Abstract

Objective: To examine the presence of borderline personality disorder (BPD) traits in Gulf War veterans, and to assess psychiatric comorbidity, health status, healthcare utilization, and quality of life (QOL) along a continuum of BPD trait severity.

Method: BPD and traits were evaluated using the Schedule for Non-Adaptive and Adaptive Personality in 576 veterans who were either deployed to the Persian Gulf (1990–1991) or were on active duty though not deployed to the Gulf. Demographic and military characteristics, personal and family history, psychiatric comorbidity, and QOL were also assessed.

Results: One or more BPD traits were present in 247 subjects (43%), and BPD (>5 traits) was identified in 15 subjects (3%). The number of traits was significantly associated with age and level of education. Lifetime psychiatric comorbidity was significantly associated with the number of BPD traits present, and level of functioning, health status, healthcare utilization, social functioning, self-injurious tendencies, and military/behavioral problems.

Conclusion: BPD and traits identified in Gulf War veterans were associated with significant psychiatric morbidity, poorer QOL, and increased utilization of healthcare resources. Early recognition and treatment of veterans with BPD symptoms may be warranted to minimize the burden on the healthcare system.

Introduction

Borderline personality disorder (BPD) has emerged as a major health problem, having a community prevalence of ~1%, though the disorder is more frequent in selected populations, including psychiatric inpatients and outpatients, individuals with substance misuse, and incarcerated individuals.^1-3 The disorder is associated with substantial medical morbidity, functional impairment, impaired quality of life (QOL), and increased healthcare utilization.^4,5

BPD is also associated with significant psychiatric comorbidity, particularly mood, anxiety, and substance use disorders.^6,7 The symptom of BPD that probably makes the greatest demand on mental health resources is recurrent suicidal threats and gestures, which may be a defining characteristic of the disorder.⁸ Suicidal behavior affects up to 84% of patients with BPD⁹; Soloff and colleagues¹⁰ calculated a mean of 3.4 lifetime attempts per individual. Deliberate self-harm, typically involving such behaviors as cutting, burning, or head banging, is also frequent and, while potentially life-threatening, is not generally motivated by a desire for death.

We recently had an opportunity to examine patterns of psychiatric comorbidity, healthcare utilization, and QOL variables in a well-characterized sample of Gulf War veterans. In this study, we focused on BPD and BPD traits assessed using the Schedule for Non-Adaptive an Adaptive Personality (SNAP),¹¹ a self-report instrument of known reliability. To our knowledge, this is one of the first explorations of BPD and its traits in a military population. Our purpose was to examine patterns of psychiatric comorbidity and health-related QOL and their relationships with BPD and BPD traits. We hypothesized that individual BPD symptoms would be frequent, but that BPD itself would be relatively uncommon; applicants to the military are screened to exclude those with evidence of past or present emotional instability, which could compromise one’s ability to succeed in the military. We hypothesized that BPD traits would be present along a continuum, and that there would be a positive association between the number of traits present and the frequency of psychiatric comorbidity and healthcare utilization, and a negative association between number of traits present and QOL variables. The results are reported herein.

Methods

Subjects

Subjects were participants in the second of a two-stage investigation. The first wave involved a telephone survey of military personnel on active duty during the first Gulf War, some deployed to the Gulf and others who were not, who had participated in a structured, computer-assisted telephone interview.¹² The interview was administered to 3,695 individuals, or 76% of those eligible to participate. The second wave involved extensive in-person interviews and examinations of a stratified random sample of the telephone survey participants who were living in Iowa or one of the surrounding states.¹³ The selection was made on the basis of survey status (case versus control) and deployment status (deployed versus non-deployed) in order to obtain a representative sample of the outcomes of interest. Subjects were recruited for the second wave if they had one or more of the most prevalent a priori-defined conditions of interest in the first wave (depression, chronic widespread pain, cognitive dysfunction). “Depression” included the categories of major depression, minor depression, and dysthymia; the diagnosis “chronic widespread pain” was based either on a self-report of having had fibrositis or fibromyalgia in the prior year, or a self-report of overall body pain of almost daily occurrence with a minimum duration of 3 months in the previous year. “Cognitive dysfunction” was diagnosed when a subject reported at least one of the following: severe memory impairment, moderate to significant difficulties with confusion/disorientation, or moderate to significant problems caused by at least two of the following: clear thinking, concentration, comprehension, reading comprehension, slips of the tongue, forgetfulness or memory problems. A control group that did not meet criteria for any of these conditions was also recruited. Approximately two cases were drawn for each control subject. An adaptive randomization was used to ensure comparability across cases and controls. The study was approved by the University of Iowa Institutional Review Board and written, informed consent was obtained from all subjects who volunteered to participate. Additional details about this study are reported by Ang and colleagues¹⁴ and Forman-Hoffman and colleagues.¹⁵

