Interpretation of findings and comparison with existing literature

This was a highly trauma-exposed sample, with 93% of patients reporting having experienced traumatic event(s) as defined in the ICD-11 diagnostic criteria for PTSD, most commonly sexual trauma or violence. As in previous studies of trauma-exposed individuals, C-PTSD was more prevalent than classic PTSD alone. ^{Reference Karatzias, Cloitre, Maercker, Kazlauskas, Shevlin and Hyland28} The 57% rate of C-PTSD we have identified among people with a clinical BPD diagnosis falls in-between the C-PTSD rates of 45% and 68% identified in previous UK and USA studies, which included patients who did not necessarily have a clinical diagnosis of BPD, but were assessed by an interviewer to meet the diagnostic criteria. ^{Reference Cloitre, Garvert, Weiss, Carlson and Bryant13,Reference McBride, Goulden, Barnicot, Corrigan, Shen and Guillemard29}

We found that sexual trauma was particularly strongly associated with C-PTSD, and with the much smaller subset of participants experiencing PTSD without DSO, compared with those without PTSD symptoms. This is consistent with findings from prior studies in which sexual violence was linked to a greater likelihood and severity of PTSD symptoms compared with non-sexual violence. ^{Reference DiMauro, Renshaw and Blais30–Reference Briere and Elliott32} We found no evidence of an association between C-PTSD and age at trauma onset, consistent with prior findings that severe interpersonal trauma in childhood or adulthood is equally likely to result in C-PTSD. ^{Reference Palic, Zerach, Shevlin, Zeligman, Elklit and Solomon33} We also found no evidence of an association between C-PTSD and repeated trauma. One explanation for this lies in the fact that the majority of our participants with a BPD diagnosis had a history of repeated sexual and/or violent trauma, most commonly originating in childhood, and thus this factor did not differentiate well between BPD with and without C-PTSD. However, these analyses were limited by missing data on trauma onset and frequency.

Although C-PTSD and BPD share near-identical diagnostic criteria for emotional dysregulation, they have been theorised to be differentiated by stasis versus lability in interpersonal relationship difficulties and in self-concept. ^{Reference Hyland, Karatzias, Shevlin and Cloitre12,Reference Cloitre, Garvert, Weiss, Carlson and Bryant13} Yet, in our data, indices of interpersonal instability including abandonment fears and abandonment–avoidant behaviour, and identity instability, were significantly higher among participants meeting the criteria for both BPD and C-PTSD compared with participants who met the criteria for BPD and not C-PTSD. The identified odds ratios suggest that this could be a clinically significant effect. One interpretation of this finding is that the distinction between C-PTSD as characterised by stable negative interpersonal and identity functioning, and BPD as characterised by interpersonal and identity instability, breaks down when a person meets the diagnostic criteria for both. Certainly, both C-PTSD and BPD are characterised by dissociation, which is also strongly linked to sexual abuse and thought to contribute, at least in part, to the identity difficulties seen in people diagnosed with BPD. ^{9,Reference Wilkinson-Ryan and Westen34} Alternatively, perhaps the measures used do not adequately capture the stasis versus lability aspects of these constructs. Additionally, in alignment with their increased likelihood of engaging in self-harming behaviour, perhaps endorsing diagnostic criteria for both BPD and C-PTSD is a hallmark of worse overall psychiatric symptom severity. The wide confidence intervals suggest uncertainty in ascertaining the true magnitude of this difference. The uncertainty may stem from the wide standard deviations on the relevant BPD interpersonal instability item scores, which in turn, relates to the widely recognised heterogeneity within the BPD diagnosis whereby there are 256 different ways in which one can meet five of the nine possible diagnostic criteria. ^{Reference Fowler, Carlson, Orme, Allen, Oldham and Madan35}

Implications for clinical practice and further research

The findings imply that C-PTSD symptoms should routinely be considered in people diagnosed with personality disorder with borderline pattern or BPD and treated by mental health services. On the basis of our findings, C-PTSD symptoms may be linked to higher risk behaviours such as self-harm, emphasising the importance of assessing and treating them. The clinical utility of assigning both diagnoses should be critically considered. ICD-11 recommends that, where an individual meeting the diagnostic criteria for C-PTSD also meets diagnostic criteria for a personality disorder, the latter should only be assigned where clinically useful. ^{Reference Winsper8}

Two-thirds of the patients in our sample who did not meet the diagnostic criteria for PTSD did report C-PTSD DSO symptoms, i.e. emotional dysregulation, avoidance and alienation in relationships and a negative self-concept. Thus, it seems that although virtually all of our sample had a history of trauma, and the majority had responded to this in ways that met the diagnostic criteria for C-PTSD, a substantial minority had responded in ways that are theorised to reflect trauma-related responses, but did not meet all of the C-PTSD criteria. This potentially calls into question the utility of these diagnostic constructs, which place strict definitions around classification of someone’s experiences as either a trauma-related disorder or a personality-related disorder. In clinical practice, it is important to look beyond the binary parameters of diagnosis and to formulate each individual’s particular strengths and difficulties, with an exploration of the personal relevance of trauma, adversity and other aetiological factors in shaping their development through the life course. ^{Reference Watts36}

