Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2025 Apr 22;19(2):e70019. doi: 10.1002/pmh.70019

Symptom Overlap Between Depression and Borderline Personality Disorder: A Network Analysis

Bess M Kew 1,✉, Nathan J Monk 1, Tamlin S Conner 2, Chris Frampton 1, Roger T Mulder 1, Richard J Porter 1, Janet D Carter 3, Jennifer Jordan 1, Virginia V W McIntosh 3, Katie M Douglas 1
PMCID: PMC12013464  PMID: 40260579

ABSTRACT

Major depression (MD) is a serious mental health disorder projected to become the leading cause of global disability by 2030. Borderline personality disorder (BPD) frequently co‐occurs with MD. Individuals with both conditions often experience prolonged recovery times and exhibit high levels of suicidal behaviour. Network theory and its application, network analysis, presents a novel framework for conceptualising and understanding the comorbidity between MD and BPD. This network analysis aims to identify influential symptoms within a BPD/MD network and explore the clinical relevance of these relationships. Data from 548 participants were pooled from four clinical trials run between 1994 and 2013 at the Department of Psychological Medicine, University of Otago, Christchurch, New Zealand. All participants were diagnosed with current MD (as part of major depressive disorder or bipolar II disorder). Baseline MD and BPD symptom data from trial entry assessments were entered into a cross‐sectional network analysis. A further network analysis was estimated with the addition of three covariates (age, gender and depression severity) alongside the MD and BPD symptoms. Network analyses identified several connecting symptoms between MD and BPD. After controlling for depression severity, BPD symptoms of identity disturbance and unstable relationships had unique and robust relationships with MD suicidal ideation and behaviours. Further exploration of these bridge relationships found that participants who exhibited identity disturbance were almost three times more likely to have reported a previous suicide attempt. Results from this study have implications for risk assessment and treatment of individuals with comorbid MD and BPD.

Keywords: borderline, comorbidity, major depression, mood disorder, network analysis, network model, personality disorder

1. Introduction

Major depression (MD) is a pervasive and increasingly prevalent disorder that is predicted to become the largest contributor to global disease burden by 2030 (World Health Organisation 2008). Despite decades of pharmacological and psychological research, and advances in treatment options, up to 27% of individuals with MD struggle to become well and will go on to develop a chronic depressive illness (Malhi and Mann 2018; Boschloo et al. 2014). Comorbid mental health diagnoses in individuals with MD are common and add complexities to disorder burden, treatment and recovery (Kessler et al. 2007; Steffen et al. 2020). Borderline personality disorder (BPD) is often diagnosed alongside MD. Rates of current BPD diagnoses in MD samples have been estimated at around 19% (Grant et al. 2008), and the rate of any lifetime major depressive episode (MDE) is estimated at 83% in BPD populations (Zanarini et al. 1998; Shah and Zanarini 2018).

In comorbid MD and BPD, the course of depression tends to be more severe and chronic, and individuals diagnosed are at a greater risk of suicide compared with those with MD alone (Sarhan et al. 2019; Söderholm et al. 2023). First‐option treatment efforts for MD are less effective when BPD is present, and, once well, individuals are at greater risk of relapse (Grilo et al. 2010; Gunderson et al. 2008; Hein et al. 2022; Newton‐Howes et al. 2014). Implications for prognosis and treatment of comorbid MD and BPD have prompted research efforts to delineate associations between the two disorders. There are several competing theories that attempt to explain comorbidity between personality disorders and other mental health disorders, which are relevant to understanding co‐occurring MD and BPD.

1.1. Models to Explain Comorbidity of BPD and MDD

The ‘common cause model’ views personality and depressive disorders as separate entities that share etiological factors (Klein and Bufferd 2011). This view is supported by similar predisposing neurobiological and genetic factors that may lead to co‐development of MD and BPD (Bornovalova et al. 2018; Witt et al. 2017). In addition, a third variable such as a propensity for affective instability may lead to the comorbidity (Gunderson et al. 2004). The ‘predisposition/vulnerability model’ suggests that the presence of personality psychopathology increases the risk of onset of depressive symptoms and, as such, plays a causal role in the development of a mood disorder (Kendler et al. 2004; Vinkers et al. 2014). In this model, core symptoms of BPD such as tumultuous interpersonal relationships (American Psychiatric Association 2013) may trigger or increase risk for depression as a secondary syndrome (Beatson and Rao 2012; Hepp et al. 2018). Conversely, the ‘pathoplasty model’ suggests that personality psychopathology affects the course and severity of other mental health disorders but plays no role in the onset of symptoms (Klein et al. 2011; Morris et al. 2009; Wilson et al. 2014). Importantly, the content overlap between symptoms of MD and BPD, such as fluctuations in mood and emotions, feelings of emptiness and suicidality, may lead to difficulty distinguishing between the two disorders, or to misdiagnosis, thus inflating comorbidity estimates (Beatson and Rao 2012; Rao and Broadbear 2019).

1.2. Network Approach to Psychiatric Disorders

Network theory is an alternative model of mental health disorders that proposes that disorders arise from mutually reinforcing, causal relationships between symptoms (Borsboom 2017; Borsboom and Cramer 2013). Therefore, the onset and maintenance of mental health disorders are conceptualised through a bottom‐up approach. This contrasts with classical theories in psychiatry that view disorders as separate disease entities that produce symptoms in varying, relatively distinct and characteristic patterns (Schmittmann et al. 2013). Rather, in network theory, symptoms lie within a network structure, meaning that the presence of one symptom may trigger another in the network, further maintaining the disorder (Borsboom 2017). For example, in MD, an onset of feelings of guilt may disturb sleep, leading to loss of energy and further loss of motivation to complete daily activities.

In the context of network theory, comorbid or multimorbid mental health disorders may be explained by relationships of symptoms that ‘bridge’ between individual disorders (Cramer et al. 2010). In a clinical context, network theory assumes that associated symptoms may be targeted in treatment (Jones and Robinaugh 2021; Blanken et al. 2019). For example, targeting feelings of guilt may improve sleep, further restoring energy and motivation.

Network theory in application uses a novel psychometric approach called network analysis, where cross‐sectional models are estimated using partial correlations between symptoms. For example, Kohne and Isvoranu (2021) conducted an exploratory network analysis of MD and BPD symptoms using data from Dutch adult inpatients (n = 376) diagnosed with primary substance use disorders with comorbidities and multimorbidities. Several direct links between symptoms of BPD and MD were reported, thus showing associations among and within symptom clusters of the two disorders.

The present study aims to extend the Kohne and Isvoranu (2021) network analysis by examining MD and BPD symptom clusters and overlap in a larger sample of individuals with primary mood disorders. To our knowledge, this is the first network analysis to explore links between MD and BPD in a mood disorder sample using data from clinician‐rated assessments. We were able to further examine these links with logistic regression models, providing clinical implications that may inform diagnostic methods and treatment.

1.3. Aims

For the present study, the estimation of appropriate network model(s) will serve the primary aims of (1) identifying influential symptoms within a BPD‐MD network; (2) identifying symptoms that may bridge BPD and MD; and (3) exploring the predictive nature and potential clinical relevance of identified links.

2. Methods

All participants provided written informed consent to assessment and treatment. Each study received ethics committee approval.

