Mary Zanarini and her colleagues at McLean Hospital have now followed for a decade nearly 300 former inpatients with borderline personality disorder. For their prospective study reported in this issue of the Journal (1), they reinterviewed these patients at 2-year intervals over 10 years in an effort to answer three highly important questions about borderline personality disorder: How many of them eventually become well? How long did their progress to remission take? And how many remained stably well without succumbing to another episode of illness?

The criterion Zanarini et al. used for wellness included both symptomatic remission and attainment of good social and vocational functioning for at least 2 years. For symptom remission, the results of the McLean study are most encouraging: nearly seven of every eight borderline patients achieved symptom remission lasting at least 4 years. Half the patients achieved a more total recovery, no longer meeting diagnostic criteria for borderline personality disorder as well as achieving good social and occupational functioning. Although the authors preface that proportion with "only," for half the patients to attain that level of improvement is impressive indeed, especially within a 10-year follow-up period.

Prospective studies are more elegant and accurate than the retrospective studies that inspired them, but many of the patients in the larger retrospective studies of the 1980s (the Chestnut Lodge study [2] and the New York Psychiatric Institute study [3]) showed similar levels of total remission, albeit with an extra 15 years in which to scale that height.

The tendency of the borderline patients to show symptom recovery sooner than resolution of what Zanarini et al. (4) have called "temperamental" aspects (intense anger, abandonment concerns) is quite in keeping with the results of these other investigations. Certain symptoms of borderline personality disorder—self-mutilation, suicide gestures and attempts, impulsive acts like shoplifting and careless sexual encounters—are easier to remediate with medication or psychotherapy, or a combination of the two; good results often become discernible within a year or two (5). Conceptually, borderline personality disorder is really an admixture of symptoms, true personality traits, and cognitive peculiarities, such as identity disturbance. Personality traits are largely ego-syntonic and enduring, however unwelcome and problematical they may be (e.g., introversion, extraversion, sociability, hostility, generosity, stinginess). It is no surprise that some of the traits typical of borderline patients—irritability, moodiness, demandingness, manipulativeness, mercuriality—persist long after the symptoms of the disorder have abated. In an important cautionary note from an earlier paper, however, Zanarini et al. (6) mention that substance abuse represents a symptom that does not abate so quickly and has more ominous prognostic qualities, associated as it is with a strong tendency for failure of remission in borderline personality disorder and with suicidality.

One of the most formidable impediments to outcome research in borderline personality disorder is the broad diversity of potential contributing factors. This heterogeneity of "cause" stems from such factors as family history of emotional illness; birth complications; early verbal, physical, or sexual abuse (which may vary in onset age, intensity, duration, relationship of the offending party); intelligence; talent; physical appearance; accompanying traits from other personality and symptom disorders; socioeconomic status; and sociocultural differences. We then see a similarly bewildering variety in the clinical expression of the borderline personality disorder that follows. This variety is not attributable merely to the polythetic nature of the DSM diagnosis: it is easy to find two patients with borderline personality disorder with an identical array of the defining items who differ markedly in socioeconomic status, abuse history, talent, friendliness versus hostility, and so on, and whose long-term outcomes are widely disparate.

The large number of patients in the Zanarini et al. study helps to minimize the confusion inherent in all these differences. Still, the majority of the McLean patients eventually improved enough symptomatically to lose their borderline personality disorder diagnosis, and at least half became well even in the realms of social and vocational functioning. This means that whatever the mix of contributing factors that had been at play in each individual patient, the outcomes were encouraging.

Zanarini and colleagues' caveat that there has been too much emphasis on symptom reduction in borderline personality disorder and not enough on psychosocial rehabilitation is most timely. The accepted treatment approaches—psychoanalytically informed, cognitive-behavioral, and supportive—seem to have about equal efficacy in symptom reduction. But psychosocial rehabilitation requires attention to the less readily modifiable personality trait abnormalities in borderline patients. For their therapists, this is a much lengthier task, as difficult to effect as it is rewarding to achieve.