Multiple sclerosis is associated with high trait anger: A case-control study

Abstract

Introduction

In recent years there has been a focus on health-related quality of life in multiple sclerosis (MS) and in particular the importance of non-motor problems among which fatigue, pain, depression, anxiety, and cognitive disorders.. However, little attention has been focused on other negative emotions, such as anger. Our purpose was to evaluate whether trait anger (a predisposition to experience frequent and intense episodes of anger over time) is different between persons with and without MS after controlling for depression, anxiety, and other socio-demographic variables.

Methods

157 consecutive MS patients were enrolled in the study and compared to eighty age, gender, and education-matched healthy controls. Participants were administered affective trait measures (Beck Depression Inventory, Beck Anxiety Inventory) and the trait anger measure (the Spanish adapted version of the State-Trait Anger Expression Inventory-2 [STAXI-2]).

Results

MS patients had significantly higher scores on anger intensity (state anger) and trait anger than did controls. They also had a trend to experience direct anger toward other persons or objects in the environment (higher anger expression-out score) and to hold in or suppress angry feelings (higher anger expression-in score). However, in a regression analysis that adjusted for different demographic and clinical variables, we found that diagnosis category (MS patient vs. control) was associated with none of the highest quartiles of STAXI-2 scores, except for the Trait Anger scale (odds ratios between 2.35 and 3.50).

Conclusions

The present study provides further evidence that MS is independently associated with high trait anger.

INTRODUCTION

The focus on health-related quality of life of MS in recent years has demonstrated the importance of non-motor symptoms. Among these, fatigue, pain, depression, anxiety, and cognitive disorders are among the most important.[1–4] However, although anger is one of the basic emotions, and deleterious consequences of anger on physical health have been well reported,[5] only two studies have specifically been performed on the association between anger and MS.[6, 7] In all these previous studies,[6, 7] unmeasured confounders, including demographic variables such as marital or employment status, and the effects of medications with central nervous system activity (e.g., anxiolytics, stimulants, antipsychotics, antidepressants, antihistamines, or antiepileptic drugs),[8] may have in uenced the results.

Anger is the emotion associated with attack/threat often associated with irritability and occasionally aggression. It can be measured by the degree to which people have this mood and the characteristic means by which they express it.[5, 9] The prolonged predisposition for frequent, often intense, long-lasting anger is a relatively enduring and stable personality attribute known as trait anger.[5, 9] Anger expression, on the other hand, refers to how anger is managed, that is, whether it is expressed outwardly, held in, or controlled.[5, 9]

Both psychosocial and biological factors contribute to anger. Yet the biological basis of anger is not fully known. Serotonin, and perhaps serotonergic dysregulation, may be involved in the modulation of anger and aggressive behavior.[10–12] In this sense, biochemical studies have shown that MS patients are serotonergically depleted with the extent of cerebral depletion correlating with the degree of motor disability and a chronic progressive course.[13, 14] In some studies drugs such as selective serotonin reuptake inhibitors (SSRIs) have been associated with changes in anger.[15] Given these observations, we hypothesized that modulation of anger might be affected in MS, and more specifically, that MS is associated with higher levels of trait anger. The little attention that has been focused on trait anger in MS encouraged us to conduct a case–control study to assess this relationship, using the Spanish adapted version of the State-Trait Anger Expression Inventory-2 (STAXI-2).[9, 16, 17] Our analyses adjusted for several confounders, including socio-demographic variables and medications with central nervous system activity.[8]

METHODS

Participants

One hundred and fifty seven MS patients were consecutively recruited from January 2011 to June 2013 either from the MS Clinics at the University Hospital “12 de Octubre” (Madrid, Spain) or from the MS association of Madrid, Spain. Patients studied had clinically definite MS by McDonald criteria,[18, 19] were stable (free from any exacerbation) at the time of study, and fluent Spanish speakers. A 20 minute interview was conducted on each MS patient in which information on age of MS onset and disease evolution was also obtained. Three neurologists with expertise in MS (JBL, ALF, and SMG) participated in the clinical assessment and applied the Kurtzke Expanded Disability Status Scale (EDSS) scale to rate the severity of disease (range = 0–10).[20] For clinical course, two subgroups were defined: relapsing/remitting (RR) and progressive (secondary progressive [SP] or primary progressive [PP]). The Fatigue Impact Scale for Daily Use (D-FIS) was administered to measure subjective daily experience of fatigue in MS patients.[21, 22]

Patients taking high dose of corticosteroids at the time of the recruitment were temporarily excluded, and they underwent the emotional and neuropsychological battery at least one month after the interruption of the drug treatment. Other exclusion criteria were as follows: institutionalized at the time of observation; major acute co-morbidities or any major serious chronic illness three months before inclusion (patients with a stable chronic medical conditions were included); and neurological illness other than MS, including dementia.

