Headaches and Heart Disease: The Lack of a Positive Association

Harvey J. Featherstone, M.D., M.P.H.*

*Department of Medicine, University of Washington, Seattle, Washington

Reprint requests to: H.J. Featherstone, M.D., M.P.H., Department of Medicine RG-20, University of Washington, Seattle,WA 98195

Accepted for Publication: June 10, 1985

SYNOPSIS

In order to investigate the relationship between headaches and cardiovascular disease, 200 individuals with recurrentidiopathic headaches were compared with age- and sex-matched controls for the prevalence of cardiovasculardiagnoses and electrocardiographic (ECG) abnormalities. Only hypertension was diagnosed more frequently in theheadache group; ischemic heart disease and valvular lesions were diagnosed equally in some cases and controls, aswere rhythm disturbances. Matched ECG's were available in 161 pairs. Ischemic ECG changes were more frequent inthe controls than the headache cases. There were no other differences in ECG findings. This study fails to show apredilection for cardiovascular diseases, other than hypertension, in individuals with headaches. Ischemic heart diseasemay, by ECG analysis, be less common in headache cases than nonheadache controls.

(Headache 26:39-41, 1986)

INTRODUCTION

Because migraine and other related idiopathic headaches are thought to have a vascular mechanism, they have traditionally beensuspected to be associated with an increased risk for other cardiovascular diseases, including ischemic heart disease andhypertension. Although correlations between headaches and hypertension,1-3 mitral valve prolapse,4,5 and variant angina6 havebeen described, no studies comparing the occurrence of all cardiac diseases in idiopathic headache patients and nonheadachecontrols can be found. The following study was undertaken to look for an association of headaches with ischemic heart diseaseand valvular cardiovascular problems.

METHODS

Files of life insurance applicants were referred to a physician for 12-lead electrocardiogram (ECG) interpretation and/or a medicalopinion. These files included copies of physician office records, diagnostic test results and, where appropriate, dischargesummaries and consultation reports from inpatient records. Individuals were considered to have idiopathic headaches whenreview of the records revealed a history of recurrent headaches in the absence of fever, trauma, suspected neurological diseaseor other medical illnesses associated with headaches. The headache cases were matched by age and sex with control individualsfor whom the medical records clearly stated that the person had no headaches. Cardiovascular diagnoses listed or described inthe applicant's records were recorded for cases and controls. All electrocardiograms were read blindly by one board-certifiedinternist, and the results recorded for cases and controls. (Criteria for ECG abnormalities are shown in Table 1.) The case-controlpairs were then compared for differences in the prevalence of cardiovascular diagnoses and in ECG interpretations. Statisticalanalysis was performed using the sign test on discordant pairs.7

RESULTS

Two hundred cases of idiopathic headache were matched with controls, including 100 pairs of men and 100 pairs of women. Agerange was 25 to 63

Table 1Criteria for abnormal electrocardiograms.

1. Ischemic Changes a. Pathological Q waves - Q waves of at least .03 seconds in duration either in both leads Ill and aVF or V1 and V2. b. ST segment depression - horizontal ST depression of at least 1 mm. in two or more of the 12 leads. c. T wave inversion - inverted T waves either in leads II, III and aVF or in two or more V leads from V3 to V6.2. Conduction Defects a. Left anterior hemiblock - leftward axis deviation beyond -45°, plus a Q wave in lead I and an S wave in lead III. b. Right bundle branch block - a QRS duration of at least. .12 seconds with terminal S waves in leads I and V6 and a RSR' configuration in lead V1. c. Left bundle branch block - a QRS duration of at least .12 seconds with all-positive R waves in leads I and V6, and a deep S wave in V1. d. Other - a QRS duration of at least .12 seconds without the specific patterns in b. and c. above.3. Other Abnormalities a. Left ventricular hypertrophy - QRS voltage criteria: either S wave V2 plus R wave V5 of at least 40 mm or lead I R wave greater than 14 mm and aVL greater than 10 mm. b. Nonspecific ST-T changes - flattened or absent T waves or ST segment scooping with or without biphasic T Waves.

years, and mean age was 45 for both male and female pairs. Headache diagnoses consisted of "migraine" or "vascular" in 32 cases (16%),"tension" in 38 cases (19%), both in 4 cases (2%) and no specific headache diagnosis in the remainder (126 cases, 63%).

