Cluster headache course over ten years in 189 patients
Gian Camillo Manzoni, Giuseppe Micieli, Franco Granella, Cristina Tassorelli, Carla Zanferrari, Anna Cavallini
Greppi Prize: Senior. Cephalalgia Manzoni GC, Micieli G, Granella F, Tassorelli C, Zanferrari C, Cavallini A. Cluster headache-course over ten years in 189 patients. Cephalalgia 1991;11:169-74. Oslo. ISSN 0333-1024 One-hundred-and-eighty-nine cluster headache patients, referred to Parma and Pavia Headache Centres between 1976 and 1986 with a disease duration of over 10 years, were interviewed about the course of cluster headache. They were classified as episodic (n = 140) or chronic (n = 49) cluster headache patients on the basis of course during the year of onset. Episodic patients showed the following outcome: maintenance of an episodic form (primary episodic form) in 80.7% of cases, shift towards a chronic form (secondary chronic form) in 12.9% and shift towards an intermediate pattern ("combined" form) in 6.4%. In chronic patients, cluster headache was still chronic (primary chronic form) at the moment of observation in 52.4% of cases, while it turned into an episodic form ("secondary" episodic form) in 32.6% and into a "combined" form in 14.3%. Nineteen patients (10%) had had no attacks for at least three years at the moment of examination. We can conclude from our data that: cluster headache is a disease of long duration, perhaps lifelong; episodic cluster headache tends to worsen; chronic cluster headache may easily turn into a better prognostic episodic form; prophylactic drugs are unable to induce recovery. The following factors seem related to a poor outcome: a later onset, the male gender and a disease duration of over 20 years for the episodic forms. Gian Camillo Manzoni, Franco Granella, Carla Zanferrari, University Centre for Adaptive Disorders and Headache, Department of Neurology, University of Parma; Giuseppe Micieli, Cristina Tassorelli, Anna Cavallini, Department of Neurology, University of Pavia; Correspondence to Professor Gian Camillo Manzoni, Centro Cefalee, Clinica Neurologica, Via del Quartiere, 4, 43100 Parma, Italy; Accepted 11 June 1991 In the last decade the clinical features of cluster headache have been so accurately studied and described (1, 2) that the illness can now be reasonably classified in the field of primary headaches as an autonomous nosographic disease (3). Nevertheless, we have a poor knowledge of the course of cluster headache, of the frequency and duration of its remission periods, and of possible complete recoveries and variations that therapies currently used and life habit changes may induce to the natural history of the disease. We carried out this study on patients who have been suffering with both episodic and chronic cluster headache for at least 10 years in order to clarify some of these cluster headache features. Patients and methods
Patients suffering from cluster headache over a period longer than 10 years were selected from those who had been followed up at the Headache Centres of the Parma and Pavia Neurologic Departments from 1976 to 1986. We contacted them by telephone or by letter, asking them to fill in a previously prepared questionnaire. For patients no longer alive we interviewed both the near relatives, who could tell us all about the course of the disease, and the family doctor to obtain some information on the cause of death. Patients unable to reach our Centres (because of living too far from the centres or because of work or private reasons) were interviewed by telephone. In all patients diagnosis of cluster headache had been originally made according to the recommendations of the Ad Hoc Committee of 1962 (4). In all cases a new diagnosis was made according to the criteria of the IHS Classification Committee (3). Patients in whom diagnosis could not be made in accordance with the IHS criteria were not examined. To name the painful crisis, the active periods and the headache-free intervals, we used the terms recently proposed by Ekbom (5), i.e. attacks, cluster periods and remission periods, respectively. Statistical comparisons were made by means of the chi-square test and Student's t-test. Results
Of the 221 patients who were contacted, 200 (90.5%) collaborated in the study; 146 (73%) came to the Headache Centres in person, and 54 (27%) were interviewed by telephone. The remaining 21 did not collaborate for a variety of reasons: impossible to find (9 cases; 4.1%), explicit refusal (6 cases; 2.7%), death (6 cases; 2.7%). For patients no longer alive, information on the course of the disease and on the cause of death was gathered. Eleven patients (4.9%)
