Background: Naturalistic longitudinal studies of Borderline Personality Disorder (BPD) indicate that symptoms improve over time. In the present study, we applied network theory to the question of how BPD symptom networks may differ as a function of age. Methods: In a transdiagnostic sample of 5,212 patients presenting for acute psychiatric treatment, we administered a measure of BPD symptoms and then used a novel machine learning technique to test the hypothesis that symptom networks would significantly differ across the age of participants. Results: Results supported two significant differences in the BPD symptom network that emerged at age 46. In older participants, the relationships among symptoms of non-suicidal self-injury/suicide and emptiness was weaker, yet the relationship between anger and relationship problems was stronger. No differences emerged for relationships between all other symptoms. Limitations: Given the cross-sectional nature of this study, the potential influence of cohort effects cannot be ruled out. Conclusions: These findings support the utility of network theory for elucidating potential pathways by which the relationships between symptoms of BPD may differ as a function of age in treatment-seeking individuals. In parallel, results of this study support the highly central role of strong emotions in BPD regardless of age.

Introduction

Borderline Personality Disorder (BPD) is a severe psychiatric disorder marked by symptoms including affective instability, impulsive and self-destructive behaviors, and identity disturbance (American Psychiatric Association, 2013). Although diagnostic criteria for personality disorders emphasize stability of symptoms over time, numerous studies show that individuals with BPD report significant symptomatic improvement over long-term follow up (Paris et al., 1987; Soloff and Chiappetta, 2019; for review, see Temes and Zanarini, 2018). Given the high rate of impairment linked to symptoms of BPD, researchers have sought to understand why symptoms appear to improve over time for many individuals with this diagnosis. Certain predictors such as higher IQ, less psychiatric comorbidity, and better initial psychosocial functioning appear to be associated with favorable outcomes over long-term follow-up (Soloff and Chiappetta, 2018; 2019; Zanarini et al., 2018). Longitudinal data also indicates that some symptoms such as non-suicidal self-injury (NSSI) and affective instability diminish relatively quickly, while other symptoms including anger and emptiness may persist for longer intervals (Zanarini et al., 2007). Longitudinal studies have yielded inconclusive evidence about whether improvements in symptoms can be explained by the passage of time as opposed to differences in the age of individuals in the study (Álvarez-Tomás et al., 2019; Shea et al., 2009; Soloff and Chiappetta, 2018). However, the possibility of age-driven decreases in symptoms is supported by findings indicating that BPD-relevant symptoms change as a function of age across psychopathology: for example, both impulsivity and aggression are more strongly related to completed suicides in younger vs. older individuals (McGirr et al., 2008), and the prevalence of substance use disorders decreases with older age (Compton et al., 2007). Network theory of psychopathology is one potential way to understand significant changes in the trajectory of BPD symptoms over time.