Eur opean Rev iew for Med ical and Pharmacol ogical Sci ences 2013; 17: 2065-2079 Do Akiskal & Mallya’s affective temperaments belong to the domain of or to that of normality?

L. ROVAI 1, A.G.I. MAREMMANI 1,2 , F. RUGANI 1, S. BACCIARDI 1, M. PACINI 1,3 , L. DELL’OSSO 1, H.S. AKISKAL 4, I. MAREMMANI 1,2,3

1Vincent P. Dole Dual Diagnosis Unit, Department of Neurosciences, Santa Chiara University Hospital, University of Pisa, Italy 2Association for the Application of Neuroscientific Knowledge to Social Aims (AU-CNS), Pietrasanta, Lucca, Italy 3G. De Lisio Institute of Behavioural Sciences, Pisa, Italy 4International Mood Centre, University of California, San Diego, CA, USA

Abstract. – BACKGROUND: Kraepelin and Key Words: Kretschmer hypothesized a continuum between Affective temperaments, TEMPS, Pathology inter - full-blown affective pathology and premorbid face, Normality interface, Patho-plastic role, Adaptive temperaments. More recently Akiskal proposed role, Somatic diseases, Psychiatric diseases, Substance a putative adaptive role for the four fundamen - use disorders, Professional choice, Quality of life. tal temperaments: the hyperthymic one charac - terized by emotional intensity, the cyclothymic one by emotional instability, the depressive one by a low energy level, and the irritable one by an excessive response to stimuli. Today it is Introduction widely debated whether affective temperaments belong to the domain of pathology or to that of The concept of temperament was postulated by normality. Hippocrates in the fourth century B.C. More re - PURPOSE: To make clear, by applying an inte - cently, Kraepelin and Kretschmer hypothesized a grated model, the position of affective tempera - ments within the continuum between normality continuum between full-blown affective pathology and pathology. and premorbid temperaments, referring to them as METHODS: We reviewed several papers that lifelong, early-onset, attenuated, subclinical forms explore the distribution of affective tempera - of manic-depressive psychosis 1,2 . ments among the general population, and their According to the conceptualization of Akiskal involvement both in pathological conditions (so - and Mallya, based on Kraepelin’s theory and matic and psychiatric) and in human activities (professions and other occupations). clinical observations, four fundamental affective RESULTS: Far from being intrinsically patho - temperaments do exist (Table I): the depressive logical conditions, affective temperaments seem temperament is distinguished by stably depressed to represent adaptive dispositions whose dys - mood, introversion, a low energy level and hy - regulation can lead to full-blown affective pathol - persomnia; displays ogy. All the temperamental types display some extroversion, a high energy level, emotional in - impact on people’s lives by influencing personal tensity and little need for sleep; the cyclothymic skills and professional choices over a wide field of human activities. temperament shows a central dimension that in - CONCLUSIONS: Affective temperaments are cludes rapid fluctuations in mood and emotional not problematic when they appear in a mild form, instability; the irritable temperament, which is but when they occur in extreme form we have less consistent than the others, includes a predis - observed a gap between the hyperthymic tem - position to being litigious and aggressive and to perament, which represents the most functional encountering difficulties in interpersonal rela - and desirable, and the cyclothymic, depressive, irritable and phobic anxious ones, which are tionships. A new development has been the elab - closer to mood, anxiety, and substance use dis - oration of a putative phobic-anxious tempera - orders, and imply a component of somatic dis - ment consisting of increased sympathetic activi - eases and life stressors. ty, fear of illness, hypersensitivity to separation,

Corresponding Author: Icro Maremmani, MD; e-mail: [email protected] 2065 L. Rovai, A.G.I. Maremmani, F. Rugani, S. Bacciardi, M. Pacini , L. Dell’Osso, H.S. Akiskal, I. Maremmani

Table I. Affective temperamental characteristics according states, exists because it serves, in a polygenic to Akiskal and Mallya’s formulation. model, as the genetic reservoir for adaptive tem - peraments. If we think of affective disorders as ex - Depressive temperament tremes in an oligogenic model of inheritance, • Gloomy, pessimistic, humourless or incapable of fun • Quiet, passive or indecisive adaptive temperaments represent the milder phe - • Sceptical, hypercritical or complaining notypes of their constituent traits. Depressive traits • Brooding and given to worry appear to indicate sensitivity to the suffering of • Conscientious or self-disciplining other members of the species, and to overlap with • Self-critical, self-reproaching, or self-derogatory those of the generalized anxious temperament. • Preoccupied with inadequacy, failure and negative events to the point of morbid enjoyment of one’s failures may have evolved, thanks to the role Hyperthymic temperament played by creativity in sexual seduction, as a • Cheerful, overoptimistic or exuberant mechanism in reproductive success. Hyperthymic • Naïve, overconfident, self-assured, boastful, bombastic traits seem to offer distinct advantages in leader - or grandiose ship, exploration, territoriality and mating. On the • Vigorous, full of plans, improvident, carried away by basis of this comprehensive model, creative and restless impulses • Over-talkative eminent individuals probably occupy a somewhat • Warm, people-seeking or extroverted unstable terrain that is intermediate between tem - • Overinvolved and meddlesome perament and affective disease 10,11 . • Uninhibited, stimulus-seeking or promiscuous The whole question of the nature of affective Cyclothymic temperament temperaments is still an open issue. The concept of • Biphasic dysregulation characterized by abrupt endo-re - temperament derives from the clinical active shifts from one phase to the other, each phase 6,12,13 lasting for few days at a time, with infrequent tradition , and, in the latest version of the Diag - • Marked unevenness in quantity and quality of pro - nostic Statistic Manual (DSM-IV-R), affective tem - ductivity-associated unusual working hours peraments are classified among mood disorders, as • Lethargy alternating with eutonia cyclothymic and dysthymic disorders 14 . As a mat - • Pessimistic brooding alternating with and ter of fact, the validation process of Akiskal and carefree attitudes 15 • Mental confusion alternating with sharpened and cre - Mallya’s criteria for the diagnosis of affective ative thinking temperaments, both in their self-questionnaire and • S haky self-esteem alternating between low self-confi - interview version, has confirmed that affective tem - dence and overconfidence peraments are a widespread dimension in the gen - • Hypersomnia alternating with decreased need for sleep eral population, with significant sex-, age- and lati - • Introverted self-absorption alternating with uninhibit - 4,5,16-24 ed people-seeking tude-related differences . • Decreased verbal output alternating with talkativeness In this review, we have taken into considera - • Unexplained tearfulness alternating with excessive tion several papers that explore the distribution of punning and jocularity affective temperaments among the general popu - Irritable temperament lation, and their involvement both in pathological • Indeterminate early onset conditions (somatic and psychiatric) and human • Habitually moody-irritable and choleric with infre - quent euthymia activities (professions and other occupations), so • Tendency to brood as to be able to specify, while applying an inte - • Hypercritical or complaining grated model, the position of affective tempera - • Ill-humoured joking ments along the continuum between normality • Obtrusiveness and pathology. • Dysphoric restlessness • Impulsiveness Psychometrics of Affective Temperaments difficulty in leaving familiar surroundings, a Development of TEMPS-I, TEMPS-A, marked need for reassurance, and oversensitivity TEMPS-A[P] to drugs and substances 3-9 . The Temperament Evaluation of Memphis, Currently a widely debated issue is whether Pisa, Paris and San Diego (TEMPS), both in its in - affective temperaments belong to the domain of terview and self-questionnaire versions, is based pathology or to that of normality, on Akiskal and Mallya’s criteria for affective tem - Within the general framework of evolutionary peraments, validated in an Italian population of biology it has been suggested that affective dis - 1,010 students aged from 14 to 26 years. Item ease, including and associated psychotic wording and selection have been performed