2.1. Participants

Data used in this study were compiled from four clinical trials published between 1994 and 2013 conducted in the Clinical Research Unit at the Department of Psychological Medicine, University of Otago, Christchurch, New Zealand (Carter et al. 2013; Joyce et al. 1994, 2002; Luty et al. 2007). These studies were outpatient intervention trials examining the use of either medication or psychological therapy for the treatment of MD. Total sample size across the four trials was N = 578. Cross‐sectional data from baseline assessments were used.

Study procedures such as recruitment, diagnostic assessment and inclusion and exclusion criteria were similar across the four studies. These procedures are summarised below and are described in detail in the primary outcome papers (Carter et al. 2013; Joyce et al. 1994, 2002; Luty et al. 2007).

Participants were recruited from community mental health services, psychiatric emergency services and from primary care. Self‐referrals were also accepted. A structured clinical interview was administered to confirm MD diagnosis. Studies 1–3 used the Structured Clinical Interview for DSM‐III‐R (SCID‐P) (Spitzer et al. 1992) and Studies 2 and 3 extended this interview to include DSM‐IV criteria for melancholic and atypical depression. Study 4 used the Structured Clinical Interview for DSM‐IV Axis I disorders (SCID‐I) (First and Gibbon 2004).

Participants were included in the studies if they had a primary diagnosis of major depressive episodes, either as part of major depressive disorder or bipolar II disorder (BD II). All studies excluded participants if they had trialled the respective study intervention in the year leading up to study entry.

Study 1 required no alcohol or drug abuse 1 month prior to trial entry. Studies 2–4 excluded participants with current moderate to severe drug or alcohol dependence. Exclusion criteria for Studies 2–4 included a history of mania (bipolar I disorder, BD I) or schizophrenia. Studies 2 and 3 excluded participants with severe antisocial personality disorder that was likely to interfere with compliance to research protocols. Participants were excluded from all studies if they had any major physical illness, which could interfere with assessment or treatment.

All participants underwent a wash‐out period of at least 2 weeks of all psychotropic medications, except for the occasional hypnotic for sleep, before admission into the trial and administration of baseline measures.

Study 1: Joyce et al. (1994):

  • Study design
    • ○
      Double‐blind randomised control trial (RCT) examining clinical predictors of drug response to either clomipramine of desipramine.
  • Participants
    • ○
      One hundred and four participants (18–60 years), diagnosed with current DSM‐III‐R MD.

Study 2: Joyce et al. (2002):

  • Study design
    • ○
      Double‐blind RCT examining predictors of short‐ and long‐term response to either fluoxetine or nortriptyline.
  • Participants
    • ○
      One hundred and ninety‐five participants (≥ 18 years) diagnosed with current DSM‐IV MD.

Study 3: Luty et al. (2007):

  • Study design
    • ○
      Assessor‐blind RCT comparing interpersonal psychotherapy (IPT) and cognitive behavioural therapy (CBT).
  • Participants
    • ○
      One hundred and seventy‐seven participants (≥ 18 years) diagnosed with current DSM‐IV MD.

Study 4: Carter et al. (2013):

  • Study design
    • ○
      Assessor‐blind RCT comparing schema therapy (ST) and CBT.
  • Participants
    • ○
      One hundred participants (≥ 18 years) diagnosed with current DSM‐IV MD.

2.2. Measures

The following section describes key measures used in the four trials relevant to the aims of this paper.

2.2.1. Hamilton Depression Rating Scale

Symptoms of MD were assessed using the clinician‐administered 17‐item Hamilton Depression Rating Scale (HAM‐D17) (Hamilton 1960). The HAM‐D17 is a widely‐used measure assessing depression severity in individuals diagnosed with a mood disorder (Bech 2009). A total score representing severity of depressive symptoms is calculated by summing scores from each item.

2.2.2. Structured Clinical Interviews for Diagnostic and Statistical Manual of Mental Disorders

Symptoms of BPD were assessed using relevant symptom criteria from versions of the Structured Clinical Interview for DSM personality disorder symptoms. All studies used the Structured Clinical Interview for DSM‐IV personality disorders (SCID‐II) (First et al. 1997) with the exception of Study 1 (Joyce et al. 1994), which administered the interview for DSM‐III‐R personality disorders (Spitzer et al. 1992). Borderline personality disorder symptoms are endorsed ‘present’ or ‘not present’ and are therefore binary coded for the network analyses.

2.3. Statistical Analyses

2.3.1. Network Estimation

All network analyses were performed in the open‐source R software environment (version 3.6.3) (R Core Team 2020).

Networks presented in the current study are Gaussian graphical models (Lauritzen 1996) estimated via the qgraph R package (Epskamp et al. 2012). In network models, symptoms are represented as ‘nodes’, which are connected by ‘edges’, representing partial correlations between nodes after controlling for all other nodes in the network (Epskamp and Fried 2018). Nodes in the following network models represent DSM‐5 symptom criteria for BPD and MD. Edges represent associations between symptoms within each disorder and associations that bridge between BPD and MD.

Node arrangement is controlled by the Fruchterman–Reingold algorithm (Fruchterman and Reingold 1991), which places nodes with stronger associations within the centre of models and nodes with weaker associations on the perimeter. Associations among nodes are represented by green (positive) and red (negative) edges, with the thickness representing the strength. The wider and more saturated the edge, the greater the partial correlation between two nodes (Epskamp et al. 2012).

Each model was regularised using graphical LASSO (gLASSO) (Epskamp and Fried 2018), which creates a sparse network by penalising each model for its complexity. This reduces all edge weights and removes likely spurious edges from a model, estimating a parsimonious network. The degree of penalty applied to the MD/BPD networks was selected using the Extended Bayesian information criterion (Foygel and Drton 2010). The EBIC tuning parameter (y) was set to 0.5 for all estimated networks. Gaussian graphical models assume multivariate normal distribution of variables; however, data in the present study are a combination of ordered categorical and binary variables. Recent simulations suggest the assumption of multivariate normality can be relaxed by estimating networks with rank‐transformations, such as Spearman correlation coefficients (Isvoranu and Eskamp 2021). As such, Spearman correlation coefficients were used to estimate edges in this research.

2.3.2. Centrality and Accuracy

Centrality was quantified using bridge strength; this identifies the most active bridge edges within a network (Jones et al. 2021). Accuracy analyses were implemented using the bootnet R package (Epskamp et al. 2018). Bootstrapping was performed at n = 1000 iterations.

2.3.3. MD/BPD Network

Items were selected from each measure (HAM‐D17 and SCID‐II) to best approximate DSM‐5 diagnostic criteria (American Psychiatric Association 2013). However, not all items from each measure could be included due to sample size restrictions. The MD/BPD networks were estimated using nine HAM‐D17 items: (1) depressed mood, (2) feelings of guilt, (3) suicidal ideation and behaviours, (4) early insomnia, (5) loss of interest in work/activities, (6) psychomotor retardation, (7) psychomotor agitation, (8) appetite loss and (9) general somatic symptoms. Eight SCID‐II BPD diagnostic criteria items were included: (1) frantic efforts to avoid abandonment, (2) unstable relationships, (3) identity disturbance, (4) self‐damaging impulsivity, (5) self‐harm/suicidal behaviour, (6) affective instability, (7) emptiness and (8) anger.

