Symptoms of Borderline Personality Disorder Predict Interpersonal (but not Independent) Stressful Life Events in a Community Sample of Older Adults

Abstract

Individuals with borderline personality disorder (BPD) often experience stressful life events at a higher frequency than those without BPD. It is less clear what specific types of events are involved in this effect, and it has not been determined whether some features of BPD are more important than others in accounting for this effect. The latter issue is important in light of the heterogeneous nature of this diagnostic construct. These issues were examined in a large, representative community sample of men and women, ages 55–64. Ten Diagnostic and Statistical Manual of Mental Disorders (4th ed., text rev., DSM–IV–TR, Washington, DC, American Psychiatric Association, 2000) personality disorders were assessed at baseline using the Structured Interview for DSM-IV Personality: SIDP-IV (B. Pfohl, N. Blum, & M. Zimmerman, 1997, Washington, DC, American Psychiatric Press). Life events were measured at three sequential assessments following baseline at 6-month (N = 1,294), 12-month (N = 1,070), and 18-month (N = 837) follow-ups. Stressful life events were identified using a self-report questionnaire (LTE-Q; List of Threatening Experiences Questionnaire: A subset of prescribed life events with considerable long-term contextual threat by T. Brugha, C. Bebbington, P. Tennant, and J. Hurry, 1985, Psychological Medicine, Vol. 15, pp. 189–194.) followed by a telephone interview. Only borderline personality pathology was related to an increase in the frequency of interpersonal stressful life events. Three specific symptoms of BPD were largely responsible for this connection: unstable interpersonal relationships, impulsivity, and chronic feelings of emptiness (negative association). Symptoms of avoidant and schizoid personality disorders were associated with a reduced number of stressful life events that are considered to be outside a person’s control (e.g., serious illness, injury, or death of a loved one). None of the personality disorders predicted an increase in the number of stressful financial events (e.g., major financial crisis). These findings suggest that, as individuals approach later life, certain features of BPD continue to serve as important risk factors for stressful life events of an interpersonal nature.

Background

The presence of psychopathology, including personality disorder (PD) and depression, can affect the number of stressful life events (SLEs) that a person will experience, as well as the way in which those stressors will affect the individual (Hammen, 2006; Pagano et al., 2004). Research on stress generation and depression posits that certain personality characteristics play an active role in generating these events (Harkness & Luther, 2001; Monroe & Harkness, 2005). Individuals with depressive disorders experience more SLEs than do individuals without depressive histories, and these events are at least partially dependent on the person’s behavior. PDs are another important risk factor to consider and are linked with increased experience of SLEs (Shevlin, Dorahy, Adamson, & Murphy, 2007; Gleason, Powers, & Oltmanns, 2012).

SLEs have been categorized into two main types: independent and dependent events (Brown & Harris, 1978; Paykel, 1987). Independent SLEs are presumably outside a person’s control, such as the illness or death of a loved one. Dependent SLEs may be influenced by the personality and behavior of the individual. For example, getting fired from a job is an event that could be dependent on the person’s own behavior. Interpersonal events related to conflict (e.g., separation from a spouse or major arguments with family/friends) represent one specific subtype of dependent events, and are particularly likely to be influenced by an individual’s personality.

Previous research on stress generation and PDs has shown that risk for an increased number of negative events is most likely to involve dependent SLEs, and interpersonal events in particular (Kendler, Gardner, & Prescott, 2003). PDs, by definition, are associated with problems in the way individuals interact with their environment, as well as the way individuals view themselves and those around them. Borderline PD (BPD) has emerged as a particular risk factor for negative stressors and is consistently linked with greater experience of interpersonal stressors and impaired functioning following those stressors (Pagano et al., 2004; Shevlin et al., 2007).

