Prevalence of BPD in population samples

The DSM-IV-TR and previous versions of this manual stated that there is a 3:1 female to male sex ratio for BPD, perpetuating the belief that this disorder is much more prevalent among females (Sansone & Sansone, 2011). Numerous high-quality epidemiological studies of the population prevalence of DSM-IV PDs have now been published. Among these, the National Comorbidity Survey Replication (Lenzenweger et al., 2007), based upon a representative sample of 9,282 adults from the United States, reported the prevalence of BPD to be 1.4%, with no sex difference. Torgensen and colleagues’ (Torgersen, Kringlen, Cramer, 2001) study of a representative community sample of 2,053 adult residents of Oslo, Norway and found the prevalence of BPD to be 0.7%, with no sex differences. The British National Survey of Psychiatric Morbidity (Coid et al., 2006) studied a representative sample of 8,886 16- to 74-year-olds reported the prevalence of BPD to be 0.7%, with no sex differences. Finally, Wave 2 of the National Epidemiological Survey on Alcohol and Related Conditions (NESARC; Grant et al., 2008) studied a representative sample of 34,653 American adults. The lifetime prevalence of BPD was 5.9%, with no sex differences. The NESARC has been criticized for its measurement of PD and a conservative re-analysis of the data by Trull and colleagues (2010) reported the overall prevalence of BPD to be 2.7%, with a statistically significant but slight preponderance of BPD among females (3.02% vs. 2.44% among males).

The only study to report reliably on sex differences among young people with BPD (Zanarini et al., 2011), assessed a community sample of 6,330 11-year-olds from Bristol, England and found the prevalence of BPD to be 3.2%, with no statistically significant sex difference.

Taken together, the results from these population studies do not support the assumption that BPD is much more prevalent among females. In fact, the rate of BPD is either equal or only slightly higher in females.

Biased sampling in research studies

Contrary to the aforementioned studies, clinical epidemiological studies do reveal sex differences, with a preponderance of females (around 75–80%) in clinical services. It has been suggested that this might be the reason why clinicians perceive more females have BPD (Skodol, 2003). This might be due to sampling biases associated with the clinical presentation of BPD. For example, it has been argued that males with BPD are more likely to present to drug and alcohol treatment services and are less likely to utilize pharmacotherapy and psychotherapy (Sansone & Sansone, 2011), thereby accounting for the greater proportion of women with BPD in clinical services. This is supported by data from the Collaborative Longitudinal Personality Disorders Study (CLPS), which investigated differences among their sample of 175 females and 65 males with BPD (Johnson et al., 2003). Males with BPD were more likely to present with substance use disorders, and with schizotypal, narcissistic and antisocial PDs, while females with BPD were more likely to present with post traumatic stress disorder (PTSD), eating disorders and the BPD criterion of identity disturbance.

Biased application of diagnostic criteria

Another reason for a possible disparity between the actual prevalence of BPD among females and the assumed prevalence might be biased (unequal) application of diagnostic criteria (Widiger, 1998). Anderson, Sankis, and Widiger (2001) asked 720 members of the American Psychological Association to judge the frequency of DSM-IV personality pathology in female and male patients. They cross-validated these results in a second sample of 900 additional members of the American Psychological Association. The findings of this study indicated that when clinicians applied the DSM-IV PD criteria, they rated BPD symptoms to be as infrequent in males, and as pathological in males, as they were in females. There was no significant statistical difference between sexes. This suggests that the application of diagnostic criteria might be neutral.

Sex bias in the DSM criteria

Few studies have investigated the DSM-IV criteria for sex bias. The most robust and largest study conducted used factor analytic strategies to investigate the measurement invariance for these criteria using the Norwegian Twin database (Aggen, Neale, Roysamb, Reichborn-Kjennerud, & Kendler, 2009). A total of 2,794 twins were interviewed, 1772 females and 1022 males, and the SCID-II interview was administered for DSM-IV PDs. The results of the statistical analyses confirmed the coherence of the nine criteria in a general population sample. This BPD factor differed according to sex and age in relation to affective instability and impulsivity. Females reported more affective instability and less impulsivity than males. In younger females, impulsivity was more prominent and this changed over time.

Social constructivism

The tendency to diagnose BPD more often in women than men might be partly due to sociocultural factors and the cultural history of BPD (Bjorklund, 2006). The basis of this hypothesis has been informed by cross-cultural studies where, for example, in Ethiopia and India only 1–3% of psychiatric outpatients were diagnosed with PDs compared with a rate as high as 32% in the United Kingdom. This variation might be due to different diagnostic practices and the way people seek help in these societies. Other factors might include economic status, optimism, religion, psychological awareness and medical orientation of the society. Cultural factors also influence the way mental illness manifests in relation to sex. Expectations exist concerning women’s tendency to be more emotional and relational and relationship-dependent than men (see Chapter 1). Therefore, it is possible for diagnostic constructs to be gender-biased. Sexist characterizations of women’s health and illness are affected by political, economic and cultural factors along with biomedical ones. Sex differences exist in normal behavior (Simmons, 1992). For example, excessive anger, argumentativeness and sexual promiscuity are described as pathological behaviors that might lead to BPD diagnosis. However, if displayed by men, these behaviors might be seen as acceptable, expected or even admirable. It might be rare for a man to be diagnosed as having BPD based upon these behaviors. He would be more likely to be diagnosed as having ASPD, if at all. More research is clearly needed in this area.

