Borderline personality disorder (BPD) is an impairing mental disorder that concerns 1-2% of the general population. It is characterized by a pervasive pattern of instability in affect regulation, impulse control, interpersonal relationships, and self-image 1 . Although BPD is usually diagnosed in adults, symptoms of BPD can often be traced back to childhood 2 . Several studies have shown that specific features of BPD, such as self-harm, impulsivity and emotional dysregulation, present during childhood or adolescence, are predictive of BPD diagnoses in adulthood 3 4 5 . Among these, impulsivity in particular is regarded as a core feature of BPD 1 6 . Impulsivity is associated with factors contributing to the severity of the disorder, such as suicidal/self-harming behaviours or increased risk for substance abuse 7 8 . Impulsivity in BPD has been related to dysfunction in inhibitory systems mediated by fronto-striatal circuits 9 10 11 12 .

Impulsivity, along with inattention and hyperactivity, is also one of the core symptoms of Attention-Deficit/Hyperactivity Disorder (ADHD) 13 . Impulsivity may contribute to motor (overactivity), cognitive (poor cognitive control), emotional (uncontrolled tempers) and interpersonal (social disinhibition) dysfunctions reported in patients with ADHD 14 . Meta-analytical reviews have confirmed that deficient inhibitory functions, especially executive motor inhibition, are among the most robust findings in ADHD research 15 16 . Response inhibition deficits in ADHD have been related to functional and volumetric changes in the right inferior frontal cortex (IFC) and in its associated circuitry involving projections from the basal ganglia and into the striatum 17 18 .

Thus, ADHD and BPD share dysregulation in emotional and impulse control, with a possible mediating role of a dysfunction of neuronal inhibitory systems. Interestingly, several reports concerning the greater-than-chance co-occurrence of these two disorders have been published 19 20 21 . Since ADHD, as a neuro-developmental disorder, appears earlier than BPD, it has been suggested that ADHD may contribute to the development of BPD 22 . Longitudinal prospective studies indicate that adolescents and young adults with a childhood history of ADHD are more likely than those without that history to have a personality disorder, with a higher risk for borderline and antisocial personality disorders than for others 23 24 . Stepp and colleagues recently published the first longitudinal study to examine ADHD and ODD symptom trajectories as specific childhood precursors of BPD symptoms in adolescent girls 25 . They performed a series of latent growth curve models on two cohorts of girls annually assessed between the age of 8 and 14. They found that higher levels of ADHD and ODD scores at age 8 uniquely predicted BPD symptoms at age 14; over and above depression symptoms at outcome. However, as suggested by Davids and Gastpar 26 , BPD subjects are likely to be an heterogeneous group, with some subjects characterized by prominent impulsive features, others by prominent affective or dissociative features. ADHD may thus represent a risk factor for BPD patients with a predominance of impulsivity features. However, impulsivity is not a unidimensional construct and authors regard it as composed of several dimensions, such as motor, attentional and cognitive impulsivity. Data from the literature report the existence of deficits in the three facets of impulsivity in ADHD subjects 27 . However, less is known about subjects with BPD and ADHD. As adolescence is a key period for the onset of personality disorders, focusing on BPD adolescents with persistent ADHD comorbidity since childhood can cast light on the developmental trajectory of BPD. As the literature concerning these topics is sparse, our study aimed to: 1) assess the past and current prevalence of a comorbid ADHD/BPD diagnosis and its impact on the clinical presentation of borderline personality disorder in adolescents and; 2) determine which type of impulsivity specifically characterizes adolescents with BPD-ADHD.

One-hundred and seven adolescents with a DSM-IV clinical diagnosis of BPD were referred to the study by their psychiatrists. Of these subjects, 85 fulfilled SIDP-IV criteria for a BPD and composed the final sample of the study. There were no significant differences between the recruitment centres in terms of subject age and educational level, numbers of borderline criteria and in/outpatient ratio. The mean age of the sample was 16.3 ± 1.4 yrs; 74 (87%) were girls. The sample had a severe clinical profile: 67% (n = 89) of the subjects were recruited from inpatient units and had a mean score of 17.6 ± 3.9 on the SIDP-IV borderline diagnostic criteria (minimum required score being 15 with a maximum of 27). The most frequently endorsed criteria (more than 85% of the sample) were Impulsivity, Suicidal/self-mutilating behavior, Affective instability and Inappropriate anger. The majority of adolescents met the criteria for at least one Axis-I disorder (N = 76, 89%). Mood disorders were the most frequently observed comorbidity (N = 47, 55.3%) followed by eating disorders (N = 27, 31.8%), disruptive behavior disorders (N = 22, 25.9%), and substance use disorders (N = 17, 20%). The sample showed a severe impairment in the overall level of psychosocial functioning with a mean GAF score of 47.2 ± 14. 76% of the samples were currently under medication. Antidepressants and antipsychotics were the most commonly prescribed drugs. No patients were currently taking stimulants, although 2 patients had been under methylphenidate treatment during childhood.

