Different drug classes have variable effects on blood pressure depending on the time of day:

Abstract

Background:

Blood pressure (BP) is controlled by a variety of systems, the activities of which vary throughout the day. As drugs are developed that selectively block these systems, the fall in BP may not be consistent over 24 h.

Methods:

A total of 24 patients (aged >65 years) with systolic BP (SBP; >150 mm Hg) that had not been treated entered a substudy of a larger study performed in 74 patients. In a double blind, crossover study with a balanced design, they received placebo, atenolol 50 mg, perindopril 8 mg, felodipine 10 mg, or hydrochlorothiazide 50 mg. The study periods were 2 months. Ambulatory BP monitoring was performed at the end of each period, and was divided into awake periods (9:00 AM to 10:00 PM), sleep periods (12:00 AM to 6:00 AM), and morning periods (6:00 AM to 9:00 AM). Medication was taken at 9:00 AM.

Results:

The four drug classes lowered 24-h mean SBP (P < .05), but the fall with atenolol was less than with the other drugs. The fall in awake BP with perindopril was less than with felodipine or hydrochlorothiazide. Atenolol caused no significant fall in sleep or morning SBP, and the falls with the other three drugs were significant and were greater than the fall with atenolol. The fall in sleep BP with perindopril was greater than with the other drug classes. The awake–sleep difference in SBP increased with perindopril, stayed the same with felodipine and hydrochlorothiazide, and was reduced by atenolol.

Conclusions:

In this study, the response to the different drug classes differed. The response to drugs that work relatively nonspecifically (diuretics, calcium blockers) was relatively consistent over 24 h. The response to β blockers and to angiotensin converting enzyme inhibitors reflected the activity of control systems. This finding supports the concept of multiple drug therapy that may need to be tailored to the time of day. Am J Hypertens 2003;16:46–50 @ 2003 American Journal of Hypertension, Ltd.

Chronobiology, chronotherapeutics, renin angiotensin, angiotensin converting enzyme inhibitors, β blockers, dihydropyridine calcium blockers, thiazide diuretics, 24-hour ambulatory monitoring, dippers

Blood pressure (BP) varies markedly over a 24-h interval. Some of this variation is related to activity and stress, but there is also an important circadian variation associated with sleep.^1,2 During sleep, the pulse pressure and BP fall markedly, largely because of a decrease in cardiovascular sympathetic outflow. Plasma renin increases three- to fourfold during sleep, and it is probable that BP during sleep is regulated to a large extent by the circulating level of plasma angiotensin II.^2,3 Earlier hypertensive agents were relatively nonselective, but newer drugs specifically inhibit systems that control BP. Thus, β blocking drugs block the action of components of the sympathetic nervous system. Angiotensin converting enzyme (ACE) inhibitors and angiotensin II type 1 (AT₁) receptor blocking drugs prevent the action of angiotensin II.^4,5 Endopeptidase inhibitors prevent the breakdown of atrial natriuretic peptide.⁶ With drugs that target specific systems, it would seem likely that they would be effective only while that system was regulating BP. However, there are relatively few reports as to whether different classes of drugs alter BP differently at various times of the day.

This study investigated the effect of the four major drug classes on BP at different times of the day in a group of elderly individuals with systolic hypertension.

Methods

The overall study has previously been described.⁷ The current study was conducted in a subset of the patients previously studied. Patients included in this study were all eligible to receive the four drugs to be used: atenolol, perindopril, felodipine, and hydrochlorothiazide. Patients were aged >65 years and had not been treated for hypertension. They were randomized to the study if the systolic (BP) (measured three times at the clinic while patients were supine) was >150 mm Hg after at least three visits. In a double-blind, balanced design, subjects were randomized to receive placebo, atenolol 25 mg, hydrochlorothiazide 25 mg, perindopril 4 mg, or felodipine 5 mg. They received this dose for 4 weeks and were then force titrated to receive double the dose of each agent unless this was contraindicated because of side effects or an excessively low BP. Subjects received this higher dose for 1 month and then crossed over to the other therapy. Thus, there were five treatment regimens of 8 weeks each.