2066 Do affective temperaments belong to the domain of normality? through an iterative process that incorporated manic symptoms or full-blown mood disorders. feedback from clinicians and researchers 6,12,13 . The The cyclothymic scale appears to be the best de - interview version of TEMPS (TEMPS-I) is a 20- fined; the score obtained by subjects on this scale minute interview usually administered by two psy - exceeds the second standard deviation in the Ital - chiatrists. It has been designed to quantify tem - ian population mentioned above, and the cy - perament in psychiatric patients and healthy sub - clothymic proves to be completely saturated by jects, and includes sections on emotional reactivi - its constituent traits. The depressive features ex - ty, and on cognitive, psychomotor, circadian and ceed the second standard deviation too, and are social-behavioural traits 4,5,15 . negatively correlated with the hyperthymic tem - The self-questionnaire version of TEMPS perament. The exploratory factorial analysis sup - (TEMPS-A) is a self-reporting, yes-or-no type of ports Kretschmer’s hypothesis of a central under - questionnaire formulated on the basis of the lying cycloid temperament. A four-factor struc - same diagnostic criteria as those used for affec - ture emerges, showing on one hand a stable hy - tive temperaments. The first version included 84 perthymic temperament that is negatively corre - items that had been selected to assess dysthymic, lated with cyclothymia, and, on the other, a cen - cyclothymic, hyperthymic and irritable tempera - tral cyclothymic disposition that is closely corre - ments. At a later stage, clinical and theoretical lated with the depressive traits. The irritable tem - considerations led to the addition of 26 new perament might be considered to be a variant of items that were included to account for the anx - the broader temperamental construct of cy - ious temperament, so resulting in the full, 110- clothymia, with coexisting hyperthymic attribut - item-long version of the TEMPS-A 25,26 . es, in line with Kraepelin’s conceptualization of a Subsequently TEMPS-A has been translated combination of other temperamental traits. and validated in several languages, both in a TEMPS-A has been translated into over 25 short and an extended version. Among the Italian language versions. American, Argentinian 31 , Ital - versions of TEMPS-A, there is the special case ian 17,18,28,29 , French 32 , Lebanese 23,33 , Hungarian 20 , of TEMPS-A[P], a 61-item auto-questionnaire Japanese 34 , Portuguese 21 , Brazilian 35 and Turk - derived in Pisa directly from the Italian version ish 36 versions have already been validated. A of TEMPS-I and including the four classical sub - five-factor structure including the four classical scales for dysthymic, cyclothymic, hyperthymic temperaments with the recent addition of the and irritable temperaments 27-30 . anxious subscale has been confirmed in every To deepen our understanding of the nature of version. With regard to the Lebanese-Arabic ver - affective temperaments, it would be useful to sion of TEMPS-A, the strongest correlation was know how their constitutive traits are distributed that observed between the anxious and the cy - among general populations. With respect to bio - clothymic temperament subscales 23,33 . With re - logical characteristics, a Gaussian distribution gard to the Turkish Version of TEMPS-A, the among the general population is usually support - anxious temperament has shown some peculiari - ive of their physiological nature. During the vali - ties. On one hand cognitive anxiety traits have dation process of TEMPS-I, this kind of distribu - proved to overlap with depressive ones. On the tion has been confirmed for the depressive, cy - other, a distinct “nervous”-anxious factor has clothymic and hyperthymic temperaments, but emerged too. Dominant irritable, nervous-anx - not for the irritable temperament, which appears ious and depressive temperaments turned out to to be much less consistent, as demonstrated by be the most common in this population, whereas the lack of discriminant specificity in exploratory dominant cyclothymic and hyperthymic tempera - factor analysis, where the irritable traits did not ments have been shown to be relatively uncom - saturate any function. Except for the irritable mon 36 . As to the Buenos Aires (Spanish) version scale, discriminant analysis has demonstrated the of TEMPS-A, the hyperthymic temperament ability of TEMPS-I to classify subjects. With re - proved to be very uncommon in the general pop - gard to the score obtained by subjects on the hy - ulation 31,37 . Referring now to the Japanese ver - perthymic scale, it never exceeded the second sion of TEMPS-A, the factor validity of depres - standard deviation in the Italian population of sive, hyperthymic, cyclothymic and irritable tem - 1,010 students, and all the traits included in this peraments has been confirmed, with cyclothymic construct turn out to be highly inter-correlated 4,5 . traits proving to be those best represented in the It is likely that the hyperthymic temperament is general population. Moreover Japanese Authors abnormal only in the presence of chronic hypo - have proposed a new version of the question -

2067 L. Rovai, A.G.I. Maremmani, F. Rugani, S. Bacciardi, M. Pacini , L. Dell’Osso, H.S. Akiskal, I. Maremmani naire, calling it the “Japanese Temperament and Cloninger’s Tri-dimensional Personality Personality (JTP) Scale”; this scale includes the Questionnaire (TPQ) classical subscales, but adds a melancholic and a In the Italian population of 1,010 students who schizoid type to them. Within this model the participated in the validation of Akiskal and traits of the depressive and melancholic types Mallya’s criteria for affective temperaments, emerge as rather distinct, with the depressive TEMPS-I has been compared with Cloninger’s re - type characterized by interpersonal sensitivity, vised “Tri-dimensional Personality Questionnaire” and the melancholic one characterized by a high - (TPQ) deriving from the traditions of experimental er functional efficiency, and the inclusion of oth - psychology. As was to be expected, depressive tem - ers features, such as being perfectionist and perament was correlated with a high level of harm work-oriented 34,38 . avoidance. High scores for novelty-seeking were re - lated both to the hyperthymic and to the cy - Reliability and Stability (T-retest) clothymic temperament. Cyclothymic traits were Given that affective temperaments are supposed related both to the harm avoidance and novelty to be lifelong and early-onset dimensions, the test- seeking dimensions, in line with Kretschmer’s hy - retest reliability of TEMPS was evaluated in 206 pothesis of a central cycloid temperament. On a Italian high school students (aged 14-18). The more theoretical plane, hyperthymic novelty seek - temperamental traits of students were measured at ers should be over-represented among those who T0 and T1 (two years later) by means of the inter - have reached a high level of achievement; by con - view version of TEMPS (TEMPS-I). Age, sex, trast, a moody, restless disposition (a cyclothymic- psychometric properties, raw scale score, and harm avoidant type) should be liable to negative af - weighted cut-off (calculated from a specially fective arousal. These considerations may shed weighted linear combination of items) were used some light on the origins of socially adaptive be - as predictive variables of stability. Affective tem - haviour (referred to as belonging to “sunny ” or peraments showed a low to moderate level of sta - “sanguine ” types) on one hand, and borderline con - bility, reaching 60% in the case of subjects with a ditions typical of anxious-hostile bipolarity (found dominant cyclothymic temperament. The stability in “dark ” types) on the other 39 . of the depressive temperament was primarily relat - ed to its weighted cut-off. The stability of the hy - Multiphasic Personality Inventory (MMPI) perthymic temperament appeared related to male In a sample of 693 candidates who were apply - sex, young age, and total scale score. Male gender ing to become cadets in the Italian Air Force, turned out to be the best predictor of stability for TEMPS-A[P] has been correlated with the MMPI the cyclothymic temperament as well. Unstable validity and clinical scales after a stressful chal - depressive and hyperthymic subjects were preva - lenge represented by the academy entrance trial. lently reallocated into the dominant cyclothymic MMPI is a self-reporting questionnaire designed to group. The irritable construct was the least stable. identify personality structure and psychopathology In summary, TEMPS-I showed considerable fluc - in psychiatric patients and healthy volunteers. As tuation and instability in depressive and hyper - regards MMPI validity scales, TEMPS-A[P] de - thymic temperaments in mid-adolescence; con - pressive candidates tended to report their symptoms versely, the cyclothymic temperament turned out sincerely, hyperthymic candidates tended to give to be the most stable, and for this reason, in refer - false answers, with the aim of being seen in a good ring to cyclothymia, the term “stable instability” is light. Cyclothymic and irritable candidates tended often used 16 . to exaggerate symptoms. As regards the MMPI clinical scales, a low linkage between affective tem - Correlation with Personality peraments and abnormal personality traits has been Questionnaires (Convergent validity) found. From a personological point of view tem - Considering the considerable uncertainty peraments have proved to belong to the realm of found in the current literature about the relation - normality rather than to the realm of pathology, in ship between personal traits and affective tem - line with their putative adaptive role 27 . peraments, the criteria of Akiskal and Mallya have been compared with a few correctly validat - Occupational Personality Questionnaire ed instruments for the assessment of personality (OPQ32) dimensions, which supposedly overlap with tem - In order to understand the extent to which affec - peraments. tive temperaments are adaptive, it would be useful

