A nationwide survey of migraine in France: prevalence and clinical features in adults

Patrick Henry1, Philippe Michel1,2, Bruno Brochet1, Jean François Dartigues1,2, Sylvie Tison2, Roger Salamon2, and the GRIM3

Service de neurologie, Hôpital Pellegrin, Bordeaux, France1; Unité INSERM 330, Université de Bordeaux II, Bordeaux, France2; Groupe de Recherche Interdisciplinaire sur la Migraine, Paris, France3

Cephalalgia

Henry P, Michel P, Brochet B, Dartigues JF, Tison S, Salamon R, and the GRIM. A nationwide survey of migraine in France: prevalence and clinical features in adults. Cephalalgia 1992;12:229-37. Oslo. ISSN 0333-1024 In November 1990 a nationwide survey of migraine was conducted in France on a representative sample of residents aged 15 years and older. The diagnosis of migraine was based on the International Headache Society (IHS) classification. In a previous study, we validated a diagnostic algorithm which classifies headache sufferers as IHS migraine, "borderline" migraine, possible migraine and non-migrainous headache. The overall prevalence of migraine patients with the IHS criteria in the present study was 8.1%; another 4% were classified as "borderline" migraine, which we in fact considered as definite migraine. Age, gender and occupation were found to be risk factors for migraine. Neither frequency and duration of attacks nor length of time of disease differed with gender. Expressed intensity of attacks, however, was greater in females. • Diagnosis, epidemiology, migraine, prevalence, risk factors P Henry, Service de neurologie, Hôpital Pellegrin, 33076 Bordeaux, France. Received 31 December 1991, accepted 27 March 1992

Although migraine is one of the most common pathological disorders, the prevalence of the disease remains controversial. Prevalence rates of migraine vary considerably in the literature from 3.1% (1) to 26% (2). These differences may be explained by selection bias in the studied sample and by a lack of validity of the diagnostic criteria. Population-based studies are generally void of selection bias, but the accuracy of diagnosis is often unsatisfactory (3, 4). Most of the studies use the Gowers definition (5), the Waters definition (6), or the criteria of the Ad Hoc Committee of the National Institutes of Health (NIH) (7). In 1980, we elaborated and validated a standardized questionnaire and a diagnostic algorithm which can be applied by lay interviewers in epidemiological studies (8). The prevalence of migraine was estimated in this way at 10.1% in a sample of 3,000 workers. In 1990, we adapted this algorithm, taking into account the classification of the International Headache Society (IHS) (9). The validity of the diagnosis of migraine by lay interviewers was computed and compared with that of a senior neurologist on 96 headache sufferers, including 59 migraines and 37 other headaches, detected during their systematic annual occupational medical check-up. The headache sufferers were screened by the question "Are you subject to headache?", which is understood in France as "Have you suffered repeatedly from headaches during the past few years?", to rule out the acute headache related to local disorders, and evolutive headache. The questionnaire and the algorithm (see Appendix) used the criteria A, B, C and D of the IHS criteria for migraine. Sensitivity for the diagnosis of migraine was 95% and specificity 78% (10). The aim of the present study was to estimate the prevalence of migraine, to determine the sociodemographic risk factors (age, gender, occupation and region) and to analyse the clinical features of the disease by using our diagnostic procedure on a representative sample of the general population of the whole of France. Methods

Sample The epidemiological survey was carried out by the Institut Français d'Opinion Français (IFOP), a national public opinion poll agency. A nationwide representative sample of 4,204 subjects in France aged 15 years and older was constituted according to the quota method (11), a stratified non-random sampling method. The French population was stratified by gender, age (< 25, 25 to 34, 35 to 49, 50 to 64, > 64 years), occupational categories defined by the Institut National de la Statistique et des Etudes Economiques (INSEE) (12) and the population of place of residence ( < 2,000, 2,000 to 19,999, 20,000 to 99,999, ³ 100,000 inhabitants). The quota was the number of subjects per stratification. Over the whole of France, each of the 2,000 professional interviewers of the IFOP was responsible for fulfilling the quota in their defined area, during a "random route" (13). In each household they selected and interviewed one subject of the required age, gender and occupation.