were not examined because they did not fulfil all the IHS diagnostic criteria, the attacks lasting longer than normal and the site of the pain being atypical. The remaining 189 (85.4%) submitted to the questionnaire. One-hundred-and-fifty-eight of them (83.6%) were male and 31 (16.4%) were female (male: female ratio 5.1:1). Characteristics of the population examined On the basis of cluster headache course in the first (onset) year, patients were grouped as episodic (140 cases; 74.1%-118 males, 22 females) and chronic (49 cases; 25.9%-40 males, 9 females). The mean age was 45.7 ± 11.8 (range 21-80). Age at onset and duration of the disease In our population the disease occurs on average much earlier in females than it does in males (F: 23.3 ± 9.0 years; M: 29.6 ± 10.9 years; p = 0.0006), both in the episodic (F: 23.1 ± 9.3 years, M: 28.7 ± 10.3 years; p = 0.03) and the chronic forms (F: 23.8 ± 9.7 years; M: 32.8 ± 12.6 years; ns), although the chronic form has no statistical significance because of the limited number of cases. In particular, cluster headache first appeared in only 4 females over the age of 30 (12.9%), whereas it first appeared in the 29% of males over the age of 30. No significant difference in the age of onset has been noticed between the episodic and chronic forms. Mean duration of cluster headache at the time of examination was 17.4 years (range 10-54 years); in 59 patients (31.2%) the disease had a course longer than 20 years. Twenty-nine of the patients (15.3%) were over 60 when examined. In the group of patients aged over 70 only one had developed cluster headache when over the age of 60. Outcome of episodic cluster headache Most patients with episodic cluster headache had an episodic pattern (primary episodic form) for the first year, whereas in a significant percentage of them (about 13%) there was a shift to a chronic form later on (secondary chronic form or chronic form evolved from episodic form according to IHS nomenclature) (Table 1). In addition, some cases showed a cluster headache with a course characterized by alternating active phases either longer or shorter than one year ("combined" form). The frequency of shifts towards the chronic form seemed to increase with duration of the disease. In fact, in 20.5% of patients with a course longer than 20 years there was a shift towards the chronic form, whereas the percentage in those with a course shorter than 20 years was 9.4. The mean age at onset was 27.1 ± 10.06 years in Table 1. Outcome of CH patients with an episodic form during the onset year. F M Total n (%) n (%) n (%) Maintain a primary episodic form 20 (90.9) 93 (78.8) 113 (80.7) Secondary chronic form 2 (9.1) 16 (13.6) 18 (12.9) Acquire a combined form 09 (7.6) 9 (6.4) Total 22 (100.0) 118 (100.0) 140 (100.0)
patients with primary episodic form, while in those with the secondary chronic form it was 34.9 ± 12.39 years ( p = 0.037). Moreover, a slight though not significant prevalence of secondary chronic form was noticed in males. Sixty (53.9%) of the episodic cluster headache patients always had the cluster periods in the same month ± 2 ("cyclic" form-F: 16/20; M: 44/93;p= 0.016), even though cluster periods did not always recur every year. On the other hand, in 53 cases (46.1%) the cluster periods appeared to be randomly distributed ("erratic" form-F: 4/20; M: 49/93). Characteristics of active periods The mean frequency of active periods in primary episodic cluster headache patients did not vary during the course of the disease. Mean duration was longer in the last year than in. the year of onset in 53 cases (48.2%); it remained unchanged in 48 (43.6%), shortened in 7 (6.4%), and was variable in 2 (1.8%) (longer vs shorter p = 0.0001). One-hundred-and-fourteen patients (60.3%) experienced mini-bouts (periods lasting 3-6 days) (6), while 90 patients (47.6%) also complained of isolated attacks. Outcome of chronic cluster headache Among patients with chronic cluster headache during the first year, 25 maintained a chronic pattern (primary chronic form), while in 16 there was a shift towards the episodic form ("secondary" episodic form) (Table 2). All the cases we identified as "secondary" episodic had evolved from a primary chronic form. Nine out of 16 patients with a "secondary" episodic form exhibited a "cyclic" pattern of cluster periods; 7 out of 49 chronic cluster headache patients showed a "combined" pattern. In 4 patients classified as primary chronic, cluster headache was characterized by active periods always longer than one year separated by remission periods which always lasted more than 14 days, and often months or even years ("remittent" chronic form). In all cases
