MMPI personality profiles in patients with primary headache syndromes
Volker Pfaffenrath, Josef Hummelsberger, Walter Pöllmann, Holger Kaube, Michael Rath
Cephalalgia Pfaffenrath V, Hummelsberger J, P611mann W, Kaube H, Rath M. MMPI personality profiles in patients with primary headache syndromes. Cephalalgia 1991;11:263-8. Oslo. ISSN 0333-1024 The role of psychological factors in the course of primary headache syndromes is still controversial. Using the Minnesota Multiphasic Personality Inventory (MMPI) we investigated the personality profiles of 434 headache patients (160 migraineurs, 95 with tension type headache, 30 with cluster headache and 149 with combination headache) in accordance with the IHS criteria. In the first three MMPI scales (hypochondria, depression, hysteria) there was a slight increase in T mean values to over 60, but still in the range of two standard deviations of the normal population. There were no statistically significant differences between the four headache groups and between patients with and without analgesic abuse. It was impossible to distinguish headache groups on the basis of their personality profiles by means of reclassification with discriminant analysis. In a cluster analysis, patients with cluster headache showed the highest number (20%) of abnormalities, but also the highest percentage (13%) of completely normal results. Our findings- a cross section analysis of personality profiles-contradict many other MMPI-based studies. • Cluster analysis, cluster headache, discriminant analysis, migraine, MMPL psychopathology, tension type headache Volker Pfaffenrath, Josef Hummelsberger, Holger Kaube, Michael Rath, Neurologische Praxis, Leopoldstr. 59, 8000 München 40, Germany; Walter Pö1lmann, Marianne-Strauss-Klinik, Milchberg 21, 8137 Berg, Germany; Correspondence to Volker Pfaffenrath; Accepted 4 November 1991 Nowadays, there is no longer any doubt that migraine is an organic disease. The role that psychological factors play in the triggering and course of this headache syndrome is controversial, as seen in the unfortunately contradictory-results of psychological tests. Junkerius (1), as early as in the 18th century, described suppressed anger as a cause of headache, and in this century the first systematic investigations of the personality of headache patients were undertaken during the 1930s. H. G. Wolff (2) described the characteristics of migraineurs as increased uncertainty, consciousness bordering on pedantry, inflexibility, emotional rigidity, sexual inhibitions and dissatisfaction with their personal living circumstances. Wolff himself saw the limitations of his results, as some of the patients investigated showed personality features that were neither pathognomonic for migraineurs nor restricted to this syndrome. Friedman (3), too, denied a migraine specific type, though he observed a tendency to over-anxiety, compulsiveness and depression in the over 7000 migraineurs he examined. They very often suppressed their feelings and overemphasized rational thinking. In muscle contraction headache-in today's terminology, according to IHS criteria (4), "tension type headache associated with disorder of the pericranial muscles"-anxiety and problems with the environment are considered as causes of increased tension of the pericranial muscles (5, 6). Underlying conflicts are dependency, sexual problems and a lacking of ability to cope with aggressive impulses. Dhopesh (7) found in patients with tension type headache seen in casualty: 58% with raised anxiety, 21% with depression and 10 % even with psychotic symptoms. The Minnesota Multiphasic Personality Inventory (MMPI) is the most widely used multidimensional personality test in the Anglo-Saxon world (8). An increase in the first three MMPI scales (HD = hypochondria, D = depression, HY = hysteria) was frequently seen in migraineurs (9-12), but also in patients with tension type headache (13-17) and cluster headache (18). In other studies, the existence of abnormal personality structures was doubted for migraineurs and cluster headache patients (13-15). The aim of this study was to investigate personality profiles with the MMPI in several defined headache syndromes (migraine, tension type headache, cluster headache and the combination of migraine and tension type headache), according to the IHS criteria (4), to compare the results with previous studies by means of advanced statistical analysis and thereby to evaluate the validity of the MMPI in headache research. Material and methods
Patients were included in the study prospectively when they attended a specialized outpatient department for their headache. An MMPI profile was obtained in 490 patients, out of which 56 could not be used: 29 patients had not answered or poorly answered too many questions, 27 other patients
suffered from rare headache syndromes (e.g. chronic paroxysmal hemicrania) where statistically relevant results could not be expected due to the very small group numbers. The MMPI scores of 434 patients could therefore be analysed. The mean age of 137 men and 297 women was 40.2 ± 12.3 (14-81) years. According to the IHS criteria, 160 patients (120 women and 40 men) with a mean age of 38.7 ± 11.1 (14-70) years suffered from migraine without aura, 95 patients (49 women and 46 men) with a mean age of 39.7 ± 14.5 (14-72) years suffered from tension type headache and 30 patients (12 women and 18 men) with a mean age of 46.3 ± 15.7 (20-81) years had cluster headache. We considered 149 patients (116 women and 33 men) with a mean age of 41.1 ± 10.9 (21-71) years as a separate group, in whom according to the IHS criteria a migraine without aura was accompanied by a tension type headache (referred to earlier as combination headache). In all patients a complete history was taken and a neurological examination was performed. The consumption of analgesics or ergot-amine preparations on 5-7 days a week was considered as abuse of painkillers. Patients who could not answer the test because of language or other comprehension problems, or who refused a psychological examination, were excluded from the study. We applied the German version of the MMPI as a questionnaire according to Engel and Kunze (19), in which patients rate the statements as applicable (right) or inapplicable (wrong). The answers were transferred onto machine-readable "Optical-Mark-Reader-Cards" (OMR) which, by means of a scanner, were stored in an HP9839A computer. With special software (19), first raw (R) and then transformed values (T) of the MMPI were computed, in which the T-values were calculated from the means and standard deviations of the scales; 40 to 60 T is considered the normal range, 50 is the normal mean value and every 10 T is defined as the equivalent of one standard deviation. Values between 61 and 70 T represent slight deviations of personality, although 13% of the normal population still lie in this range (2nd standard deviation). T-values over 70 are certainly pathological. The tests are considered invalid when 16 or more items were not answered or both possibilities were marked. Statistical analysis
