Negative Affect Hypothesis

Because broad measures of neuroticism have revealed no consistent relation with essential hypertension (Cochrane, 1973; Davis, 1970; Kohler et al., 1993; Sainsbury, 1964; Spiro et al., 1995), most investigators have explored the Negative Affect Hypothesis by administering trait measures of negative affect, most commonly anxiety, anger, and depression. Although measurement of the experience of negative affect can be conducted using behavioral coding of facial displays of emotion (Ekman and Friesen, 1978), most investigators of the relation between personality factors and hypertension have used self-report questionnaires. Behavioral coding of emotions is simply too labor-intensive to conduct during the large-scale epidemiologic investigations that are typically used to address these relations.

Trait Anxiety

In contrast to the acute alterations in anxiety that occur in response to exposure to a threatening stimulus, trait anxiety refers to the relatively enduring anxiousness that persons experience on a daily basis. High-trait-anxious individuals simply tend to be more nervous or 'high-strung' than low-trait-anxious individuals. Elevated trait anxiety is associated with an increased sensitivity to physiological cues of arousal as well as cognitive symptoms of chronic worry and the tendency to view the environment as threatening or dangerous. The hypothesis that high-trait-anxiety persons exhibit an increased risk for hypertension makes intuitive sense; if an individual goes through life with a characteristic hyper-vigilance for internal cues of anxiety and sources of threat from the environment, one could easily speculate that high blood pressure might be a potential medical consequence. Let's exam ine the literature that has tested the relation between trait anxiety and risk for essential hypertension.

Although increased levels of trait anxiety have been reported among hypertensive patients in comparison to normal blood pressure controls (Markovitz et al., 1991; Sullivan et al., 1981), others have reported no relation between trait anxiety and blood pressure status (Friedman et al., 2001; Shinn et al., 2001; Siegel and Leitch, 1981). Because definitive conclusions cannot be drawn from case-control research, the more convincing data associating trait anxiety with risk for hypertension come from epidemiological cohorts in which larger samples of individuals are followed for extended periods of time. For example, data from the Framingham Study, an epidemiologic study that has spanned almost half a century, revealed that there was more than a twofold increased incidence of hypertension associated with a measure of anxiety (or tension) among men but not among women (Markovitz et al., 1993). Even stronger associations between anxiety and hypertension were reported by Jonas, Franks, and Ingram (1997) in a national sample of both black and white men and women (nhanes I). Interestingly, in this study, significant risk for hypertension associated with trait anxiety was observed for black men and women of all ages, but only for older white men and women. Findings from a four-year prospective study relating psychosocial factors to onset of hypertension among white and black students at the University of Zimbabwe confirmed the association between trait anxiety and increased risk for hypertension among black students (Somova, Connolly, and Diara, 1995). However, no relation between trait anxiety and hypertension was observed among white students. Data from three other large-scale epidemiologic studies, the cardia Study (Yan et al., 2003), the Tecum-seh Study (Young et al., 1998), and a study of Australian government workers (Chapman et al., 1990), however, failed to confirm any increased risk for hypertension for persons with high trait anxiety.

Although findings regarding the relation between trait anxiety and hypertension have been mixed, there is some evidence that anxiety is related to autonomic nervous system functioning, which in turn may lead to elevated blood pressures. For example, Piccirillo et al. (1998) measured parameters of the autonomic nervous system and morphological characteristics of the heart among groups of high- and low-anxious hypertensive and normotensive adults. Congruent with their hypotheses, higher anxiety scores were associated with the increased sympathetic nervous system influence on heart rate as well as left ventricular hypertrophy. Comparable associations between anxiety and physiological parameters were observed for both hypertensive and normotensive participants, indicating that this relation was not unique to those diagnosed with hypertension. High trait anxiety has also been shown to be associated with diminished baroreceptor sensitivity, consequently resulting in increased SBP variability (Virtanen et al., 2003; Watkins et al., 1998), indicating heightened sympathetic nervous system activation and reduced parasympathetic tone. Although trait anxiety appears to be associated with alterations in autonomic nervous system activity, there is currently no evidence that hypertensive and normotensive patients differ regarding how trait anxiety is linked to autonomic nervous system activity.

