Introduction

The rostral ventrolateral medulla (RVLM) is an important center for the regulation of sympathetic and cardiovascular activities.1, 2 After the initial report by Jannetta and Gendell in 1978,3 several clinical studies have suggested an association between neurovascular compression (NVC) of the RVLM and essential hypertension (EH).4, 5, 6, 7, 8, 9, 10 We have reported that magnetic resonance imaging (MRI) with a high-resolution matrix revealed a significantly higher incidence of NVC of the RVLM in EH patients as compared with those with secondary hypertension or normotensive patients, although the degree of organ damage from hypertension was not significantly different between the two groups of hypertensive patients.11, 12 Jannetta et al.13, 14 and we15 have reported that microvascular decompression (MVD) of the RVLM decreases blood pressure (BP) in hypertensive patients with NVC of the RVLM. Therefore, it is assumed that MVD of the RVLM could be a useful therapeutic strategy to reduce BP in these patients. As MVD is an invasive therapeutic strategy, it seems worthwhile to seek useful antihypertensive agents for hypertensive patients with NVC of the RVLM. We and others have reported that EH patients with NVC of the RVLM show increased plasma levels of norepinephrine (NE).15, 16 Furthermore, these patients show an increased muscle sympathetic nerve activity (SNA).17, 18 Therefore, although sympatholytic agents are expected to lower BP in these patients, this remains to be clarified.

Cilnidipine is a unique Ca2+ channel blocker that not only inhibits vascular L-type and sympathetic N-type Ca2+ channels but is also used to treat patients with EH.19, 20, 21, 22 The N-type voltage-dependent calcium channel has an important role in sympathetic neurotransmission and regulates the release of NE from sympathetic nerve endings.23 Cilnidipine inhibits SNA and heart rate and lowers BP in hypertensive patients. Therefore, in this study, we examined the effect of cilnidipine on BP in patients with EH with NVC of the RVLM.

Methods

Case selection

Subjects were selected among patients with untreated essential hypertension who visited our hospital between April 2006 and March 2008. They were all Japanese and those with secondary hypertension (renal and endocrine sources of hypertension) were excluded by appropriate physical and laboratory evaluation. The patients had to have a systolic pressure of ⩾140 mm Hg and/or diastolic pressure of ⩾90 mm Hg on three consecutive office visits. Patients who had suffered a hemorrhagic stroke or cardiac infarction in the previous 6 months, as well as those with chronic renal failure (serum creatinine of ⩾1.5 mgdl−1), pregnant women and those with peripheral vascular or malignant disease, were excluded.

MRI evaluation

MRI studies were performed with a 1.5-T system (Signa Excite HD, GE Medical Systems, Milwaukee, WI, USA) and a neurovascular coil. The slices were axial, parallel to a line connecting the anterior commissure and the posterior commissure. T2-weighted three-dimensional fast recovery fast spin echo (3D-FRFSE) images (repletion time [TR]/echo time [TE], 1500/15 msec; echo train length, 48) were obtained. The voxel size was 1.4 × 0.6 × 0.8 mm. Two experienced radiologists (KI, MM), who were blind to the individuals' medical histories, assessed all scans independently to check for the presence of NVC of the RVLM. In instances of disagreement between the reviewers, a consensus interpretation was used for the case. NVC of the RVLM was defined as an arterial vessel or vascular loop with its convexity contacting the surface of the RVLM. The anterior border of the RVLM was defined as the transitional point of the olivary convexity to the concavity of the retro-olivary sulcus, and the posterior border was at the junction of the parenchymal brain tissue to the nerve fibers. Upper and lower borders were determined by the uppermost and the lowest end of the glossopharyngeal and vagal nerve bundles entering the medulla.

Measurement of office BP and pulse rate (PR)

Office BP and PR were measured from 1100 to 1200 hours when the patient was seated after at least 5 min of rest with the use of an automated oscillometric sphygmomanometer (MPV3301, Nihon Kohden, Tokyo, Japan). Three measurements were obtained 2 min apart at each time point. The average of the second and third readings was used for all analyses.