Data collection The data were collected in the home, during face-to-face interviews, in two stages. First, in the screening stage, the headache sufferers were detected in the total sample of 4,204 subjects by two questions: "Are you subject to headache?" and "Is the frequency of your headaches very rare, rare, frequent, or very frequent?" Age, gender, occupation and place of residence of the subjects were noted. Among the 1,371 (35%) subjects complaining of headaches, 1,320 had rare, frequent or very frequent headache and were eligible for the second stage. Subjects with very rare headaches were excluded. Among these 1,320 subjects, 1,000 (75.8%) were randomly chosen for a second interview within one month of screening. One hundred and eleven (11.1%) refused to participate and 56 (5.6%) were in fact found at the second interview not to suffer from headaches. Thus, 833 subjects were included in the final sample. They were reinterviewed at their home by the same investigator. During a 90 min interview using the validated questionnaire for the diagnosis of migraine, data were obtained on the severity of the headaches, on sociodemographic and economic status, on quality of life along with use of medical services. The validated algorithm classified the headache sufferers within four groups: (i) IHS migraine strictly corresponding to IHS criteria (ii) "borderline" migraine, where either the duration of attacks was between 2 and 4 h, or there was photophobia or phonophobia, (iii) possible migraine, where at least two IHS criteria were missing (this group consisted mainly of "mixed" headache, i.e. patients with migraine attacks and tension headache), and (iv) non-migraine headache. IHS and "borderline" migraine were considered as definite migraine, as their sensitivity and specificity for the diagnosis of migraine are similar (10). No attempt was made to divide cases into those with and those without aura as this differentiation was found by the validation study to be non-reliable (10) because migraine sufferers often cannot distinguish unilateral sensory disturbances from nocturnal acroparaesthesia, and scintillating scotoma from photophobia. Finally, clinical features of migraine were also noted: frequency of migraine attacks ( < 5 attacks per year, 5 to 10/year, 1/month, 1/fortnight, 1/week, 2 to 3/week, >/week); intensity of pain (self-defined items: mild, moderate, severe, very severe); mean duration of attacks in hours; length of time of the disease in years. Data analysis Taking into account the possible lack of representation of age and gender in the final sample, an estimated prevalence of migraine in France was computed: for each gender, the age-specific observed prevalence was multiplied by the proportion of the French population in the specific age groups. That is, mi Pi = -----M

ni * ----N

where: Pi is the age and gender-specific adjusted prevalence; mi is the number of age and gender-specific migraine sufferers interviewed in the final sample; M is the age and gender-specific number of subjects interviewed in the final sample; ni is the size of the French population in the specific age and gender groups; N is the total French population aged 15 years and older. mi was corrected for the proportion of subjects who met the screening criteria (i.e. were eligible for reinterview) and who were either not sampled (n = 320) or refused to be reinterviewed (n = 111), assuming that these subjects did not differ on their disease status from the 833 who were reinterviewed. Adjustment for age and gender of estimated prevalence by region and occupation was computed using the direct standardization method. The reference population was the population of France. Place of residence was classified into eight regions (North, Paris, Bassin Parisien, West, South-West, Mediterranean, Middle-East, East). The analysis of occupation was limited to six categories (blue-collar workers, white-collar workers, top executives and secondary school teachers, craftsmen and shopkeepers, farmers, nurses, primary school-teachers and middle executives). The standardized estimated prevalence was not computed for the occupational category "housewives, students and unemployed", because the age and gender distributions did not adequately overlap with those of the other categories. A rough estimate of the prevalence was then computed, as for the more detailed classification of occupation. Chi square and adjusted chi square, t test and ANOVA were performed to compare the different distributions. Results

Representativity of the sample The screening sample of 4,204 subjects was not strictly representative of the general population according to age (chi square = 5.5, df = 4, p = 0.02) and occupation (chi square = 118, df = 6, p < 0.001) (Table 1). The gender and residence area distribution for the general population and for the screening sample was not statistically different. The final sample was representative of the headache subjects of the screening sample with respect to age and occupation but not with respect to gender (chi square = 16, df = 1, p < 0.001) and residence area (chi square = 22, df = 7, p = 0.01) (Table 1).