A few studies have explored the relation between trait anxiety and cardiovascular response to stress. Like the results of research on state anxiety (see pp. 132 -135) and the epidemiologic research examining the role of trait anxiety as a psychological predictor of hypertension, these findings have been mixed. Although some researchers have found exaggerated cardiovascular reactions among individuals high in trait anxiety (Glass et al., 1983; Houston, 1977), others have found high trait anxiety to be associated with lower cardiovascular reactions to stress (Young et al., 1998). Anxiety has also been reported to interact with parental history of hypertension in affecting the magnitude of cardiovascular responding to mental stress (Manuck et al., 1985; Miller, 1992). In these studies, high-anxious young adults with a parental history of hypertension exhibited higher cardiovascular reactions to mental stress than either persons low in anxiety or those without a parental history of hypertension.

In summary, the evidence for a causal link between high trait anxiety and increased risk for developing essential hypertension is modest. Findings from both case-control and prospective studies examining the relation have been mixed. Likewise, anxiety does not appear to be directly linked to cardiovascular response to stress; rather it may exert its influence on the magnitude of cardiovascular reactions in conjunction with other individual difference variables, like parental history of hypertension.

Trait Anger

Another component of negative affect that has been examined in relation to risk for essential hypertension is the constellation of variables referred to as trait anger or hostility. To be quite frank, these terms are often used interchangeably, although there is good reason to make careful distinctions between them. Using a comprehensive strategy for distinguishing among anger-related variables, Spielberger et al. (1985) described three distinct anger-related elements: anger, aggression, and hostility. Anger, according to Spielberger, referred to an emotional experience that was comprised of physiological arousal and a whole range of cognitive appraisals associated with the experience of anger (perceptions of unfairness or experiencing goal blockage). As stated in Chapter 5, the experience of anger can be assessed using state measures regarding how the respondent feels "right now" as well as using trait measures which request information regarding how one "typically feels." Individuals with high trait anger scores are typically temperamental and prone to overt expressions of the intense anger that 'hovers below the surface.' Aggression, the second anger-related variable described by Spielberger, referred to a prominent type of behavioral display of anger that can occur physically or verbally. Anger suppression, or the failure to express anger overtly, represents another type of behavioral response to anger; because data linking anger expression/suppression to hypertension are more pertinent for the Suppressed Hostility Hypothesis, it will be considered in the next section. Finally, Spielberger and colleagues distinguished hostility from both the experience of anger and the observable expressions of anger (expression or suppression). Although hostility is often confused with aggressiveness, in the psychological literature hostility represents an attitude of general cynicism and mistrust. As such, hostile persons believe that others are self-serving and frequently take advantage of those around them. Naturally, this suspicious demeanor results in heightened vigilance during interpersonal interactions and a tendency to perceive angry intent among others when in fact none may be present (Larkin, Martin, and McLean, 2002). Consequently, hostile persons may be more likely to experience bouts of anger than low-hostile individuals and thus more likely to express anger aggressively.

As with research on trait anxiety, some case-control studies have found support for a link between trait anger and blood pressure status (Crane, 1982; Jern, 1982; Siegel and Leitch, 1981; Sullivan et al., 1981), while others have found no association between measures of trait anger and blood pressure (Durel et al., 1989; Friedman et al., 2001). In contrast to the copious literature relating anxiety and hypertension, only two prospective studies have examined the relation between measures of trait anger and onset of hypertension (Markovitz et al., 1991; 1993). In the study of middle-aged women reported by Markovitz et al. (1991), initial levels of trait anger did not predict onset of high blood pressure. However, change in trait anger over the study duration was associated with change in both SBP and DBP; therefore, increases in trait anger among these women were associated with increased blood pressures, even after controlling for traditional risk factors. In contrast, trait anger data from the Framingham Study were not associated with subsequent risk for essential hypertension (Mar-kovitz et al., 1993).