Eighty age, gender, and education-matched healthy controls were recruited either from relatives of patients who came to the neurological clinics for reasons other than MS (e.g., headache, dizziness) or among the relatives or friends of the health professionals working at the University Hospital “12 de Octubre” of Madrid (Spain). None of the controls recruited were consanguineous of the patients with MS involved in the present study. Each control was free of known neurological. None of the control subjects had symptoms or a history of a neurological disorder (e.g., demyelinating disorders, cognitive impairment) and none had received a neurological diagnosis from a physician.

Procedure

All procedures were approved by the ethical standards committees on human experimentation at the University Hospitals “12 de Octubre” (Madrid). During recruitment, cases and controls were told that the purpose of the study was to complete a battery to assess both emotional and neuropsychological status. After the study had been described to subjects, written (signed) informed consent was obtained from all enrollees.

Measurements

Symptoms of depression were assessed by the Beck Depression Inventory (BDI).[23] The BDI is a widely used self-report measure of depressive symptoms that has been validated for MS and is recommended in this population of patients.[24] The range of responses to each depression symptom is 0–3 points, and a higher score means more severe depression (range from 0 to 63).[23] Symptoms of anxiety were assessed by the Beck Anxiety Inventory (BAI). The BAI is a self-report questionnaire to assess the severity of symptoms of anxiety.[25] Respondents are asked to rate how much each of the anxiety symptoms of the questionnaire bothered them, on a scale ranging from 0 (not at all) to 3 (severely, I could barely stand it). The total score has a minimum of 0 and a maximum of 63 and higher scores indicate higher levels of anxiety.

Anger was assessed by the State-Trait Anger Expression Inventory-2 (STAXI-2).[9, 17] Participants rate themselves on 4-point scales for each item. In each case, higher scores indicate a greater level of anger and its suppression or expression.[9, 17] The Spanish adaptation of the STAXI-2 was used: it includes 49 items (range 0 to 196), and it is organized into six scales (including five subscales) and an Anger Expression Index that provides an overall measure of total anger expression. [9, 17] The Trait Anger contains two subscales, T-Anger/T, which measures the general disposition toward angry feelings (angry temperament), and T-Anger/R, which measures the tendency to express anger when one is criticized (reaction to criticism).[9, 17] State Anger (including the three subscales: feeling angry; feel like expressing anger verbally; and feel like expressing anger physically) assesses currently anger as an emotional state.[9, 17] Additional scales include Anger Expression-Out, which measures expression of anger toward other persons or objects in the environment; Anger Expression-In, which assesses the holding in or suppression of angry feelings; Anger Control-Out, which assesses the control of angry feelings by preventing the expression of anger toward other persons or objects in the environment; and Anger Control-In, which assesses the control of angry feelings by calming down or cooling off.[9, 17] The final scale, the Anger Expression Index, is a general index of the expression of anger. This index comes from the following formula: (Anger Expression-out + Anger Expression-in)–(Anger Control-out + Anger Control-in) + 36.[9, 17]

A five-test neuropsychological test battery was administered to all subjects. Cognitive tests were conducted in a single session by experienced clinical neuropsychologists (VM, VP, AB, JFP, MIH, NC and UE, see acknowledgments) during an interview on the week in what they had completed the aforementioned affective trait measures. All the neuropsychologists were blinded to STAXI-2, BDI and BAI results. The following neuropsychological tests were administered to all participants:

1. Paced Auditory Serial Addition Test (PASAT). The PASAT measures working memory and speed of information processing. Both the 3-second and 2-second versions of the PASAT were administered. [26] The score was the total number of correct responses.[26]

2. The Symbol Digit Modalities Test (SDMT). The SDMT is a test of sustained attention that measures information processing speed.[27] This test involves the substitution of digits for symbols as quickly as possible within a 90 second time frame. The score was the total number of correct items.[27]

3. The Stroop Color–Word Trial. The Stroop Color–Word Trial is a test of executive functioning that requires participants to inhibit a natural response (reading a word) and replaces it with another response (saying a color).[28] Participants completed 45-s word naming, color naming, and color–word naming trials of a computer-based Stroop task. The score for this study was the number of correct responses in the color–word trial.[28]

4. The Controlled Oral Word Association Test (COWAT) is a test that measures phonetic uency.[29] Participants were provided three letters of the alphabet (F, A, and S), one letter at a time, and instructed to say as many words as possible that begin with this letter in a 60-second interval. [25] All responses were recorded verbatim.

