This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Brondolo, E. Articles by Sloan, R. P. Search for Related Content PubMed PubMed Citation Articles by Brondolo, E. Articles by Sloan, R. P. Related Collections Blood Pressure Other Cardiovascular Medicine Personality

Hostility, Interpersonal Interactions, and Ambulatory Blood Pressure

From St. John’s University (E.Br., R.R., S.A.E.), New York, NY and Columbia University (E.B., P.A.S. R.P.S.), New York, NY.

Address correspondence and reprint requests to Elizabeth Brondolo, PhD, Department of Psychology, St. John’s University, New York, NY. E-mail: brondole{at}stjohns.edu

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHOD RESULTS DISCUSSION NOTES REFERENCES

 
OBJECTIVE: This study examined aspects of the transactional model of hostility and health by investigating relationships among hostility, interpersonal interactions, and ambulatory blood pressure in a healthy community sample.

MATERIALS AND METHODS: Participants included 65 female and 39 male healthy adults between the ages of 18 and 46 years. Ambulatory blood pressure (ABP) and diary data on mood and social interactions were obtained every 20 minutes for 1 day. Mixed models regression analyses were used to evaluate the relationships among hostility, interpersonal interactions, and ABP.

RESULTS: Trait hostility was positively associated with the frequency and intensity of negative interactions and was negatively associated with the frequency and intensity of positive interactions. Interacting with others was associated with increases in systolic blood pressure (SBP) and diastolic blood pressure (DBP). The magnitude of the increase in blood pressure was positively associated with the degree to which the interaction was perceived as negative. Hostility was not directly associated with ABP/heart rate (HR) or ABP/HR responses during any interactions or negative interactions. However, there was an interaction between hostility and negative interaction intensity for DBP, suggesting that hostility moderates the effects of negative interactions on DBP. Specifically, increases in the intensity of negative interactions were associated with increases in DBP for participants with high, but not low, hostility.

CONCLUSIONS: The results provide partial support for the notion that hostility may be associated with risk for cardiovascular disease through its effects on interpersonal interactions and their cardiovascular correlates.

Key Words: hostility, • interpersonal interactions, • ambulatory blood pressure, • reactivity.

Abbreviations: ABP = ambulatory blood pressure;; ABPM = ambulatory blood pressure monitor;; BMI = body mass index;; BP = blood pressure;; BPM = beats per minute;; CVD = cardiovascular disease;; CVR = cardiovascular reactivity;; DBP = diastolic blood pressure;; HMR = hierarchical multiple regression;; HR = heart rate;; SBP = systolic blood pressure.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHOD RESULTS DISCUSSION NOTES REFERENCES

 
Research has consistently demonstrated a relationship between hostility and the development of cardiovascular disease (CVD) (1,2). However, the mechanisms through which a personality characteristic such as hostility might influence cardiovascular risk are still not completely specified. The transactional model suggests that hostility influences health in part through its effects on social relationships and on the patterns of cardiovascular reactivity associated with interpersonal and other stressors (3). The cumulative effects of these stress-related episodes of heightened cardiovascular reactivity may contribute to the development of coronary heart disease through a number of pathophysiological mechanisms (1,4–6).

Several methodologies have been used to support different facets of the transactional model. Retrospective questionnaire studies have examined the relationship of hostility to the frequency and intensity of interpersonal conflict and suggest that hostility influences the quality of social interactions (7–13). Laboratory studies show that, in comparison with individuals with low hostility scores, those with high hostility scores rate negative interpersonal experiences in the laboratory (eg, harassment) as significantly more anger-inducing, irritation-inducing, and tension-inducing (14,15). Many (9,14,16–18), although not all, studies (19–22) have found that high levels of hostility are associated with greater blood pressure (BP) or heart rate (HR) responses to interpersonal stressors. There is also some evidence that hostile individuals display greater physiological reactivity to stress than less hostile individuals, even when they do not differ in the self-reported level of negative emotions evoked by the stressor (23).

Although the analogue situations used in laboratory studies of the cardiovascular correlates of social interaction have been innovative and capable of eliciting cardiovascular activation, there are limits to the interpretability of the data from these studies. In the laboratory, individuals may use different strategies for expressing anger than in unobserved "real-life" interactions. This may be problematic, given that a portion of the effects of anger-related traits on cardiovascular reactivity (CVR) may be a function of the cardiovascular demands of managing the expression of emotion (24–26). Neither survey nor laboratory studies permit an evaluation of the degree to which hostility influences the frequency of exposure to negative interactions and concomitant cardiovascular activation during the course of everyday events.

Ambulatory monitoring studies have been used to address these limitations by examining the relationships among hostility, negative interactions, and BP under more naturalistic conditions. Hostility has been associated with daily negative mood in several studies (27,28); however, to date, no published study has examined the relationship of hostility to the frequency and intensity of everyday negative interpersonal interactions using ecological momentary assessment methods.

Hostility and other anger-related traits have been associated with average ambulatory blood pressure (ABP) level in many studies (24,25,28–36), although not all measures of anger and hostility are associated with ABP (28,29). Early investigations did not control for posture, location, physical activity, and drug or alcohol intake (24,25,31,33). More importantly, there are only limited data available on the relationship of hostility to the magnitude of ABP during naturally occurring social interactions.

In one of the first studies to address this issue, Jamner, Shapiro, Goldstein, and Hug (25) examined paramedics and reported that hostility and defensiveness were associated with higher DBP in settings that potentially involved interpersonal stress, such as a hospital emergency room. However, the authors did not directly assess the occurrence or the nature of the interpersonal interactions. Guyll and Contrada (30) used ABP to explicitly investigate whether hostility was differentially related to ABP during social versus nonsocial interactions. The results of the repeated-measures ANOVA indicated that there was a significant interaction of hostility with talking, such that participants high in hostility displayed higher levels of SBP than those low in hostility only when talking. Participants were subdivided into high- and low-hostility groups by gender, and the analyses included only those participants with three or more readings obtained while sitting and talking, resulting in groups with varying and very small numbers of subjects. The use of mixed models regression analyses permits the examination of the effects of hostility across all participants and observations, adjusting for individual differences in the number of observations, and controlling for the effects of posture on an observation level basis.