Measurement of home BP and PR

Home BP and PR were measured in the morning using an automated oscillometric sphygmomanometer according to the Japanese Society of Hypertension Guidelines for Self-Monitoring of Blood Pressure at Home;24 within 1 h after waking, after micturiction, sitting after 1 to 2 min of rest, before drug ingestion, and before breakfast. Home BP and PR data were determined as the averages of the measurements for 3 days before the time of estimation.

Laboratory analyses

Morning blood samples were collected from the antecubital vein after an overnight fast of >12 h and after the subjects lay in a supine resting position for at least 20 min. Serum levels of urea nitrogen (BUN), creatinine, fasting blood sugar (FBS), hemoglobin A1c (HbA1c), low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C) and triglycerides (TG) were assessed by conventional methods at the clinical laboratory center of our hospital. Plasma NE concentration was measured by high performance liquid chromatography with electrochemical detection by an external laboratory (SRL Inc., Tokyo, Japan).

Echocardiography

M-mode and pulse Doppler echocardiography was performed blindly to the MRI findings using a Vivid Seven scanner (GE VingMed, Milwaukee, WI, USA). The left ventricular end-diastolic dimension, left ventricular end-systolic dimension, end-diastolic interventricular septal thickness, and posterior wall thickness were measured on the M-mode echocardiogram. The left ventricular mass was calculated using Devereux's formula,25 and left ventricular mass index (LVMI) was obtained by dividing the left ventricular mass by the body surface area. On the pulse Doppler echocardiogram, the sample volume was placed between the anterior and posterior mitral leaflets to record the left ventricular inflow patterns; the maximum amplitudes of the early diastolic transmitral flow wave (E wave) and atrial systolic transmitral flow wave (A wave) were then estimated and used to calculate the E/A ratio. The E wave deceleration time (Dct) was also estimated. In addition, the percent fractional shortening (%FS) of the left ventricle was calculated using left ventricular internal dimensions (LVDd and LVds). The ejection fraction of the left ventricle was also calculated according to the American Society of Echocardiography's criteria using the area length method.26

Study design

At enrollment we collected information on age, gender, duration of hypertension, current diseases, medication, smoking history, body mass index (BMI), waist circumference, biochemical levels (BUN, creatinine, FBS, HbA1c, LDL-C, HDL-C, and TG), plasma NE level and echocardiographic findings. All patients were treated with 10 mg of cilnidipine monotherapy (once daily in the morning) for 16 weeks. During this study, additional treatment or change of dosing of concomitantly administered anti-hyperlipidemic, anti-diabetic or anti-platelet drugs was prohibited.

BMI, waist circumference, biochemical levels, plasma NE level and echocardiographic findings were measured at enrollment and at the end of therapy. Changes between pre- and post-treatment values were compared between patients without (NVC− group) and with (NVC+ group) NVC of the RVLM.

Office and home BP and PR were obtained at enrollment, and at 8 and 16 weeks of treatment. Changes in these values at weeks 8 and 16 in groups NVC− and NVC+ were also compared.

Safety

Information on adverse events was routinely recorded at each hospital visit; investigators were asked to give their opinion on causality.

Statistical analysis

To detect a 15 mm Hg between-group difference in systolic BP change with a power of 80% and a minimum level of significance of 0.05, given a s.d. of 10 mm Hg, 20 patients per group were necessary. Taking into account that about 10% of patients could not be evaluated, a total of 45 assigned patients were required.

Paired Student's t-test was applied to determine the significance of differences between pre- and post-treatment values. Chi-squared test and a non-paired Student's t-test were performed to assess the statistical differences between the NVC(−) and NVC(+) groups. Data are shown as the mean values±s.d., and differences were considered statistically significant at P<0.05.

The study was approved by the ethics committee of Kansai Medical University. Written informed consent was obtained from all individuals before their participation in the trial. The authors had full access to the data and take responsibility for their integrity. All the authors have read the paper and agree with the content.