5. In addition, subjects were asked to generate as many different animals as possible in one minute to test semantic verbal fluency (lower scores indicate greater cognitive impairment).[30]

Data analysis and Sample Size

Statistical analyses were performed in SPSS Version 20.0 (SPSS, Inc., Chicago, IL). All tests were two sided, and significance was accepted at the 5% level (alpha = 0.05). The Chi-square (χ²) test was used to analyze categorical variables. Using a Kolmogorov–Smirnov test, we determined that comorbidity (number of diseases), BDI and BAI scores, as well as STAXI-2 scores were not normally distributed (Kolmogorov–Smirnov tests for all items, p < 0.05). Therefore, although mean and median values were reported, case-control differences were compared using a nonparametric (Mann–Whitney U) test. Correlation analysis was performed by Spearman rank correlation.

Linear regression analyses were not possible because the STAXI-2 scales and subscales scores were not normally distributed. Therefore, to assess the effects of possible confounders (socio-demographic variables, anxiety and depressive symptoms, cognitive tests, and medications that potentially affect cognitive function), we divided the STAXI-2 scales and subscales scores into two strata (highest quartile scores vs. all other scores). Logistic regression analyses were then performed, thereby allowing us to assess the possible confounding effects of socio-demographic variables, anxiety and depressive symptoms, cognitive tests, and medications that potentially affect cognitive function. In these models, the dependent variable was the highest quartile of each one of the STAXI-2 scales and subscales scores (reference = all other scores) and the independent variable was patient vs. control (reference). We began with an unadjusted model. Then, in adjusted models, we first considered all variables in univariate analyses that were associated with both MS and the highest quartile of trait anger [high levels of trait anger] (reference = all other scores) (“Model 1” [more restrictive criteria for confounding]). Subsequently, we considered all variables in univariate models that were associated with either MS or the highest quartile of trait anger [high levels of trait anger] (reference = all other scores) (“Model 2” [less restrictive criteria for confounding]). These analyses generated odds ratios (OR) with 95% con dence intervals (CI).

The targeted sample size (80 participants in each group) had 90% power to detect as little as a 15% difference in Anger Expression Index between patients and controls (assuming alpha = 0.05).

RESULTS

The study sample consisted of 157 MS patients and 80 controls who were matched for age (Student’s t-test = −1.638, p=0.09), gender (χ2=0.237, p=0.626), and education (χ2=2.884, p=0.09) (Table 1). The MS patients were more likely to be married, unemployed, and being treated with medications with central nervous system activity (e.g., anxiolytics, stimulants, antipsychotics, antidepressants, antihistamines, or antiepileptic drugs) (Table 1). In addition, they were more likely to have higher levels of depressive and anxiety symptoms, and, in general, a worse cognitive functioning (Table 1). The median EDSS disability score was 3.0. MS duration (years) was 12.1 ± 7.4. 102 patients (65.0%) were on disease-modifying treatments for MS. Of these, 16 (15.7.%) were on interferon β 1-b subcutaneous, 14 (13.7%) on interferon β 1-a intramuscular, 31 (30.4%) on interferon β 1-a subcutaneous, four (3.9%) on azathioprine, one (1.0%) on a combination of interferon β and azathioprine, 18 (17.6%) on glatiramer acetate, two (2.0%) on intravenous immunoglobulin, 14 (13.7%) on natalizumab, and 2 (2.0%) on fingolimod. Seventy-one (45.2%) of MS patients showed cognitive deficits defined as 1.5 standard deviations (SD) below the control mean on some cognitive tests.

Table 1.