2068 Do affective temperaments belong to the domain of normality? to know if they influence the work-related abilities psychiatric disorders, with the latter including of subjects. In a sample of 921 candidates apply - not only affective disorders but ing to become cadets in the Italian Navy, TEMPS- and eating disorders, too. A[P] has been correlated with Occupational Per - sonality Questionnaire, self-reported version Somatic Pathology (OPQ32), during a stressful challenge represented The idea that affective temperament may have by the entrance examination. OPQ32 is a self-re - some correlation with somatic pathology is not porting personality questionnaire designed to pro - new. In the literature, in fact, one topic of investi - vide information on an individual’s preferred be - gation has been the temperamental profile of pa - haviour, as assessed in terms of a number of work- tients suffering from several somatic diseases, es - related characteristics. Depressive temperament pecially those supposedly related to life style and was reported to imply a low level of ability to re - behavioural patterns. Somatic and psychiatric late to others; hyperthymic temperament was char - morbidity might cluster because of reciprocal ef - acterized by high levels of feelings and emotions, fects between them, but also as a result of com - and by the capability to relate to people; cy - mon underlying factors. clothymic temperament was distinguished by cre - Among infective diseases, HIV infection rep - ativity and a low level of relationships with others; resents, of course, a condition where behavioural irritable temperament displayed an overlap with risk factors (intravenous drug use and homosexu - cyclothymic temperament, the main difference be - ality) play a crucial role. In an Italian study, HIV ing the higher level of energy and the lower level patients with index major depressive episode of empathy of irritable subjects. The four affective have been compared with seronegative major de - temperaments proved to differ significantly in the pressive episode patients by systematically ex - work capacity features measured by OPQ32 fac - amining rates of DSM-III-R bipolar subtypes and tors. These observed correlations further support Akiskal’s affective temperaments. The most im - the hypothesis that temperaments belong to the portant finding was the significantly higher pro - realm of normality rather than that of patholo - portion of HIV patients who had lifetime bipolar gy 29,40 . II disorder, and associated cyclothymic and hy - perthymic temperaments, regardless of HIV risk Emotional-Affective State Rating Scale status. To formulate the question provocatively, (EAS-RS) premorbid impulsive risk-taking traits associated In order to understand the extent to which af - with cyclothymic and hyperthymic temperaments fective temperaments are adaptive, it would be may have played an important role in needle- useful to raise the question of how they influence sharing drug use and unprotected sexual behav - an individual’s emotional and behavioural re - iour, leading ultimately to infection with HIV 41 . sponse to stress. In a sample of 693 candidates Another illness, which has been supposed to applying to become cadets in the Italian Air be related to affective temperaments, especially Force, TEMPS-A[P] scales have been correlated in terms of their diagnostic and therapeutic fea - with the emotional-affective state reported by tures, is type 2 diabetes mellitus. Among diabetic subjects after a stressful situation – the entrance patients, those with excessive depressive tem - examination to the academy. It is hardly surpris - perament showed a worse psychological adjust - ing that the hyperthymic temperament on one ment to diabetes, and worse metabolic control 42 , hand, and the cyclothymic and depressive tem - so prompting the idea that depressive tempera - peraments on the other, turned out to be distin - ment may constitute a vulnerability factor to the guished by counter-polar emotional states fol - behavioural and biological outcome of this dis - lowing the test; these were desirable in the first ease, or even a potential risk factor for its later case and undesirable in the other two 27 . incidence 43 . Also, the anxious temperament proved to have an impact on the metabolic para - meters of this disease. Among diabetic patients, Affective Temperaments in Somatic and glycated haemoglobin levels at baseline and at a Psychiatric six-month follow-up were inversely associated with the presence of an anxious temperament. In the literature, affective temperaments have This metabolic outcome was most probably due been reported to correlate with a broad variety of to the impact of temperament on the glycated pathological conditions ranging from somatic to haemoglobin at baseline, rather than to the up -

2069 L. Rovai, A.G.I. Maremmani, F. Rugani, S. Bacciardi, M. Pacini , L. Dell’Osso, H.S. Akiskal, I. Maremmani take of self-management behaviours after diag - The reversal from one temperament to an nosis. Markedly anxious temperamental traits episode belonging to the ‘opposite’ polarity have also been associated with an increased like - seems to represent a fundamental aspect of the lihood of being diagnosed with a prediabetic dysregulation that characterizes bipolar disor - condition, and with a poorer quality of life. It is der. Both men and women possessing a hyper - surprising that the same temperamental profile thymic temperament appeared to be protected facilitates the early detection of the illness, but against depressive symptom formation during a not the subsequent behavioural and emotional manic episode, which, accordingly, was consid - adjustment to it 44 . ered to be relatively ‘pure’. Pure mania appears to be over-represented in men because they Psychiatric Pathology show higher scores for the hyperthymic tem - Some Authors have stated that temperamental perament, whereas pictures of mixed mania dysregulation may represent the phenotypic ex - seem to be more typical of women, consistently pression of the underlying bipolar genotype 15,45-48 ; with their well-known tendency to experience on the basis of this connection they have ex - depression. All this evidence supports the gener - plored the role of affective temperaments in al idea that “mixity ” should be conceptualized pathological conditions such as mood disorders, as an intrusion of mania into its “opposite” tem - substance use disorders, and eating disorders. perament, especially in females 52,53 . More re - With reference to mood disorders, it has been cently, affective temperaments have been ex - shown that affective temperaments influence the plored in a sample of 153 maniac inpatients, di - clinical features of Bipolar Disorders in terms vided into five groups according to the sympto - of both clinical and course characteristics. For matological subtype of mania (Depressive, Irri - example, depressions arising from a cy - table-Agitated, Euphoric-Grandiose, Accelerat - clothymic temperament, and likely to be misdi - ed-Sleepless, Paranoid-Anxious). The patients agnosed as personality disorders, presented, in who belonged to the “Euphoric-Grandiose”, 194 patients taking part in a French national “Paranoid-Anxious” and “Accelerated-Sleep - multisite study, high familial load for mood dis - less” subtypes were those most likely to present orders, so validating its putative bipolar nature. a hyperthymic temperament, while the ‘Depres - These patients, characterized as the “darker” ex - sive’ dominant group had the highest frequency pression of the more prototypical “sunny” type- among those with a depressive temperament. II bipolar phenotype, seem to represent a more The ‘Irritable-Agitated’ group was a frequent “unstable ” variant of 49 . Among finding for both these temperaments. A hyper - 106 bipolar type-I patients of a multicentric Ital - thymic temperament seemed to underlie the ian study, dominant cyclothymic and hyper - most extreme manic excitement with euphoric- thymic subjects reported important differences accelerated-paranoid phenomenology. By con - in terms of gender distribution, family history, trast, the depressive temperament seemed to number and polarity of previous episodes, hos - modify the expression of mania into a depres - pitalizations, suicidality, rates of comorbid anxi - sive-manic phenomenology 54 . Lastly, affective ety and personality disorders. These observa - temperaments have been also reported to modu - tions are consistent with the hypothesis that af - late gender-related differences among bipolar fective temperaments, in particular cyclothymia, patients. In a sample of 538 subjects with pri - could be utilized as quantitative, intermediate mary mood disorders, females, with respect to phenotypes in order to identify genes linked males, presented a lower number of hypomanic, with susceptibility to bipolar disorders 50 . Fur - and a higher number of depressive, episodes, ther evidence has supported the familial, possi - with a higher rate of comorbid anxiety disorders bly genetic, role of the hyperthymic tempera - and somatization. Moreover, they were more ment in the genesis of and likely to exaggerate symptoms and to undergo pure mania, while suggesting a higher specifici - hospitalization. These gender differences could ty of the cyclothymic temperament to the bipo - partly be explained by the higher prevalence of lar II subtype and to mixed states 49-53 . Affective the depressive temperament in women, and of temperaments have also been shown to influ - the hyperthymic temperament in men 55 . ence the pathogenesis of mixed states like With the aim of expanding the boundaries of mixed mania, as reported in a sample of 104 the bipolar spectrum disorder, the role of tem - manic patients during the acute hospital phase. peramental features has also been studied in psy -

2070 Do affective temperaments belong to the domain of normality? chiatric conditions, such as eating disorders, that sizing “sensation-seeking ” and “novelty-seeking ” are not traditionally classified among mood dis - dimensions as the main personality characteris - orders. Recent data have, in fact, indicated a sig - tics of addiction 67 . In order to understand nificant clinical, biological and treatment re - whether the prevalence of cyclothymic traits sponse overlap between eating and bipolar disor - among heroin addicts and alcoholics was due to ders, especially when the soft symptoms of either the co-occurrence of a full-blown bipolar disor - spectrum disorders are considered. When 107 der, in the two papers just mentioned the Authors consecutive patients suffering from a major de - compared the temperamental traits of dual diag - pressive episode with atypical features were in - nosis patients with those of patients without a vestigated, the patients who did meet the DSM- dual diagnosis. The cyclothymic disposition was IV criteria for bulimia nervosa, compared with unable to differentiate these two groups of pa - those who did not, were indistinguishable in tients, prompting the idea that the importance of terms of all their demographic and most of their cyclothymia in addictive disorders is unrelated to psychopathological and clinical features (includ - the presence of dual diagnosis 66,67 . ing bipolar I and II), but showed significantly Temperaments have been evaluated too in the higher frequencies in lifetime comorbidity for field of stimulant abuse, with the aim of outlining Narcissistic, Histrionic, Borderline and Depen - the nature of a hypothesized stimulant bipolar dent personality disorders as well as in that for spectrum. Among abusers of stimulants, cy - Cyclothymic temperament. They also scored clothymic and hyperthymic traits have proved to higher on reactivity of mood and interpersonal precede by years the use of stimulants, which sensitivity. Cyclothymic temperament and related seemed to serve the purpose of maintaining the mood reactivity and interpersonal sensitivity ac - subthreshold that brings the reward of mood ela - counted for much of the relationship between tion. This seems to corroborate a bipolar-stimu - Atypical Depression and Bulimia Nervosa. Nar - lant spectrum where subthreshold bipolar traits cissistic, histrionic and borderline traits, too, are complicated by the abuse of stimulants, even - seemed to be related to the presence of a cy - tually leading to new pathological features in clothymic disposition. Taken as a whole, these both disorders 68 . data support the hypothesis that places Bulimia Nervosa in the “ultra-soft ” bipolar realm 56 . Affective Temperaments and Emotional- Substance Abuse Disorders Behavioural Problems in Childhood Given the generally accepted link between substance use disorders and bipolarity 57-65 , affec - Since affective temperaments are considered tive temperaments have been studied in heroin stable, subclinical forms of the manic-depressive addicts, alcoholics, and cocaine abusers. illness, we can hypothesize that their early mani - With regard to heroin addiction, which pro - festations can be observed in childhood or ado - vides a paradigm for addictive disorders, the tem - lescence . Starting from this hypothesis, the corre - peramental traits of 59 consecutive stabilized lation between affective temperament and emo - methadone-treated heroin addicts were compared tional-behavioural problems has been assessed in with those of 58 healthy volunteers sharing simi - a juvenile population of 1,010 students without lar social and regional demographics. Cy - major psychiatric disorders. The depressive tem - clothymic temperament, and, to a lesser extent, perament turned out to be a construct that partly irritable traits (the “dark side ”), were reported to overlapped with behavioural inhibition, while ex - provide the most representative temperamental tremes of emotionality and behaviours occurred profile for heroin addicts 66 . As to alcohol abuse, preponderantly in children who showed cy - ninety-four consecutive alcoholics then respond - clothymic traits. In particular, cyclothymic sub - ing to treatment were compared, with reference jects have reported the highest rates of anxiety- to their affective temperaments, with 50 healthy sleep disorders and sensitivity to separation, in volunteers who displayed the same social charac - addition to eating disorders in females and anti - teristics and belonged to the same environment. social-aggressive behaviour in males. These data Alcoholics were distinguished from controls in support the idea that the cyclothymic disposition terms of cyclothymic traits, including a depres - is the most ‘morbid’ temperament, as it is associ - sive component. These characteristics tend to co - ated both with internalizing and externalizing here with previous conceptualizations hypothe - disorders 69 .