Sex differences in BPD symptom criteria

Studies have investigated the prevalence of individual diagnostic criteria for BPD in females and males and found a sex difference in the type of psychopathology associated with this diagnostic label. A study of 11-year-old children with BPD found that girls were significantly more likely to fulfill DSM-IV criteria for mood reactivity, frantic efforts to avoid abandonment and unstable relationships, compared with boys who were more likely to evidence impulsivity and physically self-destructive acts (Zanarini et al., 2011). The Collaborative Longitudinal Personality Disorders Study (CLPS) found females were significantly more likely to rate higher than males for affective instability and identity disturbance (Johnson et al., 2003). Whereas other studies have reported that males more commonly endorsed the item “intensive anger” (and to some degree “impulsivity”), and that females more commonly endorsed “affective instability” (Tadic et al., 2009). By contrast, some researchers have reported no sex differences in BPD criteria, except for dissociation, which is more prevalent among females (McCormick et al., 2007). Therefore, while there were few differences among the majority of BPD criteria, there was evidence of some difference in terms of affective instability, anger and impulsivity.

Sex differences in the traits and behaviors underlying BPD

Widiger (1998) suggests that trait-based models of personality disorder, in particular the five-factor model (Wiggins, 1996), might provide some assistance in resolving the sex bias controversy by providing a theoretical basis for sex prevalence rates. He hypothesizes that, if personality disorders are extreme maladaptive variants of normal personality traits, the prevalence of personality disorders would be expected to mirror any differential sex prevalence of the factors and facets of the five-factor model. In fact, in the literature on normal personality, consistent sex differences have been obtained for most of the domains and facets of the five-factor model (Widiger, 1998) and for some other trait measures (Jang, McCrae, Angleitner, Riemann, & Livesley, 1998). Furthermore, there are also significant sex differences in the magnitude of heritability of those traits and some evidence of sex-specific genetic effects (Jang et al., 1998).

A study of maladaptive personality among 681 volunteer twin pairs from the general population showed that most maladaptive traits are heritable in both sexes (Jang et al., 1998). Cognitive dysregulation, suspiciousness and self-harm were the only traits that did not have a significant heritable component in females and likewise for submissiveness in males. Sex-specific genetic effects were detected for all traits, except insecure attachment. There was no evidence to support sex-specific, non-shared environmental influences for any traits underlying BPD. It is also noteworthy that callousness and conduct problems, along with their higher order dissocial factor were not significantly heritable for females.

The DSM-5 BPD diagnostic criteria comprise a mixture of more stable traits (e.g., chronic anger, intolerance of aloneness) and unstable symptomatic behaviors (e.g., self-mutilation) (Gunderson et al., 2011; Zanarini et al., 2007). Several studies have examined sex differences among the traits that underlie BPD, using measures of normal and pathological personality. In a community sample of 263 children aged between 9 and 13 years (Gratz et al., 2009), affective dysfunction and disinhibition significantly predicted borderline personality symptoms in both girls and boys, and that there were no sex differences in the level of BPD symptoms. Among girls, affective dysfunction and sensation seeking accounted for a significant amount of additional variance in BPD symptoms, whereas among boys, only affective dysfunction accounted for a significant amount of unique variance. Likewise, a study of undergraduate university students found that women rated significantly higher on affective instability, abandonment and relationships; and men on impulsiveness (Fonseca-Pedrero et al., 2011). However, these findings are limited by their reliance on a self-report borderline personality disorder scale, not a formal DSM-IV diagnostic interview.

In a study comparing male and female adolescents with BPD (measured by the SWAP-200, which uses a Q-sort methodology), girls showed more internalizing psychopathology and were more emotionally dramatic, whereas boys were more behaviorally disinhibited, externalizing and angry (Bradley, Zittel Conklin, & Westen, 2005). The young women in this study were described as more likely to be drawn into relationships where they could be emotionally or physically abused, choose inappropriate romantic partners, be overly sexually seductive or provocative, make suicidal threats or gestures, have uncontrolled eating binges, and be preoccupied by food-related issues. In contrast, the young men were described as more likely to bully others, feel self-important, show reckless disregard for the rights of others, prefer to operate as if emotions were irrelevant, take advantage of others, have trouble making decisions, seek to dominate others, show no remorse, not need contact with other people, and promise to change but then revert to their previous maladaptive behavior.

