Discussion

To the best of our knowledge, this is the first meta review of reviews aiming to understand the developmental pathways of BD and BPD, disorders that share some phenotypic features that could imply an overlap of aetiological mechanisms. 22 eligible reviews provided significant data about the factors which might contribute to the onset of BD or BPD. The current meta-review demonstrates that there are many ‘distinct’ clinical, environmental, psychosocial, and biological variables that can be found early in the course of BD and BPD, even in at-risk stages, but the disorders share a variety of clinical and vulnerability factors too. However, since these ‘distinct’ variables are evident only either in BD or BPD at-risk reviews, their distinctive value is speculative until further systematic longitudinal studies examine these factors in both disorders.

A notable and critical limitation of the literature is there were no studies comparing BD and BPD at-risk populations at the same time, compounding the difficulty of understanding specific BD or BPD developmental trajectory. Additionally, the neurobiological data from the BD at-risk studies are currently limited and there are no comparing studies done in BPD at-risk populations. This is why despite many previous commentary pieces on this issue (Bassett, Reference Bassett2012; Bayes et al., Reference Bayes, McClure, Fletcher, Román Ruiz del Moral, Hadzi-Pavlovic, Stevenson and Parker2015; Deltito et al., Reference Deltito, Martin, Riefkohl, Austria, Kissilenko, Corless and Morse2001; Massó Rodriguez et al., Reference Massó Rodriguez, Hogg, Gardoki-Souto, Valiente-Gómez, Trabsa, Mosquera and Amann2021; Paris, Reference Paris2004; Sanches, Reference Sanches2019; Smith, Muir, & Blackwood, Reference Smith, Muir and Blackwood2004; Stone, Reference Stone2006; Zimmerman & Morgan, Reference Zimmerman and Morgan2013a, Reference Zimmerman and Morgan2013b), in reality at the current time it is not possible to answer whether these disorders should be on the same affective continuum, or they should be regarded as separate nosological conditions.

Gender, differences in the white matter, changes in the amygdala, neural reward circuit dysfunctions, DBD, subsyndromal hypomania, cyclothymia or bipolar NOS, frequency and loading of affective symptoms, and antidepressant use were factors examined only in BD studies. Only changes in the amygdala, neural reward circuit dysfunctions, subdyndromal hypomania, cyclothymia or bipolar NOS, frequency and loading of affective symptoms consistently predicted BD transition. Interestingly there was no data available for emerging BPD for cyclothymia although previous research comparing participants with BD and BPD found that participants with BPD too show similar or even higher levels of abnormal cyclothymic temperament (Eich et al., Reference Eich, Gamma, Malti, Vogt Wehrli, Liebrenz, Seifritz and Modestin2014; Nilsson, Jørgensen, Straarup, & Licht, Reference Nilsson, Jørgensen, Straarup and Licht2010). Previous evidence also shows that hypomanic days were reported frequently in both the BD and BPD subcohort (Socada, Söderholm, Rosenström, Ekelund, & Isometsä, Reference Socada, Söderholm, Rosenström, Ekelund and Isometsä2021). However, based on our findings while hypomanic symptoms predicted BD onset (Faedda et al., Reference Faedda, Marangoni, Serra, Salvatore, Sani, Vázquez and Koukopoulos2015), there was no data pertaining to BPD onset. Likewise, accumulated evidence shows BD patients demonstrated enlarged amygdala (Soares & Young, Reference Soares and Young2016) while BPD patients showed decreased amygdala volumes (Perez-Rodriguez et al., Reference Perez-Rodriguez, Bulbena-Cabré, Bassir Nia, Zipursky, Goodman and New2018). Our results are not in accordance with these results because according to Cahn et al.'s (Reference Cahn, Keramatian, Frysch, Yatham and Chakrabarty2021) findings, adolescents with mania failed to exhibit normal increases in amygdala volume. It is interesting that studies of people who are at-risk of developing BD has showed decreased in amygdala volumes while previous studies with BPD populations have also showed the same results. However, since there was no data on amygdala changes or hypomania symptoms in participants with BPD features, it is not possible to conclude at the moment whether both at-risk populations fail to exhibit normal increases in amygdala or hypomania is a shared feature.