Cardiovascular diagnoses recorded in the 200 pairs are shown in Table 2. Except for hypertension, which was seen more frequently in theheadache cases, there were no differences between the individuals with headache and the controls. There was also no difference in theprevalence of cigarette smoking (61 cases, 61 controls), diabetes mellitus (11 cases, 15 controls) or hyperlipidemias (7 cases, 5 controls).Chest pain of unknown etiology, including many instances with normal coronary arteriograms, was reported in 21 headache cases and 30controls (p = .25).

The comparison of ECG findings is featured in Table 3. Of the 200 pairs, 161 matched for the presence of an ECG in both files. Of those 39pairs missing an ECG, 33 were in the female pairs and 27 were either male or female pairs below the age of 46 years. In pairs older than 45years, 52 of 54 males (96%) and 40 of 50 females (80%) were matched for ECG availability. Most ECG's had been routinely taken and withinthree years, although there was no matching for why or how long ago the ECG was obtained. Of the ECG findings, only ischemic changes(pathological Q waves, ST segment depression and inverted T waves) were seen with different frequencies in the headache cases andcontrols, being more common in the latter. This difference was still present after controlling for the diagnosis of hypertension (12 cases, 31controls, p = .01). When pairs were grouped according to headache diagnosis, the lower

Table 2Clinical cardiovascular diagnoses in individuals with andwithout

idiopathic headaches (n = 200 matched pairs).Number ofHeadache Cases Number

of Signifi-cControls ance

Hypertension 103 76 p < .01Ischemic Heart Disease 14 20 N.S. Myocardial Infarction 3 9 Other 11 11Structural Heart Disease 26 18 N.S. Mitral Valve Prolapse 7 4 Other Valvular Disease 11 6 Congenital Heart 4 5Disease Nonspecific Murmur 4 3Rhythm Disturbances 7 12 N .S. Supraventricular Tachycardia 5 6 Premature Beats 1 4 Other 1 2Miscellaneous 4 3 N.S.

Table 3Electrocardiographic findings in individuals with and without idiopathic headaches (n = 161

matched pairs).Number ofHeadache Number of Signifi-

Cases Controls canceNormal 69 59 N.S.Ischemic Changes 30 50 p < .05 Pathological Q Waves 14 21 ST Depression 6 12 T Wave Inversion 10 17Conduction Defects 17 24 N .S. Left Anterior Hemiblock 7 11 Right Bundle Branch Block 7 11 Left Bundle Branch Block 1 0 Other 2 2Left Ventricular Hypertrophy 14 14 N.S.Nonspecific ST-T Changes 55 45 N.S.Premature Beats 4 8 N.S.Miscellaneous Abnormalities 4 8 N.S.

prevalence of ECG ischemic changes in headache cases was seen in the "migraine" (p = .05) and "no specific diagnosis" groups (p = .10),but not in the "tension" group.

DISCUSSION

The results of this study did not support the widely held but rarely published supposition that migraine predisposes to ischemic heart disease.Rather, fewer ischemic electrocardiographic changes were seen in headache patients, especially those with the diagnosis of migraine, than inthe nonheadache controls. Reported myocardial infarction was also less common, although not significantly so. Could migraine somehowhave a protective effect? Perhaps the proposed migraine sequence of vasoconstriction and ischemia leading to reactive vasodilatation8 beforepermanent ischemic damage occurs in the intracranial circulation also takes place in other organs, including the heart. Is "migraine" abeneficial response to nonspecific ischemic threats? More study is needed in this area of migraine and the heart.

The apparent protective effect of headache against ischemic heart disease was seen despite an increased prevalence of the diagnosis ofhypertension in the headache cases. The relationship of headaches to hypertension has been studied repeatedly with the publication ofconflicting results. Gardner et al.1 reported a higher prevalence of migraine in hypertensive individuals than in normotensive controls. Walker2

noted an increased occurrence of migraine with increasing blood pressure values. However, several epidemiological studies of blood pressureand headache have failed to show a direct

relationship between the two.9-11 The current study examined the relationship of physician diagnoses ofhypertension to headache; no attempt was made to correlate headache to actual blood pressure values, whichwould have been difficult since many patients were under treatment for high blood pressure. The danger of anartifactual association because blood pressure may be more likely to be measured in patients with headaches hasbeen mentioned by Walker2 and was a potential bias in this study.