There is also some evidence to suggest there are sex differences in behavioral symptoms in BPD. In a sample of internal medicine patients screened (but not formally assessed) for BPD, women were more likely to be engaged in a sexually abusive relationship, whereas men were significantly more likely to bang their head or to lose a job on purpose (Sansone, Lam, & Wiederman, 2010). In a sample of low socioeconomic status middle- and high-school students, males were significantly more likely to burn or punch themselves (Gratz et al., 2012). Another study compared female and male offenders and found that, in females, the combination of interpersonal-affective traits and impulsive-antisocial traits combined to increase the risk of self-directed violence, whereas in males, interpersonal-affective traits were associated with reduced symptoms of BPD, which in turn reduced the risk of self-directed violence (Verona, Sprague, & Javdani, 2012).

Sex differences in co-occurring disorders

Aside from differences in how the BPD criteria are manifest, it is possible that sex differences occur in regard to co-occurring mental state (aka axis I) and trait (aka axis II) pathology.

The Methods to Improve Diagnostic Assessment and Services (MIDAS) study used semi-structured DSM-IV diagnostic interviews to investigate the pattern of comorbidity among 130 outpatients (71 women, 30 men) with BPD (Zlotnick, Rothschild, & Zimmerman, 2002). They found that there were sex differences in the pattern of lifetime impulse-related disorders. Women were more likely to have a history of eating disorder, but in contrast it was more common to find substance abuse disorders, intermittent explosive disorder and ASPD histories among males. There were no differences in rates of comorbid PTSD.

A study of comorbidity among 484 consecutively admitted outpatients with a diagnosis of BPD, where most participants were female (83%), found 74% met criteria for at least one co-occurring DSM-IV PD, 33.5% had two or three PDs, and 7.6% had between four and eight PDs (Barrachina et al., 2011). Women tended to have more disorders than men. In women, the most frequent co-occurring disorders (from highest to lowest) were depressive, paranoid, passive-aggressive, dependent, avoidant and obsessive-compulsive PDs. In contrast, among men the most frequent (from highest to lowest) were paranoid, antisocial, passive-aggressive, depressive and avoidant PDs. Of particular relevance were the significantly higher rates of dependent PD in women, supporting their tendency to internalize symptoms and among males with BPD the significantly higher prevalence of antisocial PD, which gives further support to the suggestion that males manifest impulsivity through externalising behaviors.

Another study investigated DSM-IV axis I and axis II comorbidity among patients with BPD (Tadic et al., 2009). Of the 110 women and 49 men studied using a standardized diagnostic interview, they found women were significantly more likely to have a co-occurring eating disorder and affective instability. Males were significantly more likely to have a co-occurring substance use disorder, disorder of social behavior, antisocial PD and intensive anger. Their findings further support the hypothesis that women and men have more similarities than differences in the number of comorbid axis I disorders, axis II disorders and BPD criteria. However, women tend to have more internalizing psychopathology and men more externalizing psychopathology associated with BPD. With regard to mental state disorders, females have higher rates of anxiety, affective and eating disorders, whereas males have higher rates of substance use disorders. With regard to personality disorder, compared to males, females show significantly lower rates of behavioral problems in childhood and adolescence and ASPD.

When standard personality tests and structured clinical interviews were used to investigate sex differences among 114 female and 57 male patients with BPD (Banzhaf et al., 2012), they found that women with BPD were significantly more likely to meet criteria for PTSD, panic disorder with agoraphobia and bulimia nervosa. Men with BPD were more likely to meet criteria for binge eating disorder, a disorder of social behavior in childhood, and antisocial and narcissistic PDs. These findings are comparable with those reported by other studies, except for the absence of substance use disorders among males with BPD.

Another study investigated comorbidity in BPD and found no differences in the frequency of axis I disorders, except that women were significantly less likely to have alcohol use disorders (Barnow et al., 2007). Patients with BPD had, on average, three additional PDs. Women were more likely to have an avoidant PD, whereas men had significantly higher rates of antisocial personality disorder.

These studies suggest that there are sex differences in clinical presentation of patients with BPD in terms of criteria, patterns of co-occurring PD and mental state pathology. In all these aspects of presentation, females tend to have more internalizing symptoms and males tend to have more externalizing symptoms (Eaton et al., 2011). Most striking, however, is that females and males with BPD generally display more similarities than differences, a fact that often seems to be overlooked in clinical practice and which might contribute to the misdiagnosis of BPD among males (Johnson et al., 2003). Where differences occur, the studies above indicate that females with BPD tend to have greater affective instability and to engage in a variety of methods of self-harm and may place themselves in situations where they are vulnerable to being mistreated by others. In contrast, males with BPD pathology tend to be more impulsive and to physically harm themselves using greater self-directed violence.

Clinical studies indicate there are sex differences in the way BPD presents, with females having more affective instability and males being more impulsive. This in turn is reflected in sex differences in co-occurring disorders, functioning and traits. Women with BPD tend to have greater rates of internalizing psychopathology, whereas males have a greater number of externalizing behaviors.