2.3.4. Covariate Network

An additional network was estimated using the nine HAM‐D17 items and eight SCID‐II BPD items as in the MD/BPD network, with the addition of three covariates: age, gender and the HAM‐D17 total score. These factors have been shown to impact clinical presentations of MD and BPD (Gunderson et al. 2004; Labaka et al. 2018; Qian et al. 2022). Gender was self‐reported by participants as female or male in the entry assessments. Participants were asked to self‐identify gender in binary form given the time‐period of the original studies.

2.3.5. Exploratory Analyses

Bridge edges identified between MD and BPD nodes were further explored. A series of logistic regression analyses were applied examining main effects and interactions associated with these bridge edges to determine significance of these relationships. Analyses were performed using IBM SPSS V28.0 software (IBM Corp, 2021).

3. Results

Demographic characteristics including age, gender and ethnicity are reported in Table 1. Table 2 highlights pertinent clinical characteristics across the four studies. After removing cases due to missing data, a total of N = 548 participants' data were included in the network models. A median HAM‐D17 score of 19 indicated participants' mood was rated by mental health clinicians within the moderately severe range (Zimmerman et al. 2013). Supporting Information S1 shows the number of participants who exhibited each BPD symptom criteria, with the three most prevalent BPD symptoms being emptiness (n = 191), affective instability (n = 167) and identity disturbance (n = 166).

TABLE 1.

Demographic characteristics of participants.

Variable Labels Study 1 (n = 100) Study 2 (n = 183) Study 3 (n = 168) Study 4 (n = 97) All (n = 548)
Joyce et al. (1994) Joyce et al. (2002) Luty et al. (2007) Carter et al. (2013)
Age (years) Mean (SD) 31 (10.3) 32 (11.3) 36 (10.2) 38 (11.4) 34 (11.1)
Gender

Female n (%)

Male n (%)

51 (51.0)

49 (49.0)

107 (58.5)

76 (41.5)

123 (73.2)

45 (26.8)

66 (68.0)

31 (32.0)

347 (63.3)

201 (36.7)
Ethnicity

NZ European n (%)

Māori n (%)

Other n (%)

95 (95.0)

2 (2.0)

3 (3.3)

176 (96.2)

4 (2.2)

3 (1.6)

148 (88.1)

8 (4.8)

12 (7.1)

83 (85.6)

4 (4.1)

10 (10.3)

502 (91.6)

18 (3.3)

28 (5.1)
Education (years) a

Secondary education, mean (SD)

Tertiary education,

Mean (SD)

 — — —

4 (1.0)

3 (1.95)

 —
Open in a new tab
a

Years of education only documented in Study 4 (Carter et al. 2013).

TABLE 2.

Clinical characteristics of participants.

Variable Labels Study 1 (n = 100) Study 2 (n = 183) Study 3 (n = 168) Study 4 (n = 97) All (n = 548)
Joyce et al. (1994) Joyce et al. (2002) Luty et al. (2007) Carter et al. (2013)
Primary diagnosis

MDD n (%)

BD II n (%)

BD other n (%)

Dysthymia n (%)

87 (87.0)

4 (4.0)

8 (8.0)

1 (1.0)

165 (90.2)

18 (9.8)

0 (0.0)

0 (0.0)

162 (96.4)

6 (3.6)

0 (0.0)

0 (0.0)

89 (91.8)

8 (8.2)

0 (0.0)

0 (0.0)

503 (91.8)

36 (6.6)

8 (1.5)

1 (0.1)
Depression onset (age in years) Median (IQR) a 20 (16.8, 28.3) 20 (15.0, 30.0) 19 (14.0, 28.3) 18 (13.0, 29.0) 19 (15.0, 29.0)
Number of depressive episodes

1 episode n (%)

2–9 episodes n (%)

10+ episodes n (%)

39 (39.0)

34 (34.0)

27 (27.0)

67 (36.6)

56 (30.6)

60 (32.8)

44 (26.2)

88 (52.4)

36 (21.4)

25 (25.8)

45 (46.4)

27 (27.8)

175 (31.9)

223 (40.7)

150 (27.4)
BDI Median (IQR) — — 26 (22.0, 35.0) 25 (21.0, 30.0) —
HAM‐D17 Median (IQR) 21 (18.0, 24.0) 19 (17.0, 23.0) 16 (13.0, 20.0) 16 (13.0, 19.0) 19 (15.0, 22.0)
MADRS Median (IQR) — 31 (27.0, 35.0) 24 (20.0, 27.3) 24 (19.0, 27.0) —
Diagnosed with BPD n (%) 19 (19.0) 30 (16.4) 20 (11.9) 9 (9.3) 78 (14.2)
Any anxiety disorder b n (%) 40 (40.0) 77 (42.1) 72 (42.9) 60 (61.9) 249 (45.4)
Any eating disorder b n (%) 6 (6.0) 30 (16.4) 15 (9.0) 12 (12.4) 63 (11.5)
Any substance dependence or abuse b n (%) 36 (36.0) 64 (35.0) 53 (31.5) 39 (40.2) 192 (35.0)
Open in a new tab

Abbreviations: BD = bipolar disorder, BDI = Beck Depression Inventory, BPD = borderline personality disorder, HDRS‐17 = Hamilton Depression Rating Scale 17, IQR = interquartile range, MADRS = Montgomery–Asberg Depression Rating Scale.

a

IQR was calculated when distributions identified as nonnormal by Shapiro–Wilk test.

b

Lifetime or current diagnoses.

3.1. MD/BPD Networks

The MD and BPD network model is illustrated in Figure 1. Several direct, positive edges are shown to bridge between BPD and MD nodes. A strong bridging edge was identified between suicidal ideation and behaviour (D3) and self‐harm/suicidal behaviour (B5). Both BPD nodes unstable relationships (B2) and identity disturbance (B3) showed independent bridge edges with suicidal ideation and behaviour (D3). Notably, the positive bridge edges between suicide (D3) and self‐harm/suicide (B5), suicide (D3) and unstable relationships (B2), and suicide (D3) and identity disturbance (B3) remained after later inclusion of covariates (as seen in Figure 2).

FIGURE 1.

FIGURE 1

Open in a new tab

Network model of major depression and borderline personality disorder symptoms. Note: Regularised partial correlations are represented by the red (negative) and green (positive) edges between nodes (e.g., D3 and B5), whilst controlling for all other nodes in the network. The thickness of each edge indicates the strength of the correlation: the thicker the edge, the stronger the association between two nodes.

FIGURE 2.

FIGURE 2

Open in a new tab

Network model of major depression and borderline personality disorder symptoms with three covariates included. Note: Covariate network models are interpreted in the same way as ordinary network models (as seen in Figure 1). However, the addition of covariates provides three important factors to this analysis. First, partial correlations between covariates and symptoms are identified. Second, all previously identified partial correlations between symptoms are now controlled for the added covariates. Third, the process of adding nodes (variables) into a network allows for identification of more robust associations, given that the gLASSO algorithm penalises models as they increase in complexity.