Comparison of demographic and clinical characteristics of multiple sclerosis patients vs. controls

	Multiple sclerosis patients (N = 157)	Controls (N = 80)	p Value
Age in years	41.7 ± 9.2 (41)	39.4 ± 10.6 (37.5)	0.09 *
Gender (female)	105 (66.9%)	56 (70.0%)	0.626
Education &			0.09
Primary studies	36 (23.1%)	11 (13.8%)
≥Secondary studies	120 (76.9%)	69 (86.2%)
Marital status &			0.041
Single	37 (23.7%)	30 (38.0%)
Married or cohabitant	99 (63.5%)	36 (45.6%)
Widowed	2 (1.3%)	3 (3.8%)
Separated or divorced	18 (11.5%)	10 (12.7%)
Occupational status &			<0.001
Employed	56 (35.9%)	56 (71.8%)
Unemployed	100 (64.1%)	22 (28.2%)
On medications with central nervous system effect	65 (41.4%)	9 (11.2%)	<0.001
Comorbidity (number of diseases)	0.5 ± 0.8 (0)	0.4 ± 0.8 (0)	0.183 **
Beck Depression Inventory total score	12.5 ± 8.8 (10)	6.3 ± 5.6 (5)	<0.001 **
Beck Anxiety Inventory total score	17.0 ± 10.9 (15)	9.2 ± 10.2 (6)	<0.001 **
Paced Auditory Serial Addition Test (2-second version) (total number of correct responses	27.5 ± 12.6 (28)	31.8 ± 10.2 (31)	0.008 *
Paced Auditory Serial Addition Test (3-second version) (total number of correct responses)	37.5 ± 14.9 (28)	40.2 ± 10.7 (31)	0.510 **
Symbol Digit Modalities Test (total number of correct items)	44.9 ± 15.4 (48)	55.1 ± 13.9 (55)	<0.001 *
Stroop Color–Word Trial (total number of correct responses in the color–word trial)	37.4 ± 13.7 (38)	46.9 ± 12.1 (48)	<0.001 *
Controlled Oral Word Association Test (total number of correct items)	33.9 ± 10.7 (33)	39.9 ± 11.9 (40.5)	<0.001 *
Total number of animals as possible in one minute	20.5 ± 6.4 (20)	23.5 ± 5.9 (24)	0.001 *
Disease duration (years)	12.1 ± 7.4 (11.4)	-	-
Type of multiple sclerosis
Relapsing-remitting	105 (66.9%)	-	-
Progressive forms	52 (33.1%)
EDSS score	3.5 ± 2.5 (3)	-	-
Fatigue Impact Scale for Daily Use total score	14.1 ± 8.6 (15)	-	-

Mean ± SD (median) and frequency (%) are reported.

^*Student’s t tests or

^**Mann-Whitney U test were used for comparisons of continuous data, and X2 test or Fisher’s exact test for proportions for proportions. & In these cells, there are missing data so that the total number of participants does not add up for cases or for controls.

Among controls, STAXI-2 scores correlated with the BDI score (with only three exceptions, all Spearman’s correlation coefficients > 0.22 and p values < 0.05). Among MS patients, these correlations were more robust (all coefficients had p values < 0.05).

Of the 237 participants, 60 scored in the higher quartile of trait anger, and 177 in the remaining quartiles (Table 2). Participants who scored in the highest quartile of trait anger were slightly younger, and treated with more medications with central nervous system activity than those who scored in the remaining quartiles (Table 2). In addition, they had higher levels of depressive and anxious symptoms, and were more likely to have lower verbal fluency (Table 2).

Table 2.

Demographic and clinical characteristics of cohort stratified by trait anger quartiles

