Diffuse cerebrovascular dilation: Case report of amezinium metilsulfate-induced reversible cerebral vasoconstriction syndrome

Introduction

Reversible cerebral vasoconstriction syndrome (RCVS) usually presents with thunderclap headache attacks (i.e. sudden-onset severe headaches reaching a pain zenith in less than a minute) over one to two weeks, frequently accompanied by moderate headaches between the attacks (1). Cerebral angiography often reveals no abnormality within one week after headache onset, followed by widespread segmental vasoconstriction (1–4). The vasoconstriction typically resolves in three months (5). Another important characteristic is that this syndrome is precipitated by various factors, including the postpartum state and vasoactive medications (e.g. illicit drugs, antidepressants, and sympathomimetics) (1–5). Here, we report a rare case of a patient with RCVS induced by amezinium metilsulfate (amezinium, an indirect sympathomimetic) in whom magnetic resonance angiography (MRA) revealed diffuse dilation of cerebral arteries before segmental vasoconstriction. In addition, MRAs were performed serially and mean sectional areas of cerebral arteries were measured for quantitative analysis.

Case report

A 34-year-old woman with no history of headache was prescribed amezinium 20 mg/day, difenidol hydrochloride 75 mg/day, and adenosine triphosphate disodium hydrate 0.3 g/day for hypotension and dizziness. Twelve days later, when she was cleaning house, a severe headache occurred suddenly, reaching a peak pain level in a few seconds. The headache was too excruciating for her to sit still but resolved after three hours, leaving her with a residual moderate headache. Although she stopped taking the prescribed medications, similar severe headaches recurred on days 2 and 5 after the headache onset. In addition, the moderate headache persisted all day. She was referred to our department; no neurological deficits were found. Nonenhanced brain computed tomography, routine blood tests, and vasculitis screening were normal. However, magnetic resonance imaging (MRI) performed on day 9 with a 3-Tesla Siemens Verio system revealed linear hyperintensities in a wide range of sulcal spaces, with a predominance in the temporo-occipital lobes on fluid-attenuated inversion recovery (FLAIR) images (Figure 1(a), left, arrows). Furthermore, MRA (three-dimensional time-of-flight MRA obtained from original images with the maximum intensity projection method) showed diffusely dilated anterior, middle, and posterior cerebral arteries (Figures 1(b), left and 2(a)) with vasoconstriction in the posterior cerebral arteries (Figures 1(b), left and 2(a), arrowheads). Mild vasoconstriction was also observed in the second segment of the middle cerebral artery and basilar artery (Figure 1(b), left, arrowheads). The first segments of middle cerebral arteries (Figures 1(b), left and 2(a), arrows) were as thick as internal carotid arteries. The patient was tentatively diagnosed with RCVS; treatment with oral verapamil 120 mg/day was started on the same day. Lumbar puncture was not performed because the clinical manifestations and magnetic resonance findings suggested a low probability of other diseases. After treatment, the thunderclap headache did not recur, and the moderate headache disappeared on day 12. On day 15, follow-up MRA revealed widespread segmental vasodilation and vasoconstriction (Figures 1(b), right and 2(b), arrowheads), whereas the linear hyperintensities on FLAIR images disappeared (Figure 1(a), right). Repeated MRA indicated mild vasoconstriction on day 37 (Figure 2(c), arrowheads) and no abnormality on days 65 (Figure 2(d)) and 91, leading to a final diagnosis of RCVS (6). After we confirmed cerebrovascular normalization and stabilization on day 91, verapamil was tapered off without recurrence. OPEN IN VIEWER

OPEN IN VIEWER

Figure 2.

Magnetic resonance angiographies (axial view) performed on days 9 (a), 15 (b), 37 (c), and 65 (d) after symptom onset. (a) On day 9, anterior, middle, and posterior cerebral arteries were diffusely dilated with vasoconstriction in the posterior cerebral arteries (arrowheads). The first segments of the middle cerebral arteries (arrows) were approximately the same size as internal carotid arteries. (b) The dilated cerebral arteries were constricted prominently (arrowheads) and partially changed to segmental vasodilation and vasoconstriction on day 15. (c) Mild vasoconstriction (arrowheads) remained on day 37. (d) Cerebral arteries were normalized on day 65.

We performed a quantitative analysis on the original MRA images (repetition time, 22 ms; echo time, 3.64 ms; flip angle, 20 degrees; image size, 384 × 384; thickness, 650 µm) using a DICOM viewer OsiriX 6.0. Vascular volumes were measured and vascular lengths were calculated by summing distances of the vascular central positions on two consecutive original images. Mean vascular sectional areas were calculated by dividing the volumes with lengths. Mean sectional areas of the first segments of the middle cerebral arteries (M1; Figures 1(b), left and 2(a), arrows) and precavernous portions of the internal carotid arteries (pICA; Figure 1(b), left, asterisks) were measured serially (Table 1).

