L’objectif était d’évaluer les taux de surveillance métabolique biologique après initiation d’un antipsychotique de première (APG) ou de second génération (ASG) chez des personnes initialement traitées par lithium ou anticonvulsivants. Les données concernant les bilans glycémique et lipidique et les traitements médicamenteux ont été extraites de la base de données nationale 2004–2006 du Régime social des travailleurs indépendants. Ont été inclus les patients (n=3170) traités pendant trois mois consécutifs par lithium ou anticonvulsivants sans traitement antipsychotique concomitant. Une surveillance métabolique à l’initiation et au bout de trois mois d’un traitement antipsychotique à été réalisée chez 14 % des nouveaux usagers d’APG et 12 % des nouveaux usagers d’ASG. Les taux de surveillance étaient plus bas pour les bilans lipidiques que pour les bilans glycémiques. Les usagers d’ASG étaient plus susceptibles d’avoir une surveillance métabolique que les personnes sans antipsychotique. Les taux étaient similaires entre les usagers d’APG et d’ASG. Cette étude souligne l’écart entre les recommandations et les pratiques et la sous-évaluation par les prescripteurs des risques métaboliques liés à la prescription des ASG notamment en début de traitement.

To assess the rate of metabolic testing after initiation of second-generation antipsychotics (SGA) prescription in persons initially treated by conventional mood-stabilizers (lithium or anticonvulsants, as a proxy of bipolar disorder diagnosis) and to compare the rates of metabolic testing in these persons with those in persons with initiation of first-generation antipsychotics (FGA) prescription or with no antipsychotic prescription.

Data were anonymously extracted from the 2004 to 2006 French national health database of the Régime Social des Travailleurs Indépendants (RSI). Patients aged 18years and over were included in the cohort if they fulfilled the following criteria over a three-month inclusion period: refunding of lithium or anticonvulsant over the 3months without discontinuation (as a proxy of bipolar disorder diagnosis), no concomitant refunding over the 3months of antipsychotic, and no concomitant refunding over the 3months of an anti-diabetic drug (as a marker of diabetes) or a lipid-lowering drug (as a marker of hyperlipidemia). Metabolic testing was assessed using information collected in the RSI database on the reimbursement of glucose-specific serum tests (glycaemia) and lipid-specific serum tests (total cholesterol). Serum glucose and lipid testings were assessed at baseline and at 12-week follow-up for the first episode of antipsychotic dispensing. Multivariate analyses were performed to compare the rate of metabolic testing in users of SGA to those of users of FGA and to those of non-users of antipsychotics.

Three thousand one hundred and seventy patients were included. Of the 490 (15.4%) persons with a first episode of antipsychotic dispensing after the index date, 138 (4.3%) were dispensed only FGA over the first episode and 352 (11.1%) SGA (including 37 patients with both SGA and FGA dispensing). Metabolic testing at baseline and at 12-week follow-up was performed for 14% of persons with initiation of FGA and 12% with initiation of SGA. Almost no patient had both baseline and follow-up testing. Testing rates were lower for lipid testing than for glucose testing. Compared to persons with no antipsychotic, persons with SGA were significantly more likely to have metabolic testing at baseline and at follow-up, independently from other characteristics (adjusted OR=0.24, 95% CI 0.16 to 0.36). No difference was found between persons with SGA and those with FGA (adjusted OR=1.12, 95%CI 0.62 to 2.0). Regarding the other characteristics associated with likelihood of metabolic testing (irrespective of the treatment group), women were more likely than men to have metabolic testing at baseline but not at follow-up. Elderly persons and persons with low occupational status were more likely to have metabolic testing at follow-up.

From a public health point of view, such findings indicate that the metabolic risks associated with SGA use in real-life conditions are widely underestimated. Regarding the temporal trends of antipsychotic prescription, with the dramatic rise of SGA use observed in most countries, it is a public health priority to improve metabolic monitoring in SGA users, irrespective of the underlying diagnosis. Since it is more complex to modify pre-existing inadequate practices than to initiate correct ones in new prescribers, great attention should be paid to the need for delivering strong messages regarding the metabolic risks associated with SGA prescription during the initial training of physicians.