By means of the programs of Berkeley Medical Department (BMD) (20) the following computations were performed on the raw (R) and transformed (T) values: Comparison of the scale mean values A comparison of the scale mean values of the single headache groups with formulation of an MMPI group profile was performed. Patients with and without analgesic abuse were analysed separately. Discriminant analysis A discriminant analysis was performed with the BMD program 7.M (20) to reconstruct the real group membership with as few variables as possible. Variables were combined and for each of them an F-ratio as discriminant criterion was computed that would allow evaluation of its discriminant power. Then, depending on the size of F, the most discriminant items were chosen. F-values higher than 3.996 were neglected. In the discriminant analysis (D1) the grouping variable was the kind of headache, the dependent variables were all transformed (T) values of the MMPI standard scales. An evaluation of the quality of the discriminant analysis was performed by so-called reclassification, in which, depending on the calculated scores, diagnostic groups are redefined and compared with the original ones. In the second discriminant analysis (D2) new variables S1 and S2 derived from the MMPI scales (10) were defined as: S1=(HY+PT)/2 and S2= (HD + D + MF)/3 - (SC + SI)/2 where HY is defined as hysteria, PT as psychasthenia, HD as hypochondria, D as depression, MF as male/female, SC as schizophrenia and SI as social introversion. With the help of these scale combinations the authors (10) found specific psychological abnormalities in a group of 50 migraineurs in comparison to 42 patients with psychosomatic disorders and 45 healthy subjects. A distinction between different groups can be particularly successful if the numbers of cases are similar and if the group differences with respect to items are marked. In the group of patients with cluster headache the first condition is not fulfilled because of the small number of cases. The latter criteria are not met in the group with combination headache due to the similarity of symptoms. Therefore the discriminant analysis was computed separately for the groups with migraine and tension type headache (group A) and also for all groups together (group B). Cluster analysis With the aid of the BMD program 2.M (20) we attempted to identify patients who scored abnormally in the psychological testing. As criterion for abnormality in the MMPI scales a new variable S3 = (HD + D + HY)/3 was defined here. Headache patients with test-psychologically normal scores (S3 < 50) were compared with patients with marked
deviations (S3 > 75). It was the aim of the cluster analysis to demonstrate possible clusters of abnormal patients in the different headache groups or to identify subgroups. Results
Comparison of the scale mean values The T mean values of the four headache groups examined are shown in Fig. 1, the means, standard deviations according to sex in Table 1. Mean values of the validity scales (?, L, F, K) are not presented; they were completely within normal limits (? was below 30, others below 55). For patients with migraine, tension headache, cluster headache and combination headache there is a slight increase in the T-values in the hypochondria (HD) and hysteria (HY) scale to values over 60, which still lies within the range of two standard deviations of the normal population. The most marked increase is found in patients with tension type headache with T-values in hypochondria of 67.4 and 67.7 in hysteria, followed by the group with combination headache with T-values of 66.9 in hypochondria as well as 66.9 in hysteria. In migraine and cluster headache the T-values for hypochondria and hysteria lie between 62.1 and 65.2. The T-values of the depression scale lie in all four headache groups in the upper limit of normal; in the remaining scales the values are normal and exclusively under 60. In the hypochondria and hysteria scale the T-values are elevated by 2-4 for men compared to women. Men with tension type headache have T-values of 68.8 for hypochondria and 69.4 for hysteria, women of only 66.3 and 66.0 respectively. Only in the cluster headache group is the T-value of hysteria (with 