Hu et al. (Reference Hu, Stavish, Leibenluft and Linke2020) indicated that differences in white matter integrity between high-risk individuals and control could occur in earlier childhood. This is consistent with a SR (Serafini et al., Reference Serafini, Pompili, Borgwardt, Houenou, Geoffroy, Jardri and Amore2014) which found reduced corpus collosum volume and increased rates of deep white matter hyperintensities were more specific to paediatric BD in comparison to unipolar depression. There is however a need to replicate these findings with future longitudinal follow-up studies in both at-risk populations.

Amongst the factors examined in relation to transition to BD, the greatest amount of evidence was for family history of BD. Although inconsistencies in results were present in family history of BD studies (Keramatian et al., Reference Keramatian, Chakrabarty, Saraf and Yatham2021; Lau et al., Reference Lau, Hawes, Hunt, Frankland, Roberts and Mitchell2017; Narayan, Allen, Cullen, & Klimes-Dougan, Reference Narayan, Allen, Cullen and Klimes-Dougan2013; Rasic, Hajek, Alda, & Uher, Reference Rasic, Hajek, Alda and Uher2013; Ratheesh et al., Reference Ratheesh, Davey, Hetrick, Alvarez-Jimenez, Voutier, Bechdolf and Cotton2017), some studies suggest that there might be a relative specificity of family history of BD to predicting later BD in MDD samples (Ratheesh et al., Reference Ratheesh, Davey, Hetrick, Alvarez-Jimenez, Voutier, Bechdolf and Cotton2017; Vandeleur, Merikangas, Strippoli, Castelao, & Preisig, Reference Vandeleur, Merikangas, Strippoli, Castelao and Preisig2013). Likewise, the evidence in the current review was also inconsistent as some of the individual studies found a significant relationship between a family history of other mental illnesses (i.e. affective disorder or depression) and later BD conversion while the others did not. This is not surprising because although high-risk studies can be informative about transition to BD (DelBello & Geller, Reference DelBello and Geller2001; Duffy et al., Reference Duffy, Doucette, Lewitzka, Alda, Hajek and Grof2011; McGuffin et al., Reference McGuffin, Rijsdijk, Andrew, Sham, Katz and Cardno2003), these studies still have not supported the validity of the pre-pubertal BD phenotype and not all children of parents with BD develop BD or a mood disorder (Duffy, Carlson, Dubicka, & Hillegers, Reference Duffy, Carlson, Dubicka and Hillegers2020; Malhi, Moore, & McGuffin, Reference Malhi, Moore and McGuffin2000; Malhi, Morris, Hamilton, Outhred, & Mannie, Reference Malhi, Morris, Hamilton, Outhred and Mannie2017).

Parenting behaviour/style, parent-child relationship quality, maternal characteristics, attachment, impulsivity, experiential avoidance, disturbances in self representation, dissociation, comorbid oppositional defiant disorder, comorbid conduct disorder, somatisation, general psychosocial functioning, and social and physical aggression in childhood were examined only in BPD studies. Attachment, impulsivity, experiential avoidance, disturbances in self representation, dissociation, comorbid oppositional defiant disorder, somatisation, general psychosocial functioning, social and physical aggression in childhood consistently predicted later BPD symptoms. Interestingly again, none of the BD at-risk reviews mentioned impulsivity although it is commonly found both in BD and BPD patients (di Giacomo et al., Reference di Giacomo, Aspesi, Fotiadou, Arntz, Aguglia, Barone and Zaccheroni2017; Pauselli, Verdolini, Santucci, Moretti, & Quartesan, Reference Pauselli, Verdolini, Santucci, Moretti and Quartesan2015; Reich, Zanarini, & Fitzmaurice, Reference Reich, Zanarini and Fitzmaurice2012).