Assessments

Demographic data, including age, sex, race, education, income, and marital status, were obtained along with military variables of interest such as military status, branch of service, rank, deployment to the Gulf, and combat during the first Gulf War.

Personality was assessed using the SNAP,¹¹ a factor-analytically derived, self-report instrument that assesses three major trait dimensions important in personality disorders, including negative temperament, positive temperament, and disinhibition. The scales assessing these dimensions are internally consistent and have acceptable test/retest reliability. The SNAP also assesses the presence of the Diagnostic and Statistical Manual of Mental Disorders, Third Edition-Revised (DSM-III-R)¹⁶ personality disorders and their individual criteria.

Through structured questions, personal, medical, and psychiatric history, items were obtained, including information regarding depression, anxiety, drug abuse, use of psychiatric medication, psychiatric hospitalization, attempted suicide, and incarceration. Family history items assessed included anxiety, alcoholism, drug addiction, psychiatric medications, psychiatric hospitalization, psychotherapy or counseling, suicide attempt, or completed suicide.

The Structured Clinical Interview for the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (SCID)¹⁷ was administered to all study subjects by a trained rater in order to collect psychiatric diagnostic information. This widely used instrument has shown reliability compared with that of other major diagnostic instruments used to assess Axis I disorders employing the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) criteria. Interviewers were trained to administer the SCID by one of the investigators. The inter-rater reliability of SCID results was compared on three separate occasions using audio tapes from randomly selected cases. On these occasions, the kappa coefficient was consistently >0.8. Assessments based on the trained interviewer’s use of the SCID and blinded assessments by the psychiatrist based on all available data were nearly always identical. Raw data included self-report questionnaires and the SCID interview and were reviewed by one of the psychiatrist investigators who then assigned a “best estimate diagnosis” for each subject, taking all data into account.¹⁸ The study psychiatrist was blind to case versus control status and deployment status.

Subjects were regarded as having been involved in combat if they were in the Gulf during the war and reported at least three of nine combat-related exposures: Scud missile explosion; artillery, rocket, mortar explosion; small arms fire; combat-related injury; saw dead body; witnessed dying; exposed to nerve gas; exposed to mustard or other gas; caused the death of another.¹⁹

Level of exposure to combat was assessed with the Expanded Combat Exposure Scale (CES).²⁰ This is a 34-item scale designed to assess the frequency of major combat-related stressors. It is a modified version of the Laufer Combat Scale, originally developed to assess traditional combat experiences, such as whether an individual has received friendly fire or incoming fire, whether his/her unit had been ambushed, attacked, or received sniper fire, or whether he/she had seen Americans or other troops killed or wounded. The psychometric properties of the original Laufer Combat Scale were satisfactory.²¹ The modifications include assessing exposures to events encountered during deployment to the Gulf, such as being placed on alert for chemical or biological attacks, exposure to poison gas or germ warfare, or loss of communication among units. The Social Provisions Scale was used to assess social support.²² This scale contains 24 statements concerning current relationships that call for responses on a four-point scale (ranging from strongly agree to strongly disagree). The scale provides a reliable measure of general support. It has been shown to be related to stress, depression, and symptoms of illness, and has been validated in a wide range of settings.²³