67.1) in women higher than in men (with 63.0). In the other scales, there is no sex-specific difference in T-values which are in the normal range. Altogether the differences between the four investigated headache groups are extremely small and are of no statistically significant relevance. Table 2 shows the age dependency of the T-values of the four headache groups. It can be seen that the T-values of the scales hypochondria, depression, hysteria and psychasthenia rise with age. Thus, the T-mean for 41-50-year-old migraineurs for hypochondria is 61.0, for depression 57.6 and for hysteria 60.8; in the over 60-year-olds 75.8, 64.8 and 74.8, respectively. Apart from a rise in the T-mean in the social introversion scale to 65.0 in patients with combination headache, the other scales show no age dependence. Group comparison of patients with and without analgesic abuse is shown in Fig. 2. Values of validity scales not presented were completely normal. In 434 patients 52 abused analgesics or ergot preparations. The mean age of these patients of 43.5 ± 10.5 (26-69) is somewhat higher than in patients without analgesic abuse with a mean age of 39.6 ± 12.4 (14-81) years. The percentage of women with 71% is as high as in the total group with 68%. The most pronounced abuse is found in patients with tension type headache (15.8%) followed by combination headache (13.4%) and migraine (10.6%). Analgesic or ergot abuse is not seen in patients with cluster headache. In the comparison of both groups there is a slight increase in the T-mean values among the 52 patients
Table 1. Sex specific difference of MMPI-T mean values (standard scales). Diagnosis M TH TH + M CH T-value (SD) m/f m/f m/f m/f HD 62.1 (10.1) 67.4 (12.1) 66.9 (10.8) 63.9 (14.1) 63.6/61.5 68.8/66.3 68.8/64.1 64.5/62.4 D 57.8 (10.9) 59.9 (11.5) 59.9 (10.8) 58.1 (12.9) 56.2/58.4 59.9/60.1 59.6/59.9 57.0/58.9 HY 63.1 (11.5) 67.7 (11.2) 66.9 (10.2) 65.2 (13.2) 63.3/63.0 69.4/66.0 68.5/65.8 63.0/67.1 PP 52.5 (9.4) 52.9 (11.0) 52.2 (11.2) 55.3 (13.2) 52.2/52.5 52.3/53.5 58.8/55.3 53.4/51.7 MF 44.7 (9.8) 45.7 (10.8) 45.7 (9.7) 50.4 (10.0) 49.5/43.2 49.3/42.4 50.5/44.1 53.5/40.2 PA 47.5 (10.5) 50.6 (12.2) 58.7 (11.2) 53.2 (12.5) 47.6/47.5 47.5/53.5 46.0/54.3 50.9/44.0 PT 53.2 (9.6) 57.2 (11.7) 55.9 (11.5) 53.0 (12.2) 57.3/51.8 59.9/55.7 59.6/54.1 54.3/57.9 SC 51.5 (8.4) 55.3 (11.1) 53.4 (11.4) 54.0 (10.4) 53.9/50.6 52.7/55.6 54.3/57.2 52.7/53.4 MA 47.4 (10.4) 48.3 (10.1) 48.4 (10.8) 52.7 (9.1) 47.6/47.3 46.5/49.9 48.6/43.4 51.7/42.3 SI 53.3 (10.0) 55.0 (11.8) 55.0 (11.0) 50.1 (10.0) 54.7/52.8 54.8/55.2 56.6/57.3 51.9/57.9
Mean 65.1 (11.1) 59.0 (11.6) 65.1 (10.3) 52.7 (10.4) 45.9 (10.0) 48.9 (11.3) 55.1 (11.0) 52.9 (10.3) 48.3 (10.4) 54.0 (10.8)
HD = hypochondria, D = depression, HY = hysteria, PP = psychopathia, MF = male/female, PA = paranoia, PT = psychasthenia, SC = schizophrenia, MA = mania, SI = social introversion. M = migraine, TH = tension type headache, CH = cluster headache. Table 2. Age dependency of T mean values in migraine, tension type headache, cluster headache and migraine + tension type headache. Age (yrs) Scales 14-17 18-30 31-40 41-50 5160 >60 Mean HD M 62.0 60.3 61.5 61.0 64.1 75.8* 62.1 TH 49.6 60.3 66.0 66.1 76.4* 71.6 67.4 CH 61.0 58.9 65.0 61.0 71.8* 63.9 M + TH 65.8 64.0 66.3 71.3 76.0* 66.9 D M 63.5 55.2 57.4 57.6 60.2 64.8* 57.8 TH 53.6 55.4 58.3 59.8 67.2* 58.0 59.9 CH 60.0 51.9 59.4 51.7 59.4 58.1 M + TH 54.4 58.1 59.1 64.3 76.7* 59.9 HY M 69.2 63.2 62.2 60.8 65.7 74.8* 63.1 TH 59.0 63.2 68.4 64.1 75.3* 69.6 67.7 CH 74.8* 60.4 63.2 60.3 69.4 65.2 M + TH 66.8 64.6 66.7 64.5 68.7* 66.9 PT M 57.0 51.1 54.8 52.8 51.9 62.2* 53.2 TH 56.7 51.1 55.9 57.4 60.2* 53.4 57.2 CH 52.5 47.1 57.2 50.0 58.8* 53.0 M + TH 53.5 54.6 55.2 59.8 63.3* 55.1 *
= Maximum values in headache group.