2071 L. Rovai, A.G.I. Maremmani, F. Rugani, S. Bacciardi, M. Pacini , L. Dell’Osso, H.S. Akiskal, I. Maremmani

Affective Temperaments and Professional Choice in Student Populations “Goodness of Fit” In order to assess the role of affective tempera - ments in determining personal choices and aspi - The concept of matching temperament with rations, the temperamental profile of 1,386 stu - learning styles to create a “goodness of fit ” as pro - dents aspiring to enter different professional posed by Thomas and Chess (1977) 70 is summa - fields has been described. Future physicians rized by the Authors as follows: “Stated briefly, failed to show any predominant temperament, there is a goodness of fit when the person’s tem - whereas future lawyers and artists predominantly perament and other characteristics such as motiva - revealed a cyclothymic or irritable temperament, tion and levels of intellectual and other abilities, future engineers showed a hyperthymic tempera - are adequate to master the successive demands, ex - ment and, future psychologists and nurses dis - pectations, and opportunities of the environment ”. played predominantly depressive and anxious In order to understand the extent to which af - temperaments 73 . fective temperaments are adaptive, the first issue that arises is how they influence the quality of Professional Choice in Applicants life. A multicentric Argentinian study has as - We have recently studied the temperamental sessed, among clinically unaffected relatives of traits of 1,548 candidates applying to become a bipolar patients, the prevalence of affective tem - cadet officer in the Italian Air Force. Extremely peraments and their impact on the quality of life. high scores on the hyperthymic scale combined Compared with controls, the relatives of bipolar with extremely low ones on the cyclothymic scale patients showed higher scores on all TEMPS-A seemed to correspond to the specific temperamen - subscales, except for the hyperthymic subscale. tal profile of young applicants, and to the highest Only in the case of hyperthymic subjects was the likelihood of success. Those who took the en - quality of life equal to that of controls, a finding trance examination proved to be more hyper - that further qualifies hyperthymia as the most thymic than their peers, and the specificity of this adaptive temperamental subtype 71 . correlation was confirmed by the fact that appli - cants who made a second attempt to pass the en - trance examination after an initial failure were Affective Temperaments and more hyperthymic than first-time applicants. In Professional Choice addition, success in specific psychological admis - sion tests was related to the same temperamental Affective temperaments have also been ex - profiles, since those who proved to be psychologi - plored in a broad range of human activities, with cally fit were more hyperthymic 27 . In another the purpose of identifying the temperamental study carried out on the Italian Air Force we in - profile of specific professional fields. quired into the question whether gender differ - ences in temperament have continued unchanged Professional Choice in Outpatients in a field – specifically, that of a military career – Temperamental traits have been assessed in a that has been historically characterized by a male sample of 263 psychiatric outpatients involved in identity. In the general population, males and fe - several professional fields. Dysthymic and obses - males have shown different temperamental pro - sional attributes were notable in lawyers and physi - files. As a matter of fact, over the past fifty years cians. Cyclothymia seemed to be the dominant af - in the Western world the professions and activities fective temperament of artists and architects, while that were once practised only by males, are now hyperthymic temperament appeared to play a cen - available to females, too. Among our aspiring tral role among managers, self-made businessmen cadets, both males and females showed high and journalists. The role of cyclothymic and hyper - scores on the hyperthymic scale (in general, a typ - thymic temperaments turned out to be distin - ically male temperament) and very low ones on guished by obsessional traits. In particular, artists’ the cyclothymic scale (generally, a typically fe - creativity was ‘liberated’ by low levels of obsessive male temperament). Low scores were also ob - traits, whereas among architects, relatively high served on the depressive and the irritable scales. levels of obsessive traits contributed to the execu - These observations support the idea that the differ - tion of their work. Journalists, as a group, showed ences observed in gender-related temperaments that they possess the broadest representation of af - are functional to the differentiation between the fective temperaments 72 . roles played by males and females during the bio -

2072 Do affective temperaments belong to the domain of normality? logical and social evolutionary process 74 . In anoth - wide range of diseases, extending far beyond the er study we extended our investigation to another field of affective illness. In particular, tempera - military service – the Italian Navy. We compared mental traits have been shown to influence not temperaments between those who had applied to only psychiatric but also somatic diseases, and become a cadet officer in the Italian Air Force or this tendency is further confirmed when tempera - in the Italian Navy, with special reference to gen - ments are measured with instruments other than der differences and the ability of the two types of TEMPS. applicants to pass the psychiatric examination for In connection with somatic diseases, an inter - admission. Hyperthymic traits were well repre - esting study carried out on a very large cohort of sented in both these armed services. Navy appli - psychiatric patients suffering from a wide range cants differed from Air Force applicants in obtain - of somatic diseases, concluded that neuroticism, ing higher depressive, cyclothymic and irritable which is suggestive of a cyclothymic tempera - scores. Navy applicants who passed the psychi - ment, has proved not only to raise the risk of psy - atric entrance examination showed the same chiatric disorders but also, regardless of whether prevalence of hyperthymic temperament as their manifest psychiatric disorders have developed, of Air Force counterparts, but higher depressive, cy - a broad spectrum of chronic somatic diseases 75 . clothymic and irritable scores 30 . Hyperthymic tem - Taking into account surgical pathologies, tem - perament turns out to represent the temperamental peramental traits were reported to influence chil - profile of those who aim to become a cadet officer dren’s behavioural recovery after surgery, and in two of the Italian armed forces (the Air Force adults’ capability to cope with traumatic injuries. and the Navy). Among children undergoing tonsillectomy and To enter a particular professional field, personal adenoidectomy, an inhibited temperament was aspirations and attitudes are not the only features found to predict postoperative sleep disorder, that are required, and other professional skills which seems to affect recovery following might be necessary. To date, there is an unjustified surgery 76 . Among adults admitted to a hospital lack of data on the correlation between tempera - less than 8 h after a car accident, lower scores mental profiles and the ability to pass examination than controls reported on a harm avoidance scale, tests. In order to fill this gap we tested correlations together with high scores for self-directedness between affective temperaments and step-by-step and cooperativeness – a temperamental profile results during the entrance examination of 921 ap - suggestive of a hyperthymic temperament – cor - plicants to become a cadet in the Italian Navy. Hy - responded to a high capability to cope with so - perthymic temperamental traits proved to be im - matic problems 77 . Neuroticism and negative emo - portant not only in choosing a profession, but also tionality have also been shown to represent a risk in passing entrance examinations. Nevertheless, factor for the development of the metabolic syn - affective temperaments (high scores for hyper - drome. Essential parameters of the metabolic thymic and low ones for cyclothymic, depressive, syndrome have been examined in healthy chil - irritable traits) successfully predicted only the re - dren. As a result, mental vitality and positive sults of psychiatric examinations and, to a lesser emotionality, which seem to be supportive of a extent, the results of medical exams and aptitude hyperthymic temperament, were related to a low tests. Passing the high school-leaving exam and somatic risk level, whereas hyperactivity, nega - passing mathematical exams proved to be skills tive emotionality, responsiveness to others, and independent of temperamental traits, and seemed cooperativeness, which seem to reflect a cy - to be influenced by personal skills that were not clothymic or depressive disposition, were related strictly related to temperaments (unpublished da - to a high level of somatic risk 78 . The relationship ta). These data corroborate the hypothesis that hy - between affective temperaments and psoriasis perthymic traits bring distinct advantages in a pro - has been investigated too, with similar results. fessional field such as a military career, which is Patients suffering from psoriasis have shown a closely related to leadership. distinctive temperamental profile, represented by significantly higher scores for harm avoidance and lower scores for self-directedness, compared Discussion with controls 79 . These observations, consistently with those Affective temperaments have been reported to carried out on infective and metabolic diseases display an important pathoplastic action on a by means of TEMPS criteria, support the idea