The aim of this study was to explore the prevalence of a comorbid ADHD-BPD diagnosis and its impact on the clinical presentation of borderline personality disorder adolescents, and to explore which type of impulsivity is specifically associated with BPD-ADHD adolescents. To our knowledge, this is the first study investigating ADHD in BPD in this specific age group.

Concerning the prevalence of ADHD diagnosis in our sample, we found a current rate of 11%. This result is close to the 16% rate found by Philipsen and colleagues in a sample of adult BPD female patients 19 , notwithstanding some methodological differences between the studies. In the Philipsen's study, current ADHD was diagnosed by self-assessment using the short version of the WURS (for childhood ADHD symptoms) and the adult ADHD-Checklist, whereas in our study diagnosis was ascertained by experienced clinicians using a valid and reliable diagnostic interview integrating all relevant data from the clinical records of the patients, including parental reports. Although the current prevalence observed here may appear not very high, up to 46% of the subjects presented at least one symptom with a clinical or subclinical significance and some impact on functioning in the ADHD screening, eventually qualifying for a diagnosis of ADHD-NOS. It is interesting to note that symptoms of inattention, hyperactivity and impulsivity were evenly distributed across the sample. This points to the fact that all types of ADHD symptoms, not solely impulsivity, are frequently found in BPD adolescents. Moreover, comorbidity rates did not change when diagnosis was made without including impulsivity, thus reducing the criticism of an overestimation of ADHD diagnosis in BPD due to symptom overlap.

The results of this study also show that the presence of a comorbid ADHD diagnosis influences the clinical presentation of BPD in adolescents. ADHD in BPD was significantly associated with a greater likelihood of disruptive disorders (particularly ODD) and with a trend for a greater likelihood of other cluster B personality disorders (histrionic, narcissistic and antisocial personality disorders). This result is not surprising since in longitudinal studies, ODD in childhood as well as antisocial behaviours in adolescence and adulthood have been frequently observed as main outcomes for ADHD children 41 42 . Impulsivity has been suggested as an important mediator of this negative outcome among ADHD children 43 44 . The role played by impulsivity in the relationship between ADHD and outcome was indirectly suggested in our study by the observation of higher levels of impulsivity on all Barratt subscales (although significant only for the Attentional/Cognitive subscale) and in the specific domain of impulsivity on the DIB-R in the BPD-ADHD group. The impulsivity dimension of the DIB-R includes several externalizing behaviours, driven by impulsivity, such as substance abuse, promiscuous sex, reckless driving or self-harming/suicidal behaviours. A reverse tendency on the DIB-R was observed in the domain of cognition, with borderline adolescents without ADHD showing a clinical profile characterized by more internalising symptoms such as odd thinking, unusual perceptual experiences or paranoid/quasi-psychotic experiences. This dual dissociation on the DIB-R indices between BPD and BPD-ADHD adolescents moderates the conclusions reached by Philipsen and colleagues 19 , suggesting that this association might not be equivalent to a more severe form of the borderline disorder, but could correspond to a specific subtype of BPD with high impulsivity associated with an ADHD profile. This hypothesis is in line with recent conclusions drawn by Ferrer and colleagues 21 who have suggested that BPD patients should be distinguished in two subgroups according to the presence or absence of ADHD, with the former subgroup showing a specific profile of impulsive comorbidity. Moreover, these results recall the ICD-10 conceptualization of the emotionally unstable personality disorders, which specifically includes an impulsive sub-type alongside the typical borderline profile 45 . Our study suggests that the ICD-10 impulsive sub-type could be more developmentally driven, with ADHD symptoms persisting since childhood. This proposal could be of interest for the possible inclusion of a developmental perspective in the DSM classification of personality disorders. A similar proposal for differentiating borderline patients according to specific developmental features has already been suggested by Andrulonis 46 who, in a sample of DSM-III BPD adults, identified a separate group of patients showing severe hyperactivity, distractibility and/or learning disabilities and episodes of behavioral dyscontrol. This group reported hyperactive and aggressive behaviours during childhood and antisocial acting-out with drug/alcohol abuse during adolescence but, like our sample, did not show any micro-psychotic episodes. This association also supports one of the developmental routes to BPD suggested by Nigg 12 , which he has termed as the primary impulsivity route, as opposed to the traumatogenic route more related to severe disruptions in early caregiving experiences and mainly affecting the development of affect regulation. For this author, this impulsive BPD subgroup could arise from weak executive response inhibition mechanisms, leading to extremes of impulsivity, behavioural disturbances during childhood, inappropriate interpersonal relations, and a cascade of negative socialization experiences leading to personality disturbances. From a temperamental perspective, specific features related to impulsivity in ADHD children, such as Novelty Seeking, have also been found to increase the risk of development of BPD in adulthood 47 . Data supporting this theoretical perspective have been reported by Lampe and colleagues 48 who assessed various motor and cognitive inhibitory functions in adult ADHD patients, with and without BPD, compared to subjects with BPD alone and controls. In this study, ADHD subjects (whether or not comorbid with BPD) had higher scores than BPD subjects on all behavioural subscales of the BIS and showed impaired inhibition on the Attentional Network Task (Stop and Interference). Conversely, BPD subjects (without comorbid ADHD) did not differ from their matched controls, a result which led the authors to conclude that an impairment of inhibitory control could be a core deficit of BPD only when associated with ADHD. This result suggests that the cognitive component of inhibitory control may play a specific role in the phenomenology of the impulsive/developmental sub-type of BPD. Results from the regression analysis of our study showed a specific association between Barratt's Attentional/Cognitive Impulsiveness and ADHD diagnosis in borderline adolescents. The Attentional/Cognitive impulsivity of the BIS-11 involves several clinical features in the domain of attention and of cognitive stability: the inability to inhibit irrelevant information held in working memory and to focus on the task at hand leading to distractibility 49 ; and an excessive cognitive speed in decision-making 50 with an aversion to externally imposed delays 51 leading to cognitive and behavioural mistakes or acting-out behaviours, especially under emotional conditions 52 . Attentional impulsivity has been linked to the the dorsolateral prefrontal cortex 53 whereas cognitive impulsivity has been correlated to the orbitofrontal/ventromedial areas of the prefrontal cortex, especially the more anterior sector of this region, the frontal pole 49 . Some preliminary results support the hypothesis that orbitofrontal/ventromedial prefrontal dysfunction may underlie some of the behavioural manifestations of BPD-ADHD patients 54 55 , but more data are needed, especially in adolescent samples.

