Le trouble de la personnalité borderline (TPB) est le trouble de la personnalité le plus fréquent parmi les patients psychiatriques. Pendant longtemps, les patients borderline ont été considérés comme particulièrement résistants aux traitements et prédestinés à une évolution chronique. Cette vision pessimiste demeure toujours assez répandue chez les cliniciens, même si les données empiriques les plus récentes la contredisent en montrant qu’une évolution favorable du TPB est non seulement possible, mais même fréquente. L’objectif de ce travail est de proposer une synthèse des études longitudinales disponibles en les présentant dans l’ordre chronologique de leur parution, ce qui permet de voir l’évolution des idées sur le TPB. Les premiers travaux ont cherché à savoir si le TPB pouvait évoluait vers la schizophrénie, ce qui a été démenti. Après l’introduction des critères TPB dans le DSM en 1980, les études longitudinales se sont intéressées principalement à la stabilité et à la spécificité du diagnostic TPB. Enfin, les recherches prospectives les plus récentes, appliquant une méthodologie rigoureuse, ont exploré plusieurs aspects évolutifs en mettant en évidence l’évolution des différents symptômes. Malgré une réduction importante de la symptomatologie avec le temps, certains symptômes se résolvent relativement rapidement, tandis que d’autres ont tendance à persister.

Borderline personality disorder (BPD) is a serious mental disorder associated with severe emotional, behavioral, cognitive and interpersonal dysfunction, extensive functional impairment and frequent self-destructive behaviour, including deliberate self-harm and suicidal behaviour. For quite some time, BPD has been viewed as a chronic disorder and borderline patients as extremely difficult to treat, doomed to a life of misery. However, those views are changing and there is an increasing recognition that BPD has a far more benign course than previously thought. The purpose of this study is to show how those views changed over time by reviewing longitudinal studies of the course of BPD.

We have reviewed the literature published from 1968 to March 2009, using the following key words: borderline personality disorder, outcome, follow-up studies with some additional references.

The aim of the longitudinal studies conducted prior to the DSM definition of BPD criteria was to determine whether borderline patients could become psychotic over time, but no such evidence was found even though their functioning was at a relatively low level. The studies conducted after the introduction of BPD in the DSM in 1980 tested the stability and the specificity of BPD diagnosis, concluding that the criteria were relatively stable in the short run since the majority of patients continued to meet them at the follow-up assessments. However, those studies had many methodological drawbacks which limited their generalizability such as small sample sizes, high attrition rates, the absence of comparison groups, etc. Four retrospective studies of the 15-year outcome of borderline patients obtained virtually identical results despite methodological differences, showing that the global functioning of borderline patients improved substantially over time with mean scores of the GAF scale falling within a mild range of impairment. One 27-year retrospective study showed that borderline patients continued to improve as they grew older, only 8% of the cohort still meeting criteria for BPD. Two recent carefully designed prospective studies showed that the majority of BPD patients experienced a substantial reduction in their symptoms far sooner than previously expected. After six years, 75% of patients diagnosed with BPD severe enough to be hospitalized achieve remission by standardized diagnostic criteria and after 10 years, the remission rate raises up to 88%. Recurrences are rare, no more than 6% over six years. The dramatic symptoms (suicidal behaviour, self-mutilation, queasy psychotic thoughts) resolve relatively quickly, but abandonment concerns, feeling of emptiness and vulnerability to dysphonic states is likely to remain in at least half the patients.

This contrasts with the natural course of many Axis I disorders, such as mood disorders, where improvement rates may be somewhat higher and more rapid but recurrences are more frequent. The findings of longitudinal studies raise doubts about the validity of the definition in the DSM, which implies that personality disorders must necessarily be chronic. However, it should be noted that even the most encouraging findings do not show full recovery since the majority of patients seem to suffer from some residual symptoms.

These findings have very important clinical implications and borderline patients should be told that they can expect improvement, no matter how intense their current emotional pain. However, we still lack evidence-based findings on mechanisms that lie behind the recovery process in BPD. Future research should explore the mechanisms of recovery in BPD.