2073 L. Rovai, A.G.I. Maremmani, F. Rugani, S. Bacciardi, M. Pacini , L. Dell’Osso, H.S. Akiskal, I. Maremmani that a hyperthymic temperament might increase If we consider the wide range of psychiatric subjects’ ability to cope with somatic diseases, disorders, it can be stated that, while both hyper - while a cyclothymic disposition, which is situat - thymic and cyclothymic temperamental dysregu - ed closer to neuroticism, might impair emotion - lation are involved in the pathogenesis of mood al and behavioural adaptation to the same dis - disorders, in eating disorders, especially the bu - eases. Affective temperaments seem to be in - limic subtype, a central role is played by the cy - volved with the course of somatic illness at dif - clothymic temperament. ferent levels, as they influence diagnostic and With regard to substance abuse disorders, we therapeutic features and have an impact on emo - previously mentioned the crucial role of Akiskal tional and behavioural adjustment after diagno - and Mallya’s cyclothymic temperament, because sis, ultimately leading to differences in inci - of its feature of proneness to addictive disorders. dence and outcome. Among urban, northern Italian high school stu - With regard to psychiatric diseases, the bound - dents, higher sensation-seeking, the impairment aries of bipolarity have been expanding over the of the ability to cope socially, direct aggressive - past decade, with affective temperaments playing ness and poor school achievements – usually typ - a central role in defining the new, softer expres - ical of a cyclothymic disposition – have been sions of bipolarity. On the basis of the demon - found to be associated with illicit drug use and strated clinical validity of these expressions, a alcohol abuse 85 , in line with what has been ob - broader concept of soft bipolarity has been pro - served by applying TEMPS criteria. The influ - posed, of which an increasing portion consists of ence of temperament on substance abuse has also the softest expressions of bipolarity, which are been investigated among schizophrenics, who re - intermediate between bipolar disorder and nor - vealed a significant correlation between mality 80 . When temperamental traits have been Cloninger’s novelty-seeking dimension and the measured among euthymic bipolar patients by use of alcohol, cannabis, and nicotine, so sug - applying Cloninger’s criteria, persistence, self-di - gesting that novelty-seeking type behaviours rectedness, and cooperativeness had significantly contribute to substance use in psychotic patients, lower scores, and harm avoidance and reward de - too 86 . One result has been that the association be - pendence had significantly higher ones than tween substance use disorders and temperamen - those recorded for controls 81,82 . We previously tal characteristics has been investigated among mentioned that, according to Akiskal and hospitalized patients with mood disorders or non- Mallya’s criteria, two different subtypes of bipo - affective psychotic disorders. A lifetime history lar disorder can be considered separately, one of alcohol and cannabis misuse was independent - arising from the cyclothymic temperament, and ly associated with having higher scores for sensa - the other from the hyperthymic one. This tem - tion-seeking traits. These findings suggest that peramental dichotomy is only partly confirmed sensation-seeking and are tempera - by Cloninger’s criteria, which, with regard to the mental characteristics that probably favour sub - novelty-seeking dimension, did not reveal any stance use, independently of categorical diag - significant differences between patients and con - noses 87 . After adopting the hypothesis of a con - trols, in line with the adaptive nature of the hy - tinuum between use, abuse and addiction, we perthymic temperament. Conversely, when TPQ wondered if the link bipolarity-substance abuse dimensions are used to differentiate depressed could be extended from illegal substances of unipolar patients from bipolar patients, the latter abuse to recreational ones, with special reference proved to be significantly more novelty-seeking to social drugs (caffeine, nicotine and chocolate). and less harm-avoidant than the former – a tem - We evaluated the social drug habits of patients peramental pattern that appears to be closer to suffering from mood disorders, according to that reported by applying Akiskal and Mallya cri - DSM-IV-R criteria (major depressive episode, re - teria 83 . In addition, in the field of eating disor - current depression, bipolar type 1 and type 2 dis - ders, temperaments have been investigated with orders, and depression not otherwise specified). instruments other than TEMPS, and, especially The same patients were also divided into bipolar for bulimia nervosa, the current evidence indi - or not bipolar subgroups according to inter- cates a personological profile that shows a good episodic clinical features assessed by means of fit with the fundamental traits of the cyclothymic Angst criteria, a correctly validated instrument temperament, including impulsiveness, anxiety which proposes a broader concept of soft bipolar - and a tendency towards dyscontrol 84 . ity, comprising both the spectrum of bipolar dis -

2074 Do affective temperaments belong to the domain of normality? orders proper, and other soft expressions of bipo - dependence and persistence (“disengaged sub - larity intermediate between bipolar disorder and jects”) – personality traits supportive of a cy - normality 80,88,89 . Only Angst criteria were able to clothymic temperament – showed the opposite predict a greater use of social drugs, so confirm - associations 99 . Temperament and character traits ing that the attenuated phenotypes of bipolarity and health-related quality of life, have also been are not of secondary importance among the en - evaluated in patients with cancer, by three Italian tire range of bipolar expressions 90 . oncology departments. Lower levels of harm If we take into account the wide field of sub - avoidance and higher levels of self-directedness stance abuse disorders, the cyclothymic disposi - – features suggestive of a hyperthymic tempera - tion seems to represent a crucial dimension in ment – were significantly correlated with a better determining proneness to addictive diseases, quality of life, in line with the adaptive nature of thus favouring exposure to substances of abuse this temperamental type 100 . (the novelty-seeking and sensation-seeking di - If we take into consideration the global im - mensions), and facilitating addictive processes pact of affective temperaments on the quality of (impulsiveness, motivation and reward alter - life of subjects, a dichotomy can be confirmed ation, anxiety, lack of control, lack of resiliency between the hyperthymic temperament, which to stress). Moving now from drug addiction to has been shown to improve the quality of life, drug abuse, a minor but still considerable role of and the cyclothymic temperament, which has the hyperthymic temperament emerges in the been shown to decrease it. These effects have field of stimulant drugs. In the case of social been observed both in healthy subjects and in pa - drug habits, too, a greater use of coffee, choco - tients suffering from a wide range of somatic dis - late and tobacco seems to be favoured by sub - eases, and could, to some extent, be mediated by threshold affective traits, independently of the the vulnerability to stress that is related to the hy - temperamental type involved. We propose that, pothalamic-pituitary-adrenal axis reactivity, beyond the relationship between cyclothymia which has been reported to be lower in subjects and addiction, with regard to use and abuse be - with low scores on harm avoidance, and higher in haviours, both cyclothymic and hyperthymic subjects who have high scores on self-directed - temperaments, especially when extreme, seem to ness and harm avoidance 101 . display a favouring effect. In summary, it can be stated that affective tem - We mentioned earlier that, within a juvenile peraments have been shown to influence people’s population, affective temperaments, in particular lives, both in the field of pathology and in that of the cyclothymic temperament, have been associ - normality. ated with emotional-behavioural problems in In the field of pathology there is growing evi - both genders. This association has been widely dence to suggest that affective temperaments not confirmed by the current literature, which, ac - only display a crucial effect on the pathogenesis cording to models other than TEMPS, has de - of mood disorders, but also have an impact on fined temperamental characteristics as predictors proneness to non-affective diseases, even in pa - of behaviour disorders in children. In particular, a tients who are not strictly bipolar. Among full- longitudinal relationship has emerged between blown affective disorders the central role of cy - temperament traits in the third year of life and clothymia should be recognized; it has, in fact, behavioural adjustment at school entry 91-98 . been proposed that quantitative, intermediate Affective temperaments have been reported to phenotypes of cyclothymia allow the identifica - influence the quality of life of bipolar patients’ tion of genes that indicate susceptibility to bipo - relatives, with a positive impact displayed only lar disorders. It is also worth mentioning the in - by hyperthymic temperament. When the associa - volvement of the hyperthymic temperament in tion of temperaments with psychopathology and the pathogenesis of pure mania. However, out - wellness was investigated in children of the side the boundaries of mood disorders as such, North-Eastern United States, by applying the hyperthymic temperament seems to be an Cloninger’s criteria, subjects with a low level of adaptive and desirable condition, which allows novelty-seeking but a high level of persistence for hyperadaptation to somatic diseases, whereas (“steady subjects”) showed better functioning the cyclothymic temperament seems to increase and lower levels of psychopathology, whereas the risk of substance abuse and eating disorders, subjects with higher scores for novelty-seeking and to imply difficulty in adapting emotionally and harm avoidance, but lower ones for reward and behaviourally to somatic diseases.