Results

MRI evaluation and study subjects

The reviewers considered that out of the 48 MRI studies they analyzed, that of one patient with EH (2.1%) was inadequate for assessment because of poor images and was excluded from further evaluation. NVC was observed on the left side in 13 patients, right side in 7, and on both sides of the RVLM in 4 of 46 patients; there were 22 patients in the NVC(−) group and 24 in the NVC(+) group. The contacting artery was the vertebral artery (VA) in four patients, posterior inferior cerebellar artery (PICA) in seven, and the anterior inferior cerebellar artery in 10 of the 20 patients with NVC of the ipsilateral RVLM. As for the three patients with NVC of the bilateral RVLM, the contacting arteries were VA and PICA in one, PICA and anterior inferior cerebellar artery in another and two PICAs in the third patient.

Changes in baseline characteristics

Baseline characteristics were well matched in the two groups; as summarized in Table 1, age, gender, duration of hypertension, smoking status, BMI, waist circumference and biochemical parameters did not differ significantly between the two groups. The number of patients taking anti-hyperlipidemic drugs (three in NVC− and four in NVC+), anti-diabetic drugs (two and two) and anti-platelet drugs (three and two) did not differ significantly either.

Table 1 Characteristics of the subjects
Full size table

No subjects changed their smoking status during the study period. No patients had their anti-hyperlipidemic, anti-diabetic, or anti-platelet drugs changed during the study. BMI (24.2±3.5 and 24.6±4.2 kg m−2 in the NVC(−) and NVC(+) groups, respectively, after treatment), waist circumference (84.2±9.5 and 83.8±9.8 cm), BUN (15.2±5.3 and 14.8±4.9 mg dl−1), creatinine (0.74±0.12 and 0.72±0.17 mg dl−1), FBS (113±7 and 112±6 mg dl−1), HbA1c (5.4±0.7 and 5.3±0.8%), LDL-C (132±40 and 134±38 mg dl−1), HDL-C (65±15 and 63±19 mg dl−1), and TG (147±83 and 141±78 mg dl−1) did not change significantly because of treatment. Changes in these factors did not differ significantly between the two groups.

Changes in office BP and PR

Office systolic and diastolic BPs were not significantly different between the two groups (Figure 1a). Office systolic and diastolic BPs were significantly decreased at 8 and 16 weeks of treatment with cilnidipine in the NVC− and NVC+ groups (Figure 1a). The net changes and percentage changes in office systolic and diastolic BPs were significantly greater in the NVC+ group than in the NVC− group at 16 weeks of treatment (Table 2). Office PR did not differ significantly between the two groups before treatment and did not change significantly in either group at 8 and 16 weeks of treatment (Figure 1a). Neither the net change nor the percentage change in office PR differed significantly between the two groups (Table 2).

Figure 1
figure 1

Time courses of changes in blood pressure (BP) and pulse rate (PR) during treatment with cilnidipine. (a) Office BP and PR. (b) Home BP and PR. Dotted line, hypertensive patients without neurovascular compression of the rostral ventrolateral medulla (NVC−) (n=22); solid line, hypertensive patients with neurovascular compression of the rostral ventrolateral medulla (NVC+) (n=24). sBP, systolic blood pressure; dBP, diastolic blood pressure. *P<0.001 vs. before treatment; †P< P<0.05 vs. NVC− group.

Full size image
Table 2 Net changes in office and home blood pressure and pulse rate (PR) at 8 and 16 weeks of treatment with cilnidipine
Full size table

Changes in home BP and PR

Before cilnidipine treatment, home diastolic BP did not differ significantly between the two groups, whereas home systolic BP was significantly higher in the NVC+ group than in the NVC− group (Figure 1b). Home systolic and diastolic BPs were significantly decreased at 8 and 16 weeks of treatment with cilnidipine in both groups (Figure 1b). The net changes and percentage changes in home systolic and diastolic BPs were significantly greater in the NVC+ group than in the NVC− group at 16 weeks of treatment (Table 2). Home PR did not differ significantly between the two groups before treatment and did not change significantly in either group at 8 and 16 weeks of treatment (Figure 1b). Neither the net change nor the percentage change in home PR differed significantly between the two groups (Table 2).

