Systemic Hypertension, Headache, and Ocular Hemodynamics: A New Hypothesis

Abstract

The association between systemic hypertension and headache remains controversial and its pathophysiologic basis is uncertain. A rather characteristic early-morning pulsating headache is commonly seen in hypertensive patients, and a recent meta-analysis supports the link between these 2 entities. Epidemiologic evidence has paradoxically suggested a negative association between hypertension and headache. Unpredictable clinical association between severe hypertension and headache indicates that another cranial perfusion-related variable exerts a critical role. Neuroanatomically, head and neck pain primarily involves the ophthalmic division of the trigeminal nerve (V1). A link between systemic hypertension, pulsatile choroidal blood flow (CBF), and intraocular pressure (IOP) has been established. I propose that a trait ocular sympathetic hypofunction permits rapid episodic ocular choroidal overperfusion that stretches the ocular globe in the cohort of hypertensive patients with headache. Rapid distension of the pain-sensitive corneoscleral envelope can stimulate corneoscleral and iridial pain-sensitive V1 nerve endings and generate headache. Ocular tamponade function physiologically limits choroidal overperfusion. A higher basal IOP in some patients with moderate-to-severe hypertension may dampen pulsatile CBF and account for the negative epidemiologic link between sustained systemic hypertension and headache. Besides activation of the baroreceptor reflex, the association of hypalgesia with hypertension probably involves activation of the vasopressin-endorphin adaptive system consequent to mechanical stimulation of V1. The analogy between hypertensive headache and angle-closure glaucoma is rather limited because typical ocular and visual signs and symptoms of angle-closure glaucoma are not seen in hypertension-related headache. Hypertensive crises, including those associated with pheochromocytoma, are not accompanied by attacks of angle-closure glaucoma. Glaucoma is not associated with ocular choroidal congestion, but with reduced pulsatile CBF. The predisposition to develop angle-closure glaucoma is theoretically not associated with ocular autonomic hypofunction and should be conceptually dissociated from this hypothesis. The hypothesis can be evaluated by establishing significant circadian elevations of blood pressure, including nondipping nighttime pattern as well as circadian and periheadache measurements of IOP in patients with attacks of hypertension-related headache.

Introduction

Almost a century ago, Janeway[1] described the classic “hypertensive” headache: a throbbing (pulsating) painful discomfort that is present on waking and wears off during the morning. The impression has prevailed that such hypertension-associated headache is a typical and common hypertensive symptom that is reported by approximately 50% of patients and is a reason for measuring blood pressure (BP).[1,2] Additionally, a close relationship has been shown between BP and headache in mild-to-moderate essential hypertension.[3] Somewhat unexpectedly, both systolic and diastolic BP were found closely related to headache, regardless of prior therapy status or the type of antihypertensive treatment being used at the first visit; also, as noted in several studies, treatment of hypertension reduces the frequency of headache.[3–5] Moreover, in an unselected study in a primary care setting, headache was associated with hypertension for both treated and nontreated patients.[6] Pietrini and colleagues[7] recently concluded that hypertension might worsen the frequency and severity of headache attacks, both in migraine and tension-type headache. Both acute and chronic hypertension can be associated with headache.[8]

Conversely, and contrary to the long-standing general expectation, the link between hypertension and headache was subsequently relegated to a chance occurrence or a “sociopsychological” or “functional” phenomenon.[9] Cross-sectional surveys of unselected populations do not show a relation between headache and hypertension.[10,11] Second, the first prospective study of BP and the risk for headache in a large unselected population intriguingly shows that high systolic and diastolic pressure is associated with reduced risk for nonmigrainous headache,[12] thereby raising the intriguing issue of hypertension-associated analgesia in humans.[13,14] The inverse association between pulse pressure and headache in moderate-to-severe hypertension in a cross-sectional study at a tertiary care center further supports these concepts.[15] Moreover, 24 hours of ambulatory BP monitoring in mild hypertensive patients with headache found no association between the occurrence of headache and variation of BP.[16] Finally, at the extreme of such contrary perception is the opinion that hypertensive headache is a myth.[17] Currently, the consensus of opinion of the International Headache Society is that chronic arterial hypertension of mild or moderate degree does not cause headache.[18]

Headache, nevertheless, is indeed associated with certain forms of hypertension, the pathophysiologic basis of which is uncertain. Headaches in severe hypertension are believed to be linked to hypertensive vascular dilatation or stretching.[19,20] Despite the wide availability of advanced neuroimaging techniques, there is no definitive evidence to support a pathogenic role for sustained intracranial arterial dilatation in hypertension-associated headache. The inverse relationship between higher levels of BP and headache as well as hypertension-related hypalgesia cannot be explained on the basis of cranial arterial dilatation. The absence of predictable headache in severe hypertension, including that associated with pheochromocytoma-related crises,[21] also challenges the cranial arterial dilatation hypothesis.