Suls, Wan, and Costa (1995) conducted a meta-analysis of all studies conducted from 1963 to 1993 to clarify the inconsistent findings observed between hypertension and measures of trait anger. Overall effect sizes were quite small, indicating that even among those anger-related variables that were correlated with blood pressure, the magnitude of the association was not remarkable. Parameters most closely associated with blood pressure included measures of anger expression, not measures of trait anger.

Congruent with findings relating trait anger to risk for hypertension, findings relating trait anger to cardiovascular reactivity to stress have also been mixed. While some studies have reported a relation between increased trait anger and exaggerated cardiovascular reactions to stress (Durel et al., 1989; Johnson, 1989a), others have failed to find such a relation (Laude et al., 1997). Therefore, as with trait anxiety, even if trait anger is associated with increased risk for essential hypertension, it is unclear whether it influences cardiovascular reactivity to stress.


Given the substantial body of evidence linking measures of hostility to coronary heart disease (see Smith, 1992), several researchers have examined the association between hostility and blood pressure. As with trait anxiety and anger, several studies have reported a significant association between hostility and blood pressure levels (Benotsch, Christensen, and McKelvey, 1997; Durel et al., 1989; Jamner et al., 1991; 1993; Raikkonen et al., 1999), but other studies have found no relation between measures of hostility and blood pressure (Friedman et al., 2001; Pasic et al., 1994). Guyll and Contrada (1998) reported a significant association between DBP and hostility among men, but not women. Unfortunately, very few case-control studies comparing hypertensive and normotensive patients on measures of hostility or prospective studies linking hostility to hypertension have been conducted; among those that have been conducted, no relation between hostility and essential hypertension has been observed (Steptoe, Melville, and Ross, 1982). Findings from two prospective studies examining the relation between hostility and onset of hypertension have likewise been mixed. Although Siegler et al. (1992) found no prospective relation between hostility and onset of hypertension, Yan et al. (2003) reported a significant relation between high hostility and increased risk for essential hypertension in the cardia Study.

Although very little work has been conducted examining the relation between hostility and hypertensive status, a considerable amount of research has investigated the relation between hostility and cardiovascular response to stress. Suls and Wan (1993) conducted a metaanalysis of all studies between 1965 and 1992 examining the relation between the magnitude of cardiovascular reactions to stress and various measures of hostility. Although no consistent relations were observed between measures of cardiovascular response to stress and the various indices of hostility, elevated DBP reactivity to interpersonal challenges was associated with high hostility as measured by the Cook-Medley Hostility Scale (Cook and Medley, 1954), but only under conditions of provocation or harassment. Subsequent studies of the relation between cardiovascular reactivity to stress and hostility have tended to use interpersonal tasks rather than the traditional laboratory mental stress tasks. Congruent with the findings of Suls and Wan, these studies have confirmed the association between greater blood pressure reactions and high hostility during interpersonal tasks involving confrontation, marital interaction, harassment, or disclosure of a personally troubling issue (Brondolo, Rieppi, Erickson, et al., 2003a; Christensen and Smith, 1993; Powch and Houston, 1996; Smith and Gallo, 1999; Suarez et al., 1993; Suarez, Kuhn, Schanberg, et al., 1998a).

