Introduction

Recent studies have suggested that increased blood pressure (BP) variability predicts a poor prognosis for hypertensive patients independent of the elevated absolute value of BP [1,2,3,4,5,6,7,8,9]. While the development of antihypertensive drugs in the last few decades has made it possible to control the average BP in the majority of hypertensive patients, an optimal strategy to decrease BP variability in these patients remains under investigation. One possible strategy to decrease BP variability is to use a calcium channel blocker (CCB) instead of the other classes of drugs. [5, 8, 10,11,12,13] Rothwell et al. analysed the data of the ASCOT–BPLA trial in which amlodipine-based regimens were superior to atenolol-based regimens in reducing cardiovascular diseases and found that amlodipine reduced within-visit variability, as well as variability in 24-h ambulatory BP monitoring (ABPM) compared to atenolol [11]. It was also reported that the combination of an ARB with a CCB was superior to that of an ARB with diuretics in reducing home BP variability in patients with type II diabetes [9] and essential hypertension [8]. Given the different mechanisms to reduce BP among CCBs, β-blockers, and diuretics, it is suggested that the capacity of CCBs to reduce vascular resistance by blocking the L-type Ca channel contributes to the reduction in the BP variability.

Cilnidipine has a similar capacity to reduce BP compared with the most commonly prescribed CCB, amlodipine, but there is a distinct difference between these drugs because of the additional capacity of cilnidipine to block N-type calcium channels [14,15,16]. Some studies demonstrated that cilnidipine has protective effects on baroreflex sensitivity [17] or vascular function [18] through suppression of the sympathetic nerve activity with inhibition of N-type calcium channels. As fluctuations in sympathetic nerve activity are associated with BP variability, we hypothesised that inhibition of N-type calcium channels by cilnidipine would have additional effects on BP variability. To clarify the hypothesis, elderly hypertensive patients were subjected to intensive home BP monitoring and the change in BP variability was assessed during repeated change-over from amlodipine to cilnidipine. We also investigated whether arterial stiffness, assessed by the brachial-ankle pulse wave velocity (baPWV), and urinary catecholamine excretion as an index of the sympathetic nerve activity influenced the effects of the two drugs on BP variability.

Methods

Study subjects

We performed a hospital-based prospective study between December 2014 and January 2016 at Osaka University medical hospital. We recruited patients with essential hypertension aged 65 or older who received amlodipine and other antihypertensive drugs, if any. Patients were excluded if their prescription for antihypertensive drugs except the dose of amlodipine had changed within 6 months before recruitment or if they had a medical history of myocardial infraction or stroke within 6 months, heart failure (the New York Heart Association class III or IV), arterial fibrillation or severe liver dysfunction. We enroled 31 patients, and 26 of them completed the study protocol (Fig. 1a). The ethics committee in Osaka university hospital approved the study protocol, and all patients provided written informed consent (Approval number; 13423-2). We performed the study according to the ethical declaration of Helsinki and Good Clinical Practice standards.

Fig. 1
figure 1

a Flow chart presenting the study subjects. b Study design. HT hypertension, BP blood pressure, CV coefficient of variation

Full size image

Study design

In this repeated change-over study, amlodipine and cilnidipine were sequentially given to the same patient as shown in Fig. 1b. After confirmation of eligibility and providing written informed consent, patients who took amlodipine started to measure their home BP as instructed and were requested to visit the hospital at an interval of 2 months (the first amlodipine period). At the next visit, 2.5 mg, 5 mg or 10 mg of amlodipine were changed to 5 mg, 10 mg or 20 mg of cilnidipine, respectively, because there was an equivalent antihypertensive effect between 5 mg of amlodipine and 10 mg of cilnidipine in a previous randomised study [19] (the first cilnidipine period). Thereafter, physicians sequentially changed the prescription of the test drug from cilnidipine to amlodipine and from amlodipine to cilnidipine in the same fashion at an interval of 2 months (the second cilnidipine period and second amlodipine period). At each visit, physicians measured office BP in a seated position, blood and urinary samples were collected, and a medical interview was performed to confirm correct measurement of home BP and high compliance with the medication. Antihypertensive medications, except the test drug, were unchanged throughout the study. We excluded the data of one subject from the analysis because a physician needed to increase the dose of the test drugs to treat uncontrolled BP.