3.2. Covariate Network

Figure 2 illustrates the MD/BPD network model with the addition of three covariates (age, gender and HAM‐D17 total). Several negative edges were identified between age and nodes of MD and BPD. The covariate gender formed two negative edges with psychomotor retardation (D6) and weakly with emptiness (B7).

3.3. Centrality and Accuracy

Supporting Information S2 details the bootstrap analyses and Supporting Information S3 presents centrality analyses. The strength and rank‐order of edge weight estimates should be interpreted with caution due to wide and overlapping confidence intervals identified by the bootstrapping procedure.

3.4. Exploratory Analyses

Logistic regression analyses were conducted to explore bridges between suicidal ideation and behaviour (D3) and unstable relationships (B2), and suicidal ideation and behaviour (D3) and identity disturbance (B3), identified by the network model. Data on previous suicide attempts were obtained from Studies 2, 3 and 4 (Study 1 did not collect these data) (Carter et al. 2013; Joyce et al. 2002; Luty et al. 2007). These data were binary coded (attempt yes/no attempt) and entered into a binary logistic regression model.

Table 3 summarises odds ratios when previous suicide attempts and the two BPD symptoms were tested separately, in addition to the odds ratios when both were entered into a logistic regression model. In unadjusted regression models with this subset of data (n = 448), there were more previous suicide attempts in participants with unstable relationships (39.5%) than in those without unstable relationships (23.7%). Similarly, participants with identity disturbance were more likely to have had previous suicide attempts (43.9%) than those without identity disturbance (20.0%). The association between unstable relationships and previous suicide attempt was attenuated after adjusting for identity disturbance, whereas identity disturbance remained a significant predictor of previous suicide attempts after adjusting for unstable relationships. The association between any previous suicide attempt and the interaction of unstable relationships (B2) and identity disturbance (B3) was not significant (p = 0.49).

TABLE 3.

Logistic regression output from unstable relationships and identity disturbance predicting previous suicide attempts.

Univariate Multivariate Unstable Relationships (B2) combined with Identity Disturbance (B3)
p Odds ratio Lower CI Upper CI p Odds ratio Lower CI Upper CI
Unstable Relationships (B2) 0.004 2.102 1.267 3.487 0.606 1.167 0.648 2.102
Identity Disturbance (B3) < 0.001 3.120 2.000 4.900 < 0.001 2.937 1.765 4.888
Open in a new tab

Note: B2 and B3 refer to the symptom nodes in the network model (see Figures 1 and 2). Significant associations are bolded for readability.

Abbreviation: CI = confidence interval.

4. Discussion

This study examined the association between two highly co‐occurring disorders, MD and BPD, using network analysis. Broadly, bridge analyses showed several direct links between symptoms of MD and symptoms of BPD. Bridge edges between the two suicidality nodes from both disorders, and those between MD suicidal ideation and behaviour (D3) and BPD unstable relationships (B2), and MD suicidal ideation and behaviour (D3) and BPD identity disturbance (B3) were the only remaining bridge edges after addition of the three covariates (age, gender and depression severity). This is a notable finding given that the gLASSO algorithm used for each estimated network penalises models as they increase in complexity (Epskamp and Fried 2018). This algorithm reduces all edge sizes and removes likely false positive, or spurious, edges from the model, indicating that edges that survive covariate control signify more robust associations between nodes (Eskamp et al. 2017). The preservation of the edges between MD and BPD is of further interest as depression symptom severity (HAM‐D17 total score) was included as a covariate and therefore all edge estimates were conditioned on overall depression severity. Thus, the preservation of the BPD bridge relationships with suicidal ideation and behaviour (D3), described above, suggests that suicidality in and of itself had a unique and robust relationship with unstable relationships (B2) and identity disturbance (B3) and is not simply a marker of severity of depression.

To further explore the associations identified between suicidality, identity disturbance and unstable relationships, logistic regression models were developed utilising retrospective data on suicide attempts. Individuals who experienced BPD identity disturbance (B3) at baseline were almost three times more likely to have a lifetime history of a suicide attempt, regardless of whether they also experienced unstable relationships (B2). This aligns with previous longitudinal research reporting significant associations between baseline identity disturbance and subsequent suicide attempt(s) in a sample of individuals with personality disorders (Yen et al. 2021). Other studies have reported associations between borderline personality features, including identity disturbance, and suicidality in adolescent samples (Yalch et al. 2014; Sekowski et al. 2022). However, no known studies have examined the connection between identity disturbance and suicide attempts in an adult mood disorder sample.

In the only previous network analysis investigating symptom overlap between MD and BPD, Kohne and Isvoranu (34) reported several bridge edges between symptom clusters. Bridge edges were detected between MD sadness and BPD emptiness and between the suicidality nodes of each disorder. As these authors noted, relationships between sadness and emptiness, and between suicidality nodes of MD and BPD (as also identified in the current study), may be attributed to construct overlap.

The difference in identified bridge relationships in Kohne and Isvoranu (34) and the current study may be explained by differences in the clinical characteristics of the samples and utilised research measures. Kohne and Isvoranu's sample were inpatients (n = 376) diagnosed with primary substance use disorders and additional comorbidities and multimorbidities. In their sample, only a small proportion met diagnostic criteria for major depressive disorder (MDD, 21.0%), reported BPD symptoms (21.3%), or exhibited both (5.3%). It is also not reported whether any of their sample met diagnostic thresholds for BPD. In the current study, all participants were experiencing depression at baseline assessment, with the majority diagnosed with MDD (91.8%) and a small number with BD II (6.6%). Fourteen per cent of the sample met diagnostic threshold for BPD. As such, the current study builds on previous research with a larger sample of participants with primary mood disorder diagnoses, and with a greater proportion meeting diagnostic criteria for BPD.

Regarding differences in administered research measures, Kohne and Isvoranu (34) used self‐report measures of depression (Beck Depression Inventory) and BPD symptoms (self‐report version of the SCID‐II). The current study used clinician‐administered interviews for both MD (HAM‐D17) and BPD (SCID‐II clinician version), as well as diagnoses determined by treating psychiatrists and clinical psychologists, thus providing validity of observed symptomatology. This is a notable strength of our study given discrepancies identified between self‐report and clinician‐rated assessments of depression and personality symptoms (Enns et al. 2000; Tsujii et al. 2014; Schneibel et al. 2012; Stanley and Wilson 2006).

Symptoms that fall outside MD diagnostic criteria, such as identity disturbance, may be important in understanding and quantifying the risk of suicide in those with depression. Depression symptoms, such as feelings of hopelessness and guilt, have been robustly linked to suicidal ideation, behaviours and suicide attempts in samples with depression (Kealy et al. 2021; Ribeiro et al. 2018; Wolfe et al. 2019). Despite these links, there is still a need for more reliable risk assessment (Saab et al. 2022) and further research concerning non‐MD symptoms and their relationship to suicidality. The centrality of depression symptoms that fall outside the DSM‐5 MDD diagnostic criteria has been explored in previous research. Results from a network analysis of 28 depression symptoms in patients with MD found that non–DSM‐5 symptoms were among the most central within the estimated network (Fried et al. 2016). This indicates that depression symptoms, such as suicidality, may be influenced by symptoms not identified within diagnostic criteria (Fried et al. 2016). Further research has supported the understanding that non–DSM‐5 symptoms of depression (such as those included in the Beck Depression Inventory and Montgomery Åsperg Depression Rating Scale) are common and clinically relevant in individuals with MD (Fried 2017; Fried and Nesse 2015). Thus, network analysis can be used as a tool to explore transdiagnostic symptom interactions in MD. Due to sample size constraints, we limited the inclusion of HAMD‐17 symptoms to those that are best aligned with DSM‐5 MDD diagnostic criteria, and consequently, we may have missed pertinent bridge edges between BPD and non–DSM‐5 MD symptoms.

