Complex post-traumatic stress disorder (CPTSD) was initially proposed as a disorder distinct from PTSD by Herman (1992), who argued that traumatic, usually childhood experiences in CPTSD are repetitive and prolonged, with the individual feeling unable to escape (Cloitre et al., 2009). The CPTSD symptom profile includes the underdevelopment of emotional, social, cognitive and psychological competencies. CPTSD diagnostic criteria include the three core PTSD symptom clusters of re-experiencing, avoidance and a sense of threat, with three additional clusters unique to CPTSD: affect dysregulation, negative self-concept and disturbances in relationships (World Health Organization, 2018). These three additional clusters are collectively known as ‘disturbances in self-organisation’ (DSO) symptoms (Cloitre et al., 2011; Maercker et al., 2013). The additional symptoms reflect the unique impact complex traumatic experiences have on the development of interpersonal skills and the emotional regulation system (Burns et al., 2010; Shipman et al., 2005) and hippocampal function (Chetty et al., 2014) as well as the ability to form secure attachments (Karatzias et al., 2019b). CPTSD symptoms differ from the primary risk factor for PTSD which is one incident of intense trauma (Zuj et al., 2016). CPTSD is more likely to arise following a combination of the above risk factors (Karatzias et al., 2019b) as they specifically impact complexity of symptoms (Cloitre et al., 2009; Hyland et al., 2017).

Following multiple empirical investigations for the validity of construct (Cloitre et al., 2013; Karatzias et al., 2017; Perkonigg et al., 2016; Sachser et al., 2017), CPTSD was only recently recognised as distinct from PTSD in the International Classification of Diseases, 11th revision (ICD-11; WHO, 2018). The ICD-11 points to DSO symptoms in order to differentiate CPTSD from PTSD. Meanwhile, CPTSD is not currently recognised by the Diagnostic and Statistical Manual of Mental Disorders, 5th revision (DSM-5; APA, 2013), although it is possible that the next revision of the DSM will consider recognising the disorder (Cliotre et al., 2020).

The utility and discriminatory validity of separating the two disorders have been a topic of debate. Prior to the publication of ICD-11, Resick et al. (2012) argued that there is no substantial evidence to support the need for a new diagnostic category, stating that the clinical benefit had not been established and that current research lacked rigour. In a response, Herman (2012) argued that the authors had set an arbitrarily high standard for inclusion of research. Additionally, Clotire et al. (2012) argued it would be clinically beneficial to have simplicity in the overall classification structure, which would in turn improve the conceptual organisation of symptoms and causes relevant to CPTSD.

This rebuttal has particular relevance to the DSM, given that the DSM-5 diagnostic criteria for PTSD underwent revisions recognising additional DSO symptoms as part of the core symptom profile. This revision, according to Cloitre et al. (2013), reduces efficiency, as mental health providers prefer diagnoses to have a limited number of symptoms, so as to increase diagnostic goodness of fit (accuracy of description for any one patient), and thus to increase efficacy of treatment. Finally, the distinction between PTSD and CPTSD, as Cloitre et al. (2009) argue further, allows for a better understanding of treatment options, which is especially important given the prevalence of CPTSD: 2.6% in Israel (Ben-Ezra et al., 2018) and 3.8% in the USA (Cloitre et al., 2019), compared with a 6.8% prevalence of PTSD (Gradus, 2007). However, it is possible that CPTSD may be underdiagnosed as it is a novel and emerging disorder.

The International Society for Traumatic Stress Studies (ISTSS) recommends that CPTSD treatment be delivered in a phase-based approach, initially addressing DSO symptoms through an emotional stabilisation phase, prior to addressing the core PTSD symptoms through a trauma-processing phase (Cloitre et al., 2011). One third of PTSD patients with persistent DSO symptoms are less responsive to traditional clinical interventions (Winders et al., 2020; Van der Kolk et al., 2007) such as prolonged exposure and cognitive-processing therapy (Jonas et al., 2013; Watkins et al., 2018). Gilbert and Irons (2004) and Karatzias et al. (2019a) propose that these more psychologically pervasive DSO symptoms may be more responsive to compassion-based interventions, and individuals with CPTSD who engage in compassion-based interventions to address these DSO symptoms will see greater symptom reduction than those that engage in traditional clinical interventions alone.

