Effect of Caffeine and Caffeine Cessation on Cerebrovascular Reactivity in Patients With Migraine

Young-Eun Gil MD, MSc; Mi Ji Lee MD, PhD; Soohyun Cho MD, MSc; Chin-Sang Chung MD, PhD


Headache. 2022;62(2):169-175. 

In This Article

Abstract and Introduction


Objective: To investigate the effect of chronic caffeine use and caffeine cessation on vasodilatory function in the posterior circulation in patients with migraine.

Background: Studies regarding cerebrovascular reactivity (CVR) using vasodilatory stimuli in patients with migraine have yielded conflicting results. We postulated that CVR may not be static, and caffeine might negatively affect vasodilatory function via its vasoconstrictive effect.

Methods: In this prospective longitudinal observation study, we recruited patients with episodic migraine who were 18–50 years of age and free of vascular risk factors at the Samsung Medical Center between August 2015 and March 2020. Patients were classified into caffeine users and non-users at baseline, and caffeine users were instructed to discontinue caffeine intake. We measured the mean breath-holding index (BHI) of bilateral posterior cerebral arteries (PCA) using transcranial Doppler in all the included patients at baseline and followed up after 3 months. We compared breath-holding indices cross-sectionally between caffeine users and non-users and analyzed BHI changes according to caffeine cessation.

Results: In total, 84 patients completed the study protocol. Cross-sectional analysis showed that the baseline BHI of PCA was lower in caffeine users (n = 56, 1.1 [interquartile range (IQR) 0.8–1.3]) than that in nonusers (n = 28, 1.3 [IQR 1.0–1.5], p = 0.030). In the longitudinal analysis, caffeine quitters showed a significant improvement in BHI in PCA (baseline 1.1 [IQR 0.8–1.2], follow-up 1.3 [IQR 1.0–1.4], p = 0.034), whereas continuous users and non-users did not. Multivariable analysis showed an independent effect of caffeine cessation on the changes in BHI of PCA (unstandardized β = 0.27, 95% confidence interval 0.01–0.53, p = 0.044).

Conclusion: In patients with migraine, caffeine use is associated with reduced CVR in the posterior circulation, and caffeine cessation might be beneficial in improving CVR.


Migraine is associated with several vascular disorders, such as hypertension, ischemic heart disease, ischemic and hemorrhagic stroke, and cardiovascular mortality.[1–6] Several theories, including endothelial dysfunction, genetic predisposition, cortical spreading depression, arterial dissection, and coagulopathies have been proposed for the migraine–stroke association.[7,8] Among these, cortical spreading depression and a high prevalence of patent foramen ovale can only explain ischemic stroke, while more widespread vascular comorbidities such as hypertension and ischemic heart disease have not been explained by such mechanisms.

Endothelial dysfunction has been suggested as one of the possible causes of vascular comorbidities in patients with migraine.[9] Furthermore, migraine is more prevalent in patients with reversible cerebral vasoconstriction syndrome than in the general population,[10,11] suggesting that migraine can precipitate acute dysregulation of cerebral vasculature. While studies using systemic biomarkers of endothelial function have shown negative results,[12–14] several other studies used cerebrovascular reactivity (CVR) as a dynamic marker of cerebral endothelial function that is measured by the vasodilatory response to hypercapnic stimuli.[15] The higher the CVR value, the better the vascular response and the more active autoregulation. CVR is usually reduced in migraine, particularly in the posterior circulation.[16–18] However, conflicting results also exist,[19,20] and the results vary depending on the vascular distribution and presence of aura.[21,22] A recent meta-analysis concluded CVR is reduced in the posterior circulation in patients with migraine, particularly in patients with migraine without aura (MWoA), but the heterogeneity among studies was quite high.[23–25] The heterogeneity in CVR study results in migraine may be caused in part by dynamic changes in CVR or uncontrolled, unrecognized factors affecting CVR in patients with migraine. However, to the best of our knowledge, dynamic changes in CVR and factors affecting CVR changes have never been studied regarding migraine.

Caffeine is the world's most popular psychostimulant. Caffeine can be effective as an adjuvant acute treatment of migraine owing to its vasoconstrictive effects on cranial arteries.[26–28] However, the vasoconstrictive effect of caffeine may theoretically inhibit reactive vasodilation, which is the main function of the endothelium. The effect of chronic caffeine use on cerebral endothelial function has not been investigated yet. In studies of healthy chronic caffeine users, high-dose caffeine use was associated with lower cerebral blood flow (CBF), and caffeine discontinuation led to an increase in CBF.[23–25] From these findings, we postulated that the vasoconstrictive effect of caffeine may negatively affect cerebral vasodilatory function. We hypothesized that caffeine discontinuation may restore the vasodilatory function in patients with migraine.

In this study, we aimed to investigate the longitudinal changes in CVR in patients with migraine and the effect of caffeine and its discontinuation on CVR. CVR was measured in patients with episodic migraine without vascular risk factors and again at 3-month follow-up. The cross-sectional effect of chronic caffeine use and the longitudinal effect of caffeine cessation on CVR were analyzed.