	Highest quartile of trait anger (N = 60)	All other scores (N = 177)	p Value
Age in years	38.3 ± 7.9 (38.5)	41.8 ± 10.2 (42)	0.006 *
Gender (female)	44 (73.3%)	117 (66.1%)	0.300
Education &			0.129
Primary studies	16 (26.7%)	31 (17.6%)
≥ Secondary studies	44 (73.3%)	145 (82.4%)
Marital status &			0.581
Single	16 (27.1%)	51 (29.0%)
Married or cohabitant	35 (59.3%)	100 (56.8%)
Widowed	0 (0.0%)	5 (2.8%)
Separated or divorced	8 (13.6%)	20 (11.4%)
Occupational status &			0.709
Employed	27 (45.8%)	85 (48.6%)
Unemployed	32 (54.2%)	90 (51.4%)
On medications with central nervous system effect	25 (41.7%)	49 (27.7%)	0.043
Comorbidity (number of diseases)	0.5 ± 0.7 (0)	0.4 ± 0.8 (0)	0.318 **
Beck Depression Inventory total score	14.9 ± 10.2 (12)	8.9 ± 7.1 (7)	<0.001**
Beck Anxiety Inventory total score	19.0 ± 12.0 (17.5)	12.8 ± 10.6 (11)	<0.001**
Paced Auditory Serial Addition Test (2-second version) (total number of correct responses	27.0 ± 12.0 (26)	29.6 ± 12.0 (30)	0.141 *
Paced Auditory Serial Addition Test (3-second version) (total number of correct responses)	36.8 ± 13.5 (37)	39.0 ± 13.7 (42)	0.156 **
Symbol Digit Modalities Test (total number of correct items)	45.8 ± 16.4 (47)	49.2 ± 15.3 (50)	0.156 *
Stroop Color–Word Trial (total number of correct responses in the color–word trial)	39.1 ± 13.2 (41)	41.1 ± 14.2 (43)	0.328 *
Controlled Oral Word Association Test (total number of correct items)	33.7 ± 11.6 (35)	36.7 ± 11.4 (37)	0.09 *
Total number of animals as possible in one minute	20.1 ± 6.3 (20)	22.0 ± 6.4 (22)	0.045 *

Mean ± SD (median) and frequency (%) are reported.

^*Student’s t tests or

^**Mann-Whitney U test were used for comparisons of continuous data, and X2 test or Fisher’s exact test for proportions for proportions. & In these cells, there are missing data so that the total number of participants does not add up for cases or for controls.

Table 3 presents the mean and median STAXI-2 scores. Overall, MS patients had significantly higher scores on intensity (State anger) anger and trait anger than did controls, as well as a trend to direct anger toward other persons or objects in the environment (higher Anger Expression-out score) and to hold in or suppress angry feelings (higher Anger Expression-in score). MS patients also had higher Anger Expression Index scores than did controls. However, in a regression analysis that adjusted for different demographic and clinical variables, we found that diagnosis category (MS patient vs. control) was associated with none of the highest quartiles of STAXI-2 scores, except the Trait Anger scale, for which the association was robust (ORs between 2.35 and 3.50) (Table 4). In a secondary analysis in which we excluded 71 MS patients with cognitive deficits, the differences in Trait Anger between MS patients and controls were similar (adjusted OR _{Model 1} = 2.44, 95%, CI= 0.99–5.99, p = 0.051; adjusted OR _{Model 2} = 2.85, 95%, CI= 1.09–7.48, p = 0.033). In another secondary analysis, in which we divided the patients with MS into two groups (those ones taking SSRIs [N = 34] vs. those ones who were not taking them [N = 123]), the differences in Trait Anger between these two types of patients and controls were statistically significant (adjusted OR _{Model 1} = 3.90 [those ones taking SSRIs], 95%, CI= 1.0–14.07, p = 0.038; adjusted OR _{Model 1} = 2.26 [those ones who were not taking them], 95%, CI= 0.97–5.25, p = 0.058; adjusted OR _{Model 2} = 6.43 [those ones taking SSRIs], 95%, CI= 1.57–26.40, p = 0.010), and adjusted OR _{Model 2} = 3.32 [those ones who were not taking them], 95%, CI= 1.30–8.49, p = 0.012).

Table 3.

STAXI-2 scores in multiple sclerosis patients vs. controls

Scales	Multiple sclerosis patients (N = 157)	Controls (N = 80)	p Value *
State Anger	21.3 ± 8.0 (18)	18.3 ± 4.5 (16)	0.008
Feeling Angry	8.2 ± 3.9 (6)	6.3 ± 2.1 (5)	0.001
Feel Like Expressing Anger Verbally	7.3 ± 3.1 (6)	6.4 ± 2.0 (5)	0.116
Feel Like Expressing Anger Physically	5.8 ± 2.1 (5)	5.5 ± 1.5 (5)	0.595
Trait Anger	21.1 ± 7.4 (19)	18.0 ± 5.2 (17)	0.006
Angry Temperament	9.3 ± 4.4 (8)	7.2 ± 2.8 (6)	0.001
Angry Reaction	11.8 ± 3.8 (12)	10.8 ± 3.4 (11)	0.066
Anger Expression-out	11.2 ± 3.7 (11)	9.7 ± 2.9 (9)	0.002
Anger Expression-in	12.7 ± 3.9 (12)	11.2 ± 3.2 (11)	0.002
Anger Control-out	16.2 ± 4.7 (16)	16.9 ± 4.9 (17.5)	0.207
Anger Control-in	15.1 ± 4.7 (16)	15.1 ± 4.6 (16)	0.976
Anger Expression Index	28.7 ± 11.3 (28)	24.8 ± 9.6 (23)	0.009

Mean ± SD (median).