Table 2. Outcome of cluster headache patients with a chronic form during the onset year. F M Total n (%) n (%) n (%) Maintain a primary chronic form 4 (44.4) 22 (55.0) 26 (53.1) Acquire a secondary episodic form 016 (40.0) 16 (32.6) Acquire a combined form 5 (55.6) 2 (5.0) 7 (14.3) Total 9 (100.0) 40 (100.0) 49 (100.0)
these remission periods were spontaneous, uninfluenced by either pharmacological or surgical therapies. The use of a prophylactic treatment (lithium) in many cases (9/16-56.2%) favoured the evolution of a chronic form into the "secondary" episodic form, but in other cases the evolution was spontaneous. Other factors seem to relate with the evolution of a chronic form into an episodic form: earlier cluster headache onset (26.0 ± 9.0 years for "secondary" episodic); 34.9 ± 12.4 years for primary chronic: p = 0.024), the duration of the disease (the chronic form persists only in the 46.6% of patients with a course longer than 20 years) and, perhaps, male sex. Remission periods Nineteen patients (10.0%) had been free of attacks for at least 3 years at the time of examination. Remission periods longer than 3 years had been shown by 35 patients (18.5%). In 9 patients (4.8%) the remission period had lasted for at least 5 years at the time of the examination; in 3 patients (1.6%) for more than 10 years. In one case the symptoms remitted after resection of the supraorbital nerve of the side involved. Prolonged remission periods appeared to be more frequent in the episodic than in the chronic form (Table 3) and do not seem to be related to the duration of the cluster headache. In fact, of 59 patients with a course longer than 20 years, only 6 (10.2%) had been in "actual" remission for at least 3 years. Furthermore, the disease did not seem to disappear with increasing age: of 29 patients over 60, only 4 (13.8%) had been in "actual" remission for at least 3 years. Changes of life habits did not seem to favour longer remission periods; no relationship was found between the reduction in smoking (13% of cases) or its withdrawal (10% of cases) and onset of a remission period. Nevertheless, 4 out of 19 patients who stopped drinking showed an "actual" remission: all of them were moderate
Table 3. Remission periods longer than three years. Present Previous n (%)* n (%)* Form of disease Primary episodic 13 (11.5) 27 (24.0)° Secondary episodic 2 (12.5) 2 (12.5) Primary chronic 2 (7.7) 0 -° Secondary chronic 02 (11.1) Combined 2 (12.5) 4 (25.0) Total *
19 (10.0)
35 (18.5)
Percentage concerns the total number of patients suffering from the single CH forms.
° p = 0.019. drinkers (alcohol consumed was lower than 50 g/ day). Fifteen out of 35 patients (42.9%) with previous remission periods longer than 3 years were having pharmacological treatment at the beginning of the remission (13 lithium carbonate, 2 verapamil). On the other hand, only 2 out of 12 patients in actual remission had the remission starting during methysergide treatment. The duration of remission periods, both spontaneous and "induced" by treatment, were extremely variable. One patient remained attack-free for 30 years and then relapsed. The average duration of longer remission periods in primary episodic cluster headache patients was 44.5 ± 53.37 months (range 6-360 months). Personal history and cause of death The most significant data concerning the personal history of our patients refer to the high prevalence of peptic ulcer (22 cases; 11.6%). Angina pectoris occurred in 4 patients and myocardial infarction in 2. Three patients were affected by cancer (1 pulmonary, 1 laryngeal, 1 pharyngeal). No patients suffered from auto-immune diseases. Of the 6 patients who were no longer alive, 2 had died of myocardial infarction, 3 of cancer (2 pulmonary, 1 bladder) and 1 had committed suicide. Characteristics of attacks When compared to the year of onset, in the primary episodic form the number of patients with 1 attack or less per day had decreased in the last year, whereas the number of patients with more frequent attacks had increased (>2 attacks per day: 10.4% at onset, 31.8% during the last year; p= 0.0001). On the other hand, in the primary chronic form, patients with less frequent attacks increased (
