Headache and Cardiovascular Risk Factors: Positive Association With Hypertension

Massimo Cirillo, MD; Davide Stellato, MD; Cinzia Lombardi, PhD; Natale G. De Santo, MD; Vito Covelli, MD

The study analyzes the prevalence of cardiovascular risk factors in 1343 patients with severe headache (399

men and 944 women), aged 15 to 64 years; analyses were controlled for sex, age, and type and frequency of

headache. Prevalence of various forms of headache was different between men and women. Age and days per year

with headache were significantly different among various forms of headache. For men and women with headache,

age directly related to prevalence of hypertension, hypercholesterolemia, and obesity. Due to low prevalence, analy-

ses by age were not done for diabetes mellitus. For cigarette smoking, prevalence was not related to age in men, but

was inversely related to age in women. With control for age, prevalence of cardiovascular risk factors was not sig-

nificantly different among patients with different forms of headache, except for cluster headache.Among men with

cluster headache, prevalence was high for cigarette smoking, but low for hypercholesterolemia. With control for

age, days per year with headache did not relate to prevalence of cardiovascular risk factors except for cigarette

smoking in men. Compared to data for a population sample used as control, patients with headache had higher

prevalence of hypertension in both sexes, independent of age (odds ratio 1.51,95% confidence interval I .28 to 1.80);

the difference between patients with headache and the control population was lower with increasing age. The high

prevalence of hypertension among patients with headache was not due to overweight. The data indicate that

headache is significantly associated with hypertension, but not with other cardiovascular risk factors.

Key words: headache, blood pressure, cholesterol, cigarette smoking

(Headache 1999;39:409-416)

Headache is a common problem in medical practice.

The etiopathogenesis is uncertain in most cases, and the

clinical presentation varies from mild to severe forms.lJ

The possible association of headache with cardiovascular

disorders is an old issue. In 1913, T.C. Janeway reported

that hypertension and headache were associated.3 After

that classic paper, clinical and epidemiologic studies

reported conflicting data on the prevalence of hyperten-

sion and other cardiovascular risk factors in patients with

From the Schools of Medicine, Chair of Nephrology, Department of Pediatrics, Second University of Naples (Drs. Cirillo, Stellato, Lombardi, and De Santo) and Chair of Neurology, Federico II University (Dr. Covelli), Naples, Italy and the Department of Preventive Medicine, Northwestern University Medical School, Chicago, III (Dr. Cirillo).

Address all correspondence to Dr. Vito Covelli, Neurologia (Pad. 17), Nuovo Policlinico, via Sergio Pansini, 5,80131, Napoli, Italy.

Accepted for publication October 1,1998.

headache.+*0 Lack of consistency in the previous data

could reflect low statistical power in clinical studies on

small samples of patients or low precision for headache

diagnosis in population-based studies or both. The present

study analyzed the prevalence of hypertension, hypercho-

lesterolemia, cigarette smoking, diabetes mellitus, and

obesity in a large cohort of patients with headache. The

analysis was controlled for sex, age, and type and fre-

quency of headache. Moreover, data of patients with

headache were compared to data of a population sample to

test possible differences in the prevalence of cardio-

vascular risk factors between patients with headache and

the general population.

METHODS

Patients and Controls.-The patient cohort consisted

of outpatients aged 15 years or older diagnosed as having

severe headaches at the Headache Unit, Department of

Headache 409

Neurosciences, Federico II University, Naples, Italy.

Headache was defined as severe when the patient was

recurrently unable to do anything due to the pain intensity

and/or accompanying symptoms (nausea, vomiting, photo-

phobia, phonophobia, etc). Patients aged 65 years or older

were excluded from analysis as only 7 men and 25 women

with headache were in this age group. Therefore, the

cohort of patients with headache included in analysis con-

sisted of 399 men and 944 women, aged 15 to 64 years.

The cohort of controls comprised participants in the

Gubbio Population Study, an epidemiologic study on a

sample of an entire population residing in central Italy.ll-18

To match sex and age of patients with headache, only par-

ticipants aged 15 to 64 years (1775 men and 2009 women)

were included in the analysis.

