Migraine Headache Is Not Associated With Cerebral Or Meningeal Vasodilatation - A 3T Magnetic Resonance Angiography Study

G. G. Schoonman; J. van der Grond; C. Kortmann; R. J. van der Geest; G. M. Terwindt; M. D. Ferrari


Brain. 2008;131(8):2192-2200. 

In This Article


We used a well-established NTG provocation model to induce faithfully migraine attacks and a highly sensitive, non-invasive 3T MRA technique to visualize and measure even small intra-individual diameter changes of cerebral and meningeal blood vessels. Contrary to longstanding and widespread belief, we failed to detect any evidence for a clinically relevant vasodilatation of major cerebral or meningeal blood vessels during migraine attacks. This finding has important implications for the understanding of the pathophysiology of the migraine headache and the development of future anti-migraine agents. Novel anti-migraine treatments may not require vasoconstrictor activity as predicted earlier (Goadsby et al., 1990).

In our provocation experiments, we infused NTG over a 20 min period and observed a vessel-dependent 7-30% vasodilatation at 10 min after beginning of the infusion. The vasodilatatory effect is believed to be due to a direct local effect of nitric oxide on vascular smooth muscle cells (Andresen et al., 2006) or through the release of vasoactive peptides such as CGRP (Strecker et al., 2002; Wei et al., 1992). Our findings on the early vascular effect of NTG are in accordance with those of (Hansen et al., 2007). Using 1.5T MRA they found a peak vasodilatation at 10-15 min after beginning of the NTG infusion and a normalization of the vascular diameters back to baseline at 45 min after stopping of the infusion. For logistic reasons, we did not scan at 45 min after the infusion to confirm normalization of the blood vessel diameter. However, in view of the well known short duration of action of NTG (Abrams, 1985) and the observed time course of the early vascular responses by (Hansen et al., 2007), we feel confident that blood vessel diameters had returned to baseline by 1 h after the second (infusion) scan. Therefore, it seems justified to compare measurements during attacks with those obtained at baseline, before infusion.

The most important finding of the present study is that migraine headache was not associated with a clinically relevant vasodilatation of major cerebral or meningeal blood vessels, not even when controlled for headache side. We feel confident that this was not due to too low a sensitivity of the detection method. The very fact that we were able to detect an early transient vasodilatation in response to NTG of as low as 7% shows that the method we used is sufficiently sensitive to measure even small diameter changes. The clinical relevance of smaller changes is doubtful as during NTG infusion, we observed an up to 30% increase in blood vessel diameter without associated migraine headache. Our results are also in agreement with at least some older transcranial Doppler studies failing to show blood velocity changes indicative for vasodilatation during migraine attacks (Caekebeke et al., 1992; Zwetsloot et al., 1993; Limmroth et al., 1996; Gori et al., 2005) Finally, BA and ICA blood flow did also not change during migraine attacks. Cerebral blood flow is dependent on cardiac output, arterial calibre and vasomotor tone in small resistance vessels (Guyton, 2006). As blood pressure (as a measure for cardiac output; data not shown) and the BA and ICA diameters had not changed, it seems likely that there were also no changes in the intracranial resistance microvasculature during migraine attacks. In conclusion, our data seem to refute an important role of cerebral or meningeal vasodilatation in causing migraine headache. This would certainly be in accordance with observations that non-vascular mechanisms, such as exposure to sildenafil, (Kruuse et al., 2003) are capable of inducing migraine attacks.

Potential limitations of our study include that we did not measure just before or at the onset of the migraine headache. We could thus have missed a brief transient vasodilatation at the very beginning of the migraine headache. Although unlikely, we cannot exclude this possibility. Another important question is whether and to what extent NTG-provoked migraine attacks are similar to spontaneous attacks. There are strong clinical and pathophysiological arguments in favour of this notion. The clinical symptoms and features, including the occurrence of premonitory symptoms several hours before the headache (Afridi et al., 2004) and the response to anti-migraine drugs (Iversen and Olesen, 1996), are strikingly similar between spontaneous and NTG-induced attacks. Likewise, in both there is an increase of CGRP in jugular venous blood (Goadsby et al., 1990; Juhasz et al., 2003) and activation of the dorsal rostral brainstem on positron emission tomography (Weiller et al., 1995; Bahra et al., 2001). The fact that NTG provokes migraine aura's only rarely, even in patients with migraine with aura (Christiansen et al., 1999; Afridi et al., 2005a), seems to point at a trigger site of action beyond the aura triggering mechanism. We thus feel confident that our findings in NTG-provoked attacks can be extrapolated to spontaneous migraine headaches.

In this study, we did not observe significant changes in blood vessel diameter or blood flow during the headache phase of provoked migraine attacks. However, there were some (non-significant) changes in the posterior circulation that need to be discussed. First, the diameter of the BA did not return to baseline levels, unlike the other blood vessels. This was, however, true for both patients who had developed a delayed migraine headache and for those who had not. Secondly, the blood flow in the BA was decreased (although not significantly) from 174 ml/min at baseline to 129 ml/min in patients who had developed a migraine headache after GTN, whilst there was no such change in patients who had not developed a migraine headache. Whether these findings are clinically relevant, needs to be explored. A tentative correlation, for instance, could be made with previous findings of in previous studies our group has shown our group demonstrating increased prevalence of pontine hyperintensities and cerebellar infarcts in migraineurs from the general population (Kruit et al., 2004, 2006).

We conclude that, contrary to a longstanding and widespread belief, cerebral and meningeal diameter changes in migraine attacks, if at all happening, appear not to be of primary importance to the pathophysiology of the migraine headache.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.