To understand the complex pathophysiologic relation between BP and headache, certain key questions must be addressed: First, what is the physiologic mechanism that protects most hypertensive patients from headache? Second, why does a sudden rise – as opposed to sustained elevation – of BP frequently cause headache? Third, what is the mechanistic basis of hypertension-associated hypalgesia? Fourth, how might both systolic and diastolic elevations of BP be associated with the occurrence of headache? A rational conclusion that may be drawn from apparently conflicting clinical studies (as discussed above) is that a high BP itself is probably not the only critical factor in the occurrence of headache in hypertensive patients. This article presents the basis for an alternative hypothesis that rationalizes known epidemiologic, clinical, and pharmacologic tenets of the phenomenological link between headache, hypertension, and analgesia.

Hypothesis

The unpredictable occurrence of hypertensive headache with a definite circadian pattern in many affected patients suggests an idiosyncratic predisposition involving homeostatic dysfunction that is particularly peculiar to the morning hours.

The hypothesis to be developed is that headache in hypertension patients is the outcome of a complex interaction between systemic BP, cranial blood flow, ocular autonomic function, ocular choroidal perfusion, and intraocular pressure (IOP) – the many trait- and state-dependent factors that determine the mechanical properties of the corneoscleral envelope, and the endogenous pain control mechanism. I propose that an idiosyncratic aberration of ocular hemodynamics is the crucial cranial perfusion-related variable that determines the development of headache as well as hypalgesia in a cohort of hypertensive patients through mechanical activation of pressure-sensitive ocular fibers of the ophthalmic division of the trigeminal nerve (V1). Once the eyeball is distended by choroidal congestion, further pulsatile (arterial) ocular perfusion with each heartbeat can exacerbate mechanical deformation of the stretched V1 fibers and generate the painful pulsatility that is typical of hypertensive headache. This hypothesis involves a single assumption: a key role for idiosyncratic ocular autonomic innervation in the regulation of the intraocular circulation.

Discussion

The mechanistic basis for the occurrence of headache in hypertension is unknown. Two aspects of hypertension appear closely linked to headache: (1) severe degree of hypertension (BP > 180/110 mm Hg) and (2) sudden rises of BP. The association between BP and headache is most clearly manifest in patients with severe diastolic hypertension[20] or with systolic BP above 200 mm Hg.[22] Second, acute hypertension is one of the most frequent etiologies of headaches secondary to systemic disorders besides fever and sinusitis.[23] Episodic headache is a feature of pheochromocytoma-related episodic sudden hypertension in 80% of cases.[24] In one series, however, BP elevations in pheochromocytoma as high as 260/160 mm Hg were not associated with headaches.[21] A third arterial perfusion-related variable – besides severity of hypertension and rapidity of rise of BP – appears to play a critical role in the development of hypertension-associated headache.

Head and nuchal pain primarily involve the distribution of the V1 nerve. In humans, pain and temperature fibers from only the V1 nerve descend to the lower limit of the first cervical spinal segment; this long-held view is supported by sectional studies at and below the obex for severe trigeminal neuralgia.[25] Frontal and/or nuchal pain in hypertensive headaches, therefore, can be explained by neural transmission in the V1 nerve. Among the structures supplied by the V1 nerve, the eye and the IOP constitute a physiologic system that is readily responsive to changes in BP. Remarkably, the ocular choroid possesses the greatest blood supply in the human body, with a circulation volume 10-20 times that of the cerebral cortex; the sympathetically innervated choroidal circulation is capable of vigorous autoregulation.[26,27] Alterations in ocular choroidal blood flow (CBF) – either increased perfusion (due to hypertension or vasodilators) or decreased perfusion (due to vasoconstrictors) – can therefore be more pronounced than changes occurring in cerebral blood flow. Reflecting changes in a relatively low-volume, highly vascularized structure, ocular hemodynamics can mirror pressure changes in the systemic circulation. Of importance, in male patients with normal or slightly elevated BP, ocular fundus pulsation amplitude (r = 0.252, P < .001) and mean flow velocity in the posterior ciliary arteries (r = 0.346, P < .001) were significantly associated with mean arterial pressure as well as both systolic and diastolic BP.[28] The fundus pulsation amplitude is a measure of pulsatile CBF; small but significant increases in CBF with mild elevations of BP[28] suggest that severe hypertension may be associated with an early phase of larger, clinically significant increases in pulsatile CBF.