Changes in plasma NE level

Although the plasma NE level varied (coefficient of variation: NVC− group, 38.9%; NVC+ group, 30.9%), it was significantly higher in the NVC+ group than in the NVC− group before treatment (Table 1), in accordance with our previous data and those of others.15, 16 Plasma NE decreased significantly in the NVC+ group (326±125 pg ml−1, P<0.01) and slightly in the NVC− group (330±110 pg ml−1), showing no significant difference between the two groups at 16 weeks of treatment. The decrease in plasma NE was significantly greater in the NVC+ group than in the NVC− group (Figure 2).

Figure 2
figure 2

Changes in plasma norepinephrine (NE) levels after treatment with cilnidipine for 16 weeks. NVC−, hypertensive patients without neurovascular compression of the rostral ventrolateral medulla (RVLM); NVC+, hypertensive patients with neurovascular compression of the RVLM. n=22 for the NVC− group and n=24 for the NVC+ group. *P<0.05 vs. NVC− group.

Full size image

Changes in echocardiographic findings

LVMI, a marker of left ventricular hypertrophy (LVH), E/A and Dct, markers of left ventricular diastolic function, and %FS and EF, markers of left ventricular systolic function, did not differ significantly between the two groups before treatment (Table 3). LVMI was significantly decreased in the NVC+ group but not in the NVC− group at 16 weeks of treatment with cilnidipine (Table 3). The decrease in LVMI was significantly greater in the NVC+ group than in the NVC− group (Figure 3). E/A, Dct, %FS or EF did not show a significant change, and neither did the changes differ significantly between the two groups.

Table 3 Echocardiographic findings
Full size table
Figure 3
figure 3

Changes in left ventricular mass index (LVMI) after treatment with cilnidipine for 16 weeks. NVC−, hypertensive patients without neurovascular compression of the rostral ventrolateral medulla (RVLM); NVC+, hypertensive patients with neurovascular compression of the RVLM. n=21 for the NVC− group and n=22 for the NVC+ group. *P<0.05 vs. NVC− group.

Full size image

Safety

No symptoms, adverse effects or abnormal laboratory findings were seen during the study. Patients in each group showed good tolerance, with no subjects discontinuing the study.

Discussion

In this study, we found that plasma NE levels were significantly higher in the NVC+ group than in the NVC− group. In addition, chronic administration of cilnidipine led to regression of the left ventricular mass and to a significantly greater decrease in office and home BPs and plasma NE in the NVC+ group compared with the NVC− group, indicating that cilnidipine is effective, especially in hypertensive patients with NVC of the RVLM thanks to its inhibitory effect on enhanced SNA.

Since the initial report by Jannetta and Gendell3 in 1978, several clinical studies have indicated a possible association between NVC of the RVLM and EH.4, 5, 6, 7, 8, 9, 10 Using MRI, we found that the incidence of NVC of the RVLM in patients with EH was significantly higher than that in patients with secondary hypertension or in normotensive subjects, although the stage of hypertension did not differ significantly between the two hypertension groups.11, 12 We and others have reported that hypertensive patients show increased plasma levels of NE.15, 16 Therefore, it is assumed that NVC of the RVLM may be, at least in part, causally related to EH.

In the clinical report mentioned above,10 Jannetta et al. described that of 51 patients with cranial neuralgia and hypertension, 42 underwent MVD of both the cranial nerves and the artery, and in 36 the results of the surgical procedure were adequate. The BP normalized in 32 of these patients and improved in the remainder, allowing a reduction of the hypertensive medication. Of the 36 patients whose decompression was deemed adequate, 30 were still available for follow-up 7 years later, and 26 were either normotensive or had improved the control of their hypertension.13 They also performed MVD of the RVLM in 12 patients with EH without cranial neuralgia and observed a reduction in BP in 10 of 12 patients.14 We reported a patient with EH and hemifacial nerve spasms who showed NVC of the RVLM and facial nerve. MVD of the RVLM successfully reduced BP and indices of SNA such as plasma and urine NE levels, low-frequency to high-frequency ratio obtained by power spectral analysis and muscle SNA.27 Therefore, MVD of the RVLM is expected to be a useful therapeutic strategy to reduce BP in these patients.