In our previous research, we used mixed models to examine the relationship of hostility to ABP responses shown by traffic agents during potentially negative encounters with the public. Communication with the public was associated with increased BP relative to most other workday activities (37). Trait hostility was associated with the magnitude of SBP response during communication with the public, but the effects were small and significant only when all agents were included in the analyses, including some taking antihypertensive medication (38).

Analyses of the relationship of hostility to ABP are further complicated by the fact that hostility is a multifaceted construct (39). Researchers have suggested that measures of hostility, including the Cook-Medley Ho contain cognitive, affective, and behavioral dimensions. Analyses have revealed that the dimensions of hostility are differentially associated with risk for CHD (40). Studies have yet to investigate the relationship of these different hostility subcomponents to negative interpersonal interactions and their cardiovascular correlates.

This study investigated the relationship of hostility to naturally occurring social interactions and to ABP in a community sample of adults. The analyses addressed four questions: (1) is hostility associated with the frequency and intensity of naturally occurring negative social interactions?; (2) are the frequency and/or intensity of naturally occurring negative interactions associated with the magnitude of ambulatory BP or HR?; (3) is hostility associated with the magnitude of BP or HR during social interactions?; and (4) which components of hostility are most closely associated with risk for conflict, with the intensity of distress during conflict, and with the magnitude of ABP response to conflict?

	MATERIALS AND METHOD

TOP ABSTRACT INTRODUCTION MATERIALS AND METHOD RESULTS DISCUSSION NOTES REFERENCES

 
Participants
Volunteers were recruited through advertisements to participate in a longitudinal study of the effects of exercise on cardiovascular function. In total, 287 individuals agreed to participate and 104 individuals met inclusion criteria. Inclusion criteria included oxygen consumption (VO₂) maximum (max) levels that were no greater than 43 and 39 ml/kg per minute for men and women, respectively, and a body mass index less than 31.3. Exclusion criteria included the following: a current program of active exercise, a history of or current depression or anxiety disorder, or the current use of any type of antihypertensive or psychoactive medications. Participants were also excluded if they reported a medical condition (eg, exercise-induced asthma, orthopedic problems, etc.) that would restrict their ability to safely participate in a demanding exercise program. Of the original 287 individuals who initially volunteered for the project, 109 decided to leave the project before the first testing session because of the magnitude of the commitment; 55 were excluded because they were too fit and did not meet the VO₂ max criteria, 10 did not complete all psychological testing, and nine had medical conditions that warranted exclusion.

The 104 volunteers accepted into the program included 65 women and 39 men. The mean age of the sample was 30 years (standard deviation [SD]: 7.70; range: 18–46 y). The sample was 36% white, 19% black, 25% Latino, 16% Asian, and 5% "other" ethnicity. Participants were employees of a major New York City medical center or students at the affiliated schools.

Equipment
ABP measurements were made with the Suntech Accutracker II (Suntech Medical Instruments, Raleigh, North Carolina), a light-weight (2.25-lb) ABP monitoring device that uses R-wave-gated Korotkoff sounds for BP determination (41). The BP cuff was programmed to automatically inflate at 20-minute intervals throughout the testing period. On each reading, the Accutracker provided information on the reliability of measurement. Excessive movement or insufficient Korotkoff sounds are examples of problems that would produce a warning message.

Psychological Measures
Demographics
A brief questionnaire inquired about age, sex, race, education, and health risk variables (eg, weight, height, alcohol and tobacco use, and history of hypertension).

Hostility
Hostility was measured using the Cook and Medley Hostility Scale (42), ie, Ho Scale, a 50-item true or false questionnaire derived from the MMPI. Psychometric properties of the scale have been well-established in other research (12,43), with the scale demonstrating good internal consistency and moderate test–retest reliability over extended periods (3,7,44). We used strategies recommended by Barefoot et al. (1989) (40) to divide the Ho Scale into rationally derived subscales including: cynicism, hostile attribution, hostile affect, aggressive responding, and social avoidance.

Social interactions
A paper-and-pencil ambulatory monitoring diary was used to collect information about social interactions and activities at the time of cuff inflation. One item inquired if the participants were interacting with others at the time of cuff inflation, and another asked if the participant was talking at the time of cuff inflation. Eight additional items required participants to rate the quality of the interaction, with each item rated on a 1- to 5-point Likert scale. Three items described positive qualities of the interaction (ie, "pleasant," "friendly," and "agreeable"). These items were internally consistent and had a Cronbach’s alpha of 0.94. The mean of these three items served as the positive intensity rating (ie, the measure of the degree to which the interaction was perceived as positive). Five items referred to negative qualities of the interaction (ie, "uncomfortable," "tense," "confrontational," "openly angry," or "about something upsetting"). These items were also highly interrelated with a Cronbach’s alpha of 0.84. The mean of these five items served as the negative intensity rating.

The total number of interpersonal interactions was computed by counting the number of times the participants indicated they were interacting at the time of the cuff inflation. The total number of negative interactions was calculated by counting the number of interactions for which the average negative intensity score was greater than 1, indicating at least some experience of anger, discomfort, upset, or tension associated with the interaction. The total number of positive interactions was calculated using a different strategy, because a low score on a positive rating can indicate that the interaction was actually negative. Therefore, interactions were only considered positive if the mean positive intensity score for the particular interaction met or exceeded the individual’s average positive intensity score. Baseline observations were not included in any of these calculations, because interactions during baseline occurred between the participant and the experimenter under laboratory (and therefore, not random or participant-driven) conditions.

