-EPI cystatin C 2012 and CKD-EPI creatinine-cystatin C 2012. Sera kappa FLC was 22.38 [6.39 - 68.91] mg/L and sera lambda FLC was 15.66 [7.29 - 50.81] mg/L. Both levels of Kappa and Lambda were correlated with renal function. Median K / L ratio was 1.34 [0.57 to 2.67] leaving 19.6% of renal transplant patients outside the normal reference range ([0.26-1.65]). However, using a reference interval normalized by renal function ([0.37-3.1]), none of our patients was classified as abnormal. K / L ratio should be cautiously analyzed in renal transplantation and renal function, estimated by creatinine/cystatin C equations should be taken into account.
    Clinical biochemistry 09/2013; 46(16-17). DOI:10.1016/j.clinbiochem.2013.08.024 · 2.28 Impact Factor
  • Source
    Journal of diabetes science and technology 09/2013; 7(5):1408-1409. DOI:10.1177/193229681300700534
  • [Show abstract] [Hide abstract]
    ABSTRACT: High prevalence of monoclonal gammopathy (MG) is observed in HIV-infected patients. We explored the conditions associated with long term persistence of serum monoclonal protein (M protein) in HIV-infected patients on antiretroviral therapy (ART). Of 21 patients with monoclonal gammopathy, M Protein disappeared in 12 patients (58%) over 5 years of ART. Higher level of serum gamma-globulin and higher percentages of circulating plasmablasts and plasma cells were observed in patients with persistent MG in comparison to patients with transient MG. MG persistence was associated with the cumulative time of detectable plasma HIV RNA after ART initiation, detection of EBV DNA in plasma and high level of EBV DNA in B cells. Poor control of HIV replication and detectable EBV replication in plasma were both associated with long-term MG persistence in patients on ART. In case of viral control, MG associated with HIV infection is usually transient.
    Blood 08/2013; 122(17). DOI:10.1182/blood-2012-12-470393 · 10.45 Impact Factor
  • Clinica chimica acta; international journal of clinical chemistry 07/2013; 425. DOI:10.1016/j.cca.2013.06.026 · 2.82 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: We sought to evaluate the added value of ultrasensitive copeptin (us-copeptin) for early rule out of acute myocardial infarction in a prospective cohort of emergency department (ED) patients with acute chest pain. This was a prospective study including consecutive patients with acute chest pain presenting to the ED within 12 hours of symptom onset. High-sensitivity cardiac troponin T (hs-cTnT, Roche Diagnostics, Meylan, France) and us-copeptin (ThermoFisher Scientific, Clichy, France) were blindly assayed from venous blood samples obtained at admission. Diagnosis was made by 2 ED physicians using all available data and serial cardiac troponin I as the biochemical standard. Diagnostic performances of us-copeptin combined with hs-cTnT were assessed using logistic regression. Analysis was conducted in all patients and in patients without ST-elevation myocardial infarction. A total of 194 patients were included (age, 61 [48-75] years; male sex, 63%). Acute myocardial infarction occurred in 52 (27%) patients, including non-ST-elevation myocardial infarction (NSTEMI) in 25 (13%). Patients with acute myocardial infarction had higher levels of hs-cTnT (50 [95% confidence interval, 19-173] ng/L) and us-copeptin (30 [13-113] pmol/L) at admission compared with those without (P < .05). Combination of markers significantly improved receiver operating characteristic area under the curve (from 0.89 [0.85-0.92] for hs-cTnT alone to 0.93 [0.89-0.97], P = .018). Sensitivity and negative predictive value were increased, particularly for NSTEMI diagnosis (sensitivity, 76% [54.9-90.6] to 96% [79.6-99.9]; negative predictive value, 95% [90.4-98.3] to 98.9% [94.2 to 100]). Assessment of us-copeptin combined with hs-cTnT on ED admission could allow safe and early rule out of NSTEMI for patients with negative results on both markers and help identify patients who may be suitable for discharge.