Previous studies comparing BD and BPD patients showed that BPD patients had significantly more difficulties in interpersonal relationships, endorsed negative and distressing beliefs about themselves and their relationships, and had dysfunctional maternal relationships as compared to BD patients (Bayes et al., Reference Bayes, McClure, Fletcher, Román Ruiz del Moral, Hadzi-Pavlovic, Stevenson and Parker2015; Fletcher, Parker, Bayes, Paterson, & McClure, Reference Fletcher, Parker, Bayes, Paterson and McClure2014; Nilsson et al., Reference Nilsson, Jørgensen, Straarup and Licht2010). Our findings also indicate that relational difficulties with the self and others, such as disturbances in self representation, negative experiences in current relationships and insecure attachment, are evident in people with BPD features. Conflictive interpersonal relationships could distinguish BPD from BD (Massó Rodriguez et al., Reference Massó Rodriguez, Hogg, Gardoki-Souto, Valiente-Gómez, Trabsa, Mosquera and Amann2021). However, to be able to support this, these factors should also be studied in BD at-risk populations.

Relatively less evidence has accumulated about precursors related to the BPD development. The reason might be ascribed to the fact that the BPD phenotype is less clearly identified compared to the BD prodromal phase, although its underlying dimensions are evident in the reviews included in this study. This might be attributable to the short follow-up periods and not integrating contemporary methods for defining biological, psychological, and social precursor signs for the development of BPD (Chanen & Kaess, Reference Chanen and Kaess2011). Staging models, like in BD or psychosis, could be utilised to help predict the course of prognosis with external validation through biomarkers (Hutsebaut & Aleva, Reference Hutsebaut and Aleva2021; Videler, Hutsebaut, Schulkens, Sobczak, & van Alphen, Reference Videler, Hutsebaut, Schulkens, Sobczak and van Alphen2019). However, early stages of most of these symptoms are non-specific and overlap with other disorders (Berk et al., Reference Berk, Post, Ratheesh, Gliddon, Singh, Vieta and Dodd2017).

Most of the precursors and vulnerability factors evident in the reviews were shared in both disorders, but some factors were either more evident in BD at-risk or BPD at-risk or they differed in phenomenological aspects. For example, BD at-risk patients had decreased need for sleep, hypersomnia, low social rhythm regularity and high energy whereas BPD patients had chronic nightmares and it was mediated by emotional and behavioural problems. BD at-risk patients showed ‘bipolar depression’ rather than unipolar depression, but in emerging BPD the course was unipolar depression. In BPD at-risk, there was risk of self-harm, but it was not stable after post-hospitalisation. As DSM criteria state that BPD traits are chronic and pervasive (American Psychiatric Association, 2013), the nature of BPD as a personality disorder thereby is doubtful. Further, previous research has observed similar frequency in self-harm in BD patients (Joyce, Light, Rowe, Cloninger, & Kennedy, Reference Joyce, Light, Rowe, Cloninger and Kennedy2010). Therefore, although self-harm is evident only in BPD studies, it does not distinguish these disorders diagnostically. Importantly, subjects at-risk for attempting suicide usually approach it through searching information and news regarding self-harm and suicidal behaviours on the Internet (Solano et al., Reference Solano, Ustulin, Pizzorno, Vichi, Pompili, Serafini and Amore2016). Better insight and understanding of suicide and suicidal risk in these at-risk populations may ultimately help clinicians to adequately detect and prevent suicidal acts.