In general, the linkage between blood pressure reactivity to provocation and hostility has been demonstrated more consistently among samples of men than women, suggesting that gender may play an important role in impacting this association. It is important to remember, however, that males may be more likely to be threatened by the sorts of interpersonal challenges used in this body of literature, due to their focus on achievement, confronting the source of harassment, or winning at all costs. In fact, high-hostile women have been shown to exhibit comparable elevated cardiovascular responses to stress to those seen among high-hostile men when a less competitive task is used (Davis, Matthews, and McGrath, 2000). Additionally, there is evidence that parental history of hypertension interacts with hostility to determine magnitude of cardiovascular response to stress (Miller et al., 1998). In this study, high-hostile and low-hostile young adult males with and without a parental history of hypertension engaged in either a harassment or non-harassment condition. High-hostile offspring of hypertensives exhibited significantly greater cardiac output and forearm blood flow responses than individuals in the other groups, but only during the harassment condition. In this regard, both dispositional and situational characteristics influenced whether exaggerated cardiovascular responses were observed.

The most consistent cardiovascular parameter associated with hostility during provocation stressors has been DBP, although increased forearm blood flow among high-hostile persons has also been reported (Miller et al., 1998; Suarez et al., 1998a). Differences in heart rate response to these stressors among high- and low-hostile persons are rarely, if ever, observed. Based upon this constellation of findings, both cardiac (lack of heart rate response differences) and vascular (differences in increased forearm blood flow response) involvement could be hypothesized to be associated with the commonly observed differ ence in DBP response to provocation between high- and low-hostile individuals. To examine the nature of the hemodynamic factors underlying this blood pressure response difference, a few studies employing impedance cardiography have revealed that blood pressure increases during provocation for high-hostile persons are associated with greater total peripheral resistance and lesser cardiac output than those for low-hostile persons (Bongard, al'Absi, and Lovallo, 1998; Davis et al., 2000). In an examination of nervous system influences upon this exaggerated peripheral resistance response, Suarez, Sherwood, and Hinderliter (1998b) contrasted alpha- and beta-adrenergic receptor responsiveness to pharmacological agonists of high- and low-hostile men. Although no differences in alpha-adrenergic receptor responsivity was observed between high- and low-hostile men, decreased beta-adrener-gic receptor responsiveness was observed among the high-hostile participants. These findings were recently replicated using a sample of middle-aged women (Sherwood et al., 2004). This pattern of results indicates that the vascular response observed among high-hostile individuals is more closely associated with a diminished vasodilatory response of the beta-adrenergic nervous system than an increased vasoconstrictive response of the alpha-adrenergic nervous system. Additionally, hostility has been shown to be associated with reduced barore-ceptor sensitivity (Virtanen et al., 2003) and lower parasympathetic regulation of the heart (Sloan et al., 2001), suggesting that the linkage between hostility and cardiovascular response to stress is not entirely mediated by peripheral vascular influences. It appears that both beta-adrenergic and parasympathetic nervous systems have a role in generating the increased blood pressure response to provocation observed among high-hostile participants in laboratory studies of the cardiovascular reactivity to stress-hostility relation.

Although a considerable amount of research has supported the hypothesis that hostility is related to an increased magnitude of cardiovascular reactivity to stress, in particular in response to situations involving confrontation, harassment, or self-disclosure, it is important to consider alternative hypotheses as well. For example, risk for essential hypertension associated with hostility has been hypothesized to be associated with differential lifestyle habits, including smoking, alcohol use, and eating and exercise habits (Leiker and Hailey, 1988). In their examination of the relation between hostility and lifestyle factors, Musante et al. (1992) found some support for this alternative hypothesis. Increased hostility was related to poorer health habits, including consumption of a less healthy diet (increased fat and sugar consumption accompanied by decreased fiber intake) and a greater tendency to smoke cigarettes. Data from the cardia Study support these conclusions as well, with hostility being associated with higher frequency of smoking cigarettes and marijuana as well as consumption of a higher number of calories (Scherwitz et al., 1992). Finally, prospective links between measures of hostility taken during the undergraduate years and health behaviors in adulthood have provided additional evidence for this hypothesis. Data from the UNC Alumni Heart Study showed that high hostility during young adulthood was associated with increased smoking and consumption of alcohol and a high-fat diet, obesity, and low social support at midlife (Siegler et al., 2003). Therefore, there is evidence that the association between hostility and risk for essential hypertension may not reflect a simple linear relation and that multiple pathways may be involved.