    The American journal of emergency medicine 06/2013; 31(9). DOI:10.1016/j.ajem.2013.04.033 · 1.27 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Objectives: Analytical and clinical performances of the new standardized cystatin C particle-enhanced turbidimetric immunoassay (PETIA) using DiaSys reagents on Olympus AU2700® analyzer were evaluated. Design and methods: We have studied imprecision, linearity, limit of detection and limit of quantification of this new immunoassay. Method comparison was assessed in relation to results generated by the standardized Siemens-particle-enhanced nephelometric immunoassay (PENIA). In order to evaluate the clinical relevance of this assay, estimated glomerular filtration rate (GFR) was calculated using MDRD, CKD-EPI creatinine, CKD-EPI cystatin C 2012 and CKD-EPI creatinine-cystatin C 2012 equations and compared to GFR measured using urinary clearance of (99m)Tc-DTPA in 100 hypertensive patients. Results: Cystatin C measurements using DiaSys reagents have reliable analytical performances and are comparable to the standardized Siemens-PENIA method (bias of 0.01 mg/L). The mean measured GFR was 90.0±29.7 mL/min/1.73 m². Bias and accuracy of the three CKD-EPI equations were better than the MDRD. Both CKD-EPI creatinine-based and cystatin C-based formulae had similar bias, precision and accuracy. The combined creatinine-cystatin C equation was significantly more accurate and precise than the CKD-EPI creatinine equation in patients with GFR above 60 mL/min/1.73 m². Conclusions: The use of cystatin C in a combined equation with creatinine could improve the accuracy of eGFR in the reference interval.
    Clinical biochemistry 05/2013; 46(15). DOI:10.1016/j.clinbiochem.2013.05.056 · 2.28 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: To assess the frequency of glomerular and tubular proteinuria in a cohort of HIV-infected patients, and to determine the factors associated with each type of injury. Cross-sectional study of 1210 consecutive HIV-infected adults followed in HIV outpatient unit (Montpellier/France). Spot urine protein to creatinine (uPCR), albumin to creatinine (uACR) and albumin to protein (uAPR) ratios were assessed. Glomerular injury was defined as uACR at least 30 mg/g or uPCR at least 200 mg/g with uAPR at least 0.4. Tubular injury was defined as uPCR 200 mg/g or more with uAPR less than 0.4. Multivariate logistic regression identified independent factors of each type of proteinuria, in the 1158 patients with estimated glomerular filtration rate (eGFR) at least 60 ml/min per 1.73 m, using re-expressed modification of diet in renal disease equation. Frequency of proteinuria was 18.2% among patients with eGFR at least 60 ml/min per 1.73 m consisting in tubular proteinuria for 50.7% of them. Factors associated with glomerular proteinuria were age [OR 1.34/10-year increment (95%CI: 1.08-1.66)], diabetes [OR 3.37 (95%CI: 1.53-7.44)], and arterial hypertension [OR 2.52 (95%CI: 1.36-4.66)]. Factors associated with tubular proteinuria were age [OR 1.43 (95%CI: 1.14-1.79)], current tenofovir use [OR 3.52 (95%CI: 1.86-6.65)], hepatitis C co-infection [OR 1.62 (95%CI: 1.00-2.65)], AIDS stage [OR 1.83 (95%CI: 1.18-2.82)], CD4 cell count less than 200 per μl [OR 2.48 (95%CI: 1.31-4.70)]. This study distinguished risk factors for tubular injury, mainly related to HIV disease and its treatment (tenofovir), and glomerular injury, linked to non HIV-related variables (age, diabetes, hypertension). Measuring uPCR, uACR and uAPR may help with the detection and specific management of early chronic kidney disease in HIV-infected patients having normal or sub-normal eGFR.
    AIDS (London, England) 05/2013; 27(8):1295-302. DOI:10.1097/QAD.0b013e32835fac51 · 5.55 Impact Factor

Publication Stats

602 Citations
201.11 Total Impact Points


  • 2013
    • Centre Hospitalier Universitaire de Montpellier
      • Département de biochimie
      Montpelhièr, Languedoc-Roussillon, France
  • 2008–2013
    • Université de Montpellier 1
      Montpelhièr, Languedoc-Roussillon, France
    • University of Bordeaux
      Burdeos, Aquitaine, France
  • 2010–2011
    • Université Montpellier 2 Sciences et Techniques
      Montpelhièr, Languedoc-Roussillon, France
    • Institute of Research for Development
      Marsiglia, Provence-Alpes-Côte d'Azur, France
  • 2009
    • Université de Bouaké
      Bwake, Vallée du Bandama, Ivory Coast