Table 1. Comparison of the distributions of quota variables in the total screening sample (%). French population Eligible for Total aged 15 Final second screening years and sample interview sample older* Total 833 1,320 4,204 43,000,000 Gender Male 29.0 32.8 47.3 47.9 Female 71.0 67.2 52.7 52.1 Age 15 - 25 20.3 21.5 19.3 19.3 25 -34 27.6 24.8 21.8 19.6 35 - 49 26.2 27.1 23.4 23.4 50-64 16.7 16.2 19.6 21.3 ³ 65 9.2 10.4 15.9 16.4 Occupation Farmers 2.2 2.1 3.1 4 Craftsmen, shopkeepers 3.4 2.9 3.7 5 Managers, professionals 3.6 3.4 3.9 5 Nurses, primary school-teachers, middle executives 10.6 9.9 10.4 10 White-collars 21.8 20.8 16.5 15 Blue-collars 12.6 13.2 14.5 19 Inactives, housewives, students 45.9 47.5 47.7 42 Region Paris 12.5 18.5 18.4 18.7 North 7.2 6.6 7.3 7.0 East 10.3 9.7 9.0 9.0 Bassin Parisien 18.7 18.5 18.8 18.1 West 14.6 12.7 12.3 13.1 South-West 12. 1 11. 1 10.8 10.8 South-East 10.9 10.6 11.8 11.7 Mediterranean 13.3 12.5 11.6 11.6 * Institut National de la Statistique et des Etudes Economiques (1989)

Prevalence of migraine Of the 833 cases of the final sample, 230 (27.6%) were classified by the algorithm as IHS migraine, 110 (13.2%) as "borderline" migraine and 283 (34.0%) as possible migraine. Two-hundred-and-ten (25.2%) had non-migraine headache. Estimated prevalence in France of IHS migraine was 8.1% and that of "borderline" migraine 4.0% (Table 2). The prevalence of definite migraine (IHS and "borderline") was thus 12.1%. Definite and possible migraine grouped together had a total estimated prevalence of 22.6%. Prevalence of IHS migraine was 4.0% in men and 11.9% in women. Prevalence of definite migraine was 6.1% in men and 17.6% in women. The female/male ratio was 3.84. The prevalence of migraine was highest between ages 30 and 39 in both genders. Variation of prevalence according to age for females Table 2. Prevalence rates of migraine in the French population aged 15 years and over (%). Males Females Total 95% CI IHS migraine 4.0 11.9 8.1 [6.2-10.0] Borderline migraine 2.1 5.5 4.0 [2.7-5.3] IHS + borderline 6.1 17.6 12.1 [9.7-14.5] migraine (definite migraine) Possible migraine 8.2 12.6 10.5 [ 8.3 - 12.9] Non-migraine 7.3 9.0 8.2 [6.0 - 10.9] headache was found not to be significantly different to that of males, whether using IHS or definite criteria (Table 3).

Table 3. Prevalence rates of IHS and definite migraine according to age and gender (%). IHS migraine Definite migraine Age Males Females Both Males Females Both 15 - 19 4 6 5.0 5 11 8.0 20-29 5 12 8.5 8 18 13.0 30-39 7 19 13.0 9 26 17.5 40-49 3 17 10.0 6 25 15.5 50-59 5 12 8.5 6 19 12.5 60-69 2 5 3.0 2 8 5.0 70 - 79 3 7 5.5 3 8 6.0 > 80 3 3 3.0 3 6 5.0