2075 L. Rovai, A.G.I. Maremmani, F. Rugani, S. Bacciardi, M. Pacini , L. Dell’Osso, H.S. Akiskal, I. Maremmani

In the field of normality, all the temperamental 6) AKISKAL HS, A KISKAL K. Cyclothymic, hyperthymic and types display some impact on the life of subjects, depressive temperaments as subaffective variant of mood disorders. In: Tasman A, Riba M B, editors. by influencing personal skills and professional Annual Review of Vol n° 11. Washington, choices. In this case too, however, the hyper - DC: American Psychiatric Press, 1992. thymic temperament has proved to be the most 7) PHILLIPS KA, G UNDERSON JG, H IRSCHFELD RM, S MITH adaptive; it is distinguished by a better quality of LE . Review of the depressive personality. Am J life, a reduced reactivity to life stressors, and a Psychiatry 1990; 147: 830-837. greater ability to relate to others. 8) CASSANO GC, A KISKAL H S, S AVINO M, M USETTI L, P ERU - GI G, S ORIANI A. Proposed subtyper of bipolar II disorder with hypomanic episodes (or cy - clothymia) and with hyperthymic temperament. J Conclusions Affect Disord 1992; 26: 127-140. 9) PERUGI G, A KISKAL HS, G EMIGNANI A, P FANNER C, PRES - Far from being intrinsically pathological condi - TA S, M ILANFRANCHI A, L ENSI P, R AVAGLI S, M AREMMANI tions, affective temperaments seem to be adaptive I, C ASSANO GB. Episodic course in obsessive-com - pulsive disorder. Eur Arch Psychiatry Clin Neu - dispositions whose dysregulation can lead to full- rosci 1998; 248: 240-244. blown affective pathology. All the temperamental 10) AKISKAL HS, A KISKAL KK . In search of Aristotle: tem - types display some impact on the life of subjects, perament, human nature, melancholia, creativity by influencing personal skills and professional and eminence. J Affect Disord 2007; 100: 1-6. choices in the broad range of human activities. Af - 11) AKISKAL KK, A KISKAL HS . The theoretical underpin - fective temperaments are not problematic when nings of affective temperaments: implications for evolutionary foundations of bipolar disorder and they appear in a mild form, but when they take an human nature. J Affect Disord 2005; 85: 231-239. extreme form, we have observed a gap between the 12) AKISKAL HS . The temperamental borders of affective hyperthymic temperament, which must be consid - disorders. Acta Psychiatr Scand 1994; 89(Suppl ered the most functional and desirable, and the cy - 379): 32-37 . clothymic, depressive, irritable and phobic anxious 13) AKISKAL HS . Temperament and mood disorders. ones, which are closer to mood, anxiety and sub - Harv Ment Health Lett 2000; 16: 5-6. stance use disorders, and imply difficulty in adapt - 14) A.P.A. DSM-IV-TR. Diagnostic and Statistical Manu - al of Mental Disorders. Washington, DC: Ameri - ing emotionally and behaviourally to somatic dis - can Psychiatric Association 2000. eases and life stressors . 15) AKISKAL HS, M ALLYA G. Criteria for the 'soft' bipolar spectrum: Treatment implications. Psychopharma - ––––––––––––– –– ––– –– col Bull 1987; 23: 68-73. Conflict of Interest 16) PLACIDI G F, M AREMMANI I, S IGNORETTA S, L IGUORI A, None. AKISKAL HS . A prospective study of stability and change over 2 years of affective temperaments in 14-18 years-old Italian high school students. J Af - fect Disord 1998; 51: 199-208. References 17) POMPILI M, G IRARDI P, T ATARELLI R, I LICETO P, D E PISA E, TONDO L, A KISKAL KK, A KISKAL HS . TEMPS-A 1) KRAEPELIN E. Manic-Depressive Illness and Para - (Rome): psychometric validation of affective tem - noia. Edinburgh: Livingstone, 1921. peraments in clinically well subjects in mid- and 2) KRETSCHMER E. Physique and Character. New York, south Italy. J Affect Disord 2008; 107: 63-75. NY: Macmillan, 1936. 18) PRETI A, V ELLANTE M, Z UCCA G, T ONDO L, A KISKAL K, 3) PERUGI G, T ONI C, T RAVIERSO MC, A KISKAL HS . The AKISKAL H. The Italian version of the validated short role of cyclothymia in atypical depression: toward TEMPS-A: the temperament evaluation of Mem - a data-based recondeptualization of the border - phis, Pisa, Paris and San Diego. J Affect Disord line-bipolar II connection. J Affect Disord 2003; 2010; 120: 207-212. 73: 87-98. 19) BORKOWSKA A, R YBAKOWSKI JK, D ROZDZ W, B IELINSKI M, 4) PLACIDI GF, S IGNORETTA S, L IGUORI A, G ERVASI R, KOSMOWSKA M, R AJEWSKA -R AGER A, B UCINSKI A, A KISKAL MAREMMANI I, A KISKAL HS . The semi-structured af - KK, A KISKAL HS. Polish validation of the TEMPS-A: the fective temperament interview (TEMPS-I). Relia - profile of affective temperaments in a college student bility and psychometric properties in 1010 14-26- population. J Affect Disord 2010; 123: 36-41. year old students. J Affect Disord 1998; 47: 1-10. 20) ROZSA S, R IHMER A, K O N, G ONDA X, S ZILI I, S ZADOCZKY 5) AKISKAL HS, P LACIDI GF, M AREMMANI I, S IGNORETTA S, E, P ESTALITY P, R IHMER Z. [Affective temperaments: psy - LIGUORI A, G ERVASI R, M ALLYA G, P UZANTIAN VR . chometric properties of the Hungarian TEMPS-A]. TEMPS-I: delineating the most discriminant traits Psychiatria Hungarica 2006; 21: 147-160. of the cyclothymic, depressive, hyperthymic and 21) FIGUEIRA ML, C AEIRO L, F ERRO A, S EVERINO L, D UARTE PM, irritable temperaments in a nonpatient population. ABREU M, A KISKAL HS, A KISKAL KK . Validation of the J Affect Disord 1998; 51: 7-19. Temperament Evaluation of Memphis, Pisa, Paris

2076 Do affective temperaments belong to the domain of normality?

and San Diego (TEMPS-A): Portuguese-Lisbon ver - 34) AKIYAMA T, T SUDA H, M ATSUMOTO S, M IYAKE Y, K AWAMU - sion. J Affect Disord 2008; 111: 193-203. RA Y, N ODA T, A KISKAL KK, A KISKAL HS . The proposed 22) ALBANESI DE NASETTA S, V AZQUEZ GH . [Socio-demo - factor structure of temperament and personality in graphic aspects of the temperaments evaluation Japan: combining traits from TEMPS-A and MPT. according the Argentine TEMPS-A]. Vertex 2007; J Affect Disord 2005; 85: 93-100. 18: 272-279. 35) WOODRUFF E, G ENARO L T, L ANDEIRA -F ERNANDEZ J, 23) KARAM EG, M NEIMNEH Z, S ALAMOUN M, A KISKAL KK, CHENIAUX E, L AKS J, J EAN -L OUIS G, N ARDI AE, V ERSIANI AKISKAL HS . Psychometric properties of the MC, A KISKAL HS, M ENDLOWICZ MV. Validation of the Lebanese-Arabic TEMPS-A: a national epidemio - Brazilian brief version of the temperament auto- logic study. J Affect Disord 2005; 87: 169-183. questionnaire TEMPS-A: the brief TEMPS-Rio de Janeiro. J Affect Disord 2011; 134: 65-76. 24) BLOINK R, B RIEGER P, A KISKAL HS, M ARNEROS A. Factorial structure and internal consistency of the German 36) VAHIP S, K ESEBIR S, A LKAN M, Y AZICI O, A KISKAL KK, TEMPS-A scale: validation against the NEO-FFI AKISKAL HS . Affective temperaments in clinically-well questionnaire. J Affect Disord 2005; 85: 77-83. subjects in Turkey: initial psychometric data on the TEMPS-A. J Affect Disord 2005; 85: 113-125. 25) AKISKAL HS, A KISKAL KK, (E DS ). TEMPS: Temperament Evaluation of Memphis, Pisa, Paris and San Diego. 37) VAZQUEZ GH, A KISKAL H. [The temperament evalua - Special issue. J Affect Disord 2005; 85: 1-242. tion of the Memphis, Pisa, Paris, and San Diego autoquestionnaire, Argentine version (TEMPS-A 26) PERUGI G, A KISKAL HS, L ATTANZI L, C ECCONI D, M AS - Buenos Aires)]. Vertex 2005; 16: 89-94. TROCINQUE C, P ATRONELLI A, V IGNOLI S, B EMI E. The high prevalence of 'soft' Bipolar (II) features in Atypical 38) MATSUMOTO S, A KIYAMA T, T SUDA H, M IYAKE Y, K AWAMU - Depression. Compr Psychiatry 1998; 39: 63-71. RA Y, N ODA T, A KISKAL KK, A KISKAL HS . Reliability and validity of TEMPS-A in a Japanese non-clinical 27) MAREMMANI I, D ELL 'OSSO L, R OVAI L, P ACINI M, A RD population: application to unipolar and bipolar de - UINO G, M ONTAGNARI A, A BBENANTE D, M AREMMANI AG, pressives. J Affect Disord 2005; 85: 85-92. GIULIO P, A KISKAL K, A KISKAL H. TEMPS-A[p] tempera - ment profile related to professional choice A study in 39) MAREMMANI I, A KISKAL HS, S IGNORETTA S, L IGUORI A, 1548 applicants to become a cadet officer in the Ital - PERUGI G, C LONINGER R. The relationship of Krae - ian air force. J Affect Disord 2010; 124: 314-318. pelian affective temperaments (as measured by TEMPS-I) to the tridimensional personality ques - 28) MAREMMANI I, D ELL 'O SSO L, R OVAI L, A RDUINO G, M ON - tionnaire (TPQ). J Affect Disord 2005; 85: 17-27. TAGNARI A, A BBENANTE D, P OPOVIC D, M AREMMANI AGI, PERUGI G, A KISKAL K, A KISKAL HS. Discriminant and con - 40) SAVILLE P, H OLDSWORTH R, N YFIELD G, C RAMP L, M ABEY W. vergent validity of TEMPS-A[P]. Correlation with Occupational Personality Questionnaire (OPQ32), V. MMPI and the emotional-affective state following a 2. Thames Ditton: SHL Group Ltd, 2010. stressful situation. J Affect Disord 2011; 129: 27-33 . 41) PERRETTA P, A KISKAL HS, N ISITA C, L ORENZETTI C, Z A- 29) ROVAI L, L EONARDI A, B ACCIARDI S, M AREMMANI AGI, CCAGNINI E, D ELLA SANTA M, C ASSANO GB . The high RUGANI F, D ELL 'O SSO L, A KISKAL K, A KISKAL HS, M AREM - prevalence of bipolar II and associated cy - MANI I. Discriminant and convergent validity of clothymic and hyperthymic temperaments in HIV- TEMPS-A[P]. Correlation with OPQ32 following a patients. J Affect Disord 1998; 50: 215-224. stressful situation. J Affect Disord 2012; 42) GOIS C, B ARBOSA A, F ERRO A, S ANTOS AL, S OUSA F, http://dx.doi.org/10.1016/j.jad.2012.06.036. AKISKAL H, A KISKAL K, F IGUEIRA ML . The role of affec - tive temperaments in metabolic control in patients 30) ROVAI L, M AREMMANI AGI, L EONARDI A, B ACCIARDI S, with type 2 diabetes. J Affect Disord 2011; 134: RUGANI F, D ELL 'O SSO L, A KISKAL HS, M AREMMANI I. TEMPS-A[P] Temperament Profile Related to 52-58. Professional Choice. Differences between Appli - 43) GOIS C, A KISKAL H, A KISKAL K, F IGUEIRA ML. Depres - cants to Become a Cadet Officer in the Italian Air sive temperament, distress, psychological adjust - Force or Navy. J Affect Disord 2012; Aug 9. [Epub ment and depressive symptoms in type 2 dia - ahead of print], PMID: 22884231. betes. J Affect Disord 2012. 31) VAZQUEZ GH, N ASETTA S, M ERCADO B, R OMERO E, 44) HALL PA, R ODIN GM, V ALLIS TM, P ERKINS BA . The TIFNER S, R AMON MDEL L, G ARELLI V, B ONIFACIO A, consequences of anxious temperament for dis - AKISKAL KK, A KISKAL HS . Validation of the TEMPS-A ease detection, self-management behavior, and Buenos Aires: Spanish psychometric validation of quality of life in Type 2 diabetes mellitus. J Psy - affective temperaments in a population study of chosom Res 2009; 67: 297-305. Argentina. J Affect Disord 2007; 100: 23-29. 45) AKISKAL HS . New insights into the nature and het - 32) KREBS MO, K AZES M, O LIE JP, L OO H, A KISKAL K, erogeneity of mood disorders. J Clin Psychiatry AKISKAL H. The French version of the validated 1989; 50 Suppl: 6-10; discussion 11-12. short TEMPS-A: the temperament evaluation of 46) AKISKAL HS, C ASSANO GB, M USETTI L, P ERUGI G, T UNDO Memphis, Pisa, Paris and San Diego. J Affect Dis - A, M IGNANI V. Psychopathology, temperament, and ord 2006; 96: 271-273. past course in primary major depressions. 1. Re - 33) KARAM EG, M NEIMNEH ZN, S ALAMOUN MM, A KISKAL HS, view of evidence for a bipolar spectrum. Psy - AKISKAL KK . Suitability of the TEMPS-A for population- chopathology 1989; 22: 268-277. based studies: ease of administration and stability of 47) CASSANO GB, A KISKAL HS, M USETTI L, P ERUGI G, S ORIANI affective temperaments in its Lebanese version. J A, M IGNANI V. Psychopathology, temperament and Affect Disord 2007; 98: 45-53. paste course in Primary Major Depression. Toward