Whilst AI is transdiagnostic (Marwaha et al., Reference Marwaha, Gordon-Smith, Broome, Briley, Perry, Forty and Jones2016), it is also regarded as a shared feature in BD and BPD diagnosis. AI was evident in BD onset coexistent with baseline MDD, and it was defined as ‘having ups and downs’ whereas in BPD onset, it was part of negative affectivity, aggression, and impulsivity. Difficulties in relationships are a core BPD feature, manifested by idealisation and devaluation as well as by rejection sensitivity (Bayes et al., Reference Bayes, McClure, Fletcher, Román Ruiz del Moral, Hadzi-Pavlovic, Stevenson and Parker2015; Gunderson, Reference Gunderson2007). Considering the findings, they are in line with the previous cross-diagnostic studies. Saunders, Goodwin, and Rogers (Reference Saunders, Goodwin and Rogers2015) reported that patients with BPD had higher negative affect, impulsivity, aggression and reduced cooperative relationships. Likewise, Henry et al. (Reference Henry, Mitropoulou, New, Koenigsberg, Silverman and Siever2001) suggested that BPD is not simply an attenuated subgroup of affective disorders and that it could be distinguished from BD on the basis of temperament and character. Additionally, the valence, frequency and nature of mood/affect regulation or mood swings is key to both BD and BPD, and likely especially as the conditions are developing (Marwaha et al., Reference Marwaha, He, Broome, Singh, Scott, Eyden and Wolke2014). It was therefore surprising that this aspect of psychopathology has not been comprehensively assessed in people with at-risk conditions. Indeed, this is one way that the conditions could be distinguished. Advancing this field will require future comparative studies of affect/mood regulation in young people with emerging BD v. BPD. The time scale of the mood fluctuations can be a useful marker in clinical practice to differentiate BD and BPD in at-risk asymptomatic periods.

Childhood adversity was evident in both disorders (e.g. Palmier-Claus, Berry, Bucci, Mansell, and Varese, Reference Palmier-Claus, Berry, Bucci, Mansell and Varese2016; Ratheesh et al., Reference Ratheesh, Davey, Hetrick, Alvarez-Jimenez, Voutier, Bechdolf and Cotton2017; Skabeikyte and Barkauskiene, Reference Skabeikyte and Barkauskiene2021; Stepp et al., Reference Stepp, Lazarus and Byrd2016; Winsper et al., Reference Winsper, Marwaha, Lereya, Thompson, Eyden and Singh2016b). However, the evidence was much scarcer and sparser in BD studies. For example, apart from childhood sexual, physical, and verbal abuse and neglect, peer victimisation and abuse in romantic relationships were also evident in BPD. These findings are consistent with the previous literature stating that there is a higher likelihood of experiencing childhood adversity in BPD patients compared to BD (Afifi et al., Reference Afifi, Mather, Boman, Fleisher, Enns, MacMillan and Sareen2011; Cotter, Kaess, & Yung, Reference Cotter, Kaess and Yung2014). These adverse experiences might be the reason why BPD patients tend show higher aggressiveness and anger in mood shifts compared to BD patients.

In line with the recent research on the circadian rest-activity patterns in BD and BPD patients (McGowan, Goodwin, Bilderbeck, & Saunders, Reference McGowan, Goodwin, Bilderbeck and Saunders2019), sleep disturbances and difficulty falling asleep were common to both disorders. In BPD studies, chronic nightmares were significantly predictive of the onset whereas in BD, participants had decreased need of sleep and it was part of hypomanic symptoms. Vöhringer et al. (Reference Vöhringer, Barroilhet, Alvear, Medina, Espinosa, Alexandrovich and Ghaemi2016) too indicated that decreased need for sleep was part of manic symptoms and were specific to BD and not to BPD patients.