While the relationship between personality pathology and SLEs has been well established in younger adults, much less is known about this connection in later life (Oltmanns & Balsis, 2011). Negative interpersonal events typically become less frequent as people become more mature, and various forms of personality pathology are also less prevalent in older age groups. Therefore, it seemed possible that PDs may not be associated with an increased frequency of SLEs. Initial evidence from our lab, however, suggests the opposite. Features of BPD were related to increased experience of SLEs among older adults, even when concurrent depressed mood was controlled (Gleason et al., 2012). The analyses that we reported in that initial paper did not examine whether the events associated with symptoms of BPD were dependent or independent, and we did not consider whether the events were interpersonal or financial. Financial crises are particularly relevant in this sample because our research project was launched in 2007, at the start of the “Great Recession,” a time when many individuals began to suffer serious economic losses. Although interpersonal stressors, such as divorce, are generally less frequent among older adults, it is possible that they continue to be more common among those with more symptoms of BPD and other forms of personality pathology. It is also possible that the role of personality pathology in the experience of life stress may carry into other, less controllable events.

As part of an ongoing longitudinal study of middle-aged adults, we examined whether the association between personality pathology and risk for SLEs in later adulthood is affected by the type of stressor experienced. Specifically, we tested if personality pathology predicted a greater number of independent and dependent negative life events. Among the dependent life events, we also separately tested interpersonal and financial life events. We sought to determine whether PD features continue to have a particular influence on interpersonal-related events even into later adulthood.

Method

Design Overview

A representative, community-based sample of adults aged 55–64 was recruited to participate in a longitudinal study of personality and agingf from the St. Louis Personality and Aging Network (SPAN; see Oltmanns & Gleason, 2011 for a detailed description of recruitment and assessment procedures).

Participants

Participants eligible for the current analyses had completed the baseline portion of the SPAN Study (N = 1,630; Oltmanns & Gleason, 2011) as well as at least one of three 6-month follow-ups (up to 1.5 years after baseline; FU1: N = 1,294, FU2: N = 1,070, FU3: N = 837). The average age of participants at baseline was 59.6 (SD = 2.8) and 54% were women (N = 702). The sample was primarily Caucasian (69%; N = 893) or African American (29%; N = 376). Half of participants were currently married (N = 647) and 54% had a bachelor’s degree or higher (N = 699). Note: sample size numbers reflect the smaller sample of 1,294 participants.

Measures

PD features

Using the Structured Interview for DSM–IV Personality (SIDP-IV; Pfohl, Blum, & Zimmerman, 1997), a semi-structured diagnostic interview that consists of 101 questions corresponding to the criteria for the 10 DSM–IV–TR PDs, features were measured at baseline. Each criterion is rated using a scale from 0 (not present) to 3 (strongly present). Each participant’s scaled score for each PD was calculated by adding together the ratings for relevant criteria and computing the average. The scores were then rescaled to a 0–10 scale (all scales used as predictors in analyses were rescaled in order to make regression coefficients more comparable). Means of scaled scores ranged from 0.26 (for antisocial PD) to 1.20 (for obsessive–compulsive PD). Reliability tests indicate adequate reliability at ICC = .67 for the overall scale (all PD’s combined) with individual scales ranging from .53 for paranoid to .86 for avoidant; borderline had a reliability of .77.

SLEs

Every 6-months for 1.5 years after baseline, follow-ups using the List of Threatening Experiences Questionnaire (LTE-Q, Brugha, Bebbington, Tennant, & Hurry, 1985) assessed SLEs. The LTE-Q includes 12 life events that were found to have long-term negative effects on most people who experience them. These 12 items were presented in a checklist that also included three items added specifically for this study (see Table 1 for a list of items).

Table 1.