In a longitudinal study, a higher BP was correlated with a higher IOP, a feature probably related to the higher choroidal perfusion pressure that prevails in hypertension; such IOP elevations were not linked to the subsequent development of glaucoma.[29] In mild or moderate hypertension, ocular autonomic hypofunction may permit periodic or circadian choroidal overperfusion and episodic elevations of IOP. A nondipping night-day BP pattern (absence of nighttime BP decrease) possibly linked to sympathetic nervous or other aberrations[30] may be particularly relevant to the characteristic early-morning headache that is seen in a cohort of hypertensive-headache patients. Increased 24-hour BP load and different circadian BP or IOP or both patterns may be involved in the genesis of hypertensive headaches at other times during the day or night.

Acute corneoscleral stretching by experimental IOP elevations discharges impulses in iris, corneoscleral, and whole nerve V1 fibers probably involving mechanical distortion of the iris and chamber angle, suggesting the production of painful antidromic impulses.[31,32] The link between CBF-related IOP elevation and hypertension-related headache is further strengthened by the fact that most pharmacologic antihypertensive agents also lower the IOP; propranolol, atenolol, metoprolol, nadolol, clonidine, flunarizine, verapamil, diuretics, and angiotensin-converting enzyme inhibitors lower IOP.[29,31] Prominent alterations in intraocular hemodynamics have been reported in 2 primary vascular headache variants – cluster headache and paroxysmal hemicrania[33,34] – and are possibly involved in benign cough-induced headache.[35,36] An increase in stroke volume, BP, and pulsatile ocular perfusion possibly underlies immediate aggravation of headache with aura following therapeutic closure of atrial septal defect for migraine; such headaches usually subside with the passage of time possibly involving cardiovascular acclimatization and ocular homeostatic tissue creep.[37] Relatively rapid choroidal congestion and a rise in IOP can stretch the corneoscleral envelope and mechanically activate pressure-sensitive nerve fibers of V1, thereby also generating the pain of hypertension-related headache.

The tamponade function of IOP maintains ocular integrity.[31] A higher basal or steady-state IOP limits ocular choroidal hyperperfusion by intraocular tamponade, and possibly prevents the development of headache at higher levels of BP. Every physiologic function has an upper threshold; the tamponade effect of rising IOP is probably overwhelmed by paroxysmal rises of BP in severe or malignant hypertension, hypertensive encephalopathy, or pheochromocytoma.

In a seminal study on hypertension-associated hypalgesia, Ghione[13] elucidated possible mechanisms underlying increased nociceptive threshold in experimental and human hypertension. A well-coordinated adaptive response through reflex baroreceptor and/or a central neuronal activation involving endorphinergic, noradrenergic, or other neuronal mediators may be involved. The role of endogenous central nervous system opioids and its interaction with the alpha₂-adrenergic system appear to have particular relevance to hypertension-related antinociception.[13] The generation of antidromic impulses in V1 is a vasodilatory reflex or response that likely triggers the stress-related adaptive vasopressin-corticotropin-releasing hormone cascade; both vasopressin and endorphin have analgesic potential that appears relevant to headache pathophysiology.[38] This antinociceptive mechanism likely operates in conjunction with activation of central noradrenergic and serotonergic systems.[38]

A characteristic circadian pattern is seen in hypertensive headache; the headache is commonly present at awakening and wears off in the morning hours,[1] indicating involvement of a physiologic aberration that peaks at this time. BP manifests a prominent circadian variation that might contribute to the peaking of acute cardiovascular events in the morning hours.[39] Circadian variation of BP may be mediated at least in part by circadian variation of autonomic nervous system activity. In hypertensive patients with stroke, systolic BP, diastolic BP, pulse rate, and physical activity show peaks along with a peak in stroke onset in the morning hours.[40] Disturbed nocturnal decline in BP is associated with cerebral infarction, whereas a large morning pressor surge and a large nocturnal decline in BP, which are analogous to a large diurnal increase in BP, are associated with cerebral hemorrhage.[41] Remarkably, IOP readings in both male and female nonglaucomatous patients also show the highest values in the morning with a decreasing trend in the afternoon.[42] Nocturnal elevation of IOP can be detected in healthy young adults in both the sitting and the supine positions.[43] In accord with this hypothesis, both the nocturnal rise of IOP as well as morning surge of BP, CBF, and IOP may underlie the classic propensity of hypertensive headaches to manifest frequently during the morning hours.