Those with MD may experience BPD symptoms that directly affect their psychological health and risk of harm, but which do not meet the BPD diagnostic threshold and therefore may be overlooked or inadequately addressed by clinicians. These findings suggest that the BPD symptoms of unstable relationships and identity disturbance may have significance and should be assessed in determining whether those with MD may be at increased risk of suicidal ideation/behaviour, regardless of whether they meet the diagnostic threshold for BPD. Further research is needed in this area to strengthen the understanding of predictors of suicidality and suicide attempts that fall outside classical MD symptoms.

The diagnosis of BPD itself has become contentious in recent years (Mulder and Tyrer 2023). Some researchers and clinicians believe BPD does not belong within personality disorder domains due to its broad and overinclusive diagnostic criteria and markedly high rates of comorbidity with other diagnoses (Shah and Zanarini 2018). Personality disorders are characterised by inflexible differences in personality traits that create disturbance in functioning (American Psychiatric Association 2013; Ekselius 2018). However, studies of BPD report that symptoms often resolve with age or appropriate psychological therapies (Kramer et al. 2020; Gunderson et al. 2011; Paris and Zweig‐Frank 2001; Zanarini et al. 2012). Research in this area has challenged historical conceptualisations of how personality disorders are classified and treated. It may be that the symptoms described within BPD add non‐trait symptoms to major depression such as unstable relationships (which are associated with increased risk); however, adding the diagnosis of BPD increases stigma rather than aiding the clinician. The present research supports the notion that symptoms associated with BPD can have effect without a full diagnosis.

To determine whether the network model was influenced by broad and important characteristics, covariates of gender (male or female), age and depression severity were added to the network analysis. Gender was added to the covariate model as both depression and BPD are more common in females than males (American Psychiatric Association 2013; Qian et al. 2022; Smith et al. 2008). Age was included as a covariate as BPD symptoms tend to reduce across the life course, thus symptoms may be associated with younger age (Biskin 2015). Last, connections between symptoms may be associated with severity of depression, and hence, HAM‐D17 total score was included as a covariate. Key findings from the covariate model showed that gender formed two negative edges with psychomotor retardation (D6) and weakly with emptiness (B7) indicating that male gender is associated with the endorsement of these symptoms. This finding is congruent with research reporting that males are more likely than females to present with melancholic subtypes of depression, characterised by psychomotor slowing and marked loss of pleasure (American Psychiatric Association 2013; Khan et al. 2006).

It is common in mood disorder clinical trials to exclude participants with suicidal ideation or those who exhibit risk of self‐harm. All four of the trials included in the current study did not exclude these participants, and three of the studies collected data on extent and characteristics of prior suicide attempts of participants. This made exploratory analyses concerning suicidality possible and is a strength of the current study, allowing for the exploration of symptoms that may relate to assessed suicidal history. Although further replication is needed, the MD‐BPD bridge reported between identity disturbance and suicidality may be an important focus in treating these individuals.

4.1. Limitations

Although the sample size of the current study was larger than similar studies, the covariate control removed many bridge edges, suggesting it was not sufficiently powered to detect a large number of robust bridging symptoms between MD and BPD. Further network analyses should focus on replicating findings from the current paper with even larger samples. Second, the study was not sufficiently powered to include all MD and BPD symptoms from the administered measures (HAM‐D17 and SCID‐II). As mentioned previously, this limited the inclusion of depression symptoms to those that best aligned with DSM‐5 MD criteria, and consequently, we were not able to assess bridge edges between BPD and non–DSM‐5 MD symptoms. Last, the current sample included a relatively small proportion of participants meeting diagnostic thresholds for BPD. Further research should examine these symptom relationships in a larger sample of participants with comorbid MD and BPD.

5. Conclusion

This network analysis highlights several connecting symptoms between two highly co‐occurring disorders, MD and BPD. A key finding is that, after controlling for severity of depression, BPD symptoms of identity disturbance and unstable relationships exhibited unique and robust relationships with suicidality in MD. Notably, individuals experiencing identity disturbance were almost three times more likely to report a previous suicide attempt.

These results emphasise the importance of exploring the clinical implications of findings derived from network models, as well as considering transdiagnostic symptoms in individuals diagnosed with primary MD. Future research would benefit from larger samples that can accommodate a broader range of symptoms and covariates, particularly within MD populations that also have higher rates of BPD. Such investigations are crucial for advancing our understanding of the relationships between these disorders and improving treatment strategies.

Author Contributions

B.M.K., K.M.D., R.T.M., R.J.P. and N.J.M. conceived the study and contributed to the design and study protocol. K.M.D. and T.S.C. provided supervision of the project. B.M.K. and N.J.M. conducted statistical analysis with supervision from C.F. J.D.C., R.T.M., J.J., V.V.W.M. and R.J.P. were investigators on the original trials described in this paper. B.M.K. drafted the paper, and all authors read, critically revised and approved the final version of the manuscript.

Ethics Statement

The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.

Conflicts of Interest

KMD and RJP use software for research at no cost from Scientific Brain training Pro. RJP received support for travel to educational meetings from Servier and Lundbeck. RTM is a member of the Personality Disorders and Mental Health editorial board; however, he did not take part in the review or decision‐making process of this paper.

Supporting information

Data S1. BPD symptoms

PMH-19-0-s003.png (193.1KB, png)

Data S2. Boostrap

PMH-19-0-s002.pdf (54.3KB, pdf)

Data S3. Bridge centrality

PMH-19-0-s001.pdf (5.5KB, pdf)

Acknowledgements

We thank all participants from the original trials for their generous time and effort. Open access publishing facilitated by University of Otago, as part of the Wiley ‐ University of Otago agreement via the Council of Australian University Librarians.

Funding: The authors received no specific funding for this work.

Data Availability Statement

The data presented in this study may be available on request from the corresponding author. The data are not publicly available due to ethical reasons.