Variations of standardized prevalence according to place of residence were not significant (chi square = 5.4, df = 7, p = 0.6) (Fig. 1). Standardized definite migraine prevalence was significantly different according to occupation categories (chi square = 54, df =5, p 3/week 0 3 2 Do not know 3 2 2 Total 100 100 100 Intensity of attacks Mild 10 4 7 Moderate 21 15 19 Severe 49 46 48 Very severe 20 35 26 Total 100 100 100 Duration of attacks 2-4 h 29 24 25 4-6 h 24 17 19 6-12 h 12 12 12 About one day 23 24 24 About two days 10 9 9 About three days 1 8 6 Do not know 1 6 5 Total 100 100 100 Length of the disease (years) 0 - 10 40 41 41 11-20 30 32 31 21-30 16 14 15 31-40 7 6 6 41-50 3 4 4 51 - 60 1 3 2 61-70 3 1 1 Total 100 100 100 their assignment, the choice of the sample units to fit into the quota is left to the investigators (13). Here, this potential selection bias was reduced by the use of professional investigators. The quota method does not need the response rate of the screening stage because its only requirement is to fulfil the quota. The quota theory indeed assumes that if the quota variables are relevant, i.e. are highly correlated with headache and migraine, there is no difference between all subjects of a defined stratum for headache and migraine status (13, 14). Here, age and gender, two of the variables of IFOP, were the main risk factors for headache and migraine (15). Last, no complete sampling database was available. The quota method was thus chosen. Eliminating very rare headache excludes subjects who have had one or two attacks in the previous few years, leading to a slight underestimation of prevalence. Nevertheless, this minor part of the migraine population cannot be considered as a Public Health problem. The 8.1% prevalence of IHS migraine is consistent with the findings of the two recent studies based on the IHS definition of migraine, although at the lower range of the results reported by Ensink (16). He found a prevalence of 12% on a 9,000-subject sample representative of the general population of Canada, Great Britain, Belgium, Sweden and Italy. However, this rate is different from one country to another (from 8% in Great Britain to 19% in Italy). Stewart et al. reported a prevalence of frequent migraine at 14.6% for females and 4.8% for males from a "9,500 household sample" (17). The prevalence of migraine reported in previous general population surveys in occidental countries and Israel is between 10 and 26%, and the female to male ratio is between two and three (2, 18-21). Abramson et al. (21) found the lowest prevalence of migraine (10.1%) but also reported a high (25.6%) prevalence of non-migraine headache. We believe that mixed headache, i.e. both migraine and tension headache, was classified by Abramson as non-migraine headache, whereas in our study it was classified in the possible migraine group. The 26% prevalence found by Waters is probably due to a broad definition of migraine (unilateral headache with nausea and warning signs) (2). In our study, the prevalence of the definite and possible migraine was 22.2%, which was not very different from Waters' finding. Our more precise diagnostic tool then gives one possible explanation for the wide variation in prevalence in the literature. In other parts of the world, the prevalence rates seem to be lower than 7% (23-27), except in the most recent study from Thailand, at 12% (28). Risk factors for migraine were found in our study to be age, gender and occupation. In contrast to widespread opinion in medical circles, migraine was found to be more frequent in the broad occupational category constituted by nurses, primary schoolteachers and middle executives than in the category constituted by top executives. Opinion may have been biased as top executives tend to consult more often. In this study, the clinical features were found to be no different with regard to gender, except intensity of pain. Because we did not adjust for differences in age, these results could be confounded by age, as migraine is strongly associated with this variable. Attacks were expressed as more intense by females than by males, which is consistent with previous studies (22, 24, 29, 30). The number of migraine attacks by duration of attacks describes a bimodal curve. This may reflect two distinct populations: the subjects who treated their pain effectively at the onset of attacks and those who did not treat their headache or whose treatment was ineffective. We observed in a validation study that the IHS criteria allowed the diagnosis of a very homoge-

