Introduction

Intracerebral arteriovenous malformations (AVMs) are caused by defects in the development of blood vessels and are characterized by an arteriovenous shunt without a capillary bed, but with the presence of an arterial nidus [4, 7]. AVMs were traditionally thought to be congenital in origin, but several recent de novo cases challenge this notion [3, 4, 7, 15]. However, the pathogenesis of AVMs is not completely understood, and many AVMs have been shown to progress as well as undergo spontaneous regression [6]. We present a patient with a de novo AVM that was found incidentally 14 years after being diagnosed with Bell’s palsy on the same side. We review this case in the context of others reported in the literature looking for common patient trends in the formation of de novo AVMs.

Case report

At the age of 16, the patient was diagnosed with Bell’s palsy. Her magnetic resonance imaging (MRI) scans from 1995 revealed no vascular abnormality, abnormal signal intensity, or any other suggestion of an arteriovenous malformation (Fig. 1).

Fig. 1
figure 1

Axial proton-density MRI (a) and T2-weighted MRI (b) obtained in 1995, at the time of patient’s diagnosis with Bell’s palsy. There is no evidence of any vascular lesion

Full size image

Fourteen years later, the patient presented with a 5-year history of complicated migraines, which were at times associated with right-sided facial weakness. Two to three weeks prior to diagnosis, she had an episode where she could not speak, which lasted for about 10 to 15 min. This episode of speech arrest was likely a seizure. In order to determine the cause of her migraines and seizure, intracranial imaging was performed. Her post-contrast computed tomography (CT) and MRI revealed a left frontoparietal arteriovenous malformation (Fig. 2). She then underwent bilateral carotid angiography, and the presence of the AVM was confirmed in her left frontoparietal region (Fig. 3).

Fig. 2
figure 2

Axial post-contrast CT (a) and T2-weighted MRI (b), obtained 14 years after initial presentation, demonstrating an irregular tangle of blood vessels with enlarged draining veins, suggestive of a left frontoparietal arteriovenous malformation

Full size image
Fig. 3
figure 3

Anteroposterior (a) and lateral (b) angiograms depicting the nidus, supplying arteries, and early draining veins of the arteriovenous malformation

Full size image

Discussion

Once thought of as static lesions, AVMs are now becoming more accepted as being dynamic lesions. To our knowledge, there have been six previously reported cases of de novo AVM formation [3, 4, 7, 11, 13, 15]. We present the seventh reported case of a de novo AVM and the first reported case occurring in a patient with a history of Bell’s palsy.

Histopathologically, AVMs are characterized by vessels with thin or irregular muscularis and elastica, endothelial thickening, and media hypertrophy in the nidus [5]. These lesions are typically thought to be congenital in nature, arising around the 28th day of intrauterine life [7], and are thought to be caused by a defect in primordial capillary or venous formation [6]. Nonetheless, very few cases are detected in utero or in children [15], in comparison to the vast majority of diagnoses made during adulthood [5, 15].

However, the dynamic nature of cerebral AVMs is attested to by multiple recent case reports demonstrating spontaneous progression and resolution [2, 6, 8]. Of the six previous cases of de novo AVM formation, four were reported in children (Table 1). Three of the four children had etiologies that may have contributed to the formation of the AVM (Table 2). The fourth child developed a de novo AVM after suffering trauma to the head.

Table 1 Summary of patient characteristics in reported cases of acquired or de novo arteriovenous malformations
Full size table
Table 2 Cerebral findings in reported cases of acquired or de novo arteriovenous malformations
Full size table

In addition, six of the seven reported cases of de novo AVMs, including our patient, were found in females. On average, these lesions developed 8 years (3–17) after initial imaging, with a mean age at diagnosis of approximately 18 years (6–32). In contrast, the mean age of patients diagnosed with congenital AVMs is 33 years [10]. Hence, the initial inflammatory or ischemic insult occurring in these patients could potentially have accelerated the development of the de novo AVM in comparison to the growth of congenital AVMs.

Critics argue that these reported cases of de novo AVMs may be attributed to the lack of sensitivity and specificity of imaging techniques, specifically angiography, and that de novo AVMs may be due to the growth of a previously small but occult AVM [7]. Angiography may fail to demonstrate arteriovenous shunting due to partial or complete thrombosis, previous hemorrhage, or adjacent edema [7]. However, most reported cases of de novo AVMs have been in the last decade, during which time the availability of high-resolution, serial cross-sectional imaging has made it possible to detect lesions that were once thought to be occult.

In our patient, the lesion developed 14 years after the initial MRI and CTA study. These initial scans revealed no vascular abnormality or abnormal signal intensity (Fig. 1). Subsequent imaging revealed lesions specific enough for the diagnosis of a de novo AVM (Figs. 2 and 3).

This case may represent the third instance in which an inflammatory insult led to the formation of an AVM [4, 13]. Several studies have shown that vascular endothelial growth factor may be the key link between the initial inflammatory injury and the growth of the AVM [9, 12, 14–17].

Recently, isolation and characterization of circulating endothelial cells from hereditary hemorrhagic telangiectasia patients has revealed a decreased expression of endoglin, impaired activin receptor-like kinase-1 (ALK-1), and ALK-5 dependent transforming growth factor-β signaling, disorganized cytoskeleton, and a failure to form cord-like structures, which may lead to fragile, small blood vessels or abnormal angiogenesis after injuries [1].

Whether similar abnormalities exist in some patients that contribute to abnormal repair processes following injury, and lead to some of these acquired arteriovenous malformations, is a provocative hypothesis.

In conclusion, this case represents only the seventh reported case of a de novo AVM, and adds to the growing body of evidence challenging the traditional view that all AVMs are congenital in nature. Acquired AVMs may be more common than traditionally believed, and it is clear that a more thorough understanding of the pathogenesis of AVMs is needed to better understand their development.