References

  1. American Psychiatric Association . 2013. Diagnostic and Statistical Manual of Mental Disorders. 5th ed. American Psychiatric Association. [Google Scholar]
  2. Beatson, J. A. , and Rao S.. 2012. “Depression and Borderline Personality Disorder.” Medical Journal of Australia 1, no. 4: 24–27. [DOI] [PubMed] [Google Scholar]
  3. Bech, P. 2009. “Fifty Years With the Hamilton Scales for Anxiety and Depression. A Tribute to Max Hamilton.” Psychotherapy and Psychosomatics 78, no. 4: 202–211. [DOI] [PubMed] [Google Scholar]
  4. Biskin, R. S. 2015. “The Lifetime Course of Borderline Personality Disorder.” Canadian Journal of Psychiatry 60, no. 7: 303–308. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Blanken, T. F. , Van Der Zweerde T., Van Straten A., Van Someren E. J. W., Borsboom D., and Lancee J.. 2019. “Introducing Network Intervention Analysis to Investigate Sequential, Symptom‐Specific Treatment Effects: A Demonstration in Co‐Occurring Insomnia and Depression.” Psychotherapy and Psychosomatics 88, no. 1: 52–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bornovalova, M. A. , Verhulst B., Webber T., McGue M., Iacono W. G., and Hicks B. M.. 2018. “Genetic and Environmental Influences on the Codevelopment Among Borderline Personality Disorder Traits, Major Depression Symptoms, and Substance Use Disorder Symptoms From Adolescence to Young Adulthood.” Development and Psychopathology 30, no. 1: 49–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Borsboom, D. 2017. “A Network Theory of Mental Disorders.” World Psychiatry 16, no. 1: 5–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Borsboom, D. , and Cramer A. O.. 2013. “Network Analysis: An Integrative Approach to the Structure of Psychopathology.” Annual Review of Clinical Psychology 9: 91–121. [DOI] [PubMed] [Google Scholar]
  9. Boschloo, L. , Schoevers R. A., Beekman A. T., Smit J. H., van Hemert A. M., and Penninx B. W.. 2014. “The Four‐Year Course of Major Depressive Disorder: The Role of Staging and Risk Factor Determination.” Psychotherapy and Psychosomatics 83, no. 5: 279–288. [DOI] [PubMed] [Google Scholar]
  10. Carter, J. D. , McIntosh V. V., Jordan J., Porter R. J., Frampton C. M., and Joyce P. R.. 2013. “Psychotherapy for Depression: A Randomized Clinical Trial Comparing Schema Therapy and Cognitive Behavior Therapy.” Journal of Affective Disorders 151, no. 2: 500–505. [DOI] [PubMed] [Google Scholar]
  11. Cramer, A. O. , Waldorp L. J., van der Maas H. L., and Borsboom D.. 2010. “Comorbidity: A Network Perspective.” Behavioral and Brain Sciences 33, no. 2–3: 137; discussion 50–93–150. [DOI] [PubMed] [Google Scholar]
  12. Ekselius, L. 2018. “Personality Disorder: A Disease in Disguise.” Upsala Journal of Medical Sciences 123, no. 4: 194–204. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Enns, M. W. , Larsen D. K., and Cox B. J.. 2000. “Discrepancies Between Self and Observer Ratings of Depression: The Relationship to Demographic, Clinical and Personality Variables.” Journal of Affective Disorders 60, no. 1: 33–41. [DOI] [PubMed] [Google Scholar]
  14. Epskamp, S. , Cramer A. O., Waldorp L. J., Schmittmann V., and Borsboom D.. 2012. “qgraph: Network Visualizations of Relationships in Psychometric Data.” Journal of Statistical Software 48, no. 4: 1–18. [Google Scholar]
  15. Epskamp, S. , and Fried E. I.. 2018. “A Tutorial on Regularized Partial Correlation Networks.” Psychological Methods 23, no. 4: 617–634. [DOI] [PubMed] [Google Scholar]
  16. Epskamp, S. , Borsboom D., and Fried E. I.. 2018. “Estimating Psychological Networks and Their Accuracy: A Tutorial Paper.” Behavior Research Methods 50, no. 1: 195–212. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Eskamp, S. , Kruis J., and Marsman M.. 2017. “Estimating Psychopathological Networks: Be Careful What You Wish for.” PLoS ONE 12, no. 6: e0179891. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. First, M. B. , and Gibbon M.. 2004. “The Structured Clinical Interview for DSM‐IV Axis I Disorders (SCID‐I) and the Structured Clinical Interview for DSM‐IV Axis II Disorders (SCID‐II).” In Comprehensive Handbook of Psychological Assessment, Vol 2: Personality Assessment, 134–143. John Wiley & Sons Inc. [Google Scholar]
  19. First, M. B. , Gibbon M., Spitzer R. L., Williams J. B., and Benjamin L. S.. 1997. Structured Clinical Interview for DSM‐IV® Axis II Personality Disorders SCID‐II. American Psychiatric Association Publishing. [Google Scholar]
  20. Foygel, R. , and Drton M.. 2010. “Extended Bayesian Information Criteria for Gaussian Graphical Models.” Advances in Neural Information Processing Systems 23: 2020–2028. [Google Scholar]
  21. Fried, E. I. 2017. “The 52 Symptoms of Major Depression: Lack of Content Overlap Among Seven Common Depression Scales.” Journal of Affective Disorders 208: 191–197. [DOI] [PubMed] [Google Scholar]
  22. Fried, E. I. , and Nesse R. M.. 2015. “Depression Is Not a Consistent Syndrome: An Investigation of Unique Symptom Patterns in the STAR*D Study.” Journal of Affective Disorders 172: 96–102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Fried, E. I. , Epskamp S., Nesse R. M., Tuerlinckx F., and Borsboom D.. 2016. “What Are ‘Good’ Depression Symptoms? Comparing the Centrality of DSM and Non‐DSM Symptoms of Depression in a Network Analysis.” Journal of Affective Disorders 189: 314–320. [DOI] [PubMed] [Google Scholar]
  24. Fruchterman, T. M. J. , and Reingold E. M.. 1991. “Graph Drawing by Force‐Directed Placement.” Software: Practice and Experience 21, no. 11: 1129–1164. [Google Scholar]
  25. Grant, B. F. , Chou P., Goldstein R. B., et al. 2008. “Prevalence, Correlates, Disability, and Comorbidity of DSM‐IV Borderline Personality Disorder: Results From the Wave 2 National Epidemiologic Survey on Alcohol and Related Conditions.” Journal of Clinical Psychiatry 69, no. 4: 533–545. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Grilo, C. M. , Stout R. L., Markowitz J. C., et al. 2010. “Personality Disorders Predict Relapse After Remission From an Episode of Major Depressive Disorder: A 6‐Year Prospective Study.” Journal of Clinical Psychiatry 71, no. 12: 1629–1635. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Gunderson, J. G. , Morey L. C., Stout R. L., et al. 2004. “Major Depressive Disorder and Borderline Personality Disorder Revisited: Longitudinal Interactions.” Journal of Clinical Psychiatry 65, no. 8: 1049–1056. [DOI] [PubMed] [Google Scholar]
