Decreased outlet angle of the superior cerebellar artery as indicator for dolichoectasia in late onset Pompe disease

Patients and controls

For each Pompe patient two controls (n = 40; age 54.0 ± 14.7 years, range 19 to 80) were recruited that matched in gender, age (maximal ±3 years) and number of cerebrovascular risk factors (maximal ±1). Relevant risk factors were defined as hypertension, hyperlipidemia, obesity, smoking and diabetes. The control group consisted of patients with other neurological diseases that underwent cerebral MRI imaging for diagnostic purpose. This included headache (11/40), TIA and minor cardioembolic stroke (7/10), peripheral neuropathies (5/40), seizures (4/40), myopathies other than Pompe disease (3/40), other diagnoses (10/40) were retrobulbar optic neuritis, anterior ischemic optic neuropathy, spinocerebellar syndrome, dementia, and psychiatric diseases.

BA characteristics in 20 LOPD patients and 40 controls

In LOPD patients, there was a positive correlation between BA anatomical length (Pearson r = 0.626, p = 0.0.003), linear length (Pearson r = 0.522, p = 0.018), distal diameter (patients: Pearson r = 0.536, p = 0.015), and volume (Pearson r = 0.516, p = 0.020) with age.

The SUCA outlet angle (Figure 2) was negatively correlating with age in LOPD patients (Pearson r = − 0.605, p = 0.006) and in controls (Pearson r = − 0.483, p = 0.002).

Open image in new window

In total, in the cohort of all subjects (LOPD patients and controls) the SUCA outlet angle correlated with BA anatomic length (Pearson r = − 0,558, p < 0.001), the BA linear length (Pearson r = − 0.546, p < 0.001), the BA volume (Pearson r = − 0,545, p < 0.001). Correlations according to the Smoker criteria were as follows: with BA diameter (distal: Pearson r = − 0.407, p = 0.001; middle: Pearson r = − 0.363, p = 0.005; proximal: Pearson r = − 0.280, p = 0.032), and the grade of the BA bifurcation height (Pearson r = − 0.653, p < 0.001) and with the extend of the lateral displacement (Pearson r = − 0.287, p = 0.028).

There was no gender difference in BA diameter, anatomic length, linear length, volume or SUCA outlet angle in LOPD patients or controls.

BA characteristics during 5 year follow-up in 9 LOPD patients

Other cerebral abnormalities

During the observation period of 5 years, there were no major cerebrovascular events in 8/9 LOPD patients. However, one of the patients, a 75 year- old female, experienced an intracranial thalamic haemorrhage on the left side that led to complete immobilisation and subsequent death. This patient presented a remarkable dolichoectasia of the BA with an extreme cranial and posteriorly shifting of the BA bifurcation (Additional file 1 Figure S1).

White matter lesions were observed in 12/19 LOPD patients (63%) and 27/40 controls (68%). According to the Fazekas score grading of the LOPD patients was as follows: 40% grade 0, 30% grade 1, 25% grade 2, 5% grade 3. Grading of controls was: 32.5% grade 0, 40% grade 1, 20% grade 2, 7.5% grade 3. On a group level, Fazekas grade was not different in the LOPD patients compared to controls (0.95 ± 0.95 versus 1.05 ± 0.96). In controls, there were no major cerebrovascular abnormalities (e.g. tumor, major stroke, aneurysms).

None of the 9 LOPD patients showed progression of deep white matter lesions in within the 5 year period according to the Fazekas criteria.

The lowest SUCA outlet angle was found in the oldest LOPD patient (P17: 64°), who had a dilated BA and an additional severe mid-basilar stenosis. Other low SUCA outlet angles were found in one patient (P8: 79°) with microbleeds that were exclusively present in the posterior circulation and in one (P14: 83°) with a bilaterally fetal-type circle of Willis.

Dolichoectasia in patients and controls

The present study on 20 LOPD patients under enzyme replacement therapy (ERT) showed a significant dilation of the BA (regarding diameter and lengths) compared to matched controls. Similar changes have previously reported in 10 of these 20 patients [7]. However, in the present study a more accurate and objective assessment of the vessel diameter was performed using FWHM, based on the pixel-wise intensity [9]. The findings in our cohort are consistent with those of previous studies that found predominant dolichoectasia in the posterior cerebral circulation in LOPD patients with and without ERT [3, 11]. The reason for these changes remains unclear. It was speculated that glycogen deposition in the smooth muscles of the intima of arterial vessels provokes their dilation [2, 12]. In this study all patients received ERT and its impact on cerebral vessels remains enigmatic. However, there is a report that suggests glycogen clearance in smooth muscle via ERT in the arrector pili muscle [13].