Data Collection.-Patients were classified by different

types of headache on the basis of the characteristics of the

symptoms self-reported including location, intensity, time-

related pattern of pain, and number of days per year with

headache.‘-2 Six subgroups were used in the analyses:

migraine without aura, migraine with aura, tension-type

headache, mixed headache, cluster headache, and other

headaches, which included all other forms.

For patients and controls, standardized protocols

included measurements of weight, height, and blood

pressure; information on cigarette smoking and use of

drugs; and automated biochemistry for determination of

plasma cholesterol and glucose after an overnight fast.

Weight and height were used to calculate body mass index

(BMI = weight/height*). An average of the second and

third blood pressure measurements were used for analysis;

hypertension was defined as a systolic blood pressure of

140 mm Hg or higher, and/or a diastolic blood pressure of

90 mm Hg or higher, and/or reported use of anti-

hypertensive drug treatment. A plasma cholesterol level of

200 mg/dL or higher was defined as hypercholesterolemia.

For cigarette smoking, individuals were classified as either

current smokers or nonsmokers. Reported treatment with

antidiabetic drugs or insulin was used to define diabetes

mellitus; for patients who did not report antidiabetic treat-

ment, a plasma glucose level of 140 mg/dL or higher was

considered elevated. Overweight was defined as a BMI of

26 kg/m2 or higher.

Statistics.-Analysis of variance, chi-square analysis,

and univariate and bivariate logistic regression analysis

were used for statistical procedures for the cohort of

patients with headache. Calculation of the odds ratio (OR)

with 95% confidence interval (CI) was used to compare

the prevalence of cardiovascular risk factors between

patients with headache and participants in the Gubbio

Population Study. The cumulative OR across different

strata of age was calculated by the Mantel-Haenszel

procedure’9 to control for differences in the distribution of

age between patients with headache and individuals of the

Gubbio population.

RESULTS

Types of Headache: Prevalence, Age of Patients,

and Days Per Year With Headache.-Table 1 shows the

prevalence of different types of headache, age of patients,

and days per year with headache by sex. Days per year

with headache are not shown for patients with cluster

headache as only five patients in this group were able to

report this information. Prevalence of various types of

headache was significantly different between men and

women (chi-square 98.7, P<.OOl); male to female ratio

was lower than 1 for all types of headache except cluster

headache. Migraine without aura was the most frequent

type in both men and women; migraine with aura was less

frequent in men, cluster headache less frequent in women.

Age and days per year with headache were significantly

different among different types of headache in men and

women (Table 1). In both sexes, patients with migraine

(without or with aura) reported fewer days per year with

headache than patients with other types of headache.

Cardiovascular Risk Factors and Age in Patients

With Headache.-Table 2 shows the prevalence of hyper-

tension, hypercholesterolemia, and cigarette smoking in

patients with headache by sex and age. Age was signifi-

cantly associated with hypertension and hypercholes-

terolemia in men and women, and with cigarette smoking

in women but not in men. Among patients with headache

who were nonsmokers at the time of examination (218

men and 677 women), the prevalence of patients who had

been smokers was 12.8% in men, 4.7% in women; in both

sexes, the prevalence of ex-smokers was significantly

associated with age (E-05). Prevalence of diabetes melli-

tus in patients with headache (not shown in Table 2) was

0.5% for men (n=2) and 1.1% for women (n=lO); among

patients who did not report antidiabetic treatment

410 June, 1999

Table l.-Type of Headache, Age of Patients, and Days Per Year With Headache

Headache Type

No. (%) of Age, Y Patients Mean f SD

Days Per Year Mean f SD

Men Migraine without aura Migraine with aura Tension-type headache Mixed headache Cluster headache* Other P (ANOVA)

Women Migraine without aura Migraine with aura Tension-type headache Mixed headache Cluster headache* Other P (ANOVA)

143 (35.8) 30 (7.5) 79 (19.8) 36 (9.0) 35 (8.8) 76 (19.0)

382 (40.5)