Affect
There were five affect items (ie, happy, relaxed, sad, nervous, and irritated) designed to permit testing of the valence (positive or negative) and intensity of the emotional state. Responses to the affect items were made by marking a 4-inch line extending from the words "not at all" to "very much." Scores on the affect items could therefore range from 1 to 100, with a mark made halfway across the line indicating a score of 50.

Control variables
The ambulatory diary inquired about other factors that could influence the BP readings, including (among other variables): location and posture at the time of cuff inflation and the intake of caffeine, alcohol, or food, or the use of cigarettes or the bathroom either at the time of cuff inflation or during the period between cuff inflations.

Procedure
Ambulatory blood pressure assessment
Participants were tested individually and outfitted with the Accutracker ambulatory blood pressure monitor (ABPM) on the morning of the testing day. A series of eight BP and HR measurements were taken before the start of the workday. Three seated BP measurements were made to allow for the calibration of the ABPM. Research assistants obtained additional BP measurements on the participant’s opposite arm using an aneroid sphygmomanometer to ensure that the monitor was obtaining accurate readings. The Accutracker was re-adjusted and additional readings were taken if the aneroid readings did not correspond within 7 mm Hg SBP.

A series of eight seated and standing baseline readings were obtained. For the remainder of the testing day, BP and HR readings were obtained every 20 minutes. If a reading was judged to be invalid by the algorithms programmed in the Accutracker, an additional reading was obtained 4 minutes after the original one.

To ensure that participants understood how to complete the diary, the experimenter role-played different situations with the participants and asked them to practice choosing diary items to reflect their feelings and behavior. Two diary pages were completed under the supervision of the experimenter during the baseline testing period.

Analytic plan
To investigate the relationship of hostility and hostility subcomponents to interaction frequency, Pearson correlations were used to separately evaluate the relationship of the full Ho score and each hostility subcomponent to the frequency of social interactions. HMR analyses were used to examine the joint effects of the subcomponents as well as the unique effect of each subcomponent controlling for all others. Subcomponents were entered individually if they met the p < 0.15 criteria for inclusion and were entered in the following order: cynicism, hostile attributions, hostile affect, aggressive responding, and social avoidance. This order was chosen to permit cognitive or attitudinal subcomponents (ie, cynicism and hostile attributions) to enter before affective subcomponents (ie, hostile affect), and for both cognitive and affective subcomponents to enter before behavioral subcomponents (ie, aggressive responding and social avoidance).

To examine the relationship of hostility and hostility subcomponents to interaction intensity or to ABP or AHR or to examine the effects of interacting on ABP/AHR, mixed models analyses were performed using PROC Mixed from SAS Institute (45). As noted elsewhere (37,46–49), in comparison with repeated measures analyses conducted using ANOVA or MANOVA models, mixed models offer a more efficient and potentially more powerful strategy for significance testing in ambulatory monitoring data. These models are a generalization of the standard linear model that estimates and tests the significance of between-person and within-person effects while handling the correlated residuals that typically occur with repeated measurements from the same subject. PROC Mixed permits control of within-person covariates (ie, mood, location, posture, activity level, emotion, etc.) on an observation by observation basis.

Analyses were performed using three variance structures: compound symmetry alone, autoregression alone, and compound symmetry plus serial autocorrelation, using the sp(pow) procedure from SAS. Because the estimated autocorrelation was statistically significant in each case, and the combination of compound symmetry and autoregressive error structures produced a better fit than compound symmetry alone, the combined compound symmetry and autoregressive error structures were used in all analyses. Therefore, the F and p values reported for the within-person analyses were those obtained after specifying the combined error structure. Nonstandardized estimates are reported for each analysis to facilitate evaluation of the size of effects.

In analyses of the relationship of hostility to interaction intensity or to ABP/AHR, total Ho and then each subcomponent of hostility were first tested individually. Next, the subcomponents were entered simultaneously to evaluate their joint effects as well as the unique effects of each subcomponent controlling for all others.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHOD RESULTS DISCUSSION NOTES REFERENCES

 
Data Reduction
A total of 5036 readings were obtained from the 104 subjects. Detection and deletion of artifactual readings followed recommendations from the manufacturer, and the suggestions outlined in Clark, Denby, and Pregibone (50). SBP, DBP, and HR readings were deleted if the difference between SBP and DBP exceeded 90 mm Hg or was less than 20 mm Hg. SBP values were deleted if they decreased to less than 85 mm Hg or increased to more than 190 mm Hg or if they decreased to less than 95 mm Hg and were accompanied by an error warning from the Accutracker. DBP readings were deleted if they decreased to less than 40 mm Hg or 137 mm Hg or if they decreased to less than 60 and were accompanied by an error message from the Accutracker. HR values were deleted if they decreased to less than 45 BPM or were more than 137 BPM. These ranges were chosen because points outside this range were uniformly accompanied by codes from the Accutracker, indicating that an error (eg, movement, inadequate Korotkoff sounds, etc.) had occurred during these periods. Readings inside this range with errors were also deleted if the error code indicated that the change in SBP, DBP, or pulse pressure from the previous reading was such that the current reading was unlikely to be valid. As suggested by Jamner, Shapiro, Goldstein, and Hug (25), if either the DBP or SBP reading was deemed invalid, then both BP readings were deleted. After deletion of artifactual readings, the data set contained 84% of the BP readings (4230 observations) and 83% of HR readings attempted (4161 observations). In some cases, both BP readings appeared valid, and the Accutracker indicated no error had occurred, but the HR values were out of the acceptable range, (ie, were less than 45 or more than 137 BPM). In these cases, the HR was deleted, but the BP readings were retained. The first eight readings were collected during a structured protocol used to obtain standing and sitting baseline readings. During the remainder of the testing day after the baseline period, 3450 observations with acceptable BP readings and 3391 observations with acceptable HR readings were collected across the sample. On average, 32 (range: 10–55) readings per person were obtained during the postbaseline workday.

