The Case for Shifting Borderline Personality Disorder to Axis IEvidence for the Validity of BPDCore Symptoms of BPDA Single Diagnostic ConstructSpecificity of BPD.Course/PrognosisHeritability and FamilialityPredictability of Treatment ResponseBiological MarkersImplications of Shifting BPD to Axis IMisconceptions about BPDSummaryReferencesREVIEWThe Case for Shifting Borderline PersonalityDisorder to Axis IAntonia S. New, Joseph Triebwasser, and Dennis S. CharneyThrough reviewing what is known about the nature, course, and heritability of borderline personality disorder (BPD), we argue for areconceptualization of this disorder that would lead to its placement on Axis I. Borderline personality disorder is a prevalent and disablingcondition, and yet the empirical research into its nature and treatment has not been commensurate with the seriousness of the illness. Wenot only review empirical evidence about the etiology, phenomenology, and course of the disorder in BPD but we also address fundamentalmisconceptions about BPD that we believe have contributed to misunderstanding and stigmatization of the disease. Finally, we suggestfuture directions for research that might permit the identification of core features of this disorder, with a focus on the importance ofnaturalistic assessments and of assessments through the course of development.Key Words: Axis I, Axis II, borderline personality disorder, nosology,specificity, validityBorderline personality disorder (BPD) is a disabling con-dition with high morbidity and mortality, yet the empiricalresearch into its nature and treatment has not beencommensurate with the seriousness of the illness. Despite recentadvances in the treatment of BPD, it remains notoriously difficultto treat effectively, with many patients responding poorly even tothe most widely accepted treatment strategies (1). In addition,because the disorder has as cardinal symptoms anger andinterpersonal disruptiveness, it is often difficult to form a thera-peutic alliance with afflicted patients. These features draw atten-tion away from evidence that BPD is a serious mental disorderthat deserves much more investigative scrutiny than it hasreceived. A logical consequence of taking this disorder seriouslyis to consider reclassifying the disorder into Axis I. This reclas-sification would, we believe, provide a stimulus to new researchinto the nature and treatment of this severe illness.Evidence for the Validity of BPDThe validity of the BPD diagnosis remains a question in theminds of many clinicians, and some doubt its existence alto-gether (2). A widely accepted approach to validating the bound-aries of psychiatric disorders is the set of guidelines establishedby Robins and Guze (1970) (3), which considers accrual ofinformation from five lines of evidence important for establishingthe validity of a mental disorder. These criteria include: 1) acareful delineation of symptoms; 2) information about the courseof illness; 3) evidence of familial clustering; 4) predictabletreatment response, especially to somatic treatments; and5) biological markers (3,4). We review each of these criteria as itrelates to BPD.Core Symptoms of BPDA Single Diagnostic Construct? The current DSM criteriawere developed from observations by experienced clinicians,and it remains a question as to whether these criteria cluster intoone syndrome or into independent symptom dimensions. Afactor analysis of symptoms in a large sample of BPD patients(n ⫽ 141) revealed three factors: disturbed relatedness (unstablerelationships, identity disturbance, and chronic emptiness), be-havioral dysregulation (impulsivity and suicidality/self-mutila-tory behavior), and affective dysregulation (affective instability,inappropriate anger, and efforts to avoid abandonment) (5).These factors were replicated in the CLPS (Collaborative Longi-tudinal Personality Disorders Study)—a prospective descriptivestudy of a large sample (n ⫽ 668) of patients with personalitydisorders, including schizotypal, borderline, avoidant, and ob-sessive-compulsive personality disorders and major depressivedisorder (MDD) with no personality disorder (6).Recent data, however, raise questions about the 3-factormodel and instead suggest a single underlying core that leads tothe diverse symptoms of BPD. Although Sanislow found that the3-factor model yielded a better fit with their data than a single-factor model, the factors identified (disturbed relatedness, behav-ioral dysregulation, and affective dysregulation) were highlyintercorrelated (r ⫽ .90, .94, and .99, respectively), lendingsupport to a single overarching BPD construct. A subsequentfactor analysis identified three similar factors but also concludedthat the factors were too highly intercorrelated to be consideredseparate factors (7). Providing even further evidence for BPD asa unified syndrome, a recent large study explored several 1-, 3-,and 4-factor models of DSM-IV BPD criteria and concluded thatthe BPD criteria describe a single construct rather than multipleco-occurring syndromes (8). Finally, a confirmatory factor anal-ysis of DSM-III-R BPD criteria in a large clinical and non-clinicalsample showed that a single factor fit the data best (9). This studyalso showed that “frantic efforts to avoid abandonment” was thecriterion with the highest specificity and positive predictivepower. Affective instability was also highly informative as to BPDdiagnosis, whereas identity disturbance and feelings of empti-ness were less informative. Even though factor analyses lendsupport to the presence of a unitary latent diagnostic construct,heterogeneity is observed in the clinical presentation of BPD.This might arise out of the fact that different aspects of thedisorder might be present at different times, making the disorderappear quite heterogeneous when observed cross-sectionally.This highlights the importance of a developmental approach tocharacterizing the core features of BPD.Specificity of BPD. The high rate of comorbidity with otherdisorders has also led to skepticism about the validity of the BPDdiagnosis. Data from CLPS showed that the Axis I comorbiditiesmost commonly seen with BPD were posttraumatic stress disor-der (PTSD) and substance abuse. Although BPD subjects showeda high rate of MDD (79% lifetime), this was not higher than theFrom the Department of Psychiatry (ASN, JT, DSC), Mount Sinai School ofMedicine; and the Bronx Veterans Affairs Medical Center (ASN, JT), NewYork, New