61 (6.5) 118 (12.5) 136 (14.4)

2 (0.2) 245 (26.0)

33.5 * 8.8 31.1 l 11.7 33.7 f 10.8 32.8 * 8.8 38.8 f 10.6 34.7 f 11.3

.041

34.3 f 11.8 35.9 f 11.7 32.1 f 12.0 35.9 & 11.8 54.0 f 14.1 37.4 f 12.8

c.00 1

92.1 f 82.7 68.7 f 40.1

215.8 f 122.4 158.0 f 106.9

-

183.9 f 128.7 -=.OOl

126.5 f 102.3 121.7 f 102.2 227.0 f 117.9 207.4 f 113.7

-

180.8 f 117.8 c.001

* Excluded from analysis due to limited sample size.

Table 2.-Prevalence of Hypertension, Hypercholesterolemia, and Cigarette Smoking in Patients With Headache

Men Women

Prevalence, % Prevalence, %

Age, Y

No. of Patients

Hyper- tension

Hypercho- lesterolemia

Cigarette Smoking

No. of Patients

Hyper- tension

Hypercho- lesterolemia

Cigarette Smoking

15-24 74 21.6 13.3 41.9 211 4.3 7.4 22.7

25-34 151 17.2 38.1 47.0 279 10.4 23.5 33.7 35-44 105 38.1 38.7 48.6 228 21.9 46.0 35.5 45-54 55 45.5 63.2 40.0 151 47.0 71.1 19.9

55-64 14 50.0 80.0 42.9 75 64.0 69.0 18.7

p (x2) c.001 .018 -802 c.001 c.00 1 c-00 1

Headache 411

Table 3.-Age-Adjusted Prevalence of Hypertension (HPT), Hypercholesterolemia (HCHO), and Cigarette Smoking (CS) by Sex and by Type of Headache

Men Women

Prevalence, % PrevaIence, %

No. of No. of Headache Type Patients HPT HCHO cs Patients HPT HCHO cs

Migraine without 143 29.9 46.6 41.2 382 22.2 39.5 29.0 aura

Migraine with 30 20.0 56.7 39.9 61 17.0 39.7 18.1 aura

Tension-type 79 29.4 44.6 49.4 118 21.6 31.8 32.9 headache

Mixed headache 36 34.7 29.4 36.1 136 18.8 54.2 30.9 Cluster headache 35 14.4 0.9 74.5 2* (50.0) (50.0) (50.0) Other 76 32.1 39.7 42.1 245 24.4 45.9 25.8

P (Age-adjusted ANOVA) .298 .I08 -010 .526 .311 -231

* Excluded from analysis due to limited sample size, data shown within parentheses is without age adjustment.

(397 men and 934 women), the prevalence of elevated

plasma glucose (140 mg/dL or higher) was 1.0% in men

(n=4) and 0.4% in women (n=4). Due to the low number

of cases of diabetes and/or high plasma glucose, data were

not analyzed by age and not used in further analyses.

Prevalence of overweight in patients with headache was

5 1.1% in men and 36.4% in women; in both sexes, over-

weight was progressively more frequent with increasing

age (W.01, data not shown).

ws of Headache, Days Per Year With Headache,

and Cardiovascular Risk Factors.-Table 3 shows

age-adjusted prevalence of hypertension, hypercholes-

terolemia, and cigarette smoking in patients with headache

by sex and type of headache. In analysis for men, differ-

ences among various types of headache were significant or

borderline significant for hypercholesterolemia and

cigarette smoking, but not hypertension. Compared to

patients with other types of headache, men with cluster

headache had low prevalence of hypercholesterolemia,

high prevalence of cigarette smoking, and slightly low

prevalence of hypertension. When the analysis for men in

Table 3 was done without inclusion of patients with clus-

ter headache, the prevalence of hypertension, hypercholes-

412

terolemia, and cigarette smoking was not significantly

different among patients with various types of headache

(B-.6). In the analysis for women, prevalence of hyperten-

sion, hypercholesterolemia, and cigarette smoking was not

significantly different among various types of headache

(patients with cluster headache were excluded from analy-

sis due to limited sample size, Table 3). Age-adjusted

prevalence of overweight was not significantly

different among various types of headache in the analyses

for men and for women (m-4).