Some limitations of the current study must be taken into consideration when interpreting the findings.

First, the main limit of the study is its cross-sectional design with data on childhood ADHD diagnosis collected retrospectively. Only longitudinal studies can directly support the identification of the developmental pathways leading from childhood to adult psychopathology. This is even more important if we consider that these diagnostic constructs tend to overlap, particularly in the realm of impulsivity. However cross-sectional studies on comorbid disorders in specific populations, such as adolescents, can shed light on their clinical presentation and help identifying their specific therapeutic needs. Moreover, although indirectly, the high diagnostic stability between past and current ADHD diagnosis found in our study supports the hypothesis of a subtype of BPD with a childhood history of ADHD, hypothesis that has been recently confirmed by Stepp and colleagues in their longitudinal study on adolescent girls 25 .

The second limit concerns the small sample size of the study and the potential sample selection bias of the screening phase conducted by the consulting clinicians without performing a systematic between-center inter-rater reliability. This may have reduced the statistical power of the analyses and the generalizability of the results.

For instance, our sample included a majority of female patients. It is commonly agreed that ADHD is less frequent in females, with a predominance of purely inattentional forms. It is possible that the high levels of impulsive features associated with ADHD could be due to a referral bias of our specific clinical sample composed of severe forms of BPD female adolescents. Although the size of the sample of BPD participants was reasonable compared to other studies, particularly since it was limited to adolescents with a well-characterized BPD diagnosis, results should be interpreted with caution as to know what the likelihood might be that the sample is actually representative of BPD adolescents.

Finally, to assess impulsivity, we used the validated adult version of the Barratt Impulsiveness Scale. Although the use of the adult version of the BIS-11 in adolescents can be found in the literature on impulsivity 56 57 , it could have been interesting to use the adolescent version of the scale which has been shown to present a different structure from the adult one 58 .

Notwithstanding these limitations, the results of this study confirm, in an adolescent sample, previous studies conducted in adult samples 19 showing that a co-occurring ADHD diagnosis influences the clinical presentation of subjects with borderline personality disorder. ADHD in BPD adolescents was associated with a specific comorbid profile of disruptive disorders, with a trend towards a greater likelihood of cluster B personality disorders, and with higher levels of impulsivity, especially of the Attentional/Cognitive type. These results suggest that BPD in a sub-group of patients could be more developmentally driven, with ADHD symptoms and impairments of the inhibitory system persisting since childhood, thus deserving specific interventions in childhood as well as in adulthood. If confirmed by further empirical evidence, this hypothesis could support the inclusion of a developmental perspective in the DSM classifications of borderline personality disorders. More longitudinal studies are needed to explore the role of these developmental features as risk factors for borderline personality disorders.

ADHD: Attention Deficit Hyperactivity Disorder; BPD: Borderline Personality Disorder; BPD-ADHD: Borderline Personality Disorder with Attention Deficit Hyperactivity Disorder; ODD: Oppositional Defiant Disorder; ASP: Antisocial Personality Disorder.

The authors declare that they have no competing interests.

All the authors listed in the manuscript have contributed sufficiently to the project to be included as authors. MS intiated and designed the protocol, collected data, participated in data analysis and interpretation and writing and revising the manuscript. ARL participated in data analysis and interpretation and revising the manuscript. SC participated in revising the manuscript. APS and MC participated in designing the protocol, and revising the manuscript. BF participated in data analysis and interpretation and in revising the manuscript. All authors read and approved the final manuscript.