^*Mann-Whitney U test.

Table 4.

Results of logistic regression with the highest quartile of each one of the STAXI-2 scales (reference = all other scores) and the independent variable being patient vs. control (reference)

Highest quartile of State Anger
	Unadjusted		Model 1		Model 2
	Odds ratio	95% CI	Odds ratio	95% CI	Odds ratio	95% CI
MS patients (N = 157)	2.20*	1.11–4.37	1.11	0.50–2.49	1.12	0.46–2.70
Highest quartile of Trait Anger
	Unadjusted		Model 1		Model 2
	Odds ratio	95% CI	Odds ratio	95% CI	Odds ratio	95% CI
MS patients (N = 157)	3.80**	1.76–8.19	2.35*	1.02–5.42	3.50**	1.36–8.81
Highest quartile of Anger Expression-out
	Unadjusted		Model 1		Model 2
	Odds ratio	95% CI	Odds ratio	95% CI	Odds ratio	95% CI
MS patients (N = 157)	1.74	0.90–3.37	1.07	0.50–2.27	1.23	0.54–2.81
Highest quartile of Anger Expression-in
	Unadjusted		Model 1		Model 2
	Odds ratio	95% CI	Odds ratio	95% CI	Odds ratio	95% CI
MS patients (N = 157)	2.44*	1.18–5.04	1.50	0.65–3.49	1.76	0.69–4.50
Highest quartile of Anger Control-out
	Unadjusted		Model 1		Model 2
	Odds ratio	95% CI	Odds ratio	95% CI	95% CI	Odds ratio
MS patients (N = 157)	0.68	0.37–1.27	1.15	0.57–2.32	1.09	0.50–2.36
Highest quartile of Anger Control-in
	Unadjusted		Model 1		Model 2
	Odds ratio	95% CI	Odds ratio	95% CI	95% CI	Odds ratio
MS patients (N = 157)	1.31	0.69–2.49	1.50	0.73–3.07	1.57	0.73–3.39
Highest quartile of Anger Expression Index
	Unadjusted		Model 1		Model 2
	Odds ratio	95% CI	Odds ratio	95% CI	95% CI	Odds ratio
MS patients (N = 157)	3.48**	1.32–5.33	1.34	0.61–2.98	2.0	0.79–5.08

Model 1: adjusted for Medications with central nervous system effect use, Beck Depression Inventory total score, Beck Anxiety Inventory total score, and total number of animals as possible in one minute.

Model 2: adjusted for age in years, marital status, occupational status, medications with central nervous system effect use, Beck Depression Inventory total score, Beck Anxiety Inventory total score, Paced Auditory Serial Addition Test (2-second version), Symbol Digit Modalities Test, Stroop Color–Word Trial, Controlled Oral Word Association Test, and total number of animals as possible in one minute.

^*p < 0.05;

^**p < 0.01.

In MS patients, the STAXI-2 scale scores did not correlate with EDSS score or disease duration (Table 5). However, there were significant correlations between the majority of the STAXI-2 scores and fatigue on the D-FIS (Table 5).

Table 5.

Matrix of correlations^* among the STAXI-2 scales and subscales, and clinical variables of multiple sclerosis patients (N = 157)

Scales	EDSS score	Disease duration	D-FIS
State Anger	0.032	−0.104	0.279 c
Feeling Angry	0.010	−0.114	0.312 c
Feel Like Expressing Anger Verbally	0.025	−0.107	0.206 a
Feel Like Expressing Anger Physically	0.057	−0.028	0.244 b
Trait Anger	−0.013	−0.046	0.263 b
Angry Temperament	0.021	0.015	0.247 b
Angry Reaction	−0.039	−0.091	0.248 b
Anger Expression-out	0.006	0.032	0.117
Anger Expression-in	0.081	−0.097	0.225 b
Anger Control-out	−0.044	−0.029	−0.222 b
Anger Control-in	0.023	0.015	−0.087
Anger Expression Index	0.034	−0.019	0.235 b

^*Spearman rank correlation coefficients.