In univariate logistic analyses, reported number of

days per year with headache related to prevalence of

hypertension in women (regression coefficient 0.0014,

F-042) but not in men (0.0004, p5.692); with control for

age, coefficients were not significant in either sex (B.2).

Days per year with headache did not relate to prevalence

of hypercholesterolemia in either sex without or with con-

trol for age (B-.4). Days per year with headache related to

prevalence of cigarette smoking in men without and with

control for age (coefficient 0.0019, F.03 1 and .034); in

women, relation of days per year with headache to preva-

lence of cigarette smoking was weak both without (coeffi-

cient 0.0009, P=. 123) and with control for age (coefficient

June, 1999

0.0010, P-. 115). Days per year with headache did not

relate to prevalence of overweight in either sex without and

with control for age (D.4).

Cardiovascular Risk Factors in Patients With

Headache and in the General Population: Age-

Controlled Odds Ratio.-Prevalence of cardiovascular

risk factors was compared between patients with headache

and individuals of the Gubbio population by sex and age

stratum with calculation of the Mantel-Haenszel odds

ratio. For patients with headache, sex-specific and age-

specific prevalence of major cardiovascular risk factors

are shown in Table 2; for individuals of the Gubbio popu-

lation, specific data are extensively described in several

previous papers. 11q13-15,17~1* Table 4 shows, in men and

women separately, the OR for hypertension, hypercholes-

terolemia, and cigarette smoking in the entire cohort of

patients with headache (all types) compared to individuals

of the Gubbio population. For hypertension, the OR was

significantly increased in patients with headache of both

sexes, but not significantly different between men and

women. For hypercholesterolemia, the OR was not signif-

icantly reduced in patients with headache of either sex. For

cigarette smoking, the OR was significantly reduced in

patients with headache of either sex, and not significantly

different between men and women. Among nonsmokers,

the prevalence of ex-smokers was low in patients with

headache compared to individuals of the Gubbio popula-

tion (Mantel-Haenszel OR for ex-smokers: men = 0.19,

95% CI, 0.12 to 0.29; women = 0.46, 95% CI, 0.3 1 to

0.69). For hypertension, hypercholesterolemia, and ciga-

rette smoking, findings were similar to those in Table 4

Odds ratio for hypertension

02 s-24 25-34 35-44 45-54 55-64 .

Age, years

Odds ratio for hypertension (* 95% Cl) of patients with headache compared to individuals of the Gubbio population. The dotted line (OR=l) indicates no difference in the prevalence of hypertension between patients with headache and individuals of the Gubbio population.

when patients with cluster headache were excluded from

analysis (not shown). For overweight, the OR was reduced

in both male and female patients with headache; the dif-

ference compared to individuals of the Gubbio population

was significant for women (OR=0.81, 95% CI, 0.67 to

0.97), but not for men (OR=O.84,95% CI, 0.66 to 1.07).

In analysis with the men and wornen combined in

Table 4.-Age-Controlled Odds Ratios of Hypertension, Hypercholesterolemia, and Cigarette Smoking in Patients With Headache Compared to Controls

Men Women

Odds Ratio 95% CI Odds Ratio 95% CI

Hypertension 1.!32* 1.39-2.38 1.34* 1.08-1.68 Hypercholesterolemia 0.78 0.52-1.18 0.93 O&8- 1.26 Cigarette smoking 0.74* 0.60-0.95 0.60* 0.50-0.72

*p<.o5. Cumulative overall odds ratios calculated by Mantel-Haenszel procedure.

Headache 413

both the patient group and in the control group, the cumu-

lative OR for hypertension in patients with headache com-

pared to individuals of the Gubbio population was 1.51

(95% CI, I .28 to 1.80). The sex-controlled OR for hyper-

tension was significantly different among different age

groups and inversely related to age (Figure). For other car-

diovascular risk factors, OR was not significantly associ-

ated with age.