Home BP measurements

During the study period, the patients measured home BP and pulse rate (PR) twice a day in the early morning and evening before sleep with an electronic sphygmomanometer that transmits the data wirelessly to database in a secure website to which access is limited to a physician in the hospital (HEM-7251G-HP, Omron healthcare Co, Kyoto, Japan) [20]. The patients measured their home BP at least three times in the sitting position after 2 mins of rest according to the guidelines of the Japanese Society of Hypertension. Considering the carry over effects of the test drugs, we used the data from 28 days during the latter half in each period (Fig. 1b). We used the averaged BP and PR among three measurements to obtain the BP and PR at each time-point, respectively. We calculated the coefficient of variation (CV) to assess the variability in the BP and PR with CV = Standard deviation/mean value.

Pulse wave velocity

The measurements of the brachial-ankle pulse wave velocity (baPWV) with the ankle brachial pressure index (ABI) were performed in the morning at the visit after the second amlodipine period with the volume-plethysmographic device (VP-2000; Omron Healthcare, Japan). The device evaluates both the brachial and ankle BP, as well as the pulse waves of the carotid and both femoral arteries at the same time in the supine positon [21, 22].

Biochemical measurements

The estimated glomerular filtration rate (eGFR) was calculated using the Japanese glomerular filtration rate equation based on serum creatinine, gender and age [23]. The urine samples in the morning for catecholamine were immediately frozen after collection, and the measurement was performed in SRL (Tokyo, Japan). The urinary catecholamine per creatinine was calculated.

Log amlodipine/cilnidipine CV

We divided patients into the smaller CV with amlodipine group and smaller CV with cilnidipine group based on the comparison of BP variability between the total amlodipine and cilnidipine periods. We calculated the log ratio of the CV of systolic BP during the amlodipine periods to the CV during the cilnidipine period of the same patient (log2 (amlodipine CV/cilnidipine CV) = log A/C CV). If the log ratio exceeded 0, we classified the patient as the smaller CV with cilnidipine group; if the ratio was less than 0, we classified the patient as the smaller CV with amlodipine group.

Statistical analysis

As shown in Fig. 1b, we planned a study of a continuous response variable from matched pairs of study subjects. Previous researches indicate that the difference in the response of matched pairs is normally distributed with standard deviation 8 [24]. If the true difference in the mean response of matched pairs is 5, we will need to study 22 pairs of subjects to be able to reject the null hypothesis that this response difference is 0 with power 0.8. The Type I error probability associated with this test of this null hypothesis is 0.05. The sample size was calculated with Power and Sample Size Calculation ver 3.1.2 software (Vanderbilt University Department of Biostatistics, TN, USA).

All data were analysed with JMP Pro ver 12.0.1 software (SAS Institute Inc., Cary, NC, USA). Values are shown as the mean ± standard deviation, median (interquartile range), or percentage. Paired t-tests were used to analyse the changes between the two CCBs in the average value of the BP and PR, CV in systolic BP (SBP) and PR. Student’s t-test and Fisher’s exact test were employed to examine the difference in several factors between two groups. Pearson’s correlation coefficient was used to examine univariate correlations. Stepwise regression analysis was used to determine variables related to log A/C CV. The logarithm was calculated to normal distribution. P < 0.05 was considered statistically significant.

Results

Comparison of values and variabilities of BP and PR between the amlodipine and cilnidipine treatments

In 26 subjects, who completed the study, the correct BP measurement of all the patients were confirmed by raw data with measurement times and an interview at each visit. Table 1 shows the baseline characteristics of the study subjects. The average office BP was less than 140/90 mm Hg, and the BP of the patients was well controlled throughout the study period. As shown in Table 2, systolic and diastolic BP were modestly, but significantly, higher in both the morning and evening in the cilnidipine periods compared to the amlodipine periods. In contrast, PR was modestly lower in the morning and evening in the cilnidipine periods compared to the amlodipine periods. Regarding variabilities in the BP and PR, we did not observe any difference between the cilnidipine and amlodipine periods in the CV of systolic BP, diastolic BP and PR in the morning and evening.