Frequency of Reclassified SLEs

Event type	Frequency of events
	FU1		FU2		FU3
	%	n	%	n	%	n
Independent Events
Total number	29.5	381	19.2	276	15.9	146
More than one event per participant	3.1	41	2.5	29	1.3	12
By event type:
Serious illness or injury to close other	13.0	169	8.6	100	5.8	53
Death of a partner, parent, or child	3.5	46	4.0	46	1.4	13
Death of a close friend or another relative	16.2	210	13.7	159	10.0	92
Dependent Events
Total number	20.9	271	16.6	195	8.8	80
More than one event per participant	3.7	48	2.9	33	0.9	8
By event type:
Separation due to marital difficulties	0.5	7	0.6	7	0.1	1
Breaking off a steady relationship	1.6	21	1.6	18	0.5	5
Serious problem with a close other	4.3	56	2.8	33	2.1	19
Unemployment for more than a month	3.9	51	3.6	42	0.9	8
Fired or laid off from job	2.1	27	1.7	20	0.5	5
Major financial crisis	4.6	60	4.0	46	2.4	22
Problems with police/courts	0.7	9	0.7	8	0.3	3
Something valuable was lost/stolen	2.9	37	1.5	17	1.2	11
Victim of a serious crime	0.7	9	0.4	5	0.3	3
Changes in family responsibilities	3.7	48	3.3	38	1.2	11

Note. SLE = stressful life event; FU = follow-up.

Structured follow-up interview and reclassification

If at least one event was checked, the participant received a telephone call from one of the study’s trained interviewers. The interviewer asked the participant a series of questions about the event in order to determine that it a) actually occurred, b) occurred in the preceding 6 months, and c) was a major and distinct event that caused considerable distress (see Gleason et al., 2012 for a full description of the follow-up interview). For those events coded as major events, three independent raters (undergraduate research assistants) reclassified all events into one of the 15 event categories to ensure that the event fit logically into the category. The raters were given written descriptions of all events and were asked to choose the category that best fit the event description (without knowledge of original placement). Out of 2,153 events, 30 (1.4%) were dropped because of insufficient information. Reclassification was done based on consensus across all three raters or two out of three raters. Only 74 (3.4%) events could not be agreed upon across the raters; those were discussed in a laboratory meeting and classified based on group consensus. Events classified as “other” were not included in analyses.

For the analyses presented in this paper, we used groupings based on Paykel’s (1987) definitions of independent and dependent life events. Independent SLEs included serious illness or injury to close other, death of a partner/parent/child, and death of a loved one. Dependent SLEs included separation due to marital difficulties, breaking off a steady relationship, serious problem with close other, unemployment for more than one month, losing a job, major financial crisis, problems with police, something valuable lost or stolen, victim of a serious crime, and changes in family responsibility. Interpersonal life events included separation, breakup, and serious problem with close other. Financial events included unemployment, losing a job, and major financial crisis.

Depressed mood

At every 6-month follow-up using the Beck Depression Inventory, II (BDI; Beck, Steer, & Garbin, 1988; BDI-II; Beck, Steer, and Brown, 1996), depressed mood was assessed. The BDI consists of 21 items rated on a 0–3 scale. Items were averaged and then rescaled to a 0–10 scale (M = 0.81, SD = 1.00).

Results

The number of independent SLEs reported over the 1.5 years after baseline assessment ranged from 0 to 3 (FU1: M = 0.33, SD = 0.54; FU2: M = 0.26, SD = 0.49; FU3: M = 0.17, SD = 0.41); approximately 30% of participants met criteria for at least one independent SLE as measured at FU1 (24% for FU2, 15% for FU3). The number of dependent SLEs reported over the entire 1.5 year period ranged from 0 to 4 (FU1: M = 0.25, SD = 0.53; FU2: M = 0.20, SD = 0.51; FU3: M = 0.10, SD = 0.32), with 21% of participants meeting criteria for at least one dependent event (16% for FU2, 9% for FU3). The number of interpersonal SLEs ranged from 0 to 2 (FU1: M = 0.07, SD = 0.26; FU2: M = 0.05, SD = 0.22, FU3: M = 0.03, SD = 0.16) and financial SLEs ranged from 0–1 (FU1: M = 0.10, SD = 0.31, FU2: M = 0.09, SD = 0.30; FU3: M = 0.04, SD = 0.19). Table 1 provides a list of percentages by specific events.