COMMENTS

The study reports a comprehensive analysis of cardio-

vascular risk factors in a large cohort of adult patients with

headache with control for sex, age, and type and frequen-

cy of headache. Among patients with headache, there were

significant differences among the various types of

headache for sex, age, and days per year with headache.

The prevalence of cardiovascular risk factors related to age

in patients with headache and the general population.14

Hypertension, hypercholesterolemia, and overweight were

directly related to age in both male and female patients

with headache; for cigarette smoking, relation to age was

negative in women and zero in men; for diabetes mellitus,

statistical analyses could not be done due to the low num-

ber of cases. In age-controlled analyses, prevalence of car-

diovascular risk factors was significantly different for

patients with cluster headache. Compared to other

patients, men with cluster headache had a greater preva-

lence of cigarette smoking, less prevalence of hypercho-

lesterolemia, and a trend toward low prevalence of hyper-

tension. The finding of a negative association between

cluster headache and hypercholesterolemia is new, while

findings for cigarette smoking and hypertension confirm

previous observations. *O-** The high prevalence of smok-

ing in men with cluster headache supports the idea that

smoking may have an etiologic role in this type of

headache. It is reasonable to propose that cluster headache

does not infer a particularly high cardiovascular risk since

the high prevalence of smoking was associated with a low

prevalence of hypertension and hypercholesterolemia.

Except for cluster headache, cardiovascular risk factors

were similarly prevalent among the various types of

headache, which were considered as a homogeneous group

in tither analyses.

Within the cohort of patients with headache, the num-

ber of days per year with headache was not a significant

correlate of hypercholesterolemia or overweight. For

hypertension and cigarette smoking, findings were not

consistent in all analyses. Days per year with headache

directly related to hypertension only in women; moreover,

the relation was not significant when controlled for age.

Days per year with headache directly related to cigarette

smoking only in men without and with control for age,

suggesting that frequency of headache is affected by

smoking. Lack of statistical significance for women might

reflect their low prevalence of cigarette smoking.

Other findings of the study were based on the com-

parison between patients with headache and a sample of

the general population. Compared to the control popula-

tion, patients with headache had more prevaIent hyperten-

sion and less prevalent cigarette smoking in analyses for

men and women. The design used in the study, ie, the com-

parison between a large cohort of patients and a sex- and

age-matched sample of an entire population has the advan-

tage of a high statistical power but could reflect the influ-

ence of several confounders. Patients with headache and

individuals of the control population were fi-om two differ-

ent cities, ie, Naples and Gubbio. Therefore, differences

between the two cohorts could reflect factors other than

headache. For hypertension, this seems not to be the case

as elevated blood pressure is similarly distributed in

Naples and Gubbio according to data from INTERSALT,

a large epidemiologic study.23924 Actually, according to

data from INTERSALT for a range of ages similar to that

in the present study,*4 hypertension prevalence is 13 .O% in

Naples and 16.5% in Gubbio (men and women combined,

aged 20 to 59 years). The low prevalence of hypertension

reported for individuals residing in Naples raises the

possibility that the association between headache and

hypertension could be stronger than reported in the study.

For hypertension and other variables in analysis, the

differences between patients with headache and controls

might have been underestimated, as individuals with

headache must also be present within the control popula-

tion. However, this bias should have affected the data only

to a minor degree, as prevalence of severe headache is low

within the general population.*s-**

The mechanisms underlying the association between

hypertension and headache are not clear at present.

Overweight, a factor which favors elevated blood pres-

sure,*9 did not explain the association between headache

414 June, 1999

and hypertension as it was less prevalent in patients with

headache than in the control population. Habitual alcohol

intake and/or activity of the sympathetic nervous system,

not investigated in the study, could have a role in the asso-

ciation between headache and hypertension as they are

involved in the pathogenesis of both disorders.3@33 The

sex-controlled difference in hypertension prevalence

between patients with headache and the control population

decreased progressively with age. This datum suggests a

primary association between headache and hypertension.