Diary page completion was triggered by cuff inflation. Participants completed diary pages during 87% of the readings on which BP data were judged to be reliable (3004 of a possible 3450). Because we used paper and pencil and not automated diaries, we cannot be certain that diaries were completed at the same time or within 1 minute of cuff inflation. However, a mixed models analysis with posture as an independent variable with four levels revealed that the pattern of mean DBP (adjusted for age, body mass index [BMI], gender; alcohol, caffeine, and cigarette use; and talking) associated with the different levels of posture is consistent with the expected pattern. Least squares mean DBP was lower when lying down (73.58 mm Hg; SE = 1.74) than when sitting (77.05 mm Hg; SE = 1.35) and lower when sitting than standing (78.82 mm Hg; SE = 1.38) or walking (78.06 mm Hg; SE = 1.49). Therefore, it is reasonable to suppose that diary pages were likely to be completed close to the time of cuff inflation.

Descriptive Statistics
Average total Cook and Medley Hostility scores were in the low-to-moderate range, with a mean of 17.21 (SD = 8.11; range: 3–46). As indicated in Table 1, the subcomponents of hostility were generally interrelated. The strength of the association varied, with the strongest relationship emerging between cynicism and hostile attributions (r = 0.61, p < 0.001) and the weakest relationship between aggressive responding and social avoidance (r = 0.18, p < 0.07). Average SBP and DBP for all participants (averaged by individual) was 122.52 (SD = 13.26) and 73.29 (SD = 7.15), respectively. Average HR was 80.62 BPM (SD = 8.58).

Hostility and the frequency of social interactions
Table 1 displays the relationships of hostility and its subcomponents to measures of the frequency and intensity of social interactions. Total Ho score was negatively correlated with the overall rate of interactions during the testing day and positively correlated with the absolute frequency of negative interactions and with the proportion of negative interactions (ie, the frequency of negative interactions controlling for the total number of interactions). Total Ho score was negatively correlated with the frequency of positive interactions; however, the effects were no longer significant after controlling for the total number of interactions, because the frequency of positive interactions is correlated with the frequency of all interactions (r = 0.79, p < 0.0001).

Hostility subcomponents and the frequency of social interactions
Pearson correlations indicated that both hostile attributions and aggressive responding were significantly negatively correlated with the total number of interactions (Table 1). Examined individually, hostile affect and social avoidance were positively correlated with the absolute frequency of negative social interactions; however, cynicism, hostile attributions, hostile affect, and social avoidance were all positively correlated with the proportion of negative interactions. Hostile attributions and hostile affect were significantly negatively correlated with the absolute frequency of positive social interactions. When the analyses were repeated controlling for the total number of interactions, none of the subcomponents was associated with the proportion of positive interactions.

HMR analyses were used to determine which subcomponents contributed significantly to the prediction of interaction frequency when other subcomponents were evaluated for inclusion in the equation. Details of the regression analyses are included in the notes section of Table 1. Across the analyses, hostility subcomponents predicted between 3% and 8% of the variance in social interaction frequency. Hostile attributions was negatively associated with total frequency of interactions and the frequency of positive interactions and positively associated with the proportion of negative interactions.¹ Hostile affect was positively associated with the absolute frequency of negative interactions.

Hostility and the intensity of interactions
PROC mixed analyses revealed that the total Ho score was positively associated with the intensity of negative interactions across the day (B = 0.014, SE = 0.003, df = 102, t = 4.65, p < 0.0001). A separate analysis revealed that total Ho score was negatively associated with the intensity of positive interactions (B = -0.022, SE = 0.01, df = 102, t = -2.89, p < 0.0047).

Subcomponent analyses
Examined independently in five separate mixed models analyses, all subcomponents except social avoidance were significantly associated with the intensity of negative interactions (Table 1). When all five subcomponents were in the model, both cynicism and hostile affect were significantly positively associated with the intensity of the negative interactions. Examined separately, cynicism, hostile affect, and hostile attributions were all negatively associated with the intensity of positive interactions. With all five subcomponents in the equations, only cynicism was negatively associated with the intensity of positive interactions.

Social Interactions and ABP and AHR
Proc mixed analyses were used to evaluate the association of interpersonal interactions to BP and HR. Interacting at the time of the cuff inflation (yes or no) served as the predictor variable and was tested as a random effect. Interactions occurred in 49% (1318) of these observations. Age, BMI, and gender served as between-person covariates. Posture, smoking, alcohol, and caffeine use served as within-person covariates.

BP, but not HR, was higher when participants were engaged in a social interaction (Estimates for SBP (B = -1.67, SE = 0.52, df = 100, t = -3.19, p < 0.001) and DBP (B = -1.37, SE = 0.43, df = 100, t = -3.17, p < 0.002)². Specifically, least squares means (adjusting for all other variables in the equation) revealed that the participants’ BPs were higher when they were interacting (mean SBP = 126.85 mm Hg; mean DBP = 74.39 mm Hg) than when they were not (mean SBP = 125.18 mm Hg; mean DBP = 73.02 mm Hg).³

Interaction Quality and ABP
Proc mixed analyses were used to examine the relationship of the proportion and intensity of negative and positive interactions to ABP and HR during the testing day. These analyses included observations in which an interaction occurred and for which complete data were available. The between-person predictors included the proportion of interactions regarded as negative (ie, number of negative interactions/total number of postbaseline interactions) and the proportion of interactions regarded as positive. The within-person predictor variables, tested as random effects, included the negative and positive intensity ratings for the interaction. Each set of analyses was performed three times, once for SBP, DBP, and HR. Age, BMI, and gender were included as between-person covariates. Posture, smoking, alcohol use, caffeine use, and talking were included as within-person covariates.