At each monthly visit, an ambulatory monitor (A & D, Takeda, Japan) was attached and BP measured 12 times over 1 h. The mean of this was recorded as the clinic BP. At the end of the 8-week treatment (or after 4 weeks if there was a contraindication to continuing that drug), the ambulatory monitor was worn for 25 to 26 h. Pressures were recorded three times/h from 9:00 AM to 10:00 PM, two times/h from 10:00 PM to 6:00 AM, and three times/h from 6:00 AM to 9:00 AM. The monitor was attached at 8:00 AM and medication taken at 9:00 AM. The values recorded from 8:00 AM to 9:00 AM were used to provide clinic BP but were discarded from the ambulatory results. The awake period was defined as from 9:00 AM to 10:00 PM, sleep as from 12:00 AM to 6:00 AM, and morning as from 6:00 AM to 9:00 AM. If the recording did not meet the preset criteria, it was repeated 1 week later.

If a patient did not progress to the higher dose of the drug, the values at 4 weeks were used in the analysis. Statistical analysis was done by one- and two-way analysis of variance and by paired t tests. With the paired t tests, the Bonferroni correction was used to account for multiple comparisons. The main statistical requirement was by analysis of variance, but the P values for the paired t tests are presented in the results.

Results

A total of 24 patients (18 men and six women) entered this substudy. Three of the patients on placebo, four on β blockers, one on calcium blockers, and one on diuretics did not progress to the higher dose because of side effects and, therefore, the ambulatory BP monitoring was performed after 4 weeks of therapy on the lower dose of the drug. Clinic BP was not further decreased with the higher dose of the drug in people on placebo, β blockers, ACE inhibitors, or diuretics. There was a decrease of 6.8 ± 1.3/1.9 ± 0.7 mm Hg (P < .05) in patients who went from 5 to 10 mg felodipine.

The four drugs significantly lowered systolic BP (SBP) at the clinic (P < .05) and on the 24-h mean (P < .01) from the ambulatory monitor. The effect on diastolic BP (DBP) was much smaller and more variable, and atenolol and perindopril did not lower clinic DBP. All four drugs lowered the 24-h mean DBP (Table 1).

Table 1

Clinic and ABPM blood pressure on different drugs

	Placebo	Atenolol	Felodipine	Perindopril	Hydrochlorothiazide
Clinic BP
SBP (mm Hg)	164 ± 2.7	157 ± 2.4*	147 ± 1.9†	155 ± 2.1*	148 ± 1.5†
DBP (mm Hg)	87 ± 1.5	84 ± 1.5	82 ± 1.6‡	85 ± 1.8	83 ± 1.8‡
PR (beats/min)	69 ± 1.4	62 ± 1.4†	72 ± 1.8*	71 ± 1.5	70 ± 1.4
n	24	24	24	24	24
ABPM
24-h mean
SBP (mm Hg)	157 ± 3.2	149 ± 2.6‡	142 ± 2.6†	145 ± 2.4†	142 ± 2.6†
DBP (mm Hg)	85 ± 1.6	82 ± 1.4*	80 ± 1.4‡	82 ± 1.3*	81 ± 1.4‡
PR (beats/min)	66 ± 1.8	59 ± 1.5†	68 ± 1.5	67 ± 1.5	67 ± 1.5
n	24	22	21	23	21
Awake mean
SBP (mm Hg)	160 ± 2.5	148 ± 3.1‡	144 ± 2.5‡	151 ± 2.4‡	145 ± 2.4†
DBP (mm Hg)	86 ± 1.5	82 ± 1.5*	82 ± 1.6*	86 ± 1.6	82 ± 1.8*
PR (beats/min)	68 ± 1.4	61 ± 1.4†	71 ± 1.4*	71 ± 1.5	68 ± 1.5
n	24	23	22	23	22
Sleep mean
SBP (mm Hg)	145 ± 2.9	141 ± 3.8	131 ± 3.1†	129 ± 2.8†	132 ± 2.9†
DBP (mm Hg)	79 ± 1.4	78 ± 1.8	76 ± 1.6*	75 ± 1.8*	76 ± 1.7*
PR (beats/min)	62 ± 2.1	57 ± 1.4†	64 ± 1.2	62 ± 1.2	63 ± 1.2
n	24	23	21	23	22
Morning mean
SBP (mm Hg)	164 ± 4.1	156 ± 2.8*	146 ± 3.3†	148 ± 3.5‡	146 ± 3.3†
DBP (mm Hg)	87 ± 1.6	84 ± 1.4	83 ± 1.4*	83 ± 1.6*	83 ± 1.5*
PR (beats/min)	71 ± 1.6	64 ± 1.8†	73 ± 1.7	72 ± 1.5	70 ± 1.3
n	24	22	21	23	21