Table 6. Main demographic and clinical features in IHS, borderline, definite, possible and non-migraine groups. Definite migraine Borderline (IHS + borderline) Possible Non-migraine IHS migraine migraine migraine migraine headache Size 230 110 340 283 210 Age 38.8±0.9* 38.8±1.4* 38.8±1.1* 38.1±1.1* 41.2±1.3* Gender % of females 79.9 78.2 79.3 67.5 61.8 Intensity of attacks (%) Mild 3 5 4 6 15 Moderate 19 25 21 44 54 Severe 57 53 55 40 26 Very severe 20 16 19 9 3 Do not know 1 1 1 1 2 Total 100 100 100 100 100 Frequency of attacks (%) < 5/year 5 2 4 6 8 5 to 10/year 13 14 13 16 24 1/month 31 35 32 27 23 1/2 weeks 22 23 23 14 17 1/week 18 14 17 13 11 2 to 3/week 9 5 8 11 7 > 3/week 1 5 2 11 7 Do not know 1 2 1 2 3 Total 100 100 100 100 100 Length of disease (in years) 18.8±1.0* 17.0±1.2* 18.1±1.0* 14.6±1.1* 14.6±1.0* * Mean standard error.

neous group, but were too restrictive for the diagnosis of all cases of migraine in the general population (10). We then developed a diagnostic tool which allows us to distinguish between several migraine groups. The IHS and "borderline" cases were all considered, by the neurologists involved, as definite migraine. This epidemiological study confirmed that the two groups may be the same clinical entity, because they were not significantly different with respect to the main clinical and demographic variables and because the distribution of these variables in the definite migraine group was different from that of the possible and non-migraine groups. In conclusion, the prevalence of migraine in France was calculated at 8.1% or 12.1%, depending on which diagnostic criteria were used (IHS or our own, respectively). The IHS criteria are certainly useful for cross-national studies and clinical trials, but an estimation of the migraine prevalence based on these criteria seems to imply an underestimation of about 33%. Acknowledgements.-This study was supported by a grant from the Glaxo Laboratories, France. References

1.

Witkowska K. Epidemiology of headache and migraine in adult population in Warsaw. Supplementary final report on research program. Psychoneurological Institute, Warsaw 1974

2.

Waters WE. Methodology, principles and analysis of epidemiological studies in headache. Headache 1975;15: I48-9

3.

Blau JN. Towards a definition of migraine headache. Lancet 1984;i:444-5

4.

Olesen J. New international headache classification. Neuroepidemiology 1989;8:53-5

5.

Cowers WR. A manual of disease of the nervous system. Vol 2. Churchill, London, 1888

6.

Waters WE, O'Connor PJ. Epidemiology of headache and migraine in women. J Neurol Neurosurg Psychiatry 1971; 34:148-53

7.

Ad Hoc Committee on the Classification of Headache: classification of headache. JAMA 1962;179:717-18

8.

Salamon R, Henry P, Dartigues JF, et al. Prévalence de la migraine dans une population active en Aquitaine. Méthode d'étude et résultats. Rev Epidemiol Santé Publ 1980;28: 395 -411

9.

Headache Classification Committee of the International Headache Society. Classification and diagnosis criteria for headache disorders, cranial neuralgias and facial pains. Cephalalgia 1988;8(suppl 7):1-96

10.

Henry P, Michel P, Dartigues JF, Tison S, Brochet B, Auriacombe S, Vivares C, Salamon R, and the GRIM.

Migraine prevalence in France. In: Rose FC ed. New advances in headache research. Smith Gordon Comp., 1991: 11-14 11.

Deroo M, Dussaix AM. Pratique et analyse des enquêtes par sondage. Ed PUF Paris, 1980

12.

INSEE. Nomenclature des professions et des catégories socioprofessionnelles. In: INSEE Ed. Paris 1983:IX

13.

Moser CA, Stuart A. An experimental study of quota sampling. J Roy Stat Soc 1953;116:349-405

14.

Gallup News, London, Social Surveys (Gallup Poll) Ltd, no 6, 1964

15.

Liner MS, Stewart WF. Migraine headache: epidemiologic perspectives. Epidemiol Rev 1984;6:107-39

16.