Limitations of the Hypothesis

A principal limitation of this hypothesis could be the absence of a direct or striking clinical or statistically strong positive link between glaucoma and hypertension. Glaucoma is a term that is applied to several states with either aberration in control of IOP due to anterior segment anatomic aberration involving defective circulation or outflow of aqueous humor or optic nerve head vascular insufficiency in the face of normal IOP. Nevertheless, because headache is a prominent feature of angle-closure glaucoma only, it is essential to explicate the phenomenological and mechanistic differences that underlie IOP elevation between this glaucoma variant and hypertensive headache. Crucially, this hypothesis for hypertension-related headache does not pertain to persistent glaucoma-related anterior segment anatomic and pathophysiologic mechanisms, but to episodic rapid-onset choroidal congestion that stretches the pain-sensitive corneoscleral envelope; increased CBF, in turn, can increase aqueous humor formation and secondarily raise IOP. Of interest, several studies have shown a consistent link between systemic BP and IOP.[29,44,45] Besides, there is no link between severe hypertension, including hypertensive crises in pheochromocytoma, and angle-closure glaucoma. Finally, increased pulsatile CBF is not a feature of glaucoma. On the contrary, pulsatile ocular blood flow is reduced stepwise with elevated IOP.[46] A basic pathophysiologic difference prevails between the 2 entities.

Although eye pain and headache are characteristic features of angle-closure glaucoma, the clinical analogy to hypertensive headache is limited; the absence of typically glaucomatous symptoms, such as visual blurring or loss of visual acuity, colored halos around lights, nausea and vomiting, corneal edema, or fixed semidilated pupils in patients with hypertensive headache, indicates a different mechanistic basis, probably involving a panocular or global corneoscleral distention rather than a predominantly anterior ocular segment aberration.

The second issue that challenges this hypothesis is that the IOP changes seen in patients with hypertension[29,44,45] are relatively small or modest and may not be sufficient to result in headache. First, casual tonometry and even more methodic measurements of IOP can miss biologically significant elevations of IOP.[47] This hypothesis envisions episodic rises of IOP that precede and possibly persist during the headache. Measurement of CBF or IOP during hypertensive headaches has never been carried out; only such aberrations are truly representative for this hypothesis. Second, a reciprocal relation prevails between CBF and relatively higher levels of IOP; the tamponade effect of choroidal congestion and higher IOP dampens CBF and ocular pulsatility. With the onset of hypertensive headache, diminished rather than increased CBF might be expected. Third, the development of headache activates the autonomic nervous system that may tend to lower CBF as well as IOP. Fourth, rapid increases of choroidal perfusion pressure in parallel with high systemic pulse pressure of hypertension can stretch the relatively low-volume ocular globe with only limited alterations of IOP. Finally, several factors affect static and dynamic distensibility of the corneoscleral envelope, particularly with rapid alterations in the volume-pressure relationship.[31] This hypothesis basically invokes rapid-onset panocular distention through choroidal vascular congestion rather than through a primary elevation of IOP. The need for conceptual dissociation between hypertension-related IOP variations and disease entities associated with glaucomatous features appears to be theoretically valid and clinically relevant.

Testing the Hypothesis

To support my hypothesis, it may be useful to measure hypertensive-headache-related ocular hemodynamic changes, including pulsatile ocular or CBF, fundus pulsation amplitude, or intraocular pressure pulse, as well as changes in IOP during such episodes. Because the onset of hypertensive headache, according to this hypothesis, follows a phase of ocular choroidal overperfusion/congestion and would definitely be accompanied by secondary autonomic activation, complex measures of ocular hemodynamics are likely to give variable results. Combined systematic study of 24-hour monitoring of BP as well as IOP (including home tonometry) in patients with hypertensive headache is likely to further elucidate the pathophysiologic basis of this secondary headache. Study of acute changes in automated perimetry following episodes of hypertensive headache might document reversible retinal barotrauma and increase understanding of the disorder.

Study of cranial/ocular autonomic function in patients with hypertensive headache might also be considered. However, studies of nonparalytic bilateral pupillary changes as a marker of cranial/ocular autonomic dysfunction are difficult to interpret and generally misleading.[48]

Conclusions

The development of relatively rapid ocular congestion through hypertension-induced choroidal overperfusion may underlie intermittent or circadian mechanical activation of pressure-sensitive ophthalmic nerve fibers that, in turn, cause headache in some hypertensive patients. Ocular autonomic function is an idiosyncratic variable that cannot be reliably measured or predicted in the laboratory but probably exerts a critical influence on ocular choroidal circulation. Future pathophysiologic studies of hypertension-associated headache should focus on combined 24-hour measurements of BP and IOP as well as indexes of acute retinal barotrauma.