However, as MVD is a highly invasive therapy, it seems quite unlikely to be applicable in all hypertensive patients with NVC of the RVLM. It seems more rational to use hypotensors with higher efficacy to treat such patients. It has been reported that SNA is elevated in patients with hypertension accompanied by NVC of the RVLM.15, 16, 17 This study endorsed this previous finding (Table 1). We have reported a patient with refractory hypertension and NVC of the RVLM, who revealed prominent BP reduction by additional treatment with clonidine, an α2 adrenergic agonist.28 Therefore, drugs capable of suppressing SNA among hypotensive agents are expected to reduce BP. In this study, we assessed the efficacy of cilnidipine, which is known to suppress SNA through N type calcium channel antagonism. Its hypotensive action and plasma NE-lowering effect were higher in the NVC+ group than in the NVC− group, suggesting that cilnidipine exerts a particularly potent hypotensive action in NVC+ patients. However, our previous study in rats revealed that BP elevation by pulsatile compression of RVLM becomes more marked in a pressure-dependent manner.12 Therefore, we cannot rule out that a decrease in arterial stimulation of the RVLM led to a further reduction in BP. At present, there is an ongoing clinical study designed to compare the pressure-lowering effects of L type calcium antagonists in NVC+ and NVC− patients. The ultimate goal of antihypertensive therapy is the suppression of cardiovascular complications.29, 30 Achieving BP reduction is quite important in hypotensive therapy, and it is known that a strict reduction of BP can improve the prognosis.31, 32 We may therefore expect cilnidipine to be a useful means of treatment, particularly in NVC+ patients, and to suppress cardiovascular complications more potently in the NVC+ group than in the NVC− group.

LVH is one of the risk factors of cardiovascular complications.16, 33 In this study, the reduction of LVMI was greater in the NVC+ group than in the NVC− group; although in the whole study population the reduction of LVMI was not statistically significant (P=0.13). Although the 16-week period was not long enough to reduce in the NVC− group, this period was adequate for obtaining such change in the NVC+ group. Thus, cilnidipine would be particularly useful in EH patients with NVC+ to reduce LVH. An association between elevated SNA and LVH has been reported.34, 35 It is probable that in addition to hypotensive activity, suppression of SNA may be partially involved in the potent reduction of a left ventricular myocardial mass induced by cilnidipine in NVC+ patients. However, there was no significant correlation between LVMI reduction and plasma NE level reduction in this study (r=0.288, P=0.2925). The effects of cilnidipine in suppressing and reducing cardiovascular complications warrant more in-depth studies using a detailed analysis of the mechanism of action of cilnidipine.

Keeping in view that in these cases BP elevation was associated with increased SNA15, 16, 17, 18 and that BP usually normalizes following decompression,13, 14, 27 we may say that the elevated BP seen in patients with NVC of the RVLM is attributable to the compression itself. This means that hypertension in these cases may be classified as secondary hypertension.36 It might be worthwhile to select hypotensive agent based on the presence/absence of NVC of the RVLM when dealing with patients with EH. However, several MRI studies showed similar prevalence of NVC of the RVLM between patients with EH and normotensives.37, 38, 39, 40, 41 In addition, the prevalence of NVC of the RVLM varied from 30 to 90% in essential hypertensives and from 7 to 55% in normotensives.4, 5, 6, 7, 9, 11, 12, 18, 37, 38, 39, 40, 41, 42, 43 The reason for the variation of the estimated prevalence of NVC of the RVLM may be due to the inability of MRI for the perfect and adequate assessment of the presence of NVC of the RVLM and to a lack of standardized classification of NVC of the RVLM. Investigation to establish adequate methods to assess whether NVC of the RVLM exists or not, and whether the compression is functional or not, will be needed.

In conclusion, our data suggest that cilnidipine lowers BP by inhibiting enhanced SNA and reduces left ventricular mass in hypertensive patients with NVC of the RVLM. It is possible that antihypertensive treatment to reduce SNA may be useful in hypertensive patients with NVC of the RVLM. However, it is also possible that the effect of cilnidipine in hypertensive patients with NVC of the RVLM is due to a blockade of vascular L-type Ca2+ channels. Similar studies examining the effects of other antihypertensive agents to inhibit SNA or other Ca2+ channel blockers that do not reduce SNA are needed to test this presumption.

Conflict of interest

The authors declare no conflict of interest.