Ensink FM. Epidemiology of migraine. Migraine Media Workshop 7-10 February 1991. Davos, Switzerland

17.

Stewart W, Lipton R, Celentano D, Reed M. The epidemiology of severe migraine headache froma national survey: implications of projections to the US population. Cephalalgia 199I;11(suppl):87-8

18.

Markush RE, Karp HR, Heyman A, et al. Epidemiologic study of migraine symptoms in young women. Neurology 1975;25:430-5

19.

Ziegler DK, Hassanein RS, Couch JR. Characteristics of the headache histories in a non clinic population. Neurology 1977;27:265-9

20.

Schnarch DM, Hunter JE. Migraine incidence in clinical vs non-clinical populations. Psychosomatics 1980;21:314-19

21.

Abramson JH, Hopp C, Epstein LM. Migraine and non-migrainous headaches: a community survey in Jerusalem. J Epidemiol Commun Health 1980;34:188-93

22.

D'Alessandro R, Benassi G, Lenzi PL, Gamberini G, Sacquegna T, De Carolis P, Lugaresi E. Epidemiology of headache in the Republic of San Marino. J Neurol Neurosurg Psychiatry 1988;51:21-7

23.

Levy ML. An epidemiological study of headache in an urban population in Zimbabwe. Headache 1982;23:2-9

24.

Sachs H, Sevilla F, Barberis P, Bolis L, Schoenberg B, Cruz M. Headache in the rural village of Quiroga, Ecuador. Headache 1984;25:190-3

25.

Cheng XM, Ziegler DK, Li SC, Dai KS, Chandra V, Schoenberg BS. A prevalence survey of incapaciting headache in the People's Republic of China. Neurology 1986;36:831-4

26.

Zhao F, Tsay JY, Cheng XM, Wong WJ, Li SC, Yao SX, Chang SM, Schoenberg B. Epidemiology of migraine: a survey in 21 provinces of the People's Republic of China, 1985. Headache 1988;28:558-65

27.

Longe AC, Osuntokun BO. Prevalence of migraine in Udo, a rural community in Southern Nigeria. East Afr Med J 1988;65:621 - 4

28.

Phanthumchinda K, Sithi-Amorn C. Prevalence and clinical features of migraine: a community survey in Bangkok, Thailand. Headache 1989:29:594-7

29.

Waters WE. Community studies of the prevalence of headache. Headache 1970;11:178-86

30.

Waters WE. The Pontypridd headache survey. Headache 1974;25:81-90

Appendix I Interview questions of the validated questionnaire used for algorithm diagnostic of migraine. 1. Are you subject to headache? [(1) Yes; (2) No] 2. Do you suffer from headache every day? [(1) Yes; (2) No; (3) do not know] 3. How long are your headaches usually, without medication? [(1) less than 4 h; (2) between 4 and 72 h; (3) more than 72 h; (4) do not know] 4. What is the usual location of your headaches? [(1) strictly unilateral; (2) alternately in the right and left part of the head; (3) other location; (4) do not know] 5. Pulsating type of headaches? [(1) Yes; (2) No; (3) do not know] 6. Do your headaches inhibit or prevent daily activities? [(1) Yes; (2) No; (3) do not know] 7. Do your headaches get worse during physical activity? [(1) Yes; (2) No; (3) do not know] 8. Are your headaches accompanied by nausea or vomiting? [(1) Yes; (2) No; (3) do not know] 9a. Are your headaches accompanied by photophobia? [(1) Yes; (2) No; (3) do not know] 9b. Are your headaches accompanied by phonophobia? [(1) Yes; (2) No; (3) do not know] 10. Have you had more than 4 attacks in your lifetime? [(1) Yes; (2) No; (3) do not know]

A nationwide survey of migraine in France: prevalence and clinical features in adults. GRIM.

In November 1990 a nationwide survey of migraine was conducted in France on a representative sample of residents aged 15 years and older. The diagnosi...
132KB Sizes 0 Downloads 0 Views