Compassion-Based Interventions as a Treatment for CPTSD

Compassion-based interventions can be effective when addressing symptomology associated with shame (Karatzias et al., 2019a). Lee et al. (2001) propose that shame can mediate CPTSD symptoms and Karatzias et al. (2019a) have shown an association between the two. The Ehlers and Clark’s (2000) cognitive model of PTSD focuses on mitigating fear; symptoms arise from event processing in a way that maintains a sense of threat. Treatment attempts to re-process the memories of the event, reducing maladaptive coping strategies. Lee et al.’s (2001) alternative model states that the presence of shame following a traumatic event can maintain a self-critical dialogue and in turn a sense of threat. Shame can become an obstacle to recovery, affecting the individual’s ability to engage in more traditional clinical interventions such as cognitive behavioural therapy (CBT). Lee et al. (2001) propose working with compassion-based interventions, such as compassion-focused therapy (CFT), to reduce shame and self-criticism in PTSD sufferers.

CFT was initially developed as an adjunct to CBT by Gilbert (2009), aiming to balance the affiliative system and reduce threat-based processing that can contribute to psychopathology. CFT seeks to help individuals respond to their distress through a compassionate lens, enhancing the capacity to self-soothe and reduce self-critical self-talk (Gilbert, 2009, 2010), thus allowing the individual to feel safer in their environment.

There are currently no randomised controlled trials evaluating the efficacy of CFT for CPTSD. However, Karatzias et al. (2019a) found significant associations between low scores on all self-compassion subscales (self-kindness, self-judgement, common humanity, isolation, mindfulness and over-identification) and the existence of emotional hypoactivation and negative self-concept. Significant negative association between four self-compassion subscales (self-kindness, self-judgement, common humanity and isolation) and disturbances in relationships were also identified. High self-judgement significantly predicted affect dysregulation and negative self-concept, which are associated with resistance to traditional clinical interventions in CPTSD (Gilbert & Irons, 2004). The most salient issue with this study is that analyses were based on correlation, and therefore could not deduce causality. Furthermore, self-report scales may involve biased results particularly with populations that experience high levels of shame (Van de Mortel, 2008), as this may lead to dishonest responses.

Despite this, these findings are an important first step in identifying psychological factors that are associated with CPTSD, which can inform future interventional research. Specifically, Karatzias et al. (2019a) found that self-compassion was not significantly associated with core PTSD symptom clusters despite several studies showing this relationship (Winders et al., 2020). This may be partially due to previous studies using DSM-5 (APA, 2013) rather than the ICD-11 (WHO, 2018) diagnostic criteria. The ICD-11 tightened its criteria for PTSD to avoid over-diagnosing (Hansen, 2017). In contrast, the DSM-5 expanded its criteria for PTSD from primarily a fear-based disorder to include ‘negative alterations in cognition and mood’. These expanded symptoms include persistent negative evaluation of self or others, elevated self-blame, negative emotional states and reckless and self-destructive behaviour (Stein et al., 2014). These symptoms are parallel with the DSO symptom criteria for CPTSD in the ICD-11. As a result, it is possible that studies using the expanded criteria will find correlates between the DSM-5 PTSD symptoms of ‘negative alterations in cognitions and mood’ and self-compassion subscales, which other studies using the tighter ICD-11 core PTSD symptom criteria will not. This compromises clinical utility and thus the development of treatment protocols. Uniformity regarding this would contribute to cohesion in future research especially where CPTSD is concerned. As the ICD-11 diagnostic criteria for CPTSD could be thought of as a division between different DSM-5 symptom clusters (Stein et al., 2014), research targeting these clusters in isolation will allow differentiation between correlates of PTSD and CPTSD, further highlighting if these sibling disorders require qualitatively different treatment protocols. In the absence of data regarding the efficacy of CFT for CPTSD, research regarding compassion-based interventions for CPTSD’s constituent parts, firstly the DSO symptom profile and secondly the core PTSD symptom profile, would provide valuable insight.

Compassion-Based Interventions as a Treatment for Disturbances in Self-organisation

Prior to a trauma-processing phase of treatment, an emotional stabilisation phase of treatment is recommended to address the DSO symptoms, ensure patient safety and avoid exacerbation of symptoms during the following trauma-processing phase (Cloitre et al., 2011). To date, there have been no studies evaluating compassion-based interventions as a treatment for DSO symptoms. In the absence of these studies, it is possible to assess the available research for significant effects of compassion-based interventions on the DSO symptom indicators of affect dysregulation, negative self-concept and disturbed relationships as described by the International Trauma Questionnaire (ITQ; Shevlin et al., 2018).