ABPM = ambulatory blood pressure monitoring; BP = blood pressure; SBP = systolic BP; DBP = diastolic BP; PR = pulse rate.

Data are given as mean ± SEM.

*

P < .05 v placebo.

†

P < .001 v placebo.

‡

P < .01 v placebo.

The fall in clinic SBP that was achieved with felodipine and hydrochlorothiazide was greater than that achieved with either atenolol or perindopril (Table 2). The fall in 24-h mean SBP was greater with felodipine and hydrochlorothiazide than with atenolol (Table 2). The fall in 24-h mean SBP with perindopril was not significantly different from that achieved with any of the other three drugs (P = .09).

Table 2

Falls in systolic and diastolic blood pressure (mm Hg) at different times, comparing effects of drugs versus placebo

	Atenolol	Felodipine	Perindopril	Hydrochlorothiazide
n	22	21	23	21
Clinic
SBP	6.5 ± 1.8	17.2 ± 2.1	8.0 ± 1.8*	16.4 ± 1.9
DBP	2.4 ± 1.7	4.8 ± 1.4	2.0 ± 1.2	4.5 ± 1.3
24-h
SBP	8.1 ± 2.5*	14.4 ± 2.5	11.5 ± 2.8	14.9 ± 2.5
DBP	3.1 ± 1.2	4.2 ± 1.2	2.4 ± 1.5	3.5 ± 1.3
Daytime
SBP	11.6 ± 2.7	14.8 ± 2.8	9.5 ± 2.9*	15.5 ± 1.9
DBP	4.2 ± 1.4	4.6 ± 1.1	0.8 ± 1.2	3.9 ± 1.2
Sleep
SBP	4.4 ± 2.3†,‡	12.7 ± 2.2	16.3 ± 2.7†	12.5 ± 1.5
DBP	1.2 ± 1.8§	3.1 ± 1.3	4.0 ± 1.2§	2.7 ± 1.2
Morning
SBP	7.2 ± 2.2‡	17.1 ± 3.7	15.8 ± 3.6	17.1 ± 2.8
DBP	2.6 ± 1.5	4.1 ± 1.1	4.1 ± 1.2	3.6 ± 1.3

Abbreviations as in Table 1.

Data are given as mean ± SEM.

Analysis based on paired t tests with Bonferroni correction.

*

P < 0.01 v felodipine and hydrochlorothiazide.

†

P < .01 v daytime value.

‡

P < .001 v other values across column.

§

P < .05 v daytime values on the same drug.

The fall of 8.1 mm Hg in 24-h mean SBP with atenolol was due predominantly to a fall of 11.6 mm Hg during the daytime, a smaller fall of 7.2 mm Hg on awakening, and an insignificant fall of 4.4 mm Hg during sleep. The fall during the sleep period was significantly less (P < .01) compared with the fall in the awake hours (Table 2).