In order to determine whether the number of independent, dependent, interpersonal, and financial events over the 1.5 years following baseline personality assessment (three separate 6-month follow-ups) were predicted by the 10 DSM–IV–TR PDs, we conducted a generalized estimating equation model (GEE) analysis with a Poisson probability distribution using PROC GENMOD in SAS. The Poisson distribution is recommended for use when the dependent variable is a count of a rare event; it is a probability distribution in that a one-unit increase in a predictor variable corresponds to the exponentiated value of that variable’s regression coefficient and “is the predicted multiplicative effect of 1 unit of change” in the predictor variable (Coxe, West, & Aiken, 2009; p. 125). Age, gender, race, between-subjects BDI score (assessed concurrent depressed mood at time of life-event reporting), within-subjects BDI score (assessed predictive value of changes within participant’s own range of depressed mood in predicting reporting SLEs over 1.5 years) were entered as adjustment variables; all predictor variables were centered or effect-coded to ease interpretation of the intercept.

Independent SLEs and Personality Pathology

Across the three follow-up assessments, the predicted number of independent life events when all variables are held constant at their mean was 0.30 (exp[−1.18] = 0.30). Schizoid and avoidant PD features were negatively associated with independent event reporting (exp[−0.19] = 0.82, χ² = 8.35, p = .007 and exp[−0.14] = 0.87, χ² = 7.33, p = .01, respectively); every point increase in schizoid PD score was associated with 0.82 times fewer events being reported. In other words, a person with a score of 2 on the schizoid PD scale is expected to report 0.82 times fewer SLEs as someone with a score of 1 (0.87 for avoidant PD). Between-person BDI score (concurrent depressed mood at time of each follow-up) was also significant; individuals who had a higher BDI score reported significantly more events from the previous 6 months (exp[0.18] = 1.19, χ² = 14.10, p < .001). Minority status was also significant; minority participants reported significantly more independent life events than Caucasian participants (exp[0.80] = 1.35, χ² = 12.32, p < .001). There was also an effect of time, showing that individuals reported significantly fewer independent events over time from FU1 through FU3 (exp[−0.44] = 0.64, χ² = 32.58, p < .001). There were no significant interactions between time and PD features and no effect of within-person BDI score on predicting reported events.

Dependent SLEs and Personality Pathology

The predicted number of dependent events was 0.16 (exp[−1.81] = 0.16); Only two of the 10 PDs were significantly associated with the number of dependent SLEs. BPD was positively associated with event reporting (exp[0.17] = 1.19, χ² = 7.21, p = .006); for every point increase in the BPD scale, 1.19 times more events are predicted. Histrionic PD was also significant (exp[0.14] = 1.16, χ² = 2.93, p = .06). Concurrent depressed mood and minority status also predicted a greater number of dependent life events reported (exp[0.36] = 1.44, χ² = 32.44, p < .001 and exp[−0.27] = 0.76, χ² = 5.38, p = .02, respectively). Again, there was a time effect showing that individuals reported significantly fewer dependent events over the 1.5 years measured (exp[−0.46] = 0.63, χ² = 28.44, p < .001). No significant interactions between time and PD features or effect for within-person BDI score emerged.

Interpersonal SLEs and Personality Pathology

The predicted number of interpersonal events was 0.04 (exp[−3.20] = 0.04). BPD was again positively associated with interpersonal event reporting (exp[−0.23] = 1.26, χ² = 3.56, p = .059); for every point increase in the BPD scale, 1.26 times more events are predicted. Concurrent depressed mood also predicted a greater number of interpersonal life events reported (exp[0.32] = 1.38, χ² = 10.80, p = .001). Again, there was a time effect showing that individuals reported significantly fewer dependent events over the 1.5 years measured (exp[−0.41] = 0.66, χ² = 5.88, p = .02). No significant interactions between time and PD features or effect for within-person BDI score emerged.

Financial SLEs and Personality Pathology

The predicted number of financial events was 0.07 (exp[−2.61] = 0.07). No personality variables were significantly predictive of reported financial events over the 1.5 years following baseline. However, concurrent depressed mood was significantly associated with the number of events reported in the previous 6 months (exp[.32] = 1.38, χ² = 5.88, p = .02). Younger age and minority status also significantly predicted more reported financial events (exp[−0.10] = 0.91, χ² = 14.35, p < .001 and exp[−0.28] = 0.76, χ² = 3.04, p = .08, respectively). No significant interactions between time and PD features or effect for within-person BDI score emerged.

