Azathioprine is commonly used in Crohn's disease. It has been administered to many pregnant women over many years without significant side effects. However, pancytopenia and severe combined immune deficiency-like disease have been reported in infants whose mothers received azathioprine throughout pregnancy. Moreover, myelotoxicity has been described in patients being treated with azathioprine and having a low or absent thiopurine S-methyl transferase [TPMT] activity.Here, we describe the case of a newborn girl found to be highly lymphopenic [< 300 CD3+ T cells] after a positive newborn screening for severe combined immuno deficiency. The clinical examination was normal. The mother was treated with azathioprine throughout her pregnancy, without any reduction of the dose. It was shown that the mother was heterozygous for the 3A [TPMT] activity mutation and that the baby was homozygous for the same mutation; 6-thioguanine nucleotides were high (744 pmol/8.108 red blood cells [RBC]) in the mother and detectable in the infant [177 pmol/8.108 RBC].Although rare, this case illustrates the potential grave consequences of unsuspected TPMT homozygosity in a newborn of a mother receiving thiopurines during pregnancy. Because of the severity of the risk for the newborn, consideration should be given to performing maternal genetic testing and newborn routine blood count in cases of thiopurine treatment during pregnancy.

Aims Polypharmacy increase the risk of drug–drug interactions (DDIs) in the elderly population living with human immunodeficiency virus (HIV). Several expert databases can be used to evaluate DDIs. The aim of the study was to describe actual DDIs between antiretroviral drugs and comedications in an elderly population and to compare grading of the DDIs in 3 databases. Methods All treatments of HIV‐infected subjects aged 65 years and older were collected in 6 French HIV centres. Summary of Product Characteristic (SPC), French DDI Thesaurus (THES), and Liverpool HIV DDI website (LIV) were used to define each DDI and specific grade. DDIs were classified in yellow flag interaction (undefined grade in SPC and THES or potential weak interaction in LIV), amber flag interaction (to be considered/precaution of use in SPC and THES and potential interaction in LIV) and red flag interaction (not recommended/contraindication in SPC and THES and do not administer/contraindication in LIV). Results Among 239 subjects included, 60 (25.1%) had at least 1 DDI for a total of 126 DDIs: 23/126 red flag DDIs were identified in 17 patients. All these 23 DDIs were identified in LIV. THES and SPC missed 6 and 1 red flag DDIs, respectively. Seven of 23 red flag DDIs were identified in the 3 databases concomitantly. Conclusion Polypharmacy is frequent in this elderly HIV population leading to DDI in a quarter of the subjects. The discrepancies between databases can be explained by differences in analysis methods. A consensus between databases would be helpful for clinicians.

No abstract

Introduction : La notification spontanée des effets indésirables médicamenteux (EIMs) est un des piliers de la surveillance en post-AMM. Une des principales limites de cette dernière concerne la qualité de l’information de ces notifications. L’objectif principal de cette étude était d’évaluer l’informativité des observations d’EIMs enregistrées dans la Base Nationale de PharmacoVigilance (BNPV) à partir des déclarations des médecins généralistes (MGs). L’objectif secondaire était d’identifier les facteurs associés à une déclaration complète. Méthode : Toutes les observations d’EIMs déclarées en France, en 2015, par les MGs ont été extraites. Après analyse du contenu de ces déclarations (EIM, date d’apparition, description clinique, médicament imputé, etc.), l’informativité était analysée à partir de plusieurs critères obligatoires (âge, genre, EIM et médicament(s) suspecté(s)) et facultatifs (antécédents médicaux, médicaments concomitants, évolution de la symptomatologie et examens complémentaires) ; cela a permis de classifier les déclarations en « bien renseignée », « moyennement renseignée » et « mal renseignée ». Résultats : Durant l’année 2015, la BNPV comportait 3 020 déclarations réalisées par les MGs français. Seulement 16,4 % des déclarations de pharmacovigilance rapportées étaient « bien renseignées ». Les critères les moins bien documentés étaient les médicaments concomitants (41,4 %) et les examens complémentaires (37,4 %). Les observations complètes étaient associées à la gravité des cas (OR = 3,02 [IC 95 % 2,44 ; 3,23], p  < 10 –3 ) et aux personnes âgées par rapport aux adultes (OR = 1,76 [IC 95 % 1,42 ; 2,18], p  < 10 –3 ) ou aux enfants (OR = 4,59 [IC 95 % 2,51 ; 8,39], p  < 10 –3 ). Conclusion : Notre travail montre que seulement un peu plus d’une déclaration sur six est « bien renseignée ». Il apparaît important de promouvoir la pharmacovigilance afin d’améliorer l’informativité des déclarations.