HD = hypochondria, D = depression, HY = hysteria, PT = psychasthenia, M = migraine, CH = cluster headache, TH = tension type headache. with drug abuse in the hypochondria, depression and hysteria scales; however, this is not statistically significant. Discriminant analysis By application of D1 in group A a reclassification was possible for migraine in 59.4%, and for tension type headache in 53.7%; i.e. it is practically impossible to distinguish between the two groups on the basis of their personality features, especially taking into consideration that similar percentages could be expected in purely random grouping. In group B, reclassification was successful in migraineurs in 54.5%, in tension type headache in 49.5%, in combination headache in 5.4% and in cluster headache in 10%. Again, it is evident that it is impossible to distinguish between migraine and tension type headache and that reclassification is incorrect for the other headache syndromes. Analysis D2, with the factors according to Huber et al. (10), leads to correct reclassification for migraine in 33.1%, for tension type headache in 41.1% and for cluster headache in 40%. A better discrimination was not obtainable with this method. Cluster analysis Altogether 51 of the 434 patients examined (11.8 %) showed abnormalities (S3 > 75) in the MMPI; 31 (7.1%) were completely normal (S3
Volker Pfaffenrath, Josef Hummelsberger, Walter Pöllmann, Holger Kaube, Michael Rath
Cephalalgia Pfaffenrath V, Hummelsberger J, P611mann W, Kaube H, Rath M. MMPI personality profiles in patients with primary headache syndromes. Cephalalgia 1991;11:263-8. Oslo. ISSN 0333-1024 The role of psychological factors in the course of primary headache syndromes is still controversial. Using the Minnesota Multiphasic Personality Inventory (MMPI) we investigated the personality profiles of 434 headache patients (160 migraineurs, 95 with tension type headache, 30 with cluster headache and 149 with combination headache) in accordance with the IHS criteria. In the first three MMPI scales (hypochondria, depression, hysteria) there was a slight increase in T mean values to over 60, but still in the range of two standard deviations of the normal population. There were no statistically significant differences between the four headache groups and between patients with and without analgesic abuse. It was impossible to distinguish headache groups on the basis of their personality profiles by means of reclassification with discriminant analysis. In a cluster analysis, patients with cluster headache showed the highest number (20%) of abnormalities, but also the highest percentage (13%) of completely normal results. Our findings- a cross section analysis of personality profiles-contradict many other MMPI-based studies. • Cluster analysis, cluster headache, discriminant analysis, migraine, MMPL psychopathology, tension type headache Volker Pfaffenrath, Josef Hummelsberger, Holger Kaube, Michael Rath, Neurologische Praxis, Leopoldstr. 59, 8000 München 40, Germany; Walter Pö1lmann, Marianne-Strauss-Klinik, Milchberg 21, 8137 Berg, Germany; Correspondence to Volker Pfaffenrath; Accepted 4 November 1991 Nowadays, there is no longer any doubt that migraine is an organic disease. The role that psychological factors play in the triggering and course of this headache syndrome is controversial, as seen in the unfortunately contradictory-results of psychological tests. Junkerius (1), as early as in the 18th century, described suppressed anger as a cause of headache, and in this century the first systematic investigations of the personality of headache patients were undertaken during the 1930s. H. G. Wolff (2) described the characteristics of migraineurs as increased uncertainty, consciousness bordering on pedantry, inflexibility, emotional rigidity, sexual inhibitions and dissatisfaction with their personal living circumstances. Wolff himself saw the limitations of his results, as some of the patients investigated showed personality features that were neither pathognomonic for migraineurs nor restricted to this syndrome. Friedman (3), too, denied a migraine specific type, though he observed a tendency to over-anxiety, compulsiveness and depression in the over 7000 migraineurs he examined. They very often suppressed their feelings and overemphasized rational thinking. In muscle contraction headache-in today's terminology, according to IHS criteria (4), "tension type headache associated