Association Between Specific BPD Symptoms and SLEs

To further understand associations between borderline pathology and dependent and interpersonal SLEs, we also examined specific BPD symptoms in relation to these types of SLEs. We ran additional Poisson regression models to assess the relation between all nine borderline symptoms and the two types of life events. For both dependent and interpersonal events, three borderline symptoms were significantly predictive of a greater number of events: “unstable interpersonal relationships”, “impulsivity,” and “chronic feelings of emptiness” (see Table 2 for coefficients) even when demographic variables, mood, and time were included in the models.

Table 2.

Significant PD Coefficients Predicting Dependent and Interpersonal SLEs

BPD features	Life events
	Coefficient	SE	χ2	p
Dependent
Unstable relationships	0.34	0.12	5.66	0.02
Impulsivity	0.27	0.08	10.21	0.001
Chronic emptiness	−0.27	0.13	5.37	0.02
Interpersonal
Unstable relationships	0.49	0.25	2.53	0.05
Impulsivity	0.50	0.15	8.59	0.00
Chronic emptiness	−0.55	0.25	5.63	0.03

Note. PD = personality disorder; BPD = borderline PD; SLE = stressful life event.

Discussion

The results of this study demonstrate that the relationship between personality pathology and the experience of SLEs differs in important ways as a function of the type of event being considered. Consistent with previous research (e.g., Pagano et al., 2004), we found that borderline pathology was significantly predictive of experiencing a “dependent” SLE in the 18 months following the baseline personality assessment. More specifically, borderline pathology was related to a higher frequency of interpersonal SLEs. Although histrionic PD features were also significantly associated with dependent SLEs, only BPD was associated with interpersonal events, suggesting that the constellation of symptoms in individuals with borderline pathology is particularly detrimental in relation to interpersonal functioning in older aged community residents. Although the overall difference in risk for interpersonal SLEs between those high and low in BPD symptoms was relatively small (1.26 times more likely to report SLE), the fact that these events were so rare within our sample suggests that it is likely very disrupting for those individuals who are experiencing them, and these experiences may have a particularly detrimental impact on later functioning. In fact, a recent study by Distel et al. (2011) regarding Gene × Environment interactions in relation to BPD and SLEs showed that exposure to certain life stressors, including divorce, can serve to exacerbate BPD symptoms. Their study was done with a sample of adult twins (average age in mid-30s). It will be important in future research to examine exactly how exposure to these kinds of interpersonal stressors in later adulthood can lead to a deterioration in symptoms of psychopathology and levels of functioning.

Of the nine BPD symptoms, three stood out as being particularly important in relation to dependent and interpersonal SLEs: unstable interpersonal relationships and impulsivity were associated with an increase in SLEs, whereas chronic feelings of emptiness showed the opposite pattern. It is not surprising that previous patterns of unstable and intense interpersonal relationships would be related to ongoing interpersonal stressors included in the LTE-Q. Similarly, impulsive behaviors, such as gambling, promiscuity, or shoplifting, are often related to relational problems; if someone engages in these negative behaviors, problems with spouses, family, or friends would likely result. It is interesting that these associations remain evident in our sample of older adults. The prevalence of divorce and serious interpersonal altercations declines with age (Jordanova et al., 2007), but our findings suggest that some individuals may remain at greater risk for those outcomes as a result of personality pathology. Consistent with expectations regarding the “burnout” of borderline pathology in later adulthood (Paris, 2003), we did not identify many individuals who met current DSM–IV–TR criteria for a categorical diagnosis of BPD. These results raise the possibility that, although certain symptoms of BPD (e.g., impulsivity) may occur less frequently in later adulthood (Gunderson et al., 2011), they continue to generate long-term negative consequences, even if they are present at sub-threshold levels of the disorder.