The intensity of negative interactions was significantly positively associated with DBP (B = 1.99, SE = 0.55, df = 1355, t = 3.60, p < 0.001) as was the intensity of positive interactions (B = 1.24, SE = 0.34, df = 1355, t = 3.65, p < 0.001). The effects of interaction intensity or frequency on SBP and HR were nonsignificant.

Because the association of interaction intensity on DBP could be an artifact of the well-documented associations of emotional state to ambulatory blood pressure (51), the analyses were repeated with variations in the five mood states included as covariates. The effects of negative and positive interaction intensity on DBP remained significant, and a significant relationship of negative interaction intensity on SBP emerged. The results of the full model (ie, including all independent variables and all covariates) are displayed in Table 2.

To determine whether hostility was associated with increased cardiovascular activation primarily during social interactions, we tested the interaction of hostility by social interaction (yes or no) on SBP, DBP, and HR. There was a marginally significant interaction of hostility and social interaction for DBP (F[1, 2596] = 3.78; p = 0.05), but not for SBP or HR. However, follow-up analyses examining the relationship of hostility to DBP revealed that the relationship of hostility to DBP was nonsignificant both when participants were not interacting (B = 0.09, SE = 0.08, df = 96, t = 1.16, p > 0.24) and when they were (B = -0.03, SE = 0.08, df = 96, t = -0.36, p > 0.72), although the size of the relationships differed. The pattern did not change when separate analyses were conducted testing the interaction of each hostility subcomponent with interacting.

To determine whether hostility is associated with the magnitude of cardiovascular activation during negative interactions, we examined the interaction of hostility and negative interaction intensity on ABP and AHR, controlling for the main effects of both negative interaction intensity and hostility and treating negative interaction intensity as a random effect. There was a significant interaction of hostility and negative interaction intensity for DBP (F[1, 1208] = 5.22; p < 0.03), but not for SBP or HR.

To facilitate interpretation of this interaction, we divided participants into high- and low-hostility groups based on the median Cook Medley Ho score (median = 16). Among those high in hostility, there was a trend toward a significant effect of negative interaction intensity on DBP (B = 1.40, SE = 0.77, df = 508, t = 1.82, p = 0.065). As the participants’ perceptions of the interaction increased in negativity, so did their DBP. In contrast, among those low in hostility, no such effects were seen and increases in negative interaction intensity were not accompanied by increases in DBP (B = 0.62, SE = 0.68, df = 696, p > 0.36).

Subcomponent analyses revealed a highly significant interaction of cynicism and negative interaction intensity on DBP (B = 0.62, SE = 0.16, df = 1208, t = 3.90, p < 0.0001). When participants were divided into high- and low-cynicism groups at the median score of 5, there was a significant relationship of negative interaction intensity to DBP for those high in cynicism (B = 2.05, SE = 0.77, df = 453, t = 2.65, p < 0.01), but not for those low in cynicism (B = -0.12, SE = 0.54, df = 748, t = -22, p > 0.82). There was also a significant interaction between negative interaction intensity and aggressive responding (B = 0.42, SE = 0.21, df = 1208, t = 2.05, p < 0.05). There was a marginally significant relationship of negative interaction intensity to DBP for those high in aggressive responding (ie, with scores above the median) (B = 1.29, SE = 0.72, df = 5.33, t = 1.78, p < 0.08), whereas there was no significant relationship among those low in aggressive responding (B = 0.41, SE = 0.65, df = 688, t = 0.63, p > 0.52). Interactions of negative interaction intensity with the other three subcomponents were all nonsignificant (all p values > 0.60).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHOD RESULTS DISCUSSION NOTES REFERENCES

 
This study tested aspects of the transactional model of hostility and health using ambulatory monitoring techniques. Specifically, we tested hypotheses about the association of trait hostility to naturally occurring social interactions and their cardiovascular correlates. The findings provide support for the notion that trait hostility is associated both with the frequency and intensity of everyday negative social interactions and the magnitude of cardiovascular response to these interactions.

Hostility was negatively correlated with the frequency of all social interactions and with the frequency and intensity of positive interactions. Hostility was positively correlated with the frequency and intensity of negative interactions. These ecological momentary assessment data corroborate the findings of retrospective, self-report, and laboratory studies and provide support for the transactional model of hostility and health (3,9,10,13).

Subcomponent analyses yielded insight into the ways in which cognitive, affective, and behavioral aspects of hostility may influence social relationships. Hostile attributions, a cognitive component of hostility reflecting suspiciousness, was negatively correlated with both the overall rate of interaction and the rate of positive interactions and was positively correlated with the rate of negative interactions. Suspiciousness may trigger hypervigilance in interactions with others, inhibiting the willingness to initiate contact and fostering a tendency to perceive existing contacts as negative. Hostile affect was also positively correlated with the frequency and intensity of negative interactions. The tendency to experience negative emotions may sensitize the individual to future experiences of negative emotion. Cynicism was positively correlated with the intensity of negative interactions and negatively correlated with the intensity of positive encounters. Cynicism is a cognitive component that may rob individuals of the pleasure of social interactions by souring their interpretation of others’ motivation.

Consistent with our previous findings (8) and those of Lynch et al. (52–54), within-person analyses indicated that engaging in social interactions (regardless of the type or intensity) was associated with acute increases in BP. The degree to which the interaction was perceived as negative was positively associated with the magnitude of the increase in DBP and, to a lesser extent, the magnitude of increase in SBP. Positive interactions also were activating, although the effects were smaller and limited to DBP. These effects are likely to be partly a function of the cardiovascular demands of talking, with the intensity of the interaction potentially influencing respiration, rate of speech, muscles movements, and so on.