Research compiled from three meta-analyses (Ferrari et al., 2019; Kirby et al., 2017; Wilson et al., 2019), one systematic review (Craig et al., 2020) and one literature review (Kirby, 2017) regarding the efficacy of compassion-based interventions yielded five interventional studies using comparison groups exploring loving kindness meditation, mindfulness-based cognitive therapy with a self-compassion element, mindful self-compassion and CFT (mindfulness-only-based interventions were excluded) as a treatment for clinical populations displaying similar symptoms to the constituent components of DSO (emotionally unstable personality disorder, depressive symptoms, mental health difficulties). The above compassion-based interventions utilise the same theoretical framework as CFT, aiming to reduce shame, balance the affiliative system and encourage self-soothing during acute stress. Of the five available studies, only two used an active control group. One of the two active control trials, Feliu-Soler et al. (2017), found significant improvements in self-concept following a 3-week loving kindness meditation relative to a mindfulness meditation intervention. This study may however lack population validity as it used a predominantly female sample. Although a lack of follow-up makes it difficult to ascertain if the results were maintained, positive outcomes were achieved in a short treatment duration. In the other active control trial, Cuppage et al. (2018) found a similar result with CFT showing significant improvements in psychopathology and a reduction in self-criticism. The transdiagnostic nature of the group suggests that CFT is addressing the underlying psychological processes giving rise to these psychopathologies, rather than just the symptoms presented. This aligns with Lee et al.’s (2001) theory of self-criticism contributing to psychopathology. Of the remaining three studies, all found significant improvements in depressive symptoms (Neff & Germer, 2013; Lee & Bang, 2010; Shahar et al., 2015).

Additional research shows significant improvements in outcome measures across domains of anxiety (Gharraee et al., 2018; Navab et al., 2019; Noorbala et al., 2013; Sommers-Spijkerman et al., 2018), substance abuse disorder (Carlyle et al., 2019) and eating disorders (Duarte et al., 2017). The available evidence suggests that CFT may be a suitable treatment to address the range of complex symptoms in the emotional stabilisation phase. In the absence of an agreed treatment protocol, it is worthwhile investigating this with actively controlled, large-scale studies to establish efficacy relative to other treatments.

Compassion-Based Interventions as a Treatment for PTSD

Because CPTSD can be resistant to conventional PTSD treatments, alternative approaches need to be evaluated. Much of the research regarding the efficacy of compassion-based interventions in PTSD treatment lacks statistical significance, control groups or causal considerations. Research compiled from a recent systematic review (Winders et al., 2020) yielded nine interventional studies exploring compassion-based interventions for PTSD. Of these nine, five used comparison groups, and of these five, one showed a greater significant reduction in PTSD symptoms in the compassion-based group (using cognitive-based compassion training) relative to the control group (Lang et al., 2017). Of the four remaining studies using comparison groups, one showed an improvement in PTSD symptoms in both the compassion-based condition (imagery re-scripting) and the traditional clinical intervention condition (imaginal exposure) with a slightly larger improvement in the traditional intervention condition (Hoffart et al., 2015). The remaining three showed a non-significant improvement (Beaumont et al., 2012, 2016; Held & Owens, 2015), although these used small sample sizes and were therefore potentially underpowered, giving rise to type II errors. Of the four studies without a control group, three found a significant reduction in PTSD symptoms following compassion-based interventions (Au et al., 2017; Kearaney et al. 2013; Muller-Engelmann et al., 2019). The study that did not find a reduction in PTSD symptoms did however find an improvement in affect (Held et al., 2018). The absence of a control group makes it difficult to ascertain if the observed effect is due to the nature of compassion-based interventions and their effect of reducing shame and improving self-soothing capacity, or due to the effect of engaging in treatment generally. Interestingly, significant correlations between PTSD symptom severity and self-compassion have been demonstrated consistently in populations with childhood abuse or interpersonal trauma (Barlow et al., 2017; Bistricky et al., 2017; Miron et al., 2015; Thomson & Waltz’s, 2008) implying the inconsistency of results in the above research may be mediated by this risk factor.

Although this research is promising, it is difficult to draw conclusions about causality from cross-sectional studies. For example, it is possible that low self-compassion is a pre-trauma risk factor for PTSD, or that self-compassion is better adopted by those with less severe symptomatology. Greater symptom reduction was found in populations with either higher self-judgement or elevated trauma-related shame (Au et al., 2017; Thomson & Waltz’s, 2008). Further research is needed to clarify if this effect is only relevant to sub-populations of PTSD, such as those with complex trauma, as the research may not generalise to all PTSD populations.