Comorbid SUD, anxiety disorders, psychotic symptoms, ADHD were common to both disorder onsets. In BD studies comorbid generalised anxiety disorder and social phobia and ADHD with and without baseline comorbid conduct disorder predicted BD onset whereas in BPD, ADHD and OCD predicted BPD. Further, psychotic symptoms in BD at-risk studies were most often linked to affective states which is in line with previous research (Bassett, Reference Bassett2012). However, the nature of the psychotic symptoms in BPD at-risk studies were not evident. Temperamental dimensions were also evident in both. Higher levels of activity and poor psychosocial functioning were common to both, but in BD onset daydreaming, cyclothymia, and temperamental instability during MDD episodes were predictors of transition. Additionally, sensitivity, hyper alertness, excessive talk or talking too loudly, somatic complaints, and impaired role in school predicted conversion to BD. In BPD studies, on the contrary, higher levels of emotionality, low levels of sociability and shyness predicted BPD symptoms. These traits again might be attributable to the fact that BPD patients having more conflictive interpersonal relationships (MacKinnon & Pies, Reference MacKinnon and Pies2006).

Depression was also predictive of BD onset and later BPD symptoms although most of the studies pertained to BD at-risk reviews. In BD at-risk studies, depressive episodes or MDD, chronicity, severity, age at onset, psychomotor retardation and frequency of depression predicted BD transition. In emerging BPD studies, only early onset of depression was related to later BPD symptoms. In BD at-risk studies, there is also coexistence of atypical depressive symptoms that are considered to be ‘bipolar depression’ and distinct in phenomenology from unipolar depression such as pathological guilt, cyclothymia, mood lability, psychotic symptoms, and subthreshold hypomania (Berk et al., Reference Berk, Conus, Lucas, Hallam, Malhi, Dodd, Yatham and McGorry2007, Reference Berk, Hallam, Malhi, Henry, Hasty, Macneil and McGorry2010). Detailed studies about depressive states in emerging BPD populations are urgently needed to be able to understand whether they can be distinguished based on the depressive symptomatology.

Family history of BD, although the results were inconsistent, was the prominent predictor of BD conversion compared to family history of depression or any affective disorders. For BPD, apart from maternal BPD symptoms (Winsper et al., Reference Winsper, Lereya, Marwaha, Thompson, Eyden and Singh2016a), paternal SUD, family history of psychiatric hospitalisation, and maternal psychopathology were significantly associated with BPD. Consistent with the previous research the family history of BD might be a prominent distinguishing feature when comparing BD and BPD (Galione & Zimmerman, Reference Galione and Zimmerman2010; Mitchell, Goodwin, Johnson, & Hirschfeld, Reference Mitchell, Goodwin, Johnson and Hirschfeld2008). However, there is paucity of research pertaining to BPD. Further, despite existing family history of BD data might support the conclusion that BD is highly heritable and unrelated to BPD, no studies have examined the familial relationship of BPD traits and conversion to BD or vice versa.

Our review has strengths. We utilised systematic search procedures to reduce risk of bias and ensure comprehensive coverage of the current literature. Inter-rater reliability was consistent with no requirement for arbitration regarding inclusion of SRs and MAs. However, several limitations of the current findings here should be considered when interpreting the results. First, we only included relevant prospective studies from the SRs and excluded primary studies with any other designs. This inhibited synthesising the articles as a whole and reporting the pooled results from eligible MA's. Second, the evidence was limited by the data that included SRs and MAs provided and some relevant prospective studies were inevitably missed. Third, there were many non-systematic literature reviews including prospective studies that the reviews we included missed out. For example, Hartmann, Nelson, Ratheesh, Treen, & McGorry's (Reference Hartmann, Nelson, Ratheesh, Treen and McGorry2018) scoping review, they provided additional evidence for family history of BD, subthreshold depression and hypomania, sleep disturbance, mood lability and later BD conversion. They also provided data for impulsivity and fun-seeking which was not investigated by the reviews included here. Fourth, we were not able to conduct an MA due to the substantial methodological and clinical heterogeneity with respect to cohort characteristics such as study design and sample size among primary studies included in the reviews. Further, not being able to pool the data precluded definitively clarifying the timing and duration of the precursors, notwithstanding the heterogeneity in at-risk populations (Radua et al., Reference Radua, Ramella-Cravaro, Ioannidis, Reichenberg, Phiphopthatsanee, Amir and Fusar-Poli2018). Fifth, because of the short follow-up periods and small number of follow-up assessments in most of the included prospective studies, the validity and utility of these factors for predicting an early prodrome of BD and BPD remains unknown. If we do not know the actual starting point of the onset of the disorder, these and any other identified factors may indicate relapse or maintenance of the disorder (Stepp & Lazarus, Reference Stepp and Lazarus2017). Sixth, none of the reviews discussed the sensitivity, specificity, and predictive value of reported precursors, important aspects that may enable better assessment of clinical utility. Therefore, again a cautious interpretation of the findings as to their generalisability is necessary. Seventh, none of the studies mentioned sub-score analyses for the examined antecedents making it challenging to compare how the two disorders presented different patterns of the shared features. Eight, the majority of the BD at-risk reviews examined youth at genetic high risk. However, most genetically high-risk individuals do not develop BD. A combination of genetic and clinical risk factors is required to optimally predict conversion to BD (Keramatian et al., Reference Keramatian, Chakrabarty, Saraf and Yatham2021). Nineth, despite our comprehensive search, we identified relatively very few studies pertaining to the BPD onset.