Gian Camillo Manzoni, Giuseppe Micieli, Franco Granella, Cristina Tassorelli, Carla Zanferrari, Anna Cavallini
Greppi Prize: Senior. Cephalalgia Manzoni GC, Micieli G, Granella F, Tassorelli C, Zanferrari C, Cavallini A. Cluster headache-course over ten years in 189 patients. Cephalalgia 1991;11:169-74. Oslo. ISSN 0333-1024 One-hundred-and-eighty-nine cluster headache patients, referred to Parma and Pavia Headache Centres between 1976 and 1986 with a disease duration of over 10 years, were interviewed about the course of cluster headache. They were classified as episodic (n = 140) or chronic (n = 49) cluster headache patients on the basis of course during the year of onset. Episodic patients showed the following outcome: maintenance of an episodic form (primary episodic form) in 80.7% of cases, shift towards a chronic form (secondary chronic form) in 12.9% and shift towards an intermediate pattern ("combined" form) in 6.4%. In chronic patients, cluster headache was still chronic (primary chronic form) at the moment of observation in 52.4% of cases, while it turned into an episodic form ("secondary" episodic form) in 32.6% and into a "combined" form in 14.3%. Nineteen patients (10%) had had no attacks for at least three years at the moment of examination. We can conclude from our data that: cluster headache is a disease of long duration, perhaps lifelong; episodic cluster headache tends to worsen; chronic cluster headache may easily turn into a better prognostic episodic form; prophylactic drugs are unable to induce recovery. The following factors seem related to a poor outcome: a later onset, the male gender and a disease duration of over 20 years for the episodic forms. Gian Camillo Manzoni, Franco Granella, Carla Zanferrari, University Centre for Adaptive Disorders and Headache, Department of Neurology, University of Parma; Giuseppe Micieli, Cristina Tassorelli, Anna Cavallini, Department of Neurology, University of Pavia; Correspondence to Professor Gian Camillo Manzoni, Centro Cefalee, Clinica Neurologica, Via del Quartiere, 4, 43100 Parma, Italy; Accepted 11 June 1991 In the last decade the clinical features of cluster headache have been so accurately studied and described (1, 2) that the illness can now be reasonably classified in the field of primary headaches as an autonomous nosographic disease (3). Nevertheless, we have a poor knowledge of the course of cluster headache, of the frequency and duration of its remission periods, and of possible complete recoveries and variations that therapies currently used and life habit changes may induce to the natural history of the disease. We carried out this study on patients who have been suffering with both episodic and chronic cluster headache for at least 10 years in order to clarify some of these cluster headache features. Patients and methods
Patients suffering from cluster headache over a period longer than 10 years were selected from those who had been followed up at the Headache Centres of the Parma and Pavia Neurologic Departments from 1976 to 1986. We contacted them by telephone or by letter, asking them to fill in a previously prepared questionnaire. For patients no longer alive we interviewed both the near relatives, who could tell us all about the course of the disease, and the family doctor to obtain some information on the cause of death. Patients unable to reach our Centres (because of living too far from the centres or because of work or private reasons) were interviewed by telephone. In all patients diagnosis of cluster headache had been originally made according to the recommendations of the Ad Hoc Committee of 1962 (4). In all cases a new diagnosis was made according to the criteria of the IHS Classification Committee (3). Patients in whom diagnosis could not be made in accordance with the IHS criteria were not examined. To name the painful crisis, the active periods and the headache-free intervals, we used the terms recently proposed by Ekbom (5), i.e. attacks, cluster periods and remission periods, respectively. Statistical comparisons were made by means of the chi-square test and Student's t-test. Results
Of the 221 patients who were contacted, 200 (90.5%) collaborated in the study; 146 (73%) came to the Headache Centres in person, and 54 (27%) were interviewed by telephone. The remaining 21 did not collaborate for a variety of reasons: impossible to find (9 cases; 4.1%), explicit refusal (6 cases; 2.7%), death (6 cases; 2.7%). For patients no longer alive, information on the course of the disease and on the cause of death was gathered. Eleven patients (4.9%)