Hostility and hostility subcomponents were not directly related to average ABP, even during social interactions. However, hostility and, in particular, cynicism and aggressive responding were associated with the magnitude of increase in DBP when individuals were engaged in a negative interaction. Unlike less hostile individuals, those higher in hostility showed increases in DBP as the interaction became more intensely negative.

These findings are consistent with laboratory studies reporting that hostile individuals respond with greater cardiovascular activation to emotional arousal (23,55). It is worth noting that there were increases in BP associated with negative interactions, despite the fact that the average intensity of these interactions was low. Consequently, the patterns of cardiovascular activation associated with hostility-driven social interactions may occur sufficiently frequently to lead to sustained changes in cardiovascular functioning over time.

Hostility may exert an effect on cardiovascular functioning through at least two pathways. First, some aspects of hostility may be associated with a propensity for aggressive action (eg, those assessed by the subscale of aggressive responding). The expression of these components may trigger increases in sympathetic nervous system activation as the individual prepares for action. Second, hostility and, in particular cynicism, may diminish the individual’s ability to use positive experiences to buffer the negative effects of social or other stressors. This may be reflected in decreased cardiac vagal modulation (55). In future studies, it would be helpful to have more sensitive measures of both sympathetic and parasympathetic nervous system activation (eg, heart period variability) to clarify the roles of these two systems in determining BP responses. We also did not measure BP recovery, but the findings are consistent with suggestions that hostility may be associated with delayed cardiovascular recovery to negative events (23,56).

Hostile individuals may avoid interpersonal exchanges because they are more likely to perceive them as negative, less likely to perceive them as beneficial, and more likely to experience potentially aversive increases in cardiovascular activation if the interaction becomes negative. This may explain variations in the findings of a relationship of hostility to average ABP. If hostile individuals are able to avoid potentially disturbing social exchanges, then the number of opportunities in which BP increases in response to social challenge may be too limited to result in an effect on overall BP load or level.

The data reported in this article were collected as part of the baseline assessment for a project investigating the cardiovascular effects of exercise, which required participants to commit to an intensive course of supervised physical training and psychophysiological testing. Consequently, the participants had to be highly motivated individuals, with at least some control over their time, and therefore, potentially over their rate of social interaction. In future research, it may be necessary to specify aspects of both the situation (ie, risk for conflict) and the person (ie, level of hostility) to fully explicate the relationship of personality traits to ABP.

In addition, the effects of hostility on average levels of BP or HR may be moderated by age or family history of hypertension, and future studies would benefit from explicit examination of these issues. The participants were all relatively young (mean age = 30) and in good health, with relatively low levels of BP and with very few participants (N = 13) displaying even a mildly hypertensive state (ie, SBP 140 or DBP 90). Finally, ABP may be a relatively insensitive measure of changes in autonomic control. Ambulatory measures of heart period variability would improve the ability to detect more subtle effects of hostility on autonomic activation (55).

In sum, this study used ecological momentary assessments and ambulatory monitoring to test aspects of the transactional model. The data indicate that trait hostility is associated with social behavior and the cardiovascular correlates of this behavior, even at relatively low levels of the trait. The findings contribute to the growing literature on the psychophysiology of social relationships (57) and support the notion that the cardiovascular system is intimately involved in the regulation of social interaction.

	NOTES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHOD RESULTS DISCUSSION NOTES REFERENCES

 
¹Because there were gender differences in the rates of positive interactions, the analyses were repeated with gender serving as a control variable and the effects for hostile attributions remained significant (ß = -0.21, t = -2.05, p < 0.05).

²Because data for covariates were missing for a substantial number of observations and these observations were deleted, we repeated the analyses with no covariates. The results are essentially identical (ie, SBP: B = -2.02, standard error [SE] = 0.50, df = 103, t = -4.06, p < 0.001; DBP: B = -1.97, SE = 0.39, df = 103, t = -5.07, p < 0.001), with adjusted mean SBP scores obtained as individuals were interacting (mean SBP = 124.69 mm Hg, SE = 1.39, mean DBP = 74.43 mm Hg, SE = 0.74) approximately 2 mm Hg higher than those obtained when individuals were not interacting (SBP: 122.67 mm Hg, SE = 1.41; DBP: 72.45 mm Hg, SE = 0.75). There were no effects on HR.

³A proportion of the variance in BP associated with interacting is likely to be a function of the cardiovascular demands associated with talking. Participants reported talking on 92% of the observations in which they indicated they were interacting. Talking itself is associated with increased SBP (F[1, 96] = 9.65, p < 0.002; mean talking = 125.87 mm Hg, mean not talking = 123.59 mm Hg) and increased DBP (F[1, 96] = 5.27, p < 0.03; mean talking = 74.59 mm Hg, mean not talking = 73.29 mm Hg).

Received for publication May 25, 2002.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHOD RESULTS DISCUSSION NOTES REFERENCES

 