2077 L. Rovai, A.G.I. Maremmani, F. Rugani, S. Bacciardi, M. Pacini , L. Dell’Osso, H.S. Akiskal, I. Maremmani

a redifinition of Bipolarity with a new semistructured 61) MAREMMANI I, M AREMMANI AGI, R UGANI F, R OVAI L, interview for depression. Psychopathology 1989; II PACINI M, B ACCIARDI S, D ELTITO J, D ELL 'O SSO L, A KISKAL 22: 278-288. HS . Clinical presentations of substance abuse in 48) AKISKAL HS . The bipolar spectrum: new concepts in bipolar heroin addicts at time of treatment entry. classification and diagnosis. In: Grinspoon L, editor. Ann Gen Psychiatry 2012; 11. Psychiatry update: the American Psychiatric Asso - 62) MAREMMANI I, P ACINI M, P ERUGI G. Addictive disor - ciation Annual Review Vol 11. Washington, DC: ders, bipolar spectrum and the impulsive link: American Psychiatric Press 1983. p. 271-292. The psychopathology of a self-regenerating 49) AKISKAL HS, H ANTOUCHE EG, A LLILAIRE JF. Bipolar II pathway. Heroin Addict Relat Clin Probl 2005; 7: with and without cyclothymic temperament: "dark" 33-46 . and "sunny" expressions of soft bipolarity. J Affect 63) MAREMMANI I, P ACINI M, P ERUGI G, A KISKAL HS . Addic - Disord 2003; 73: 49-57. tion and Bipolar Spectrum: Dual Diagnosis with a 50) PERUGI G, T ONI C, M AREMMANI I, T USINI G, R AMACCIOT - common substrate? Addictive Disorders and TI S, M ADIA A, F ORNARO M, A KISKAL HS. The influ - Their Treatment 2004; 3: 156-164. ence of affective temperaments and psy - 64) MAREMMANI I, P ACINI M, P ERUGI G, D ELTITO J, A KISKAL chopathological traits on the definition of bipolar HS . Cocaine abuse and the bipolar spectrum in disorder subtypes: A study on Bipolar I Italian Na - 1090 heroin addicts: clinical observations and a tional sample. J Affect Disord 2010. proposed pathophysiologic model. J Affect Disord 51) KESEBIR S, V AHIP S, A KDENIZ F, Y UNCU Z, A LKAN M, 2008; 106: 55-61. AKISKAL H. Affective temperaments as measured 65) MAREMMANI I, P ERUGI G, P ACINI M, A KISKAL HS . Toward by TEMPS-A in patients with bipolar I disorder a unitary perspective on the bipolar spectrum and and their first-degree relatives: a controlled study. substance abuse: opiate addiction as a paradigm. J Affect Disord 2005; 85: 127-133. J Affect Disord 2006; 93: 1-12. 52) ROTTIG D, R OTTIG S, B RIEGER P, M ARNEROS A. Tem - 66) MAREMMANI I, P ACINI M, P OPOVIC D, R OMANO A, perament and personality in bipolar I patients with MAREMMANI AG, P ERUGI G, D ELTITO J, A KISKAL K, and without mixed episodes. J Affect Disord 2007; AKISKAL H. Affective temperaments in heroin addic - 104: 97-102. tion. J Affect Disord 2009; 117: 186-192. 53) AKISKAL HS, H ANTOUCHE EG, B OURGEOIS ML, A ZORIN 67) PACINI M, M AREMMANI I, V ITALI M, S ANTINI P, R OMEO M, JM, S ECHTER D, A LLILAIRE JF, L ANCRENON S, F RAUD J P, CECCANTI M. Affective temperaments in alcoholic CHATENET -D UCHENE L. Gender, temperament, and the clinical picture in dysphoric mixed mania: find - patients. Alcohol 2009; 43: 397-404. ings from a French national study (EPIMAN). J Af - 68) CAMACHO A, A KISKAL HS . Proposal for a bipolar- fect Disord 1998; 50: 175-186. stimulant spectrum: temperament, diagnostic vali - 54) PERUGI G, M AREMMANI I, T ONI C, M ADARO D, M ATA B, dation and therapeutic outcomes with mood stabi - AKISKAL HS . The contrasting influence of depres - lizers. J Affect Disord 2005; 85: 217-230. sive and hyperthymic temperaments on psycho - 69) SIGNORETTA S, M AREMMANI I, L IGUORI A, P ERUGI G, A KISKAL metrically derived manic subtypes. Psychiatry HS . Affective temperament traits measured by Res 2001; 101: 249-258. TEMPS-I and emotional-behavioral problems in clin - 55) PERUGI G, M USETTI L, S IMONINI E, P IAGENTINI F, C ASSANO ically-well children, adolescents, and young adults. J GB, A KISKAL HS . Gender mediated clinical features of Affect Disord 2005; 85: 169-180 . depressive illness. The importance of temperamen - 70) THOMAS A, C HESS S. Temperament and develop - tal differences. Br J Psychiatry 1990; 157: 835-841. ment. New York, NY: Bunner/Mazel 1977. 56) PERUGI G, T ONI C, P ASSINO MC, A KISKAL KK, K APRINIS S, 71) VAZQUEZ GH, K AHN C, S CHIAVO CE, G OLDCHLUK A, AKISKAL HS . Bulimia nervosa in atypical depression: HERBST L, P ICCIONE M, S AIDMAN N, R UGGERI H, S ILVA the mediating role of cyclothymic temperament. J A, L EAL J, B ONETTO GG, Z ARATIEGUI R, P ADILLA E, V I- Affect Disord 2006; 92: 91-97. LAPRINO JJ, C ALVO M, G UERRERO G, S TREJILEVICH SA, 57) MAREMMANI I, C ANONIERO S, P ACINI M. Methadone CETKOVICH -B AKMAS MG, A KISKAL KK, A KISKAL HS . dose and retention in treatment of heroin addicts Bipolar disorders and affective temperaments: a with Bipolar I Disorder comorbidity. Preliminary Re - national family study testing the "endopheno - sults. Heroin Addict Relat Clin Probl 2000; 2: 39-46. type" and "subaffective" theses using the 58) MAREMMANI I, C ANONIERO S, P ACINI M, L AZZERI A, TEMPS-A Buenos Aires. J Affect Disord 2008; PLACIDI GF . Opioids and cannabinoids abuse 108: 25-32 . among bipolar patients. Heroin Addict Relat Clin 72) AKISKAL KK, S AVINO M, A KISKAL HS . Temperament Probl 2000; 2: 35-42. profiles in physicians, lawyers, managers, indus - 59) MAREMMANI I, C APONE MR, A GLIETTI M, C ASTROGIOVAN - trialists, architects, journalists, and artists: a study NI P. Heroin dependence and bipolar disorders. in psychiatric outpatients. J Affect Disord 2005; New Trends in Experimental and Clinical Psychia - 85: 201-216. try 1994; X: 179-182. 73) FIGUEIRA ML, C AEIRO L, F ERRO A, C ORDEIRO R, D UARTE 60) MAREMMANI I, L AZZERI A, L OVRECIC M, P LACIDI GF, P ERUGI PM, A KISKAL HS, A KISKAL KK . Temperament in Por - G. Diagnostic and symptomatological features in tuguese university students as measured by chronic psychotic patients according to cannabis use TEMPS-A: implications for professional choice. J status. J Psychoactive Drugs 2004; 36: 235-241. Affect Disord 2010; 123: 30-35.