Level of functioning was determined using the Medical Outcome Survey-Short Form 36 Health Survey.²⁴ This is a measure of perceived health status over the previous 4 weeks, which contains 36 items that assess dimensions of both physical and mental health, including pain. The instrument is widely used as a general health status measure of functional capacity and well-being and has excellent reliability and validity.²⁵

Military preparedness was assessed by means of six items that inquired into how adequately subjects had been prepared, how they had met fitness and height/weight requirements, and how they had been trained for their assigned tasks.²⁶ These items called for yes or no responses and, because the distribution was highly skewed, a sum was dichotomized; a score of 6 determined that subjects were fully prepared, versus a score <5, which determined that subjects were not fully prepared.
 

Statistical Analysis

Analyses were performed using Statistical Analysis System version 8.²⁷ Descriptive statistics including percentages for categorical data, means, and standard errors for continuous data were generated. Subject groups were created based on the number of BPD criteria met. The variable that defines these groups represents an ordinal, equally spaced scale, so tests of association between this variable and categorical and continuous variables of interest were performed together with trend analysis. Because of the small number of participants in the group with BPD (n=15) and low frequencies for some demographic variables and SCID mental disorders overall or in certain categories, exact non-parametric tests were used in the analysis. Fisher’s Exact Test was used to test associations between the BPD criteria group and categorical measures. Trend analysis for categorical variables was performed using Cochran Armitage trend test²⁸ for dichotomous variables or Jonkheere-Terpstra test²⁹ for multilevel variables. Associations between the BPD criteria group and continuous variables were tested using the non-parametric Kruskal-Wallis test. Presence of linear trend was assessed using regression models with number of BPD traits as a regressor variable.

Results

A total of 602 veterans were interviewed, and 576 (95.7%) completed the SNAP. As shown in Table 1, 247 subjects (42.8%) satisfied one or more criteria for BPD. For ease of analysis, we created the following subject groups: no criteria met (n=329), 1–2 criteria met (n=182), 3–4 criteria met (n=50), and >5 criteria met (n=15). Individuals in the last category met full syndromal criteria for the diagnosis of BPD.

 

As can be seen in Table 2, there was an association between number of BPD criteria met and age, level of education, and case versus control status. These results indicated that the greater the number of BPD criteria present, the younger the subjects’ age, the lower the level of education, and the greater the likelihood of fitting the case definitions of one of the three groups selected for study (depression, fibromyalgia, cognitive dysfunction). There was no association between military variables and the number of BPD criteria met. Interestingly, the number of BPD traits met was significantly associated with the CES, though there was no association between BPD traits and participation in combat itself.



Next, the association of lifetime psychiatric comorbidity and number of BPD traits was tested (Table 3). For nearly all categories of psychiatric disorders, the association was significant and showed that as the number of criteria met increased, the greater the percentage of subjects having each psychiatric disorder of interest. In fact, 80% of subjects meeting criteria for BPD (>5 traits) had a lifetime SCID disorder compared with 56.2% of those meeting no criteria.

 

Trends for health status, level of functioning, health service utilization, social functioning, self-injurious tendencies, and military/social were examined next (Tables 4 and 5). For almost all variables the association between number of BPD criteria satisfied and the variable of interest was statistically significant. Those having more BPD traits had worse health status, poorer physical functioning and increased levels of pain, more outpatient visits and inpatient stays, worse social functioning, less employment, more past suicide attempts, higher potential for self-injury, and a greater likelihood of having a family history of suicide or suicide attempts. A greater number of BPD criteria was also associated with a greater likelihood of having been incarcerated, having been court martialed, receiving some other military discipline, having used illegal drugs. There was also a negative association between number of BPD criteria met and the perception of preparedness for military service.