Depression represents another emotion that comprises negative affect. Unlike anxiety and anger, which are thought to be associated with increased physiological arousal or irritability, depression is commonly associated with under-arousal or the lack of activity. Depressed individuals often sleep more and engage in less activity than nondepressed counterparts, and this inactivity is often associated with increased complaints of fatigue and loss of energy. If we recall the defense and defeat reactions described by Henry and Stephens (1977) in Chapter 3, anger and anxiety appear to be classic defense reactions, while depression would be categorized as a defeat reaction. Based upon observations of psychiatric outpatients that revealed an increased incidence of essential hypertension associated with diagnoses of depression (Rab-kin, Charles, and Kass, 1983), speculations that perhaps a causal relation existed between depression and hypertension have been entertained (Dilsaver and Coffman, 1988). Accordingly, individuals who experience depressive episodes throughout life were hypothesized to be at increased risk for subsequent onset of essential hypertension. Let's examine the evidence supporting this hypothesis.

Given the strong association between depression and coronary heart disease (see Rozanski et al., 1999), it is certainly important to examine whether a comparable relation exists between depression and essential hypertension, a primary risk factor for coronary heart disease. The majority of case-control studies aimed at measuring symptoms of depression in hypertensive patients, however, have failed to find any relation between depression and hypertension (Edlavitch et al., 1987; Friedman and Bennet, 1977; Friedman et al., 2001; Jones-Webb et al., 1996; Kim et al., 2003). In contrast, Grewen et al. (2004) reported a relation between depression and level of blood pressure during ambulatory monitoring, but only among participants with a parental history of high blood pressure. If depression is linked with increased blood pressure or increased incidence of hypertension, it may exert this influence only among a certain subgroup of individuals.

In contrast to case-control studies, there is some prospective evidence linking indices of depression with onset of essential hypertension among blacks, but not whites (Davidson et al., 2000; Jonas et al., 1997). However, the evidence is not that consistent, as other prospective studies have failed to find any elevated risk for hypertension associated with depression at all (Levenstein, Smith, and Kaplan, 2001; Shinn et al., 2001; Yan et al., 2003). Although Raikkonen, Matthews, and Kuller (2001) did not find a relation between initial depression and subsequent risk for hypertension, they reported a significant association between increased depression and increased SBP among women participating in a nine-year prospective study. Although findings from the Kuopio Ischemic Heart Disease Risk Factor Study in Finland also failed to find a significant relation between depression and hypertension, a highly significant association was uncovered between hopelessness, an important component of depression, and risk for hypertension (Everson et al., 2000). In this study, high levels of hopelessness were associated with an increased relative risk (RR) of 3.22, indicating that persons high in hopelessness had over a three times greater risk of developing hypertension than low-hopelessness counterparts. Although the body of evidence supporting a link between depression and risk for hypertension is not large, there are some indications that depressed mood may indeed be an important psychosocial risk factor to consider in predicting onset of essential hypertension.

Despite the potential importance of depression for predicting essential hypertension, the literature examining the relation between depression and acute cardiovascular responses to stress is not substantial. Perhaps it may go without saying that because depression is typically thought to be associated with a dampened rather than an accelerated sympathetic nervous system response, fewer investigations of this hypothesis have been conducted. However, on the assumption that both depression and cardiovascular response to stress involve autonomic nervous system functioning, a few studies have examined this potentially important relation (Delehanty, Dimsdale, and Mills, 1991; Light, Kothandapani, and Allen, 1998). Although both studies found very limited support for the relation between depression and blood pressure response to stress, clear evidence was uncovered for the relation between depression and increased heart rate response to stress. The exaggerated heart rate response observed among persons with higher depression scores was accompanied by a shorter pre-ejection period (Light et al., 1998), indicating that the observed autonomic imbalance involved the sympathetic nervous system. In a recent pharmacologic investigation on depressed patients (Straneva-Meuse et al., 2004), SBP, cortisol, and epinephrine reactivity to mental stress was dampened following intervention with one of two common anti-depressant medications. Although data addressing the relation between depression and acute cardiovascular response to stress are limited, there is some indication that depression is associated with increased cardiovascular response to stress (particularly heart rate) and that this exaggerated reactivity can be attenuated with the successful treatment of depression.