Given (a) the lack of statistical significance, control groups and causal considerations in the majority of compassion-based intervention studies for PTSD and (b) the prevalence of literature reporting the efficacy of traditional clinical interventions for trauma processing such as trauma-focused cognitive-processing therapy (Jonas et al., 2013; Watkins et al., 2018), large-scale controlled studies are needed to propose CFT or other similar compassion-based interventions as a primary treatment during the trauma-processing phase of treatment for CPTSD. Future interventional and controlled research is needed to support and add causality to the results seen in the cross-sectional data. In addition to this, future research would benefit from improved statistical power. This could be achieved through use of a repeated measures design, although this can be difficult in therapeutic research; another way to achieve this is to increase sample sizes. Additionally, improvement in symptomology can be seen as a result of engaging in treatment independent of the specific intervention (Safer & Hugo, 2006), and therefore future research could benefit from including an active control group using traditional clinical interventions to account for this effect. The addition of this research would justify further investigation into CFT as a primary treatment in the trauma-processing phase of treatment of CPTSD.

Acceptability and Viability of Compassion-Based Interventions

Although efficacy of interventions can be assessed with outcome measures, their acceptability and viability are other important considerations (Moore et al., 2015). CFT has been found to have good attrition among a variety of populations including those with dementia (94%; Collins et al., 2018), perfectionism (96%; Rose et al., 2018) and psychosis (82%; Laithwaite et al., 2009) and those in an inpatient setting (95%; Braehler et al., 2013). Given the challenges faced regarding interpersonal domains as described above, this has particular relevance, as populations with CPTSD may find it difficult to commit to a programme of treatment given these disturbances (16–20%; Vogel et al., 2017). In addition, several studies have measured individual satisfaction with CFT, with consistently high levels reported (Duarte et al., 2017; Graser et al., 2016; Clapton et al., 2018), although it is worth noting that populations with high shame may feel unable to give critical feedback, so these results should be interpreted cautiously. Further research into satisfaction with the intervention would benefit from being obtained blind to account for this effect. These are particularly important factors in the evaluation of CFT for those who complete the therapy in groups, as it can be discouraging to peers for the group to reduce in size or be described negatively by peers. CFT has been effectively delivered in groups (McManus et al., 2018; Feliu‐Soler et al., 2017; Heriot‐Maitland et al., 2014), improving cost effectiveness, waitlist time and viability for treatment centres. This is particularly salient considering one implication for counsellors is the additional training they may need to undertake to effectively deliver CFT.

Further studies evaluating the extent of this for 1:1 and group delivery would inform decisions regarding the cost effectiveness versus potential changes in outcome measures. An additional issue regarding the application of CFT is an absence of implementation fidelity, as details regarding primary outcomes, who delivered the therapy, and their training were lacking (Craig et al., 2020). This can cause disparity in interpretation of results, and make it difficult for group facilitators, counsellors or psychologists to deliver CFT in the way the research has shown it to be effective. Future research would benefit from uniformity in these areas. Although CFT could benefit from this, it is considered a highly acceptable treatment option by patients and is well delivered in groups. In addition to outcome efficacy, these are important considerations when assessing its strength as an intervention.


In conclusion, the recent decision by the IDC-11 to recognise CPTSD as a disorder distinct from PTSD has raised important questions regarding risk factors, validity of construct and clinical utility. However, it has been shown (Cloitre et al., 2013) that the new diagnostic criteria promote ease of use and simplicity. It has allowed scope for studies of prevalence of CPTSD, highlighting the need to assess potential treatments. The current recommendations of a phase-based approach of treatment (Cloitre et al., 2011) entail addressing DSO symptoms separately to core PTSD symptoms as this promotes ease of treatment. Evidence supports the use of CFT as a primary intervention for DSO symptoms in the emotional stabilisation phase although there are other more suitable and well-researched interventions for the core PTSD symptoms in the trauma-processing phase. As a primary treatment for DSO symptoms, CFT could benefit from further uniformity regarding delivery, although the treatment is considered highly acceptable among patient populations and can be delivered successfully in groups. The above factors imply that CFT is a highly viable option for addressing the more psychologically pervasive symptoms of CPTSD. Future research in this population would benefit from evaluating the efficacy of a CFT intervention for emotional stabilisation prior to a trauma-processing intervention, versus a trauma-processing intervention alone. Consensus between the ICD-11 and DSM-5 regarding diagnostic criteria will be a major step in understanding these research outcomes. More high-quality and large-scale research investigating CFT as a primary intervention for individuals with CPTSD is a worthwhile movement towards an agreed treatment protocol for the disorder.