Table 1 Clinical characteristics of participants at baseline
Full size table
Table 2 Comparison of values and variabilities of BP and PR between the amlodipine and cilnidipine treatment
Full size table

Factors associated with the effectiveness of amlodipine and cilnidipine on BP variability

As we found no difference in the variabilities in the BP between the amlodipine and cilnidipine periods in all patients, we then divided patients into the smaller CV with amlodipine and smaller CV with cilnidipine groups based on the CV in systolic BP to determine the factors that are associated with the effects of each drug on BP variability. We classified a patient into the smaller CV with cilnidipine group if the CV in the amlodipine periods was higher than that in the cilnidipine period (log A/C CV > 0) and the smaller CV with amlodipine group if log A/C CV was less than 0, as described in the Methods section. Of 26 patients, 13 and 15 patients were classified into the smaller CV with amlodipine group in the morning (Table 3a) and evening (Table 3b), respectively. The background characteristics were similar between two groups. There were no differences in systolic and diastolic BP between the two groups (data not shown). We found that the urinary excretion of total catecholamines, adrenaline and dopamine were significantly higher in the smaller CV with cilnidipine group than in the smaller CV with amlodipine group in the morning. We observe this difference during amlodipine treatments, while not during cilnidipine treatments. There was no significant difference in the baPWV between the groups in the morning (Table 3a). In contrast, baPWV was significantly lower in the smaller CV with amlodipine group than in the smaller CV with cilnidipine group in the evening, while there was no significant difference in the urinary catecholamine excretion between the two groups in both the amlodipine and cilnidipine periods (Table 3b).

Table 3 Factors associated with the effectiveness of amlodipine and cilnidipine on BP variability
Full size table

Figure 2a shows that the urinary adrenaline excretion was positively correlated with log A/C CV in the morning, suggesting that cilnidipine was more effective than amlodipine in reducing BP variability in patients with higher urinary adrenaline excretion. There was no correlation between the urinary adrenaline excretion and log A/C CV in the evening (Fig. 2a). In contrast, baPWV was positively correlated with log A/C CV in the evening, which was not the case in the morning (Fig. 2b). We confirmed these correlations after adjusting for potential cofounding variables by stepwise regression analysis. As shown in Table 4, we observed a significant correlation between log A/C CV in the morning and urinary adrenaline excretion after adjusting for sex and baPWV (model 1) and for sex and age (model 2). Similarly, we observed a significant correlation between log A/C CV in the evening and baPWV after adjusting for sex and urinary adrenaline excretion (model 3), as well as for sex and age (model 4). We found that the baPWV was positively correlated with CV in the amlodipine period in the evening, which was not the case in the morning (evening: correlation coefficient (r) = 0.49, p = 0.015; morning: r = 0.036, p = 0.87).

Fig. 2
figure 2

a Correlation between Log A/C CV and urinary adrenaline excretion. b Correlation between Log A/C CV and baPWV. Pearson’s correlation coefficient was used to examine correlations. log A/C CV log2 (amlodipine CV/cilnidipine CV), r Pearson correlation coefficient, CV coefficient of variation, baPWV brachial-ankle pulse wave velocity

Full size image
Table 4 Stepwise regression analysis
Full size table

Discussion

In this study, we found that there was no difference in the effect of cilnidipine and amlodipine on home BP variability. We also performed an additional analysis to investigate the factors associated with the difference of both CCBs in BP variability for individual patients. We found that high urinary catecholamine excretion was associated with the effectiveness of cilnidipine over amlodipine in reducing morning BP variability, which was not the case in the evening. In contrast, low arterial stiffness was associated with the effectiveness of amlodipine compared to cilnidipine in reducing BP variability in the evening, which was not the case in the morning, suggesting that amlodipine is more effective in reducing evening BP variability in patients with lower arterial stiffness. Several studies suggested that CCBs are superior to other types of antihypertensive drugs in reducing home BP variabilities [9]. Asayama et al. recently reported the opposing results in which there was no difference among the effects of CCBs, angiotensin II receptor blockers, and ACE inhibitors on home BP variabilities, but they also reported that amlodipine reduced systolic BP variability in the morning compared with other CCBs [25]. Our findings suggest that cilnidipine has a similar capacity as amlodipine in reducing home BP variabilities, but the underlying mechanisms in reducing BP variabilities may be different between the two drugs.