  28. Gunderson, J. G. , Stout R. L., Sanislow C. A., et al. 2008. “New Episodes and New Onsets of Major Depression in Borderline and Other Personality Disorders.” Journal of Affective Disorders 111, no. 1: 40–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Gunderson, J. G. , Stout R. L., McGlashan T. H., et al. 2011. “Ten‐Year Course of Borderline Personality Disorder: Psychopathology and Function From the Collaborative Longitudinal Personality Disorders Study.” Archives of General Psychiatry 68, no. 8: 827–837. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Hamilton, M. 1960. “A Rating Scale for Depression.” Journal of Neurology, Neurosurgery, and Psychiatry 23, no. 1: 56–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Hein, M. , Mungo A., and Loas G.. 2022. “Risk of Relapse Within 6 Months Associated With Borderline Personality Disorder in Major Depressed Individuals Treated With Electroconvulsive Therapy.” Psychiatry Research 314: 114650. [DOI] [PubMed] [Google Scholar]
  32. Hepp, J. , Lane S. P., Wycoff A. M., Carpenter R. W., and Trull T. J.. 2018. “Interpersonal Stressors and Negative Affect in Individuals With Borderline Personality Disorder and Community Adults in Daily Life: A Replication and Extension.” Journal of Abnormal Psychology 127: 183–189. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. IBM Corp . Released 2021. IBM SPSS Statistics for Macintosh, Version 28.0. IBM Corp.
  34. Isvoranu, A. M. , and Eskamp S.. 2021. “Which Estimation Method to Choose in Network Psychometrics? Deriving Guidelines for Applied Researchers.” PsyArXiv. [DOI] [PubMed]
  35. Jones, P. J. , and Robinaugh D. R.. 2021. “An Answer to “So What?” Implications of Network Theory for Research and Practice.” Focus (American Psychiatric Publishing) 19, no. 2: 204–210. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Jones, P. J. , Ma R., and McNally R. J.. 2021. “Bridge Centrality: A Network Approach to Understanding Comorbidity.” Multivariate Behavioral Research 56, no. 2: 353–367. [DOI] [PubMed] [Google Scholar]
  37. Joyce, P. R. , Mulder R., and Cloninger C. R.. 1994. “Temperament Predicts Clomipramine and Desipramine Response in Major Depression.” Journal of Affective Disorders 30: 35–46. [DOI] [PubMed] [Google Scholar]
  38. Joyce, P. R. , Mulder R., Luty S. E., et al. 2002. “Patterns and Predictors of Remission, Response and Recovery in Major Depression Treated With Fluoxetine or Nortiptyline.” Australian and New Zealand Journal of Psychiatry 36: 384–391. [DOI] [PubMed] [Google Scholar]
  39. Kealy, D. , Treeby M. S., and Rice S. M.. 2021. “Shame, Guilt, and Suicidal Thoughts: The Interaction Matters.” British Journal of Clinical Psychology 60, no. 3: 414–423. [DOI] [PubMed] [Google Scholar]
  40. Kendler, K. S. , Kuhn J., and Prescott C. A.. 2004. “The Interrelationship of Neuroticism, Sex, and Stressful Life Events in the Prediction of Episodes of Major Depression.” American Journal of Psychiatry 161, no. 4: 631–636. [DOI] [PubMed] [Google Scholar]
  41. Kessler, R. C. , Merikangas K. R., and Wang P. S.. 2007. “Prevalence, Comorbidity, and Service Utilization for Mood Disorders in the United States at the Beginning of the Twenty‐First Century.” Annual Review of Clinical Psychology 3, no. 1: 137–158. [DOI] [PubMed] [Google Scholar]
  42. Khan, A. Y. , Carrithers J., Preskorn S. H., et al. 2006. “Clinical and Demographic Factors Associated With DSM‐IV Melancholic Depression.” Annals of Clinical Psychiatry 18, no. 2: 91–98. [DOI] [PubMed] [Google Scholar]
  43. Klein, D. N. , Kotov R., and Bufferd S. J.. 2011. “Personality and Depression: Explanatory Models and Review of the Evidence.” Annual Review of Clinical Psychology 7, no. 1: 269–295. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Klein, K. R. , and Bufferd S. J.. 2011. “Personality and Depression: Explanatory Models and Review of the Evidence.” Annual Review of Clinical Psychology 7: 269–295. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Kohne, A. C. J. , and Isvoranu A. M.. 2021. “A Network Perspective on the Comorbidity of Personality Disorders and Mental Disorders: An Illustration of Depression and Borderline Personality Disorder.” Frontiers in Psychology 12: 680805. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Kramer, U. , Beuchat H., Grandjean L., and Pascual‐Leone A.. 2020. “How Personality Disorders Change in Psychotherapy: A Concise Review of Process.” Current Psychiatry Reports 22, no. 8: 41. [DOI] [PubMed] [Google Scholar]
  47. Labaka, A. , Goñi‐Balentziaga O., Lebeña A., and Pérez‐Tejada J.. 2018. “Biological Sex Differences in Depression: A Systematic Review.” Biological Research for Nursing 20, no. 4: 383–392. [DOI] [PubMed] [Google Scholar]
  48. Lauritzen, S. L. 1996. Graphical Models. Clarendon Press. [Google Scholar]
  49. Luty, S. E. , Carter J. D., McKenzie J. M., et al. 2007. “Randomised Controlled Trial of Interpersonal Psychotherapy and Cognitive‐Behavioural Therapy for Depression.” British Journal of Psychiatry 190: 496–502. [DOI] [PubMed] [Google Scholar]
  50. Malhi, G. S. , and Mann J. J.. 2018. “Depression.” Lancet 392, no. 10161: 2299–2312. [DOI] [PubMed] [Google Scholar]
  51. Morris, B. H. , Bylsma L. M., and Rottenberg J.. 2009. “Does Emotion Predict the Course of Major Depressive Disorder? A Review of Prospective Studies.” British Journal of Clinical Psychology 48, no. Pt 3: 255–273. [DOI] [PubMed] [Google Scholar]
  52. Mulder, R. , and Tyrer P.. 2023. “Borderline Personality Disorder: A Spurious Condition Unsupported by Science That Should Be Abandoned.” Journal of the Royal Society of Medicine 116, no. 4: 148–150. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Newton‐Howes, G. , Tyrer P., Johnson T., et al. 2014. “Influence of Personality on the Outcome of Treatment in Depression: Systematic Review and Meta‐Analysis.” Journal of Personality Disorders 28, no. 4: 577–593. [DOI] [PubMed] [Google Scholar]
  54. Paris, J. , and Zweig‐Frank H.. 2001. “A 27‐Year Follow‐Up of Patients With Borderline Personality Disorder.” Comprehensive Psychiatry 42, no. 6: 482–487. [DOI] [PubMed] [Google Scholar]
  55. Qian, X. , Townsend M. L., Tan W. J., and Grenyer B. F. S.. 2022. “Sex Differences in Borderline Personality Disorder: A Scoping Review.” PLoS ONE 17, no. 12: e0279015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. R Core Team . 2020. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing. [Google Scholar]
  57. Rao, S. , and Broadbear J.. 2019. “Borderline Personality Disorder and Depressive Disorder.” Australasian Psychiatry 27, no. 6: 573–577. [DOI] [PubMed] [Google Scholar]