with disorder of the pericranial muscles"-anxiety and problems with the environment are considered as causes of increased tension of the pericranial muscles (5, 6). Underlying conflicts are dependency, sexual problems and a lacking of ability to cope with aggressive impulses. Dhopesh (7) found in patients with tension type headache seen in casualty: 58% with raised anxiety, 21% with depression and 10 % even with psychotic symptoms. The Minnesota Multiphasic Personality Inventory (MMPI) is the most widely used multidimensional personality test in the Anglo-Saxon world (8). An increase in the first three MMPI scales (HD = hypochondria, D = depression, HY = hysteria) was frequently seen in migraineurs (9-12), but also in patients with tension type headache (13-17) and cluster headache (18). In other studies, the existence of abnormal personality structures was doubted for migraineurs and cluster headache patients (13-15). The aim of this study was to investigate personality profiles with the MMPI in several defined headache syndromes (migraine, tension type headache, cluster headache and the combination of migraine and tension type headache), according to the IHS criteria (4), to compare the results with previous studies by means of advanced statistical analysis and thereby to evaluate the validity of the MMPI in headache research. Material and methods
Patients were included in the study prospectively when they attended a specialized outpatient department for their headache. An MMPI profile was obtained in 490 patients, out of which 56 could not be used: 29 patients had not answered or poorly answered too many questions, 27 other patients
suffered from rare headache syndromes (e.g. chronic paroxysmal hemicrania) where statistically relevant results could not be expected due to the very small group numbers. The MMPI scores of 434 patients could therefore be analysed. The mean age of 137 men and 297 women was 40.2 ± 12.3 (14-81) years. According to the IHS criteria, 160 patients (120 women and 40 men) with a mean age of 38.7 ± 11.1 (14-70) years suffered from migraine without aura, 95 patients (49 women and 46 men) with a mean age of 39.7 ± 14.5 (14-72) years suffered from tension type headache and 30 patients (12 women and 18 men) with a mean age of 46.3 ± 15.7 (20-81) years had cluster headache. We considered 149 patients (116 women and 33 men) with a mean age of 41.1 ± 10.9 (21-71) years as a separate group, in whom according to the IHS criteria a migraine without aura was accompanied by a tension type headache (referred to earlier as combination headache). In all patients a complete history was taken and a neurological examination was performed. The consumption of analgesics or ergot-amine preparations on 5-7 days a week was considered as abuse of painkillers. Patients who could not answer the test because of language or other comprehension problems, or who refused a psychological examination, were excluded from the study. We applied the German version of the MMPI as a questionnaire according to Engel and Kunze (19), in which patients rate the statements as applicable (right) or inapplicable (wrong). The answers were transferred onto machine-readable "Optical-Mark-Reader-Cards" (OMR) which, by means of a scanner, were stored in an HP9839A computer. With special software (19), first raw (R) and then transformed values (T) of the MMPI were computed, in which the T-values were calculated from the means and standard deviations of the scales; 40 to 60 T is considered the normal range, 50 is the normal mean value and every 10 T is defined as the equivalent of one standard deviation. Values between 61 and 70 T represent slight deviations of personality, although 13% of the normal population still lie in this range (2nd standard deviation). T-values over 70 are certainly pathological. The tests are considered invalid when 16 or more items were not answered or both possibilities were marked. Statistical analysis