In the opposite case- ie, if the association reff ected sec-

ondary mechanisms -the difference between patients with

headache and controls should be larger with increasing

duration of the disease or of its treatment. The strong asso-

ciation between headache and hypertension in youngsters

also suggests that the mechanisms underlying the associa-

tion could be related only when these disorders develop in

young adults. The absence of such an association among

the elderly could also be explained by other factors.

Theoretically, the high incidence of hypertension in older

people could dilute the association found in young adults.

A not necessarily alternative possibility is that young

patients with headache and hypertension have a high inci-

dence of premature cardiovascular lethal events. If this

were the case, the number of patients with headache and

hypertension, and thus the strength of the association,

should be progressively lower with increasing age.

The present data might have also practical implica-

tions. Hypertension, a major cardiovascular risk factor,

should be actively sought in patients with headache,

particularly among young adults. Drugs commonly used in

the treatment of headache such as beta-blockers and calci-

um antagonistss4 are also powerful antihypertensive drugs.

Therefore, if beta-blockers and/or calcium antagonists are

used as antiheadache treatment they could mask the devel-

opment of hypertension in originally nonhypertensive

individuals. In these cases, hypertension could become

evident, at times acutely, only after withdrawal of the anti-

headache treatment. Conversely, in untreated patients with

headache and hypertension, beta-blockers and/or calcium

antagonists could be used as a first-choice therapy since

these drugs could be able to control both the disorders. The

same couId be true for reduction in alcohol intake, which

could favor blood pressure decrease and control of

headache.

Headache

In summary, the present study shows that the profile

of cardiovascular risk in patients with headache is homo-

geneous except for the small percentage with cluster

headache. The prevalence of hypertension, hypercholes-

terolemia, and overweight is related to age in patients with

headache as in the general population. Hypertension is the

only cardiovascular risk factor highly prevalent in both

male and female patients with headache, particularly in

youngsters. The possible mechanisms underlying the

association between headache and hypertension are not

clear at present. However, the evidence of the association

might have practical implications for the prevention and

treatment of headache and hypertension.

REFERENCES 1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

Ad Hoc Committee on Classification of Headache.

Classification of headache. JAMA. 1962; 179: J 1 J-J 18.

Headache Classification Committee of the

International Headache Society. Classification and

diagnostic criteria for headache disorders, cranial

neuralgias and facial pain. Cephalalgia. 1988;8

(suppl 7): l-96.

Janeway TC. A clinical study of hypertensive cardio-

vascular disease. Arch Intern Med. 1913;12:755-798.

Walker CH. Migraine and its relationship to hyperten-

sion. BMJ. 1959;2:1430-1433.

Waters WE. Headache and blood pressure in the com-

munity. BMX 197 1; 1: 142- 143.

Weiss NS. Relation of high blood pressure to

headache, epistaxis, and selected other symptoms.

The United States Health Examination Survey of

Adults. N EngI J Med. 1972;287:63 I-633.

Leviton A, Malvea B, Graham JR. Vascular diseases,

mortality, and migraine in the parents of migraine

patients. Neurology. 1974;24:669-672.

Chen TC, Leviton A, Edelstein S, Ellenberg JH.

Migraine and other diseases in women of reproductive

age. The influence of smoking on observed associa-

tions. Arch Neuml. 1987;44: 1024- 1028.

Couch JR, Hassanein RS. Headache as a risk factor in

atherosclerosis-related diseases. Headache. 1989;29:

49-54.

Peroutka SJ, Price SC, Jones KW. The comorbid

association of migraine with osteoarthritis and hyper-

415

tension: complement C3F and Berkson’s bias.

Cephalalgia. 1997; 17123-26.

Il. Laurenzi M, Trevisan M. Sodium-lithium counter-

transport and blood pressure: the Gubbio Population

Study. Hypertension. 1989; 13:408-415.

12. Cirillo M, Trevisan M, Laurenzi M. Calcium binding

capacity of erythrocyte membrane in human hyper-

tension. Hypertension. 1989; 14: 152-l 55.