 

Discussion

Nearly 43% of Gulf War veterans assessed with the SNAP endorsed one or more BPD symptom, the most frequent being inappropriate displays of anger (18.4%), followed by impulsivity (17.5%), chronic feelings of emptiness (16.0%), and mood instability (13.2%). Fifteen individuals (3%) met the diagnostic threshold for BPD. This rate should not be interpreted as a prevalence estimate among all veterans because our study deliberately oversampled veterans with symptoms of depression, chronic widespread pain, and cognitive dysfunction. The study demonstrates that persons meeting criteria for BPD, or having one of its many traits, are found in military populations even though potential recruits are screened to exclude those with mental or physical disorders that could interfere with their ability to function in the military. While screening might possibly eliminate the most obvious cases of BPD, procedures are probably sufficiently imprecise to allow those with milder cases (or fewer symptoms) of BPD to enlist. Because some of these symptoms, such as anger and impulsivity, are relatively common in the general population, they may have little importance in isolation from other symptoms of BPD. Could the military culture itself be responsible for inducing these traits in recruits? While an intriguing possibility, it seems unlikely that these self-reported symptoms would reflect the influence of military-related experiences or short-term patterns of adjustment necessitated by the military lifestyle.

There are currently no studies in which BPD was examined in military personnel, although the work of Hoge and colleagues³⁰ is relevant to this subject. They found that the presence of mental illness was an important source of morbidity among active duty United States military personnel; ~9% of hospitalizations and 4% of ambulatory visits involved a diagnosis of personality disorders, though not BPD specifically.

An important finding was the association between the number of BPD symptoms and both functional impairment and utilization of healthcare resources. In short, the higher the number of traits, the greater the impairment (in multiple life domains) and the greater the use of healthcare resources, such as inpatient days, outpatient visits, and emergency department visits. Individuals with BPD are known to have high rates of healthcare utilization, and our findings are compatible with these clinical observations.^4,5 Another important finding is that comorbid psychiatric disorders are not only frequent in persons with BPD, but also in those with BPD symptoms that do not meet diagnostic threshold. If the number of criteria can be used as a proxy for severity, one can conclude that as the severity of BPD symptomatology increases, case complexity increases as well. BPD symptoms are associated with the presence of comorbid mood, anxiety, psychotic, somatoform, eating, and substance use disorders. In fact, of those meeting criteria for BPD, 66.77% met lifetime criteria for a substance use disorder, 53.3% met criteria for major depression, 46.7% met lifetime criteria for posttraumatic stress disorder, 26.7% met lifetime criteria for generalized anxiety disorder, and 13.3% met lifetime criteria each for panic disorder and a somatoform disorder. The pattern of comorbidity reflects what is seen in clinical samples, except perhaps for higher rates of substance use disorders, which may reflect the influence of having a predominantly male sample.^6,7 A finding that is likely unique to the sample is the male preponderance of those with BPD or its symptoms. While typically associated with women,^1,2 BPD occurs in men as well, as this study demonstrates.

Variables involving social and military functioning show the adverse impact of BPD symptoms. Social functioning along all its dimensions shows increasing levels of dysfunction as the number of BPD criteria increase. Employment is negatively impacted, perhaps because borderline symptoms contribute to an unstable work record⁵: for example, being fired, quitting a job out of pique, or having poor relationships with coworkers. There was an association between number of BPD symptoms and incarceration, being court martialed, having received some other form of military discipline, or having used illegal drugs. While these behaviors (or consequences of a behavior) are typically associated with antisocial personality disorder, they are reflective of the tendency of some BPD-diagnosed individuals to behave in ways that are self-defeating or damaging. These results are also indicative of the need for the military to better assess BPD and its many symptoms, and to intervene whenever possible to minimize their negative impact.

Scores on the SNAP self-harm scale and suicide proneness scale also show a relationship with number of BPD symptoms. This finding is not surprising when one considers that deliberate self-harm and suicidal behavior are common components of BPD.¹⁰ Also present was a relationship between the number of BPD traits and family history of suicidal behavior, suggesting that this particular symptom may be familial to some extent.

There was a significant relationship between BPD symptoms and the CES score. As described earlier, the scale involves the assessment of many experiences beyond combat, including perceived environmental exposures and loss of communication. The association can be interpreted in several ways. It can be interpreted to mean that exposure to wartime situations can induce BPD symptoms, in a fashion similar to what is seen with posttraumatic stress disorder; yet, because BPD traits are generally long-standing, this explanation is probably unlikely. Alternatively, because individuals with BPD symptoms are emotionally oversensitive, and often experience overly intense reactions to stressful events, these individuals might report high scores on this scale, reflecting an exaggerated perception of external events. Another explanation is that the impulsivity seen in many BPD patients leads them to select and place themselves into high-risk situations.