1. Miller TQ, Smith TW, Turner CW, et al. A meta-analytic review of research on hostility and physical health. Psych Bull 1996; 119: 322–348.[CrossRef][Medline]
2. Whiteman MC, Deary IJ, Fowkes FGR. Personality and health: Cardiovascular disease. In: Hampson SE, ed. Advances in personality psychology. Philadelphia, PA: Taylor & Francis; 2000: 157–198.
3. Smith TW. Hostility and health: Current status of a psychosomatic hypothesis. Health Psychol 1992; 11: 139–150.[CrossRef][Medline]
4. Cohen S, Kaplan J, Manuck S. Social and coronary heart disease: Underlying psychological and biological mechanisms. In: Shumaker S, Czaijkowski S, eds. Social support and cardiovascular disease. New York: Plenum Press; 1994.
5. Sloan RP, Shapiro PA, Bagiella E, et al. Cardiac autonomic control buffers blood pressure variability responses to challenge: A psychophysiologic model of coronary artery disease. Psychosom Med 1999; 61: 58–68.[Abstract/Free Full Text]
6. Williams RB, Barefoot JC, Shekelle RB. The health consequences of hostility. In: Chesney MA, Rosenman RH, editors. Anger and hostility in cardiovascular and behavioral disorders. Washington, DC: Hemisphere; 1985: 173–185.
7. Barefoot JC, Dahlstrom WG, Williams RB, et al. Incidence, and total mortality: A 25-year follow-up study of 255 physicians. Psychosom Med 1983; 45: 59–63.[Abstract/Free Full Text]
8. Brondolo E, Masheb R, Stores J, et al. Anger-related traits and response to interpersonal conflict among NYC Traffic Agents. J Appl Soc Psychol 1998; 28: 2089–2118.[CrossRef]
9. Hardy JD, Smith TW. Cynical hostility and vulnerability to disease: Social support, life stress, and physiological response to conflict. Health Psychol 1988; 7: 447–459.[CrossRef][Medline]
10. Houston BK, Kelly KE. Hostility in employed women: Relation to work and marital experience, social support, stress, and anger expression. Pers Soc Psychol Bull 1989; 15: 175–182.[Abstract]
11. Scherwitz L, Perkins L, Chesney M, et al. Cook-Medley Hostility Scale and subsets: Relationship to demographic and psychosocial characteristics in young adults in the CARDIA study. Psychosom Med 1991; 53: 36–49.[Abstract/Free Full Text]
12. Smith TW, Frohm KD. What’s so unhealthy about hostility? Construct validity and psychosocial correlates of the Cook and Medley Hostility Scale. Health Psychol 1985; 4: 503–520.[CrossRef][Medline]
13. Smith TW, Pope MK, Sanders JD, et al. Cynical hostility at home and work: Psychosocial vulnerability across domains. J Res Pers 1988; 22: 525–548.[CrossRef]
14. Suarez EC, Williams RB. Situational determinants of cardiovascular and emotional reactivity in high and low hostile men. Psychosom Med 1989; 51: 404–418.[Abstract/Free Full Text]
15. Suarez EC, Williams RB. The relationship between dimensions of hostility and cardiovascular reactivity as a function of task characteristics. Psychosom Med 1990; 52: 558–570.[Abstract/Free Full Text]
16. Christensen AJ, Smith TW. Cynical hostility, self-disclosure, and cardiovascular reactivity. Psychosom Med 1993; 55: 193–202.[Abstract/Free Full Text]
17. Powch IG, Houston BK. Hostility, anger, in and cardiovascular reactivity in white women. Health Psychol 1996; 15: 200–208.[CrossRef][Medline]
18. Smith TW, Allred KD. Blood pressure responses during social interaction in high and low cynically hostile males. J Behav Med 1989; 12: 135–143.[CrossRef][Medline]
19. Allred KD, Smith TW. Social cognition in cynical hostility. Cogn Ther Res 1991; 15 (5): 399–412.[CrossRef]
20. Diamond EL, Schneiderman N, Schwartz D, et al. Harassment, hostility, and Type A as determinants of cardiovascular reactivity during competition. J Behav Med 1984; 7 (2): 171–189.[CrossRef][Medline]
21. Glass DC, Lake CR, Contrada RJ, et al. Stability of individual differences in physiological responses to stress. Health Psychol 1983; 2 (4): 317–341.
22. Kamarck TW, Manuck SB, Jennings JR. Social support reduces cardiovascular reactivity to psychological challenge: A laboratory model. Psychosom Med 1990; 52: 42–58.[Abstract/Free Full Text]
23. Fredrickson BL, Maynard KE, Helms MJ, et al. Hostility predicts magnitude and duration of blood pressure response to anger. J Behav Med 2000; 23 (3): 229–243.[CrossRef][Medline]
24. Jamner LD, Shapiro D, Ka Kit H, et al. Hostility and differences between clinic, self-determined, and ambulatory blood pressure. Psychosom Med 1993; 55: 203–211.[Abstract/Free Full Text]
25. Jamner LD, Shapiro K, Goldstein IB, et al. Ambulatory blood pressure and heart rate in paramedics: Effects of cynical hostility and defensiveness. Psychosom Med 1991; 53: 393–406.[Abstract/Free Full Text]
26. Shapiro D, Jamner LD, Goldstein IB. Ambulatory stress psychophysiology: The study of "compensatory and defensive counterforces" and conflict in a natural setting. Psychosom Med 1993; 55: 309–323.[Abstract/Free Full Text]
27. Shapiro D, Jamner LD, Goldstein IB. Daily mood states and ambulatory blood pressure. Psychophysiology 1997; 34: 399–405.[Medline]
28. Raikkonen K, Matthews K, Flory J, et al. Effects of hostility on ambulatory blood pressure and mood during daily living in healthy adults. Health Psychol 1999; 18: 44–53.[CrossRef][Medline]
29. Ewart CK, Kolondner KB. Negative affect, and expressive style predict elevated ambulatory blood pressure in adolescents. J Pers Soc Psychol 1994; 66: 596–605.[CrossRef][Medline]
30. Guyll M, Contrada RJ. Trait hostility and ambulatory caridiovascular activity responses to social interaction. Health Psychol 1998; 17: 30–39.[CrossRef][Medline]