OPEN IN VIEWER

Table 1.

Mean sectional areas of cerebral arteries.

	Sectional area (mm2)				MRA findings on MIP images
Interval from	Right		Left
onset (day)	M1	pICA	M1	pICA	M1	pICA
9	10.1	8.4	9.9	11.2	Diffuse dilation	Normal
15	6.6	8.6	6.2	12	Segmental dilation and constriction	Normal
37	4.6	8.2	5.4	12.1	Mild constriction	Normal
65	5.6	7.6	6.0	12.2	Normal	Normal
91	5.5	7.6	6.0	12.4	Normal	Normal

On day 9 after symptom onset, the mean sectional areas of M1 (right, 10.1 mm²; left, 9.9 mm²) and pICA (right, 8.4 mm²; left, 11.2 mm²) were approximately the same size on average (M1, 10 mm²; pICA, 9.8 mm²). The M1 sectional areas became smaller on days 15 and 37, followed by mild rebound and stabilization on days 65 and 91 (right, 5.5 mm²; left, 6.0 mm²), whereas the pICA sectional areas did not change remarkably. Although three months may be insufficient for complete normalization of cerebral arteries (1–5), the M1 sectional areas on day 9 were obviously larger than those on day 91, and about the same size of the pICA sectional areas, indicating that the M1 segments were dilated initially. The vasodilation disappeared on day 37 with residual mild vasoconstriction, corresponding to the smallest M1 sectional areas on the same day.

Discussion

We have presented an adult female patient with RCVS induced by amezinium in whom MRA revealed diffuse cerebrovascular dilation before segmental vasoconstriction. Typically in patients with RCVS, MRA reveals no abnormality before vasoconstriction, although severe headaches are present (1–4).

In the present case, MRA revealed diffuse vasodilation in the anterior, middle, and posterior cerebral arteries on day 9, followed by widespread segmental vasoconstriction on day 15. In addition, the linear FLAIR hyperintensities on day 9 probably suggested slow blood flow in small cerebral surface arteries or their leptomeningeal anastomotic collaterals in RCVS patients (1,7). The diffuse cerebrovascular dilation and FLAIR abnormalities were apparent simultaneously and possibly caused by a similar mechanism in RCVS. Although adenosine triphosphate, which was prescribed before headache onset, has a short-lived vasodilatory effect on cerebral arteries, the effect could not have lasted for nine days after discontinuation. In contrast to segmental arterial dilation on cerebral angiogram (1,2,4,5), the diffuse dilation observed in this case has not been pointed out before, but may have pathophysiological implications.

The diffuse vasodilation may be caused by vasoconstriction of small cerebral arteries. In RCVS, it has been proposed that the pathological process affects distal arteries initially, and subsequently progresses toward the branches of the circle of Willis (1,4). Therefore, in the initial stage of RCVS, small arteries that cannot be visualized on MRA images may be constricted selectively, leading to blocked blood flow and dilated arteries proximal to the constricted segments. The proximal artery dilation accompanied by blocked and slow blood flow in small arteries may correspond to the diffuse cerebrovascular dilation on MRA images and linear FLAIR hyperintensities.

Amezinium probably induced RCVS in our case, although other drugs could not be neglected. The initial three drugs our patient took presumably have not been reported as etiologies of RCVS; however, amezinium was the most likely cause because of its sympathomimetic effect. Sympathomimetics such as phenylpropanolamine, pseudoephedrine, ephedrine, adrenaline, and noradrenaline are known to cause RCVS (1,8–10). Amezinium is an indirect sympathomimetic that inhibits monoamine oxidase and noradrenaline uptake with increased noradrenaline release at the terminals of peripheral sympathetic nerves, resulting in the enhanced effect of noradrenaline (11); therefore, it is logical to posit that amezinium can induce RCVS as well as noradrenaline. Although adenosine triphosphate, which also has a vasoactive effect, could trigger or cooperate together to trigger RCVS, we consider it probable that amezinium mainly induced RCVS in our case because of the above-mentioned sympathomimetic effect.

In general, MRA is very sensitive to vascular narrowing and turbulent blood flow, thus the measured sectional areas may be different from actual ones; however, 3-Tesla MRA is employed as a useful and noninvasive method (12). Therefore, we consider the measured areas can at least be surrogate markers; the vasodilation is confirmatory.

We described a patient with RCVS whose MRA revealed diffuse cerebrovascular dilation before typical vasoconstriction. The dilation may be associated with RCVS pathophysiology. Moreover, in clinical practice, physicians should bear in mind that amezinium can potentially induce RCVS.

Clinical implications

Early cerebral angiogram may reveal diffuse cerebrovascular dilation in patients with reversible cerebral vasoconstriction syndrome (RCVS).