In most previous clinical studies and actual medical practice, a half dose of amlodipine has been considered to be equivalent to a full dose of cilnidipine in reducing BP [19]. Therefore, we used this relationship when switching the test drugs in the present study. However, we found that the home BP during cilnidipine treatment was modestly, but significantly, higher than the BP with amlodipine treatment. The BP lowering effect of CCBs is determined by their capacity to reduce vascular resistance. The capacity of CCBs to reduce vascular resistance is also proposed to play a primary role in their ability to reduce BP variability [9, 13]. We found that amlodipine reduced evening BP variability more effectively than cilnidipine in patients with lower arterial stiffness determined by the baPWV. We also found that higher arterial stiffness was associated with higher evening BP variability during amlodipine treatment, which is consistent with the current concept that high arterial stiffness is associated with BP variability in the elderly [26]. Altogether, it is conceivable that the capacity of amlodipine to reduce vascular resistance is somewhat superior to cilnidipine, and the small difference makes amlodipine more effective than cilnidipine in reducing evening BP variability preferentially in patients with low arterial stiffness.

Several clinical and basic studies support the unique capacity of cilnidipine to reduce sympathetic nerve activity by blocking N-type calcium channels [15, 27]. In this study, we found that the pulse rate was significantly smaller during cilnidipine treatment than during amlodipine treatment, implying the effect of N-type calcium channel blockade by cilnidipine. The N-type calcium channel is expressed in nerve terminals, and it is involved in catecholamine secretion [15, 27]. Therefore, it is conceivable that patients with high catecholamine secretion receive a higher benefit from cilnidipine than those with low catecholamine secretion. Consistent with this notion, we found that high urinary catecholamine excretion is associated with the effectiveness of cilnidipine in reducing morning BP variability, compared to amlodipine.

Of note, our results suggest that measuring the time of the home BP influences the association between arterial stiffness and BP variability. We found a significant correlation between arterial stiffness and BP variability in the evening, which was not the case in the morning, suggesting that arterial stiffness is a potent determinant of evening BP variability. We also observed the preferential effect of cilnidipine in patients with high catecholamine excretion on BP variability in the morning, which was not the case in the evening. Interestingly, it was recently reported that evening BP variability predicted cardiovascular disease, while morning BP variability did not [25]. They showed that variability, independent of the mean and average variability, in the evening BP before sleep significantly predicted cardiovascular events independent of the BP level, while these parameters in the morning did not have an independent predictable ability for cardiovascular outcomes. It is also well-known that arterial stiffness predicts cardiovascular events in hypertensive elderly, which is independent of the BP level [28]. Taken together, it is conceivable that BP variability in the evening is a better prognostic factor than that in the morning because arterial stiffness is more purely reflected in BP variability in the evening than in the morning.

There are several limitations in the study. First, while the number of participants for evaluating the primary outcome was sufficient as described in the Method section, the number was insufficient to provide full reliability for some analyses. For example, the lack of a correlation between two variables in some of the figures needs to be validated in a larger sample size. Second, we found a difference in the BP level between amlodipine and cilnidipine treatments, suggesting that these drugs do not equivalently block L-type channels. Therefore, our results failed to draw a conclusion on whether blocking N-type calcium channels has an additional effect on home BP variability. Third, the subjects in the study are relatively healthy and have a low frequency of diabetes and chronic kidney disease. Further investigation will be required to compare the BP variability between amlodipine and cilnidipine in high-risk hypertensive patients. Fourth, the subjects were 65 years of age or older in this study, and did not include non-elderly patients. As non-elderly hypertensive patients tend to have low arterial stiffness and high sympathetic nerve activity compared to elderly patients, different findings could be withdrawn from the non-elderly patients.

In conclusion, we did not observe any difference in home BP variabilities between amlodipine and cilnidipine treatments in elderly hypertensive patients. Analysis based on the difference in the BP variability between the two drugs in individual patients provided results supporting the following hypothesis: High sympathetic nerve activity with increased catecholamine secretion is associated with the effectiveness of cilnidipine in reducing morning BP variability, and low arterial stiffness is associated with the effectiveness of amlodipine in reducing evening BP variability.