were not examined because they did not fulfil all the IHS diagnostic criteria, the attacks lasting longer than normal and the site of the pain being atypical. The remaining 189 (85.4%) submitted to the questionnaire. One-hundred-and-fifty-eight of them (83.6%) were male and 31 (16.4%) were female (male: female ratio 5.1:1). Characteristics of the population examined On the basis of cluster headache course in the first (onset) year, patients were grouped as episodic (140 cases; 74.1%-118 males, 22 females) and chronic (49 cases; 25.9%-40 males, 9 females). The mean age was 45.7 ± 11.8 (range 21-80). Age at onset and duration of the disease In our population the disease occurs on average much earlier in females than it does in males (F: 23.3 ± 9.0 years; M: 29.6 ± 10.9 years; p = 0.0006), both in the episodic (F: 23.1 ± 9.3 years, M: 28.7 ± 10.3 years; p = 0.03) and the chronic forms (F: 23.8 ± 9.7 years; M: 32.8 ± 12.6 years; ns), although the chronic form has no statistical significance because of the limited number of cases. In particular, cluster headache first appeared in only 4 females over the age of 30 (12.9%), whereas it first appeared in the 29% of males over the age of 30. No significant difference in the age of onset has been noticed between the episodic and chronic forms. Mean duration of cluster headache at the time of examination was 17.4 years (range 10-54 years); in 59 patients (31.2%) the disease had a course longer than 20 years. Twenty-nine of the patients (15.3%) were over 60 when examined. In the group of patients aged over 70 only one had developed cluster headache when over the age of 60. Outcome of episodic cluster headache Most patients with episodic cluster headache had an episodic pattern (primary episodic form) for the first year, whereas in a significant percentage of them (about 13%) there was a shift to a chronic form later on (secondary chronic form or chronic form evolved from episodic form according to IHS nomenclature) (Table 1). In addition, some cases showed a cluster headache with a course characterized by alternating active phases either longer or shorter than one year ("combined" form). The frequency of shifts towards the chronic form seemed to increase with duration of the disease. In fact, in 20.5% of patients with a course longer than 20 years there was a shift towards the chronic form, whereas the percentage in those with a course shorter than 20 years was 9.4. The mean age at onset was 27.1 ± 10.06 years in Table 1. Outcome of CH patients with an episodic form during the onset year. F M Total n (%) n (%) n (%) Maintain a primary episodic form 20 (90.9) 93 (78.8) 113 (80.7) Secondary chronic form 2 (9.1) 16 (13.6) 18 (12.9) Acquire a combined form 09 (7.6) 9 (6.4) Total 22 (100.0) 118 (100.0) 140 (100.0)
patients with primary episodic form, while in those with the secondary chronic form it was 34.9 ± 12.39 years ( p = 0.037). Moreover, a slight though not significant prevalence of secondary chronic form was noticed in males. Sixty (53.9%) of the episodic cluster headache patients always had the cluster periods in the same month ± 2 ("cyclic" form-F: 16/20; M: 44/93;p= 0.016), even though cluster periods did not always recur every year. On the other hand, in 53 cases (46.1%) the cluster periods appeared to be randomly distributed ("erratic" form-F: 4/20; M: 49/93). Characteristics of active periods The mean frequency of active periods in primary episodic cluster headache patients did not vary during the course of the disease. Mean duration was longer in the last year than in. the year of onset in 53 cases (48.2%); it remained unchanged in 48 (43.6%), shortened in 7 (6.4%), and was variable in 2 (1.8%) (longer vs shorter p = 0.0001). One-hundred-and-fourteen patients (60.3%) experienced mini-bouts (periods lasting 3-6 days) (6), while 90 patients (47.6%) also complained of isolated attacks. Outcome of chronic cluster headache Among patients with chronic cluster headache during the first year, 25 maintained a chronic pattern (primary chronic form), while in 16 there was a shift towards the episodic form ("secondary" episodic form) (Table 2). All the cases we identified as "secondary" episodic had evolved from a primary chronic form. Nine out of 16 patients with a "secondary" episodic form exhibited a "cyclic" pattern of cluster periods; 7 out of 49 chronic cluster headache patients showed a "combined" pattern. In 4 patients classified as primary chronic, cluster headache was characterized by active periods always longer than one year separated by remission periods which always lasted more than 14 days, and often months or even years ("remittent" chronic form). In all cases