However, vertebrobasilar dolichoectasia is not a specific phenomenon in Pompe disease: In general the posterior circulation has less sympathetic innervation than the anterior circulation leading to a predisposition to deformity, when exposed to increased blood flow and pressure [14]. Subsequently, there is physiological progressive luminal dilation, fragmentation of the internal elastic lamina, and elastin loss in aging individuals, that mainly affect the vertebrobasilar arteries [15, 16, 17]. Besides a higher age, also a male gender and cardiovascular risk factors (hypertension, smoking, obesity, diabetes, or dyslipidemia) have been found to be associated with evolution of dolichoectasia [18]. Therefore LOPD patients were individually matched with controls of equal epidemiological and cardiovascular parameters.

Since signs of dolichoectasia were observed in both groups, the LOPD patients and the matched controls, additional factors for the development of dilative arteriopathy have to be assumed that result from the comorbidities. However, the BA ectasia was more pronounced in the LOPD cohort compared to controls. It is therefore suggested that Pompe disease acts as independent factor for the development of BA dolichoectasia.

Until now, the most validated parameters for BA evaluation are the semi-quantitative Smoker’s criteria [8]: An abnormal height of BA bifurcation implies that the BA bifurcation is cranial shifted to the level of third ventricle floor or higher. Normally, the SUCA originates shortly under the BA bifurcation at the height of the free edge of the tentorium cerebelli [19] and the SUCA outlet angle is nearly 180°. If the SUCA origin and the adjacent basilar bifurcation shift cranial, then the SUCA outlet angle will consequently decrease. The SUCA is most consistent and the angle is easily obtained from coronal CT- or MR angiography images, independently from different imaging protocols. Therefor in this study the outlet angle of the SUCA was introduced as a novel parameter for a quantitative analysis of this cranial shift of BA bifurcation. It was demonstrated that the SUCA outlet angle in LOPD patients was significantly reduced compared to controls together with other indicators of a prominent BA ectasia in these patients. Additionally, there was a significant correlation of the SUCA outlet angle with all parameters of the Smoker’s criteria: BA diameter, bifurcation height and lateral displacement.

Analysis of progression of dolichoectasia in LOPD patients

In the present study, according to the Smoker’s criteria [8], only 3/9 LOPD patients showed an abnormal height of the BA bifurcation at baseline but 5/9 LOPD patients after 5 years follow up. In the 5 year follow up analysis there was a significant decline of the SUCA outlet angle in the cohort of 9 LOPD patients on ERT that suggests a progressive dolichoectasia in these patients. The higher slope of the regressionline of the LOPD cohort compared to controls in the age dependent analysis of the SUCA outlet suggests that these differences result from disease specific mechanism rather from the aging of the patients.

However, other parameters that define BA dolichoectasia (length, diameter, and lateral displacement of the BA) remained nearly unchanged. Similar results were reported in a recent study showing stable vertebrobasilar diameters over a period of up to 10 years in 5 LOPD patients [20]. This might be due to the relatively small diameter and lengths of the BA. Changes of these parameters may need time to evolve and are therefore not easily detectable in the 5 year observation period. However, presumed that all vertebrobasilar arteries are affected by a weakening of the vessel wall, and then the cranial migration of the BA bifurcation might be a result of a summation effect of the elongation of the BA plus the adjacent vertebral arteries. Since vasodilation results in reduced flow velocity, the dilation of cerebral arteries might be underestimated in TOF-based MRA. Limitation of this study might be the use of slight differently TOF-MRI protocols in the follow up cerebral imaging. However, this has no impact on the SUCA outlet angle. It is suggested that the SUCA outlet angle is a sensitive surrogate marker for BA dolichoectasia, which detects the elevation of the height of basilar bifurcation in a more precise and quantitative way than the Smoker’s grading.

Clinical relevance

It is well known that dolichoectasia can lead to brainstem compression, hydrocephalus microbleeds and stroke [18, 21, 22]. There are several reports on strokes [6, 23] and intracranial mirco – and macrohemorrhages [4, 5, 24] in Pompe patients. In the actual LOPD cohort a fatal thalamic hemorrhage occurred in one patient with the most prominent basilar dolichoectasia. It can be speculated, that the vertebrobasilar pathology might at least has contributed to the hemorrhage in this patient. However, in all 9 patients there was no progression of white matter lesions after the follow up of 5 years.

It seems therefore appropriate, that patients without cerebrovascular abnormalities undergo a follow up of MRI analysis after 5 years. In patients with cerebrovascular abnormalities, we recommend the strict treatment of cerebrovascular risk factors (e.g. hypertension) and follow up imaging after 1 to 2 years.