The fall of 14.4 mm Hg in 24-h mean SBP with felodipine was due to a relatively even fall in BP that was spread over the entire 24 h. The slightly smaller fall during sleep was similar to that when awake, and reflected the lower SBP (145 mm Hg) for the sleep period compared with the awake SBP (160 mm Hg) on placebo (Table 2).

The fall of 11.5 mm Hg in 24-h mean SBP with perindopril was due to a daytime fall of 9.5 mm Hg, a significantly greater fall (16.3 mm Hg, P < .01) at nighttime, and a fall of 15.8 mm Hg in the morning (Table 2).

The fall of 14.9 mm Hg with hydrochlorothiazide was due to a relatively even fall in BP that was spread over the entire 24 h.

The falls in daytime SBP with felodipine and hydrochlorothiazide were greater than those achieved with perindopril (P < .01) but were not significantly different from the fall in daytime SBP with atenolol (P = .14). The fall in daytime SBP with atenolol and perindopril did not differ significantly (P = .12).

The fall in sleep SBP with atenolol was not significant and was less than the fall in sleep SBP that was achieved with the other three drugs (P < .001). The fall in sleep SBP with perindopril was significantly greater than that achieved with either felodipine or hydrochlorothiazide (P < .01).

The fall in morning SBP with atenolol was significant (P < .05) but was significantly less than that achieved with the other three drug classes. The falls with the other three drugs did not differ significantly from each other.

The change in DBP with the different drugs at the different times of the day were overall similar to those seen in SBP; but because of the smaller falls, many of the changes did not reach significance (Table 2).However, the fall in daytime DBP with atenolol was greater than the nonsignificant nighttime fall (P < .05). The fall in nighttime DBP with perindopril was greater (P < .05) than the nonsignificant fall in daytime DBP.

The difference between daytime and nighttime SBP is indicated in Fig. 1. The day night difference was decreased by atenolol compared with placebo and was increased by perindopril. The values achieved with felodipine and hydrochlorothiazide did not differ significantly from those on placebo (Fig. 1). Similar trends were seen with DBP, but these did not reach significance. On placebo, 11 of 26 patients had a fall of less than 10% of the daytime SBP (ie,these patients were nondippers). While on atenolol, 17 of 26 patients had a fall of less than 10% of the daytime SBP (ie, atenolol had converted six patients to a nondipping status). On felodipine and hydrochlorothiazide, 11 of 24 and 12 of 25 subjects were nondippers, which was similar to the prevalence on placebo. On perindopril, six of 25 patients had a fall of less than 10% of the daytime SBP (ie, perindopril had converted five patients to dipper status). The number of dippers on perindopril (19 of 25) was greater than the number of dippers on atenolol (nine of 26).

Difference between daytime and nighttime systolic blood pressure (SBP) on placebo, atenolol, felodipine, perindopril, and hydrochlorothiazide. This alteration meant that the dipper status was maintained with felodipine and hydrochlorothiazide compared to placebo. β blockers increased the number of nondippers while perindopril increased the number of dippers.

Figure 1.

Pulse pressure was reduced at all times of the 24-h period by felodipine (10 ± 3 mm Hg), perindopril (9 ± 3 mm Hg), and hydrochlorothiazide (11 ± 3 mm Hg) (P < .001). Atenolol reduced the pulse pressure during the day and morning (5 ± 2 mm Hg) (P < .05) but caused no significant effect during the nighttime (3 ± 2 mm Hg).

Pulse rate was not altered by perindopril or hydrochlorothiazide; however, during some phases of the 24-h period, felodipine caused a marginal increase in pulse pressure (P = .11 to .03). Atenolol caused a significant decrease in pulse pressure during all time periods, but the fall during the sleep period (4.7 beats/min) was less than during the awake period (7.3 beats/min).