By means of the programs of Berkeley Medical Department (BMD) (20) the following computations were performed on the raw (R) and transformed (T) values: Comparison of the scale mean values A comparison of the scale mean values of the single headache groups with formulation of an MMPI group profile was performed. Patients with and without analgesic abuse were analysed separately. Discriminant analysis A discriminant analysis was performed with the BMD program 7.M (20) to reconstruct the real group membership with as few variables as possible. Variables were combined and for each of them an F-ratio as discriminant criterion was computed that would allow evaluation of its discriminant power. Then, depending on the size of F, the most discriminant items were chosen. F-values higher than 3.996 were neglected. In the discriminant analysis (D1) the grouping variable was the kind of headache, the dependent variables were all transformed (T) values of the MMPI standard scales. An evaluation of the quality of the discriminant analysis was performed by so-called reclassification, in which, depending on the calculated scores, diagnostic groups are redefined and compared with the original ones. In the second discriminant analysis (D2) new variables S1 and S2 derived from the MMPI scales (10) were defined as: S1=(HY+PT)/2 and S2= (HD + D + MF)/3 - (SC + SI)/2 where HY is defined as hysteria, PT as psychasthenia, HD as hypochondria, D as depression, MF as male/female, SC as schizophrenia and SI as social introversion. With the help of these scale combinations the authors (10) found specific psychological abnormalities in a group of 50 migraineurs in comparison to 42 patients with psychosomatic disorders and 45 healthy subjects. A distinction between different groups can be particularly successful if the numbers of cases are similar and if the group differences with respect to items are marked. In the group of patients with cluster headache the first condition is not fulfilled because of the small number of cases. The latter criteria are not met in the group with combination headache due to the similarity of symptoms. Therefore the discriminant analysis was computed separately for the groups with migraine and tension type headache (group A) and also for all groups together (group B). Cluster analysis With the aid of the BMD program 2.M (20) we attempted to identify patients who scored abnormally in the psychological testing. As criterion for abnormality in the MMPI scales a new variable S3 = (HD + D + HY)/3 was defined here. Headache patients with test-psychologically normal scores (S3 < 50) were compared with patients with marked
deviations (S3 > 75). It was the aim of the cluster analysis to demonstrate possible clusters of abnormal patients in the different headache groups or to identify subgroups. Results
Comparison of the scale mean values The T mean values of the four headache groups examined are shown in Fig. 1, the means, standard deviations according to sex in Table 1. Mean values of the validity scales (?, L, F, K) are not presented; they were completely within normal limits (? was below 30, others below 55). For patients with migraine, tension headache, cluster headache and combination headache there is a slight increase in the T-values in the hypochondria (HD) and hysteria (HY) scale to values over 60, which still lies within the range of two standard deviations of the normal population. The most marked increase is found in patients with tension type headache with T-values in hypochondria of 67.4 and 67.7 in hysteria, followed by the group with combination headache with T-values of 66.9 in hypochondria as well as 66.9 in hysteria. In migraine and cluster headache the T-values for hypochondria and hysteria lie between 62.1 and 65.2. The T-values of the depression scale lie in all four headache groups in the upper limit of normal; in the remaining scales the values are normal and exclusively under 60. In the hypochondria and hysteria scale the T-values are elevated by 2-4 for men compared to women. Men with tension type headache have T-values of 68.8 for hypochondria and 69.4 for hysteria, women of only 66.3 and 66.0 respectively. Only in the cluster headache group is the T-value of hysteria (with 67.1) in women higher than in men (with 63.0). In the other scales, there is no sex-specific difference in T-values which are in the normal range. Altogether the differences between the four investigated headache groups are extremely small and are of no statistically significant relevance. Table 2 shows the age dependency of the T-values of the four headache groups. It can be seen that the T-values of the scales hypochondria, depression, hysteria and psychasthenia rise with age. Thus, the T-mean for 41-50-year-old migraineurs for hypochondria is 61.0, for depression 57.6 and for hysteria 60.8; in the over 60-year-olds 75.8, 64.8 and 74.8, respectively. Apart from a rise in the T-mean in the