^ap < 0.05,

^bp < 0.01,

^cp < 0.001.

EDSS = Expanded Disability Status Scale.

D-FIS = Fatigue Impact Scale for Daily Use - Spanish Version.

DISCUSSION

The present study provides further evidence that MS is strongly associated with high trait anger. After adjustment for possible confounders, MS was significantly associated with between a 2.3 to 3.5-fold increases in the odds of high levels of trait anger. Another finding of the current study was that STAXI-2 scores did not correlate with the EDSS stage or with disease duration, suggesting that either severity is not a good marker of anger or that this “personality profile” could be a manifestation of initial brain changes from early MS. Prospective studies of at-risk individuals could further address whether STAXI-2 scores differ prior to the onset of disease. Also, family studies could be used to look for endophenotypic differences in this trait.

Although studies about anger in MS are limited, two investigations have produced comparable results with the present study.[6, 7] Nocentini et al[7] studied the anger profile of 195 cognitively unimpaired MS patients using the STAXI and found that levels of withheld and controlled anger were respectively higher and lower than what is expected in the normal Italian population. In addition, mean anger severity scores were not higher in more severe patients.[7] In another study, involving 38 patients with clinically definite MS (median EDSS = 7.5), admitted to a neurorehabilitation unit, Langdon and Thompson[6] found lower than expected State and Trait Anger and higher than expected In and Out Anger.

The biological bases of anger are still unknown. The neuroanatomic substrates of anger and aggressive behaviour have been studied by positron emission tomography, transcranial magnetic stimulation and functional magnetic resonance imaging.[31–33] The cortical regions involved in the control and expression of anger are medial frontal cortex, left inferior frontal and left temporal pole regions, right posterior temporal/parietal and right superior frontal cortex, and subcortical structures like basal ganglia and amygdala.[31–33] Serotonin is involved in the modulation of anger and aggressive behaviour.[10] Some studies suggest that depressed patients with anger attacks have a greater central serotonergic dysregulation than depressed patients without such attacks.[11] In line with this, a polymorphism of the gene coding for tryptophan hydroxylase, the rate limiting enzyme in serotonin biosynthesis, was found to be associated with anger-related traits of personality.[12] MS patients have an impairment of serotonin metabolism.[13, 14] We hypothesize that serotonergic dysregulation may play a role in the abnormalities in trait anger in MS patients. However, the mechanism by which MS is associated with trait anger is unknown. A first hypothesis may be a useful conceptual framework from which to understand the link between trait anger and MS; that is, anger is associated with perceived stress during free living,[34] and the latter is possible linked to MS.[35] A second hypothesis suggests that persons who have higher levels of anger, compared with those who have lower levels, may be more likely to engage in adverse health behaviors (eg, sedentary lifestyles, cigarette smoking, greater consumption of alcohol, and overeating) that place them at increased risk for MS.[36–39] In our study, those who were taking SSRIs were associated with higher odds of high levels of trait anger compared to those who were not taking them. This would support the hypothesis that serotonin dysregulation is involved in the pathogenesis of anger and that SSRIs (or conditions linked with their prescription) may be associated with anger in some circumstances.

We acknowledge several limitations of this study. First, we recruited a group of patients with MS (i.e., patients seen in a clinic)) and, as a consequence, our results might not be generalized to population-dwelling MS patients. Therefore, we acknowledge the need to replicate these findings in a population-based survey of unselected patients. Second, while we matched patients and controls based on age, gender, and education, we were not able to adjust for other potential unmeasured confounders.

In conclusion, MS patients showed higher levels of trait anger when compared to controls of the same age, gender, and education. Anger may have a significant impact on health-related quality of life and engagement with care and possibly MS outcomes. Longitudinal studies of anger in MS are needed to characterize the stability and evolution of this characteristic within the natural history of the disease process.

Highlights.

We provide further evidence that multiple sclerosis is strongly associated with high trait anger. Those multiple sclerosis patients who were taking selective serotonin reuptake inhibitors (SSRIs) were associated with higher odds of high levels of trait anger compared to those who were not taking them. This would support the hypothesis that serotonin dysregulation is involved in the pathogenesis of anger and that SSRIs (or conditions linked with their prescription) may be associated with anger in some circumstances.