  58. Ribeiro, J. D. , Huang X., Fox K. R., and Franklin J. C.. 2018. “Depression and Hopelessness as Risk Factors for Suicide Ideation, Attempts and Death: Meta‐Analysis of Longitudinal Studies.” British Journal of Psychiatry 212, no. 5: 279–286. [DOI] [PubMed] [Google Scholar]
  59. Saab, M. M. , Murphy M., Meehan E., et al. 2022. “Suicide and Self‐Harm Risk Assessment: A Systematic Review of Prospective Research.” Archives of Suicide Research 26, no. 4: 1645–1665. [DOI] [PubMed] [Google Scholar]
  60. Sarhan, Z. A. E. , El Shinnawy H. A., Eltawil M. E., Elnawawy Y., Rashad W., and Saadeldin Mohammed M.. 2019. “Global Functioning and Suicide Risk in Patients With Depression and Comorbid Borderline Personality Disorder.” Neurology Psychiatry and Brain Research 31: 37–42. [Google Scholar]
  61. Schmittmann, V. D. , Cramer A. O. J., Waldorp L. J., Epskamp S., Kievit R. A., and Borsboom D.. 2013. “Deconstructing the Construct: A Network Perspective on Psychological Phenomena.” New Ideas in Psychology 31, no. 1: 43–53. [Google Scholar]
  62. Schneibel, R. , Brakemeier E.‐L., Wilbertz G., Dykierek P., Zobel I., and Schramm E.. 2012. “Sensitivity to Detect Change and the Correlation of Clinical Factors With the Hamilton Depression Rating Scale and the Beck Depression Inventory in Depressed Inpatients.” Psychiatry Research 198, no. 1: 62–67. [DOI] [PubMed] [Google Scholar]
  63. Sekowski, M. , Gambin M., Sumlin E., and Sharp C.. 2022. “Associations Between Symptoms of Borderline Personality Disorder and Suicidality in Inpatient Adolescents: The Significance of Identity Disturbance.” Psychiatry Research 312: 114558. [DOI] [PubMed] [Google Scholar]
  64. Shah, R. , and Zanarini M. C.. 2018. “Comorbidity of Borderline Personality Disorder: Current Status and Future Directions.” Psychiatric Clinics of North America 41, no. 4: 583–593. [DOI] [PubMed] [Google Scholar]
  65. Smith, D. J. , Kyle S., Forty L., et al. 2008. “Differences in Depressive Symptom Profile Between Males and Females.” Journal of Affective Disorders 108, no. 3: 279–284. [DOI] [PubMed] [Google Scholar]
  66. Söderholm, J. J. , Socada J. L., Rosenström T. H., Ekelund J., and Isometsä E.. 2023. “Borderline Personality Disorder and Depression Severity Predict Suicidal Outcomes: A Six‐Month Prospective Cohort Study of Depression, Bipolar Depression, and Borderline Personality Disorder.” Acta Psychiatrica Scandinavica 148, no. 3: 222–232. [DOI] [PubMed] [Google Scholar]
  67. Spitzer, R. L. , Williams J. B., Gibbon M., and First M. B.. 1992. “The Structured Clinical Interview for DSM‐III‐R (SCID). I: History, Rationale, and Description.” Archives of General Psychiatry 49, no. 8: 624–629. [DOI] [PubMed] [Google Scholar]
  68. Stanley, B. , and Wilson S. T.. 2006. “Heightened Subjective Experience of Depression in Borderline Personality Disorder.” Journal of Personality Disorders 20, no. 4: 307–318. [DOI] [PubMed] [Google Scholar]
  69. Steffen, A. , Nübel J., Jacobi F., Bätzing J., and Holstiege J.. 2020. “Mental and Somatic Comorbidity of Depression: A Comprehensive Cross‐Sectional Analysis of 202 Diagnosis Groups Using German Nationwide Ambulatory Claims Data.” BMC Psychiatry 20, no. 1: 142. [DOI] [PMC free article] [PubMed] [Google Scholar]
  70. Tsujii, N. , Akashi H., Mikawa W., et al. 2014. “Discrepancy Between Self‐and Observer‐Rated Depression Severities as a Predictor of Vulnerability to Suicide in Patients With Mild Depression.” Journal of Affective Disorders 161: 144–149. [DOI] [PubMed] [Google Scholar]
  71. Vinkers, C. H. , Joëls M., Milaneschi Y., Kahn R. S., Penninx B. W., and Boks M. P.. 2014. “Stress Exposure Across the Life Span Cumulatively Increases Depression Risk and Is Moderated by Neuroticism.” Depression and Anxiety 31, no. 9: 737–745. [DOI] [PubMed] [Google Scholar]
  72. Wilson, S. , DiRago A. C., and Iacono W. G.. 2014. “Prospective Inter‐Relationships Between Late Adolescent Personality and Major Depressive Disorder in Early Adulthood.” Psychological Medicine 44, no. 3: 567–577. [DOI] [PMC free article] [PubMed] [Google Scholar]
  73. Witt, S. H. , Streit F., Jungkunz M., et al. 2017. “Genome‐Wide Association Study of Borderline Personality Disorder Reveals Genetic Overlap With Bipolar Disorder, Major Depression and Schizophrenia.” Translational Psychiatry 7, no. 6: e1155. [DOI] [PMC free article] [PubMed] [Google Scholar]
  74. Wolfe, K. L. , Nakonezny P. A., Owen V. J., et al. 2019. “Hopelessness as a Predictor of Suicide Ideation in Depressed Male and Female Adolescent Youth.” Suicide & Life‐Threatening Behavior 49, no. 1: 253–263. [DOI] [PMC free article] [PubMed] [Google Scholar]
  75. World Health Organisation . 2008. The Global Burden of Disease: 2004 Update. World Health Organization. [Google Scholar]
  76. Yalch, M. M. , Hopwood C. J., Fehon D. C., and Grilo C. M.. 2014. “The Influence of Borderline Personality Features on Inpatient Adolescent Suicide Risk.” Personality Disorders, Theory, Research, and Treatment 5, no. 1: 26–31. [DOI] [PubMed] [Google Scholar]
  77. Yen, S. , Peters J. R., Nishar S., et al. 2021. “Association of Borderline Personality Disorder Criteria With Suicide Attempts: Findings From the Collaborative Longitudinal Study of Personality Disorders Over 10 Years of Follow‐Up.” JAMA Psychiatry 78, no. 2: 187–194. [DOI] [PMC free article] [PubMed] [Google Scholar]
  78. Zanarini, M. C. , Frankenburg F. R., Reich D. B., and Fitzmaurice G.. 2012. “Attainment and Stability of Sustained Symptomatic Remission and Recovery Among Patients With Borderline Personality Disorder and Axis II Comparison Subjects: A 16‐Year Prospective Follow‐Up Study.” American Journal of Psychiatry 169, no. 5: 476–483. [DOI] [PMC free article] [PubMed] [Google Scholar]
  79. Zanarini, M. C. , Frankenburg F. R., Dubo E. D., et al. 1998. “Axis I Comorbidity of Borderline Personality Disorder.” American Journal of Psychiatry 155, no. 12: 1733–1739. [DOI] [PubMed] [Google Scholar]
  80. Zimmerman, M. , Martinez J. H., Young D., Chelminski I., and Dalrymple K.. 2013. “Severity Classification on the Hamilton Depression Rating Scale.” Journal of Affective Disorders 150, no. 2: 384–388. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Data S1. BPD symptoms

PMH-19-0-s003.png (193.1KB, png)

Data S2. Boostrap

PMH-19-0-s002.pdf (54.3KB, pdf)

Data S3. Bridge centrality

PMH-19-0-s001.pdf (5.5KB, pdf)

Data Availability Statement

The data presented in this study may be available on request from the corresponding author. The data are not publicly available due to ethical reasons.

Articles from Personality and Mental Health are provided here courtesy of Wiley

RESOURCES