13. Trevisan M, Laurenzi M. Correlates of sodium-

lithium countertransport. Findings from the Gubbio

Epidemiological Study. Circulation. 199 1;84:

201 l-2019.

14. Laurenzi M, Cirillo M, Angeletti M, et al. Gubbio

population study: baseline findings. Nutr Metab

Cardiovasc Dis. 1991;1:Sl-S18.

15. Cirillo M, Laurenzi M, Trevisan M, Stamler J.

Hematocrit, blood pressure, and hypertension. The

Gubbio Population Study. Hypertension. 1992;20:

3 t 9-326.

16. Cirillo M, Laurenzi M, Panarelli W, Stamler J.

Urinary sodium to potassium ratio and urinary stone

disease. Kidney Int. 1994;46: 1133- 1139.

17. Cirillo M, Laurenzi M, Panarelli W, et al. Sodium-

lithium countertransport and blood pressure change

over time: the Gubbio study. Hypertension.

1996;27:1305-1311.

18. Laurenzi M, Cirillo M, Panarelli W, et al. Baseline

sodium-lithium countertransport and &year incidence

of hypertension. The Gubbio Population Study.

Circulation. 1997;95:581-587.

19. Robins J, Greenland S, Breslow NE. A general

estimator for the variance of the Mantel-Haenszel

odds ratio. Am J Epidemiol. 1986; 124:7 19-723.

20. Manzoni GC, Terzano MG, Bono G, Micieli G,

Martucci N, Nappi G. Cluster headache--clinicai

findings in 180 patients. Cephalalgia. 1983;3:2 l-30.

2 1. Littlewood JT, Glover V, Sandler M, Petty R, Peatfield

R, Clifford Rose F. Migraine and cluster headache:

links between platelet monoamine oxidase activity,

smoking and personality. Headache. 1984;24:30-34.

22. Millac P, Akhtar N. Cigarette smoking and cluster

headaches. Headache. 1985;25:223.

23. Intersalt Cooperative Research Group. Intersalt: an

international study of electrolyte excretion and blood

pressure. Results for 24 hour urinary sodium and

potassium excretion. BM.. 1988;297:3 19-328.

24. The Intersalt Cooperative Research Group. The

Intersalt Study. An international co-operative study of

electrolyte excretion and blood pressure: further

results. J Hum tlyperrens. 1989;3:279-407.

25. Bruyn GW. Epidemiology of migraine ‘a personal

view.’ Headache. 1983;23:127-133.

26. D’Alessandro R, Benassi G, Lenzi PL, et al.

Epidemiology of headache in the Republic of San

Marino. J New-01 Neurosurg Psychiatry. 1988;5 1:

21-27.

27. Duckro PN, Tait RC, Margolis RB. Prevalence of very

severe headache in a large US metropolitan area.

Cephalalgia. 1989;9: 199-205.

28. Rasmussen BK, Jensen R, Schroll M, Olesen J.

Epidemiology of headache in a general population-

a prevalence study. J Clin Epidemiol. 199 1;44: 1147-

1157.

29. Dustan HP Obesity and hypertension. Ann Intern

Med. 1985;103:1047-1049.

30. Peatfield RC. Relationships between food, wine, and

beer-precipitated migrainous headaches. Headache.

1995;35:355-357.

3 1. Blau JN. Migraine: theories of pathogenesis. Lancet.

1992;339: 1202- 1207.

32. Klatsky AL, Friedman GD, Sigelaub AB, Gerard MJ.

Alcohol consumption and blood pressure. Kaiser-

Pennanente Multiphasic Health Examination data.

N EngZ J Med. 1977;296: 1194- 1200.

33. Oparil S, Chen Y-F, Berecek KH, Calhoun DA, Wyss

JM. The role of the central nervous system in hyper-

tension. In: Laragh JH, Brenner BM, eds.

Hypertension: Pathophysiology, Diagnosis, and

Management. New York, NY: Raven Press Ltd;

1995:713-740.

34. Welch KM. Drug therapy of migraine. N Engl J Med.

1993;329: 1476-1483.

416 June, I999