Table 2. Outcome of cluster headache patients with a chronic form during the onset year. F M Total n (%) n (%) n (%) Maintain a primary chronic form 4 (44.4) 22 (55.0) 26 (53.1) Acquire a secondary episodic form 016 (40.0) 16 (32.6) Acquire a combined form 5 (55.6) 2 (5.0) 7 (14.3) Total 9 (100.0) 40 (100.0) 49 (100.0)
these remission periods were spontaneous, uninfluenced by either pharmacological or surgical therapies. The use of a prophylactic treatment (lithium) in many cases (9/16-56.2%) favoured the evolution of a chronic form into the "secondary" episodic form, but in other cases the evolution was spontaneous. Other factors seem to relate with the evolution of a chronic form into an episodic form: earlier cluster headache onset (26.0 ± 9.0 years for "secondary" episodic); 34.9 ± 12.4 years for primary chronic: p = 0.024), the duration of the disease (the chronic form persists only in the 46.6% of patients with a course longer than 20 years) and, perhaps, male sex. Remission periods Nineteen patients (10.0%) had been free of attacks for at least 3 years at the time of examination. Remission periods longer than 3 years had been shown by 35 patients (18.5%). In 9 patients (4.8%) the remission period had lasted for at least 5 years at the time of the examination; in 3 patients (1.6%) for more than 10 years. In one case the symptoms remitted after resection of the supraorbital nerve of the side involved. Prolonged remission periods appeared to be more frequent in the episodic than in the chronic form (Table 3) and do not seem to be related to the duration of the cluster headache. In fact, of 59 patients with a course longer than 20 years, only 6 (10.2%) had been in "actual" remission for at least 3 years. Furthermore, the disease did not seem to disappear with increasing age: of 29 patients over 60, only 4 (13.8%) had been in "actual" remission for at least 3 years. Changes of life habits did not seem to favour longer remission periods; no relationship was found between the reduction in smoking (13% of cases) or its withdrawal (10% of cases) and onset of a remission period. Nevertheless, 4 out of 19 patients who stopped drinking showed an "actual" remission: all of them were moderate
Table 3. Remission periods longer than three years. Present Previous n (%)* n (%)* Form of disease Primary episodic 13 (11.5) 27 (24.0)° Secondary episodic 2 (12.5) 2 (12.5) Primary chronic 2 (7.7) 0 -° Secondary chronic 02 (11.1) Combined 2 (12.5) 4 (25.0) Total *
19 (10.0)
35 (18.5)
Percentage concerns the total number of patients suffering from the single CH forms.
° p = 0.019. drinkers (alcohol consumed was lower than 50 g/ day). Fifteen out of 35 patients (42.9%) with previous remission periods longer than 3 years were having pharmacological treatment at the beginning of the remission (13 lithium carbonate, 2 verapamil). On the other hand, only 2 out of 12 patients in actual remission had the remission starting during methysergide treatment. The duration of remission periods, both spontaneous and "induced" by treatment, were extremely variable. One patient remained attack-free for 30 years and then relapsed. The average duration of longer remission periods in primary episodic cluster headache patients was 44.5 ± 53.37 months (range 6-360 months). Personal history and cause of death The most significant data concerning the personal history of our patients refer to the high prevalence of peptic ulcer (22 cases; 11.6%). Angina pectoris occurred in 4 patients and myocardial infarction in 2. Three patients were affected by cancer (1 pulmonary, 1 laryngeal, 1 pharyngeal). No patients suffered from auto-immune diseases. Of the 6 patients who were no longer alive, 2 had died of myocardial infarction, 3 of cancer (2 pulmonary, 1 bladder) and 1 had committed suicide. Characteristics of attacks When compared to the year of onset, in the primary episodic form the number of patients with 1 attack or less per day had decreased in the last year, whereas the number of patients with more frequent attacks had increased (>2 attacks per day: 10.4% at onset, 31.8% during the last year; p= 0.0001). On the other hand, in the primary chronic form, patients with less frequent attacks increased (