In conclusion, although the findings of this review may lead to support the view of BD and BPD as two distinct disorders, there is scant evidence from existing studies to either indicate that BD and BPD are separate nosological entities or that BPD should be considered as an extension of BD disorders. In clinical practice, these differences can be subtle, especially between BPD and BD-II (Massó Rodriguez et al., Reference Massó Rodriguez, Hogg, Gardoki-Souto, Valiente-Gómez, Trabsa, Mosquera and Amann2021).

Whilst the comparative literature is in its infancy there are several implications from this meta-review. On an etio-pathological level, our findings corroborate the notion that there is a prodromal stage in BD and BPD. There are overlapping risk factors in the young people with at-risk BD and BPD, these being family history of psychopathology, AI, ADHD, anxiety disorders, depression, sleep disturbances, substance abuse, psychotic symptoms, suicidality, childhood adversity and temperament. However, there are risk factors specific to the at-risk BD and BPD states. Gender, differences in the white matter, changes in the amygdala, neural reward circuit dysfunctions, DBD, subsyndromal hypomania, cyclothymia or bipolar NOS, frequency and loading of affective symptoms, and antidepressant use were evident only in people with at-risk BD. Parenting behaviour/style, parent-child relationship quality, maternal characteristics, attachment, impulsivity, experiential avoidance, disturbances in self representation, dissociation, comorbid oppositional defiant disorder, comorbid conduct disorder, somatisation, general psychosocial functioning, and social and physical aggression in childhood were evident only in at-risk BPD. These factors could form the basis of initial prediction modelling approaches which could improve clinical staging and clinical interventions. Clinicians should be aware of the high degree of comorbid psychopathology in young people developing BD and BPD and should consider both conditions in young people presenting with one. From a transdiagnostic perspective, the current review may provide a benchmark for comparing the magnitude of association of these factors with other mental health disorders. The results can also substantially advance our ability to prognosticate the onset of BD and BPD in populations at-risk, who ultimately may benefit from preventative interventions.

To be able to reliably identify target populations with greater specificity, future research is required to increase our understanding of the development of BD and BPD onset and their complex interplay by conducting prospective studies which concurrently examine multiple measures including biological, environmental, psychosocial, and clinical factors in BD and BPD at-risk populations. Systematic longitudinal studies investigating genetically and clinically high-risk youths in a structured multifactorial approach can help us understand whether both these disorders belong to the affective spectrum or not as well as their development over time (e.g. Brietzke et al., Reference Brietzke, Mansur, Soczynska, Kapczinski, Bressan and McIntyre2012). Greater predictive validity could be provided by future research identifying potential BD and BPD biomarkers whilst charting these along the illness trajectory. It is also important that future research studies use consistent recruitment criteria to ensure that findings are comparable and generalisable to other studies as far as is practicable (Malhi et al., Reference Malhi, Morris, Hamilton, Outhred and Mannie2017).