Other cardiac abnormalities reported to occur more commonly in migraine are mitral valve prolapse,4,5 paroxysmalsupraventricular tachycardia,12 and variant angina.6 None of these entities was seen more frequently in theheadache group, although numbers in each category were small. The association of MVP has been found withmigraine specifically and not with "tension" headaches4; this study did not differentiate well between migraine and"tension" headaches, which may be a reason why no correlation with MVP was seen. Mitral valve prolapse wasslightly more common in the headache cases but at a ratio that would have required four times as many cases toapproach a significant difference. In addition, a similar trend was seen for other types of valvular heart disease,making one wonder if the MVP-migraine association is specific. There was no suggestion in this study that theprevalence of variant angina or unexplained chest pain was higher in the headache or migraine group. The oneprevious report6 suggesting a correlation between variant angina and migraine was flawed by the failure to controlfor medication use; "migraine" was most common in the group of patients with the highest nitrate to beta blockerratio and was least common in the group most frequently using beta blockers. An association between variantangina and migraine remains to be proven.

Weaknesses of this study deal mainly with the heterogeneity of the collected data. Headache and cardiacdiagnoses were those of the applicants' physicians; there were no standardized diagnostic criteria. ECG's mayhave been collected routinely or for symptom evaluation. Matching case pairs eliminated some of the biasregarding cardiac diagnoses or heterogeneity of ECG collections. However, the reliability of the headache history, acritical factor, was unknown. Concern for the specific headache diagnosis (i.e., migraine vs. "tension") was lessimportant in view of the recent reports of some headache investigators that clinical classification of headaches intotypes is difficult13; "tension" and migraine may not represent separate headache entities.14,15 Notwithstanding theseissues, the data reported here suggest that not only was there no evidence of an increased association betweenheadaches and heart disease, but that patients with headaches may be protected in some way from ischemic heartdisease, even in the face of a higher prevalence of hypertension in headache cases.

REFERENCES

1. Gardner JW, Mountain GE, Hines EA: The relationship of migraine to hypertension and to hypertensionheadaches. Am J Med Sci 200:50-53, 1940.

2. Walker CH: Migraine and its relationship to hypertension. Brit Med J 2 : 1430-1433, 1959.

3. Atkins JB: Migraine as a sequel to infection by L. icterohaemorrhagiae. Brit Med J 1:1011-1012, 1955.

4. Gamberini G, D'Alessandro R, Labriola E, Poggi V, Manzoni GC, Carpeggiana P, Sacquegna T: Furtherevidence on the association of mitral valve prolapse and migraine. Headache 24:39-40, 1984.

5. Amat G, Louis PJ, Loisy C, Centonze V, Pelage S: Migraine and the mitral valve prolapse syndrome. AdvNeurol 33:27-29, 1982.

6. Miller D, Waters DD, Warnica W, Szlachcic J, Kreeft J, Theroux P: Is variant angina the coronarymanifestation of a generalized vasospastic disorder? N Engl J Med 304:763-766, 1981.

7. Armitage P: Statistical Methods in Medical Research. New York: John Wiley and Sons, 1974.

8. Saper JR: Migraine: I. Classification and pathogenesis. JAMA 239:2380-2383, 1978.

9. Waters WE: Headache and blood pressure in the community. Brit Med J 1:142-143, 1971.

10. Korczyn AD, Carel RS, Pereg I: Correlation of headache complaints with some physiological parameters ina healthy population. Headache 20:196-198, 1980.

11. Weiss NS: Relation of high blood pressure to headache, epistaxis, and selected other symptoms. N Engl JMed 287:631-633, 1972.

12. Johansson BW: Migraine and the heart. Acta Med Scand 213:241-243, 1983.

13. Ziegler DVC: Headache syndromes: problems of definition. Psychosornatics 20:443-447, 1979.

14. Cohen MJ: Psychophysiological studies of headache: is there similarity between migraine and musclecontraction headaches? Headache 18:189-196, 1978.

15. Featherstone HJ: Migraine and muscle contraction headaches: a continuum. Headache 25:194-198, 1985.