There are a number of limitations to the study. First, BPD is more frequent in women and, because the military sample is mostly male, the results cannot be generalized to all veterans. Second, while recall bias could have altered reports of symptoms, the potential for this confound is probably reduced by the use of multiple validated self-report measures, including the SNAP and the SCID, both widely used psychiatric diagnostic instruments that have each been subject to validation studies. Third, SCID raters were personally trained by one of the investigators and reliability checks during the study showed excellent concordance of the raters with one of the study investigators. Finally, the best estimate diagnosis method was used so that all relevant information would be taken into account.

Conclusion

While the current study was not developed as an epidemiologic study and represents an over-sampling of specified disorders of interest (depression, chronic widespread pain, cognitive dysfunction), the findings suggest that BPD and its many symptoms do occur in a military population. Using the number of BPD traits as a proxy for severity suggests their importance due to the positive association between number of symptoms and presence of psychiatric comorbidity and evidence of psychosocial dysfunction. Thus, one of the important results of this study is the finding that as BPD severity increases, subjects not only show greater psychiatric comorbidity, but show behavioral and demographic differences as well. Another interpretation of these findings is that the military might wish to revisit the issue of enlistment screening to identify persons with BPD traits because even individuals with subclinical cases of BPD have a greater likelihood of not succeeding in the military. Alternatively, the military could use the data to encourage the early recognition and treatment of those with BPD symptoms to minimize their impact, for example, referring such individuals for cognitive-behavioral therapies known to be effective in the treatment of BPD.^31,32 One possibility is to use a quick screen for personality disorders and traits, such as one described by Langbehn and colleagues.³³  CNS