Current evidence-based psychological interventions for BPD, such as dialectical behaviour therapy and mentalisation-based therapy, are present-focused and do not directly address trauma or PTSD symptoms. ^{Reference Bateman and Fonagy37,Reference Linehan38} By contrast, multicomponent trauma-focused and exposure-based interventions to directly address PTSD symptoms, are recommended for PTSD and C-PTSD. ^{Reference Karatzias, Murphy, Cloitre, Bisson, Roberts and Shevlin39–41} Randomised controlled trials in the USA and Germany have found that trauma-focused interventions can safely and effectively be combined with dialectical behaviour therapy for dual-diagnosed individuals, achieving high PTSD remission rates. ^{Reference Harned, Korslund and Linehan42–Reference Bohus, Kleindienst and Hahn44} The effectiveness of psychological interventions in addressing C-PTSD DSO is currently unclear. Although interventions such as dialectical behaviour therapy and mentalisation-based therapy are known to help with the behavioural consequences of emotional dysregulation, such as self-harm and substance misuse, little research has evaluated their effectiveness for ameliorating internal experiences of emotional instability. ^{Reference Karatzias, Murphy, Cloitre, Bisson, Roberts and Shevlin39,Reference Cristea, Gentili, Cotet, Palomba, Barbui and Cuijpers45} Cognitive–behaviour therapy, exposure-based therapies and eye movement desensitisation reprocessing therapy have been found effective for improving C-PTSD-aligned interpersonal difficulties and negative self-concept. ^{Reference Karatzias, Murphy, Cloitre, Bisson, Roberts and Shevlin39} It is unclear whether the identity and interpersonal instability associated with BPD requires a different treatment approach, nor how treatment effectiveness differs for patients experiencing both sets of difficulties. In the UK, trauma survivors given a diagnosis of BPD are campaigning for trauma-focused treatment pathways that are entirely separated from the personality disorder construct. ^{Reference Lomani, Alyce, Aves, Chevous, Clayton and Conway7} It is important to take into account the diagnostic and treatment preferences of each individual. Beyond specific interventions, trauma-informed care is an approach that is thought to be helpful for trauma survivors transdiagnostically, and is widely used in the UK and USA. ⁴¹ It is a set of principles that can run alongside both present-focused and trauma-focused interventions. Key aspects include understanding, where relevant, how people’s experiences and behaviours have been shaped by coping with trauma and adversity, and using this understanding to formulate a compassionate approach to their difficulties. ^{Reference Sweeney, Filson, Kennedy, Collinson and Gillard46,Reference Elliott, Bjelajac, Fallot, Markoff and Reed47} Other key aspects are enabling access to trauma-focused interventions where this is needed and wanted, and creating an environment that prevents re-traumatisation by avoiding coercion, increasing trustworthiness and transparency, offering collaboration and empowerment, and creating a sense of safety. ^{Reference Sweeney, Filson, Kennedy, Collinson and Gillard46,Reference Elliott, Bjelajac, Fallot, Markoff and Reed47}

Further research on the overlap between BPD and C-PTSD should draw on both clinical and non-clinical samples, where the extent of overlap may differ. Additionally, the ICD-11 has only recently been implemented in the World Health Organization member states. In contrast to the DSM-IV/DSM-5 criteria for BPD used in the present study, the ICD-11 has further blurred the distinction between BPD and C-PTSD by stating in the diagnostic guidelines for personality disorder with borderline pattern that ‘some individuals with a borderline pattern may also believe themselves to be inadequate, bad, guilty, disgusting, and contemptible’ (akin to the stable negative self-concept associated with C-PTSD), and that ‘some individuals with a borderline pattern may feel profoundly different and isolated from other people, may feel a painful sense of alienation and pervasive loneliness…..and may have problems in establishing and maintaining consistent and appropriate levels of trust in interpersonal relationships’ (akin to C-PTSD stable difficulties in sustaining relationships and feeling close to others). ⁹ It will be important to determine how the availability of the new ICD-11 C-PTSD diagnosis affects patients with a BPD diagnosis; in particular, how many people are given an additional C-PTSD diagnosis or are re-diagnosed altogether. Further research is required to better understand the impact of classifying a person’s difficulties as BPD versus C-PTSD. Assignation of the latter diagnosis may arguably be associated with less stigmatising responses by health professionals and with better access to trauma-focused interventions, where these are wanted and deemed clinically appropriate. By contrast, concerns have been raised that trauma survivors with a BPD diagnosis who do not fit within the narrow parameters of the C-PTSD diagnostic criteria may encounter a two-tier system, where access to trauma-based formulations of their experiences and/or trauma-focused interventions is desired but denied. ^{Reference Watts36} Finally, it will be helpful to evaluate whether taking a trauma-informed approach alongside offering evidence-based interventions, where relevant, can improve patient satisfaction and treatment outcomes.

Strengths and limitations

The findings are strengthened by the high consent rate and representative sample of patients referred to multiple UK community personality disorder services. The use of validated semi-structured diagnostic interviews to ascertain BPD and PTSD diagnoses is a further strength. The main weakness was the use of an analogue self-report measure to assess C-PTSD DSO symptoms. Additionally, information on trauma timing and frequency was not available for the full sample.

In conclusion, a high proportion of patients with a BPD diagnosis in mental health services are likely to also meet the diagnostic criteria for C-PTSD. A further substantial minority are likely to have experienced trauma and to meet some, but not all, diagnostic criteria for C-PTSD. The findings imply that C-PTSD symptoms should be routinely considered in people diagnosed with BPD. The implications of co-occurring C-PTSD symptoms for case formulation and treatment should be carefully considered, in line with the recommendations in the National Institute for Health and Care Excellence guidelines that patients with PTSD should be offered evidence-based interventions that directly address PTSD symptoms. Beyond arbitrary diagnostic bifurcations, it is possible that a transdiagnostic trauma-informed approach to working with trauma survivors, alongside evidence-based interventions, may be helpful.