The fact that “chronic feelings of emptiness” was related to a significant decrease in the number of interpersonal events is somewhat surprising. Feelings of emptiness are related to low positive affect and social isolation (Klonsky, 2008), perhaps leading to fewer opportunities for interpersonal stress. This result was independent of depressed mood at the time of follow-up assessment, suggesting that there is something important about “feelings of emptiness” above and beyond low mood alone. In this case, it seems to serve as a protective factor against interpersonal SLEs, in spite of the overall negative impact of other BPD symptoms.

PD features were not significantly related to the reported experience of financial events. Instead, financial events were related only to depressed mood at the time of each follow-up assessment. It is not surprising that financial stressors would be associated with higher levels of depression. Our initial follow-up assessments began in early 2008, just as the “Great Recession” began to have a dramatic impact in the United States. Unemployment rates doubled over a period of several months, and stock market declines threatened the financial security of millions of people. Most of the participants in our study were either approaching retirement or already retired as these events unfolded, making them particularly vulnerable to negative financial events and perhaps, subsequently depression.

We found that depressed mood at the time of event reporting was significantly associated with all types of SLEs. It is possible that this association is the result of overreporting of events due to the depressive symptoms, but it also suggests that people who experience a major negative event are more likely to become depressed. A recent longitudinal study by Conway, Hammen, and Brennan (2012) examined associations between internalizing and externalizing symptoms and the process of stress generation in a sample of young adults. Internalizing symptoms (including depressed mood) predicted higher levels of interpersonal SLEs. Our results suggest that certain aspects of personality pathology, in this case, symptoms of BPD, are also relevant in the discussion of stress generation across the life span.

Our results also indicate that people with more features of schizoid and avoidant PD reported fewer independent SLEs. These events included the serious illness, injury, or death of a loved one. This result, in combination with previous findings from our lab (Gleason et al., 2012) indicates that certain PD features—schizoid, paranoid, and avoidant PD – may actually lead to either underreporting or experiencing fewer life events. We cannot determine which of these alternative explanations is correct because only participants who reported events were contacted for a follow-up interview. One mechanism that may account for this pattern involves social isolation. Because people who exhibit features of schizoid and avoidant PD tend to have limited social networks, they may be less likely to be confronted with the illness or death of a loved one. Although it could be seen as beneficial for individuals not to suffer the stress associated with someone else’s sufferings, many other negative effects of low social support have been found. For example, social support plays an important protective role against the onset and progression of chronic conditions such as cardiovascular disease or cancer and helps individuals better manage daily stressors (Uchino, 2009). This is particularly relevant in later adulthood, when individuals are more likely to face health stressors. It is possible that while individuals with these features are facing fewer stressors related to loved ones, they may be more likely to have a harder time managing stress related to their own health.

Limitations

This research was conducted with people in a specific age group (adults aged 55–64) and the experience of SLEs in older adults may differ in important ways from other age groups. Research on life events across the life span (Jordanova et al., 2007) has shown that younger adults tend to experience relational difficulties in greater frequency, whereas older adults more often report health problems experienced by their loved ones and themselves. This difference is consistent with our results. Our results do suggest, however, that the connections that we observed between personality pathology and the reported experience of life events are similar in many ways to connections found with younger adults.

Because we studied a representative sample of adults living in the community rather than a sample of clinical patients, our sample does not include a large number of individuals with extreme levels of personality pathology. However, roughly 10% of our sample met DSM–IV–TR criteria for at least one PD (Oltmanns & Gleason, 2011), as expected based on previous epidemiological studies (9.1%; Lenzenweger, Lane, Loranger, & Kessler, 2007). In addition, almost half of the people in our sample (>40%) had received some kind of mental-health treatment, demonstrating that individuals with varying levels of psychopathology were represented in this dataset (Lawton & Oltmanns, in press).

Conclusions

Very little research has been conducted on the role of personality pathology in SLEs in later adulthood. This sample provides the opportunity to gain an initial understanding of how personality pathology relates to stressors in late middle age. The results of the present study suggest that, even if the relative frequency of different types of stressors changes over the life span, the risk for negative interpersonal functioning associated with borderline pathology persists. Certain symptoms, such as impulsivity, continue to have a negative impact on interpersonal functioning as people approach the challenges of later life.