References

1. Gunderson J. Borderline Personality Disorder: A Clinical Guide. 2nd ed. Washington, DC: American Psychiatric Publishing, Inc.; 2002.
2. Widiger TA, Frances AJ. Epidemiology, diagnosis, and comorbidity of borderline personality disorder. In: Tasman A, Hales RE, Frances AJ, eds. American Psychiatric Press Review of Psychiatry. vol 8. Washington, DC: American Psychiatric Press; 1989:8-24.
3.  Zimmerman M, Coryell W. DSM-III personality disorder diagnoses in a non-patient sample: demographic correlates and comorbidity. Arch Gen Psychiatry. 1989;46:682-689.
4. Bender DS, Dolan RT, Skodol AE, et al. Treatment utilization by patients with personality disorders. Am J Psychiatry. 2001;158:295-302.
5. Skodol AE, Gunderson JG, McGlashan TH, et al. Functional impairment in patients with schizotypal, borderline, avoidant, or obsessive-compulsive personality disorder. Am J Psychiatry. 2002;159:276-283.
6.  Fyer MR, Frances AJ, Sullivan T, Hurt SW, Clarkin J. Comorbidity of borderline personality disorder. Arch Gen Psychiatry. 1988;45:348-352.
7. Zanarini MC, Frankenburg FR, Dubo ED, et al. Axis I comorbidity of borderline personality disorder. Am J Psychiatry. 1998;155:1733-1739.
8. Paris J. Social Factors in Personality Disorders: A Biopsychosocial Approach to Etiology and Treatment. New York, NY: Cambridge University Press; 1996.
9. Fyer MR, Frances AJ, Sullivan T, Hurt SW, Clarkin J. Suicide attempts in patients with borderline personality disorder. Am J Psychiatry. 1988;145:737-739.
10. Soloff PH, Lynch KG, Kelly TM, Malone KM, Mann JJ. Characteristics of suicide attempts of patients with major depressive episode and borderline personality disorder: a comparative study. Am J Psychiatry. 2000;157:601-608.
11. Clark LA. Schedule for Non-Adaptive and Adaptive Personality (SNAP): Manual for Administration, Scoring, and Interpretation. Minneapolis, Minn: University of Minnesota Press; 1996.
12. Self-reported illness and health status among Gulf War veterans. A population-based study. The Iowa Persian Gulf Study Group. JAMA. 1997;277:238-245.
13. Black DW, Carney CP, Forman-Hoffman VL, et al. Depression in veterans of the first Gulf War and comparable military controls. Ann Clin Psychiatry. 2004;16:53-61.
14. Ang DC, Peloso, PM, Woolson RF, Kroenke K, Doebbeling BN. Predictors of incident chronic widespread pain among veterans of the first Gulf War. Clin J Pain. 2006;22:554-563.
15. Forman-Hoffman VL, Carney CP, Sampson TR, et al. Mental health comorbidity patterns and impact on quality of life among veterans serving during the first Gulf War. Qual Life Res. 2005;14:2303-2314.
16. Diagnostic and Statistical Manual of Mental Disorders. 3rd ed rev. Washington, DC: American Psychiatric Association; 1987.
17. Spitzer RL, Williams JBW, Gibbon M, First MB. The Structured Clinical Interview for DSM-III-R (SCID). I: History, rationale, and description. Arch Gen Psychiatry. 1992;49:624-629.
18. Leckman JF, Sholomskas D, Thompson WD, Belauger A, Weismann MM. Best estimate of lifetime psychiatric diagnosis: a methodological study. Arch Gen Psychiatry. 1982;39:879-883.
19. Carney CP, Sampson T, Voelker M, et al. Women in the Gulf War: combat experience, exposure and subsequent health care use. Milit Med. 2003;168:654-661.
20. Rosenheck R. War Zone Stress Among Returning Persian Gulf Troops: A Preliminary Report. West Haven, Conn: VA New England Program Evaluation Center; 1991.
21. Gallops M, Laufer RS, Yager T. Revised Combat Exposure Scale. In: Laufer RS and Yager T, eds. Legacies of Vietnam: Comparative Adjustments of Veterans and Their Peers. vol 3. Washington, DC: Government Printing Office; 1994:125.
22. Cutrona CE. Social support and stress in the transition to parenthood. J Abnorm Psychol. 1984;93:378-390.
23. Cutrona CE, Russell D. The provisions of social relationships and adaptation to stress. Advances in Personal Relationships. 1987;1:37-67.
24. Ware J. Appendix C: script for personal interview SF-36 administration. In: SF-36 Health Survey Manuals and Interpretation Guide. Boston, Mass: Nimrod Press; 1993.
25. Wells KB, Stewart A, Hays RD, et al. The functioning and well-being of depressed patients. Results from the Medical Outcomes Study. JAMA. 1989;262:914-919.
26. Voelker MD, Saag KG, Schwartz DA, et al. Health-related quality of life in Gulf War era military personnel. Am J Epidemiol. 2002;155:899-907.
27. SAS software. ver 8.2. Carey, NC: SAS Institute Inc.; 2003.
28. Syndecor GW, Cochran WG. Statistical Methods. 6th ed. Ames, IA: Iowa State University Press; 1967.
29. Jonckheere AR. A distribution-free k-sample test against ordered alternatives. Biometrika. 1954;41:133-145.
30. Hoge CW, Lesikar SE, Guevara R, et al. Mental disorders among U.S. military personnel in the 1990s: association with high levels of health care utilization and early military attrition. Am J Psychiatry. 2002;159:1576-1583.
31. Linehan MM. Cognitive-Behavioral Treatment for Borderline Personality Disorder. New York, NY: Guilford Press; 1993.
32. Blum N, Pfohl B, John DS, Monahan P, Black DW. STEPPS: a cognitive-behavioral systems-based group treatment for outpatients with borderline personality disorder–a preliminary report. Compr Psychiatry. 2002;43:301-310.
33. Langbehn DR, Pfohl BM, Reynolds S, et al. The Iowa Personality Disorder Screen: development and preliminary validation of a grief interview. J Personal Disord. 1999;13:75-89.