2078 Do affective temperaments belong to the domain of normality?

74) MAREMMANI I, D ELL 'O SSO L, R OVAI L, A RDUINO G, M AREM - 89) ANGST J, C UI L, S WENDSEN J, R OTHEN S, C RAVCHIK A, MANI AGI, S CHIAVI E, P ERUGI G, A KISKAL K, A KISKAL HS . KESSLER RC, M ERIKANGAS KR . Major depressive dis - Temperamental traits of women applying for a type order with subthreshold bipolarity in the National of job that has been characterized historically by Comorbidity Survey Replication. Am J Psychia - male identity. The military career as case study. J Af - try 2010; 167: 1194-1201. fect Disord 2011; 130: 275-279. 90) MAREMMANI I, P ERUGI G, R OVAI L, M AREMMANI AGI, 75) NEELEMAN J, S YTEMA S, W ADSWORTH M. Propensity to PACINI M, C ANONICO PL, C ARBONATO P, M ENCACCI C, psychiatric and somatic ill-health: evidence from a MUSCETTOLA G, P ANI L, T ORTA R, V AMPINI C, A KISKAL birth cohort. Psychol Med 2002; 32: 793-803. HS . Are social drugs (tobacco, coffee and choco - 76) FORTIER MA, D EL ROSARIO AM, M ARTIN SR, K AIN ZN . late) related to the bipolar spectrum? J Affect Perioperative anxiety in children. Paediatr Disord 2011; 133: 227-233. Anaesth 2010; 20: 318-322. 91) EARLS F. Temperament characteristics and behav - 77) PETTERSSON K, B RANDSTROM S, T OOLANEN G, H ILDINGS - ior problems in three-year-old children. J Nerv SON C, N YLANDER PO . Temperament and character: Ment Dis 1981; 169: 367-373. prognostic factors in whiplash patients? Eur 92) GRAHAM P, R UTTER M, G EORGE S. Temperamental Spine J 2004; 13: 408-414. characteristics as predictors of behavior disor - 78) RAVAJA N, K ELTIKANGAS -J ARVINEN L. Temperament ders in children. Am J Orthopsychiatry 1973; 43: and metabolic syndrome precursors in children: 328-339. a three-year follow-up. Prev Med 1995; 24: 518- 93) RENDE RD . Longitudinal Relations between Tem - 527. perament Traits and Behavioral Syndroms in 79) KILIC A, G ULEC MY, G UL U, G ULEC H. Temperament Middle Childhood. J Am Acad Child Adolesc and character profile of patients with psoriasis. J Psychiatry 1993; 32: 287-290. Eur Acad Dermatol Venereol 2008; 22: 537-542. 94) EARLS F, J UNG K. Temperament and home environ - 80) ANGST J, G AMMA A, B ENAZZI F, A JDACIC V, E ICH D, ment characteristics as causal factors in the early ROSSLER W. Toward a re-definition of subthreshold development of childhood psychopathology. J Am bipolarity: epidemiology and proposed criteria for Acad Child Adolesc Psychiatry 1987; 26: 491-498. bipolar-II, minor bipolar disorders and . 95) GARRISON W, E ARLS F, K INDLON D. Temperament J Affect Disord 2003; 73: 133-146. characteristics in the third year of life and behav - 81) OSHER Y, C LONINGER CR, B ELMAKER RH. TPQ in eu - ioral adjustment at school entry. J Clin Child Psy - thymic manic-depressive patients. J Psychiatr chol 1984; 13: 298-303. Res 1996; 30: 353-357. 96) MAZIADE M, C APERAA P, L APLANTE B, B OUDREAULT M, 82) ENGSTROM C, B RANDSTROM S, S IGVARDSSON S, C LONINGER THIVIERGE J, C OTÉ R, B OUTIN P. Value of difficult R, N YLANDER PO. Bipolar disorder: I. Temperament temperament among 7-years-old in the original and character. J Affect Disord 2004; 82: 131-134. population for predicting psychiatric diagnosis at 83) JANOWSKY DS, M ORTER S, H ONG L, H OWE L. Myers age 12. Am J Psychiatry 1985; 142: 943-946. Briggs Type Indicator and Tridimensional Person - 97) MAZIADE M, C OTÉ R, B ERNIER H, B OUTIN P, T HIVIERGE ality Questionnaire differences between bipolar J. Significance of extreme temperament in infan - patients and unipolar depressed patients. Bipolar cy for clinical status in pre-school years: I. Value Disord 1999; 1: 98-108. of extreme temperament at 4-8 months for pre - 84) CLAES L, B IJTTEBIER P, M ITCHELL JE, DE ZWAAN M, dicting diagnosis at 4,7 years. Br J Psychiatry MUELLER A. The relationship between compulsive 1989; 154: 535-543. buying, eating disorder symptoms, and tempera - 98) MAZIADE M, C OTÉ R, B ERNIER H, B OUTIN P, T HIVIERGE ment in a sample of female students. Compr Psy - J. Significance of extreme temperament in infan - chiatry 2011; 52: 50-55. cy for clinical status in pre-school years: II. Pat - 85) GERRA G, Z AIMOVIC A, R IZZI O, T IMPANO M, Z AMBELLI terns of temperament change and implications U, V ENTIMIGLIA C. Substance abuse among high- for the appearance of disorders. Br J Psychiatry school students: relationship with social coping 1989; 154: 544-551. and temperament. Bulletin of Narcotics UNDCP 99) RETTEW DC, A LTHOFF RR, D UMENCI L, A YER L, H UDZIAK 1999; LI: 75-96. JJ . Latent profiles of temperament and their rela - 86) VAN AMMERS EC, S ELLMAN JD, M ULDER RT . Tempera - tions to psychopathology and wellness. J Am Acad ment and substance abuse in schizophrenia: is Child Adolesc Psychiatry 2008; 47: 273-281. there a relationship? J Nerv Ment Dis 1997; 185: 100) BONACCHI A, M ICCINESI G, G UAZZINI M, R OSSI A, B ACCI 283-288 . S, T OCCAFONDI A, M ARTIRE M, B ELLOTTI L, P ERFETTO R, 87) LIRAUD F, V ERDOUX H. Which temperamental char - CATANZARO P, A DAMO V, C ASCINU S, D ONI L, D I COSTAN - acteristics are associated with substance use in ZO F, R OSSELLI M. Temperament and character traits subjects with psychotic and mood disorders? associated with health-related quality of life in can - Psychiatry Res 2000; 93: 63-72. cer patients. Tumori 2012; 98: 377-384. 88) ANGST J, A DOLFSSON R, B ENAZZI F, G AMMA A, H AN - 101) RADEMAKER AR, K LEBER RJ, G EUZE E, V ERMETTEN E. TOUCHE E, M EYER TD, S KEPPAR P, V IETA E, S COTT J. The Personality dimensions harm avoidance and HCL-32: towards a self-assessment tool for hypo - self-directedness predict the cortisol awakening manic symptoms in outpatients. J Affect Disord response in military men. Biol Psychol 2009; 81: 2005; 88: 217-233. 177-183.

2079