Negative Affect and Mood

A few studies have examined the Negative Affect Hypothesis by lumping together measures of anxiety and depression into an amalgamated measure of negative affect or mood. For example, analysis of psychosocial risk factors of hypertension in the follow-up study associated with nhanes i was conducted by deriving a single negative affect factor comprised of symptoms of both anxiety and depression. As with nhanes i, the follow-up data showed a significant association between negative affect and subsequent risk for hypertension for both men and women (Jonas and Lando, 2000). The relative risk for hypertension was particularly elevated among black women (RR = 3.12). Ewart and Kolodner (1994) also employed a measure of negative affect based upon combined scores on standardized measures of depression and anxiety in a study examining ambulatory blood pressures of adolescents. Congruent with previous work on measures of negative affect, they found that negative affect was a significant predictor of ambulatory blood pressures, even after controlling for resting blood pressure, body size, and substance use. In an 18-month study of the relation between mood and blood pressure, Pollard and Schwartz (2003) reported an association between increases in negative affect and increases in SBP; however, Brondolo et al. (1999) reported no relation between negative affect and blood pressure change in a similar study. Like the findings relating both anxiety and depression to measures of blood pressure or hypertensive status, results of research examining the relation between negative affect and blood pressure status are mixed. However, this should not be that surprising, as measures of negative affect are generally comprised of combining scores from traditional measures of anxiety and depression. Furthermore, no empirical work has examined the relation between trait measures of negative affect and cardiovascular reactivity to stress.

Summary of Findings Supporting the Negative Affect Hypothesis The Negative Affect Hypothesis stipulates that persons who experience a range of negative emotions more frequently or more intensely will exhibit a greater risk for developing essential hypertension than persons who experience such emotions less often or with less intensity. Studies designed to examine this hypothesis have approached it by defining negative affect broadly to include a number of different types of negative emotions or more narrowly to focus on an individual source of negative affect like anxiety, anger, hostility, or depression. Although findings across these studies have been mixed, there is enough prospective evidence available to suggest that at least for some individuals the frequent, intense exposure to negative emotions may play a role in the etiology of essential hypertension. In particular, prospective evidence linking both anxiety and depression (as well as their combination, negative affect) with subsequent risk for developing essential hypertension suggests that these variables may be important individual difference factors to consider. However, it is important to recognize that the specific type of negative affect associated with increased risk for hypertension varies among these prospective studies. For example, Markovitz et al. (1993) found prospective evidence linking anxiety to onset of hypertension, but not anger. Yan et al. (2003) found no linkage for either anxiety or depression with onset of hypertension, but found a relation between hostility and subsequent hypertension. Therefore, although there is some support for the Negative Affect Hypothesis within each prospective study, the parameter related to onset of hypertension differs from study to study. Finally, because evidence linking both anxiety and depression to acute cardiovascular response to stress is generally limited, it is possible that their influence may exert its effect through an alternate pathway.

Prospective evidence linking trait anger and hostility to hypertension is not as convincing as the data on anxiety and depression. However, evidence linking hostility to the magnitude of cardiovascular response to stress is quite strong. Therefore, it appears that if anger or hostility is related to risk for hypertension, it is likely that their influence occurs as a result of the elevated cardiovascular reactions to stress associated with these variables.

Kicking the Habit

Kicking the Habit

Stop Thinking About How You're Going To Quit Smoking When You Can Instantly Stomp It In Less Than 30 Days With A Proven Set of Techniques. When your

Get My Free Ebook

Post a comment