Large, multilevel data sets will enable deep phenotyping and distinguish pathophysiological pathways (Phillips & Kendler, Reference Phillips and Kendler2021). For example, remote monitoring can complement symptom monitoring and capture signals more representative of the underlying pathophysiology of BD and BPD (Gillett et al., Reference Gillett, McGowan, Palmius, Bilderbeck, Goodwin and Saunders2021; Gillett & Saunders, Reference Gillett and Saunders2019). One of the ways to conduct remote monitoring is Experience Sampling Methodology (ESM). The temporal pattern in mood may be captured by ESM (Larson & Csikszentmihalyi, Reference Larson and Csikszentmihalyi2014). Researchers have widely used ESM to assess the temporal patterns of regulations of mood/affect in individuals with mood disorders as the method is more suited to capturing momentary temporal fluctuations in affect (e.g. Dubad, Elahi, and Marwaha, Reference Dubad, Elahi and Marwaha2021; Merikangas et al., Reference Merikangas, Swendsen, Hickie, Cui, Shou, Merikangas and Zipunnikov2019; Schwartz, Schultz, Reider, and Saunders, Reference Schwartz, Schultz, Reider and Saunders2016; Tsanas et al., Reference Tsanas, Saunders, Bilderbeck, Palmius, Osipov, Clifford and De Vos2016). Further, some types of ESM are not subject to recall biases as the studies do not rely on retrospective memory recall of events between assessments (Myin-Germeys et al., Reference Myin-Germeys, Oorschot, Collip, Lataster, Delespaul and Van Os2009). The data gained through could also identify new behavioural biomarkers which may lead to the identification of novel phenotypes in these disorders (Gillett & Saunders, Reference Gillett and Saunders2019). Additionally, identifying common criteria such as AI is easy but focusing on differential symptoms is a complex task (Massó Rodriguez et al., Reference Massó Rodriguez, Hogg, Gardoki-Souto, Valiente-Gómez, Trabsa, Mosquera and Amann2021). ESM could also be useful to be able to achieve this. These prospective studies may also help identifying a validated BPD prodrome criteria, despite previous resistance to the diagnosis of BPD in adolescents due to the fears of stigmatisation (Chanen, Reference Chanen2015; Laurenssen, Hutsebaut, Feenstra, Van Busschbach, & Luyten, Reference Laurenssen, Hutsebaut, Feenstra, Van Busschbach and Luyten2013; Stepp & Lazarus, Reference Stepp and Lazarus2017).

Making an accurate diagnosis of BD and BPD is further complicated by comorbidity with various other conditions such as ADHD and unipolar depression (Asherson et al., Reference Asherson, Young, Eich-Höchli, Moran, Porsdal and Deberdt2014; Mneimne, Fleeson, Arnold, & Furr, Reference Mneimne, Fleeson, Arnold and Furr2018). ADHD has been reported to coexist in around 20% of adult patients with BPD or BD (Asherson et al., Reference Asherson, Young, Eich-Höchli, Moran, Porsdal and Deberdt2014; Philipsen et al., Reference Philipsen, Feige, Hesslinger, Scheel, Ebert, Matthies and Lieb2009; Skirrow, Hosang, Farmer, & Asherson, Reference Skirrow, Hosang, Farmer and Asherson2012), while rates of co-occurrence between BPD and current major depressive disorder MDD or BD range from as low as 4% to as high as 48% (Mneimne et al., Reference Mneimne, Fleeson, Arnold and Furr2018). Since deficits in affect regulation such as AI are also strongly linked to the hyperactive/impulsivity symptoms of ADHD (Skirrow & Asherson, 2013) and unipolar depression (Balbuena, Bowen, Baetz, & Marwaha, Reference Balbuena, Bowen, Baetz and Marwaha2016), it is imperative to include these groups too for comparison to better understand the symptom profiles between at-risk BD and BPD.