En dépit du caractère récidivant et invalidant du trouble bipolaire, peu d'études récentes ont évalué le poids économique de cette affection. L'objectif de cette étude a été d'évaluer la prévalence annuelle des épisodes maniaques au cours du trouble bipolaire I en France et le coût des hospitalisations effectuées pour ce motif. La fréquence annuelle des épisodes maniaques a été estimée à partir des données épidémiologiques publiées. Les données hospitalières (pourcentage des épisodes de manie hospitalisés et durée moyenne des séjours associés) ont pu être estimées à partir des statistiques du Département d'Information Médicale de l'Hôpital psychiatrique Sainte-Anne à Paris qui représente l'un des plus importants établissements psychiatriques de la région parisienne en termes d'activité. L'extrapolation des données internationales à la France a permis d'estimer le taux moyen de prévalence du trouble bipolaire I, le pourcentage de patients cycleurs rapides et une durée moyenne de cycles. Ces estimations ont permis d'évaluer à environ 265 000 le nombre total d'épisodes maniaques survenant chaque année en France. Selon les données de l'Hôpital Sainte-Anne à Paris pour l'année 1999, 63 % des épisodes de manie ont motivé une hospitalisation. La durée moyenne de ces séjours a été de 32,4 jours. Le coût annuel des hospitalisations associées s'élève à 1,3 milliard d'euros environ. Cette revue de la littérature montre que l'impact médico-économique du trouble bipolaire est peu documenté, notamment en France où les études épidémiologiques de portée nationale font défaut. Ces études apparaissent, cependant, nécessaires pour un trouble dont la prévalence vie entière peut atteindre 3 % de la population adulte.

Bipolar disorder is a chronic, highly disabling illness. However, few studies have evaluated the economic impact of this illness. The objective of this study was to estimate : 1) the annual number of manic episodes in patients with bipolar I disorder, and 2) the costs of hospitalisations related to manic episodes in France. We only included data on bipolar I disorder, as there is greater consensus and better documentation for this subgroup of patients with bipolar di[shy]sorder. The prevalence of manic episodes was estimated using published epidemiological data. A computerised litera[shy]ture search was performed using the traditional scientific and medical databases. Additional epidemiological references were identified from published studies and textbooks. For hospitalisation data, we used the statistics of the Medical Information Department of a large psychiatric hospital in Paris for the year 1999. We estimated the annual number of manic episodes in France based on : 1) the lifetime prevalence of bipolar I disorder, 2) the average cycle duration, 3) the proportion of rapid cycling patients, and 4) the proportion of depressive vs. manic episodes for patients with bipolar I di[shy]sorder. In order to estimate the prevalence of bipolar I disorder, we conducted a random effects meta-analysis using published international data. Results of the meta-analysis, which was based on a total of 62 736 patients, showed the lifetime prevalence of bipolar I disorder to be 0.82 % [95 % CI : 0.42, 1.21]. Applied to the adult population in France, this prevalence implies that the number of persons who have ever experienced a bipolar I [shy]disorder is approximately 390,000 [95 % CI : 200,000, 575,000]. Few studies provide information on the duration of cycles in patients with bipolar I disorder. Available estimates suggest the cycle duration to be approximately 12 months. Regarding the proportion of rapid cyclers, data from the meta-analysis by Tondo et al. show that 18 % of patients with bipolar disorder experience at least four episodes of mood disorder per year. Finally, based on findings provided by cohort studies, the number of depressive episodes appears to be roughly equal to the number of manic episodes during the course of bipolar disorder. A rapid cycling rate of 18 % and a cycle duration of 12 months imply that, on average, among 100 bipolar patients, 18 will have a 3-month cycle duration and 82 a 14-month cycle duration. Given an equal proportion of manic and depressive episodes, the annual number of manic episodes would then be 68 for a cohort of 100 bipolar patients (0.68 episode per patient per year). Applying this figure to the estimate of the total number of patients with bipolar I disorder in France suggests that the annual number of manic episodes in France is 265,000 [95 % CI : 136,000, 391,000]. Based on data from a psychiatric hospital in Paris, the proportion of manic episodes that require hospitalisation was estimated to be around 63 % with an average length of stay of 32.4 days. Hence the annual number of hospitalisations for manic episodes in France is estimated to be 167 000 [95 % CI : 86 000, 246 000] and the hospitalisation-related costs 1,3 billion euros approximately. Our review of literature highlights the lack of medical and economic data at the national level on the frequency and hospitalisation-related costs of manic episodes in patients with bipolar I disorder in France. Given the lifetime prevalence of bipolar I disorder which may be as high as 3 % among adults, further studies are required in order to provide representative national data and to allow economic evaluations of costs related to bipolar disorder in France.