social introversion scale to 65.0 in patients with combination headache, the other scales show no age dependence. Group comparison of patients with and without analgesic abuse is shown in Fig. 2. Values of validity scales not presented were completely normal. In 434 patients 52 abused analgesics or ergot preparations. The mean age of these patients of 43.5 ± 10.5 (26-69) is somewhat higher than in patients without analgesic abuse with a mean age of 39.6 ± 12.4 (14-81) years. The percentage of women with 71% is as high as in the total group with 68%. The most pronounced abuse is found in patients with tension type headache (15.8%) followed by combination headache (13.4%) and migraine (10.6%). Analgesic or ergot abuse is not seen in patients with cluster headache. In the comparison of both groups there is a slight increase in the T-mean values among the 52 patients
Table 1. Sex specific difference of MMPI-T mean values (standard scales). Diagnosis M TH TH + M CH T-value (SD) m/f m/f m/f m/f HD 62.1 (10.1) 67.4 (12.1) 66.9 (10.8) 63.9 (14.1) 63.6/61.5 68.8/66.3 68.8/64.1 64.5/62.4 D 57.8 (10.9) 59.9 (11.5) 59.9 (10.8) 58.1 (12.9) 56.2/58.4 59.9/60.1 59.6/59.9 57.0/58.9 HY 63.1 (11.5) 67.7 (11.2) 66.9 (10.2) 65.2 (13.2) 63.3/63.0 69.4/66.0 68.5/65.8 63.0/67.1 PP 52.5 (9.4) 52.9 (11.0) 52.2 (11.2) 55.3 (13.2) 52.2/52.5 52.3/53.5 58.8/55.3 53.4/51.7 MF 44.7 (9.8) 45.7 (10.8) 45.7 (9.7) 50.4 (10.0) 49.5/43.2 49.3/42.4 50.5/44.1 53.5/40.2 PA 47.5 (10.5) 50.6 (12.2) 58.7 (11.2) 53.2 (12.5) 47.6/47.5 47.5/53.5 46.0/54.3 50.9/44.0 PT 53.2 (9.6) 57.2 (11.7) 55.9 (11.5) 53.0 (12.2) 57.3/51.8 59.9/55.7 59.6/54.1 54.3/57.9 SC 51.5 (8.4) 55.3 (11.1) 53.4 (11.4) 54.0 (10.4) 53.9/50.6 52.7/55.6 54.3/57.2 52.7/53.4 MA 47.4 (10.4) 48.3 (10.1) 48.4 (10.8) 52.7 (9.1) 47.6/47.3 46.5/49.9 48.6/43.4 51.7/42.3 SI 53.3 (10.0) 55.0 (11.8) 55.0 (11.0) 50.1 (10.0) 54.7/52.8 54.8/55.2 56.6/57.3 51.9/57.9
Mean 65.1 (11.1) 59.0 (11.6) 65.1 (10.3) 52.7 (10.4) 45.9 (10.0) 48.9 (11.3) 55.1 (11.0) 52.9 (10.3) 48.3 (10.4) 54.0 (10.8)
HD = hypochondria, D = depression, HY = hysteria, PP = psychopathia, MF = male/female, PA = paranoia, PT = psychasthenia, SC = schizophrenia, MA = mania, SI = social introversion. M = migraine, TH = tension type headache, CH = cluster headache. Table 2. Age dependency of T mean values in migraine, tension type headache, cluster headache and migraine + tension type headache. Age (yrs) Scales 14-17 18-30 31-40 41-50 5160 >60 Mean HD M 62.0 60.3 61.5 61.0 64.1 75.8* 62.1 TH 49.6 60.3 66.0 66.1 76.4* 71.6 67.4 CH 61.0 58.9 65.0 61.0 71.8* 63.9 M + TH 65.8 64.0 66.3 71.3 76.0* 66.9 D M 63.5 55.2 57.4 57.6 60.2 64.8* 57.8 TH 53.6 55.4 58.3 59.8 67.2* 58.0 59.9 CH 60.0 51.9 59.4 51.7 59.4 58.1 M + TH 54.4 58.1 59.1 64.3 76.7* 59.9 HY M 69.2 63.2 62.2 60.8 65.7 74.8* 63.1 TH 59.0 63.2 68.4 64.1 75.3* 69.6 67.7 CH 74.8* 60.4 63.2 60.3 69.4 65.2 M + TH 66.8 64.6 66.7 64.5 68.7* 66.9 PT M 57.0 51.1 54.8 52.8 51.9 62.2* 53.2 TH 56.7 51.1 55.9 57.4 60.2* 53.4 57.2 CH 52.5 47.1 57.2 50.0 58.8* 53.0 M + TH 53.5 54.6 55.2 59.8 63.3* 55.1 *
= Maximum values in headache group.
HD = hypochondria, D = depression, HY = hysteria, PT = psychasthenia, M = migraine, CH = cluster headache, TH = tension type headache. with drug abuse in the hypochondria, depression and hysteria scales; however, this is not statistically significant. Discriminant analysis By application of D1 in group A a reclassification was possible for migraine in 59.4%, and for tension type headache in 53.7%; i.e. it is practically impossible to distinguish between the two groups on the basis of their personality features, especially taking into consideration that similar percentages could be expected in purely random grouping. In group B, reclassification was successful in migraineurs in 54.5%, in tension type headache in 49.5%, in combination headache in 5.4% and in cluster headache in 10%. Again, it is evident that it is impossible to distinguish between migraine and tension type headache and that reclassification is incorrect for the other headache syndromes. Analysis D2, with the factors according to Huber et al. (10), leads to correct reclassification for migraine in 33.1%, for tension type headache in 41.1% and for cluster headache in 40%. A better discrimination was not obtainable with this method. Cluster analysis Altogether 51 of the 434 patients examined (11.8 %) showed abnormalities (S3 > 75) in the MMPI; 31 (7.1%) were completely normal (S3