31. Linden W, Chambers L, Maurice J, et al. Sex differences in social support, self-deception, hostility, and ambulatory cardiovascular activity. Health Psychol 1993; 12: 376–380.[CrossRef][Medline]
32. Polk DE, Kamarck TW, Shiffman SM. Hostility explains some of the discrepancy between daytime ambulatory and clinic blood pressure. Health Psychol 2002; 21: 202–206.[CrossRef][Medline]
33. Shapiro D, Jamner LD, Goldstein IB. Effects of cynical hostility, anger out, anxiety, and defensiveness on ambulatory blood pressure in Black and White college students. Psychosom Med 1996; 58: 354–364.[Abstract/Free Full Text]
34. Benotsch EG, Christensen AJ, McKelvey L. Hostility, social support, and ambulatory cardiovascular activity. J Behav Med 1997; 20: 163–177.[CrossRef][Medline]
35. Ohira T, Tanigawa T, Iso H, et al. Effects of shift work on 24-hour ambulatory blood pressure and its variability among Japanese workers. Scand J Work Environ Health 2000; 26: 421–426.[Medline]
36. Helmers KF, Baker B, O’Kelly B, et al. Anger expression, gender, and ambulatory blood pressure in mild, unmedicated adults with hypertension. Ann Behav Med 2000; 22: 60–64.[Medline]
37. Brondolo E, Karlin W, Alexander K, et al. Workday communication and ambulatory blood pressure: Implications for the reactivity hypothesis. Psychophysiology 1999; 36: 86–94.[CrossRef][Medline]
38. Brondolo E. Anger-related traits and ambulatory blood pressure responses among NYC Traffic Agents. Chaired symposium at the Eighteenth Annual Meetings of the Society of Behavioral Medicine. April 1997, San Francisco, CA.
39. Contrada RJ, Jussim L. What does the Cook-Medley hostility scale measure? In search of an adequate measurement model. J Appl Soc Psychol 1992; 22: 615–627.[CrossRef]
40. Barefoot JC, Dodge KA, Peterson BL, et al. The Cook-Medley Hostility scale. Item content and ability to predict survival. Psychosom Med 1989; 51: 46–57.[Abstract/Free Full Text]
41. White WB, Schulman P, McGabe EJ, et al. Clinical validation of the Accutracker, a novel ambulatory blood pressure monitor using R-wave gating for Korsotkoff sounds. J Clin Hypertens 1987; 3: 500–509.[Medline]
42. Cook WW, Medley DM. Proposed hostility and Pharasaic-virtue scales for the MMPI. J Appl Psychol 1954; 38: 441–418.
43. Shekelle RB, Gale M, Ostfelf AM, et al. Hostility, risk of coronary heart disease and mortality. Psychosom Med 1983; 45: 109–114.[Abstract/Free Full Text]
44. Houston BK, Vavak CR. Hostility. Developmental factors, psychological correlates, and health behaviors. Health Psychol 1991; 10: 9–17.[CrossRef][Medline]
45. Littell RC, Milliken GA, Stroup WWW, et al. SAS system for mixed models. Cary, NC: SAS Institute Inc 1996;.
46. Bagiella E, Sloan R, Heitjan DM. Mixed-effects models in psychophysiology. Psychophysiology 2000; 37: 13–20.[CrossRef][Medline]
47. Brondolo E, Rosen RC, Kostis JB, et al. Relationship of physical symptoms and mood to perceived and actual blood pressure in hypertensive men: A repeated-measures design. Psychosom Med 1999; 61: 311–318.[Abstract/Free Full Text]
48. Schwartz J, Stone A. Strategies for analyzing ecological momentary assessment data. Health Psychol 1998; 17: 6–16.[CrossRef][Medline]
49. Wolfinger RD, Chang M. Comparing SAS GLM and MIXED procedures for repeated measures. SAS User’s Group International Conference Proceeding SUGI 1995;20:1172–1182.
50. Clark L, Denby L, Pregibon D. Data analysis of ABP readings: Before p values. In: Schneiderman N, Weiss S, Kaufman P, eds. Handbook of research methods in cardiovascular behavioral medicines. New York: Plenum Press; 1989: 333–345.
51. Schwartz GE, Warren K, Pickering TG. Mood, location, and physical position as predictors of ambulatory blood pressure and heart rate: Applications of a multi-level random effects model. Ann Behav Med 1994; 16: 210–220.
52. Lynch JJ, Thomas SA, Long JM, et al. Human speech and blood pressure. J Nerv Ment Dis 1980; 168: 526–534.[CrossRef][Medline]
53. Lynch JJ, Long JM, Thomas SA, et al. The effects of talking on blood pressure of hypertensive and normotensive individuals. Psychosom Med 1981; 43: 25–33.[Abstract/Free Full Text]
54. Lynch JJ. The language of the heart: The body’s response to human dialogue. New York: Basic Books; 1985.
55. Sloan RP, Bagiella E, Shapiro PA, et al. Hostility, gender, and cardiac autonomic control. Psychosom Med 2001; 63: 434–440.[Abstract/Free Full Text]
56. Linden W, Earle T, Gerin W, et al. Physiological stress reactivity and recovery: Conceptual siblings separated at birth? J Psychosom Res 1997; 42: 117–135.[CrossRef][Medline]
57. Lepore SJ. Problems and prospects for the social support-reactivity hypothesis. Ann Behav Med 1998; 20: 257–269.[Medline]

This article has been cited by other articles:

				E. Brondolo, D. J. Libby, E.-g. Denton, S. Thompson, D. L. Beatty, J. Schwartz, M. Sweeney, J. N. Tobin, A. Cassells, T. G. Pickering, et al. Racism and Ambulatory Blood Pressure in a Community Sample Psychosom Med, January 1, 2008; 70(1): 49 - 56. [Abstract] [Full Text] [PDF]

				J. W. Burns, S. Bruehl, and P. J. Quartana Anger management style and hostility among patients with chronic pain: effects on symptom-specific physiological reactivity during anger- and sadness- recall interviews. Psychosom Med, September 1, 2006; 68(5): 786 - 793. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Brondolo, E. Articles by Sloan, R. P. Search for Related Content PubMed PubMed Citation Articles by Brondolo, E. Articles by Sloan, R. P. Related Collections Blood Pressure Other Cardiovascular Medicine Personality