Discussion

This study in a subset of patients confirmed the observation from the previous study in a larger population⁷ that, in older individuals with systolic hypertension, dihydropyridine calcium channel blocking drugs and thiazide diuretics lower BP to a greater extent than do ACE inhibitors or β blocking drugs. This observation was confirmed in the daytime (awake) measurements of the 24-h ambulatory BP measurements, but was modified in the 24-h mean systolic value because of different responses to the drugs at night. During the sleep period, ACE inhibition caused a larger fall in BP than occurred during the day, whereas β blocking drugs had a smaller effect (ie, no significant fall) during sleep. The net result of this change is that although the fall in 24-h mean SBP followed the same pattern as observed in the clinic, the difference between the drugs was less, although β blocking drugs were less effective than the other three drug classes.

β blocking drugs caused a greater fall in daytime BP (11.6 mm Hg) than was observed in the clinic BP (6.7 mm Hg) (P < .05). On the ambulatory BP monitor, this was confirmed in the response of morning BP, which was less than with the other three agents.

Hydrochlorothiazide and felodipine appeared to have a relatively consistent effect on BP. Although the fall during sleep was less than during the awake period, BP on placebo was greater during the awake period (160 mm Hg) compared to the sleep period (143 mm Hg). Blood pressure response depends, to a large extent, on the initial BP⁷ before therapy; and, if corrected for this, the response was consistent.

The responses to atenolol and perindopril were highly dependent on the time of day. Atenolol worked most effectively during the awake hours but had no significant effect during the sleep period. This was probably not due to too low a dose of atenolol because the fall in clinic BP 24 h after atenolol was similar with 25 and 50 mg.⁷ The poor response probably relates to relative absence of cardiovascular sympathetic activity during the sleep period; and as BP at that time is not under sympathetic nervous system control, it is not surprising that there is little effect.^2,8 Atenolol also did not appear to reduce the morning BP. The data at this time is less robust because of different times of awakening and arising. This is a time of increased sympathetic activity, and we postulate that the failure in effectiveness was related to insufficient blockade of the cardiovascular sympathetic response and that either a higher dose of atenolol or administration at nighttime may have increased the effectiveness of this drug in the morning. However, it should be noted that 50 mg of atenolol had no greater effect on clinic BP than did 25 mg.

Perindopril, in contrast to atenolol, lowered BP more during sleep than during the awake hours and, in this group of patients, was relatively ineffective during the daytime. If patients are maintained in a supine position, plasma renin³ and, we postulate, plasma angiotensin II rises three- to fourfold during sleep. In the absence of cardiovascular sympathetic responses controlling BP, angiotensin II becomes critical. Thus, even small reductions in circulating angiotensin II will cause large falls in BP. During normal activity, plasma renin rises during the awake hours. However at this time, the activity of the cardiovascular sympathetic responses (such as baroreflexes) are the important determinants of BP. Thus, the same reduction in circulating angiotensin II will have a lesser effect on BP. This has been observed in a study of the same dose of perindopril administered at morning or night⁹ and in the response over 48 h after trandolapril is stopped.¹⁰

Sleep BP may be very important in determining whether end organ damage, particularly cardiac hypertrophy, results.^11,12 The failure of atenolol to reduce sleep BP and its failure to reduce central aortic SBP as much as brachial artery SBP may provide an explanation for the relative ineffectiveness of β blocking drugs in reducing cardiac hypertrophy¹³ and in preventing cardiac morbidity and mortality when these drugs are used as monotherapy to treat hypertension.^14,15 In contrast, the greater effectiveness of ACE inhibitors at night may explain, in part, their greater effectiveness in reducing cardiac hypertrophy¹³ and deaths due to coronary artery disease.¹⁶

In conclusion, BP is controlled by a large number of variables. We recognize this by the individual variation that occurs in response to specific drugs. The evidence from this study that different drugs work more effectively at different times of the day may require us to rethink our therapeutic strategies. It is probably an additional reason to use multiple drugs rather than relying on high dose monotherapy.

The help of Olive Morgan, Denise Bertram, and Jann Lauri in undertaking the patient care is acknowledged. The secretarial help of Janet Lovett is appreciated.