J. Endocrinol. Invest. 13: 559-566, 1990
Nocturnal oscillations in plasma renin activity during sleep in hypertensive patients: the influence of perindopril G. Brandenberger*, J.L. Imbs**, J.P. Libert*, J. Ehrhart*, C. Simon*, J.Ph. Santoni***, and M. Follenius* * Laboratoire de Physiologie et de Psychologie Environnementales UMR 32, CNRS/INRS, 67087 Strasbourg Cedex, ** Service d'Hypertension Arterielle et des Maladies Vasculaires du CHR de Strasbourg, 67000 Strasbourg, *** Institut de Recherches Internationales Servier 22, 92201 Neuilly sur Seine, France.
ABSTRACT. In previous studies, we established a strong concordance between nocturnal oscillations in plasma renin activity (PRA) and REMNREM sleep cycles. To determine whether this relation persists in the case of moderate essential hypertension and if it is influenced by antihypertensive therapies affecting renin release, six normal subjects and six hypertensive patients were studied. The normal subjects underwent one control night. The hypertensive patients were studied during a first night when a ptacebo was given. Four of them underwent a second night following a single dose of an angiotensin-converting enzyme (ACE) inhibitor, perindopril; and a third night, 45 days later, with the antihypertensive treatment. In addition, two of the patients underwent two nightstudies, after a single and repeated doses of a beta-blocker, atenolol, to see whether preventing renin release modified the sleep structure. The relationship between the nocturnal PRA oscillations and the sleep stage patterns persisted in
hypertensive patients receiving placebo. In patients who had low PRA levels, the increases associated with NREM sleep were small. However, the mean relative amplitude of the oscillations, expressed as a percentage of the nocturnal mean, was about 600/0, which was similar to that in normotensive subjects. Active renin and PRA oscillations were closely coupled. ACE activity profiles displayed damped fluctuations and no systematic relationship with sleep stages. Perindopril, in single or repeated doses led to striking increases in PRA and amplified the nocturnal oscillations without disturbing their relationship to specific sleep stages. Atenolol almost supressed PRA fluctuations, while regular REM-NREM sleep cycles persisted. These results indicate that the relation between PRA oscillations and sleep stage alternation persists in moderate essential hypertension, and is preserved during perindopril therapy which increases the oscillation amplitude.
INTRODUCTION
sleep phases invariably coincide with increasing PRA levels, and REM sleep phases always occur when PRA levels are decreasing. Spontaneous and provoked awakenings blunt the rise in PRA normally associated with NREM sleep. So, PRA curves exactly reflect the patterns of sleep stage distribution: when sleep cycles are regular, PRA levels display a strong ultradian rhythm. For irregular sleep cycles, PRA curves reflect all disturbances in the sleep structure (4). The renin-angiotensin system plays an important role in the control of arterial blood pressure, so, in recent years great attention has been paid to its inhibition, with different drugs, in an attempt to normalize blood pressure (5-9). This study was designed to determine whether the previously described relation between nocturnal
Sleep is organized in cycles of different stages, grossly divided into Rapid Eye Movement (REM) sleep and Non-REM (NREM) sleep, each characterized by specific electrophysiological, autonomic and motor changes. Extending the results ot Mullen (1), we have reported similar cycles averaging 100 min both for REM-NREM sleep alternation and for nocturnal levels in plasma renin activity (PRA) (2, 3). Both phenomena are strongly linked. NREM Key-words: Renin activity, sleep, ultradian rhythm, hypertension, betablocker, angiotensin-converting-enzyme inhibitor. Correspondence: Dr. G. Brandenberger, Laboratoire de Physiologie et de Psychologie Environnementales, 21 rue Becquerel, 67087 Strasbourg Cedex, France.
Received October 6, 1989; accepted April 19, 1990.
559
G. Brandenberger, J.L. Imbs, J.P. Libert, et al.
PRA levels and the sleep stage patterns persists in moderate essential hypertension before and after the administration of a new ACE inhibitor, perindopril, well-known to increase renin release by removing feedback from angiotensin II.The nocturnal profiles, obtained from blood collected at 10-min intervals, were compared to the concomitant sleep stage patterns in untreated patients, after the first and after repeated doses of the antihypertensive drug. In addition, as it has been shown that disturbing sleep modifies renin release from the kidneys (4), a beta-blocker, atenolol was given to see whether, conversely, any alteration in renin release modifies the sleep structure. MATERIALS AND METHODS Subjects
Studies were performed in 6 male patients with moderate essential hypertension and in six normal subjects matched for age (43.2 ± 3.8 yr vs 41.6 ± 3.5 yr) and weight (82.7 ± 5.1 kg vs 81.1 ± 4.8 kg). The control subjects were normal volunteers, in good health, with a mean systolic pressure of 119 ± 9 mmHg and a mean diastolic pressure of 78 ± 4 mmHg. The hypertensive patients had a documented evidence of hypertension for about 8 years prior to the study. At the end of a 4-week placebo period, their mean blood pressure was 152 ± 4 mmHg for systolic and 99 ± 1 .5 mmHg for diastolic (± SE). Cases of heart or lung disease, as well as neurological, renal, endocrine or metabolic disorders, and secondary forms of hypertension were excluded. They all gave their informed consent. The experiments were approved by the Strasbourg Hospital Ethics Committee.
perindopril. Since oral perindopril doses act within 4h (1 0), the drug was given exceptionally at 2200 h. Mean blood pressure in the morning was 139 ± 4 mmHg for systolic and 88 ± 2 mmHg for diastolic. The patients underwent a third night (N3) between the 43rd and 47th days of the antihypertensive treatment. The patients received daily 4 mg perindopril taken, as usual, in the morning. Blood pressure was then 137 ± 3 mmHg for systolic and 81 ± 4 mmHg for diastolic. In addition, two of the patients were treated with a beta-blocker atenolol (ICI-Pharma, Cergy, France) at doses of 50 mg once a day. They underwenttwo night-studies after a single dose (50 mg at 22:00 h) and repeated 50 mg morning doses, 45 days later. Blood pressure was then normalized (140 and 138 mmHg for systolic and 95 and 90 mmHg for diastolic). The studies were performed in a sound-proof, airconditioned sleep chamber. Thesubjects were kept ina thermoneutral environment(20 C). A standard meal (900 kcal; carbohydrate: 45%; fat: 34%; protein: 21 %) was served at 19:15 h. Before the subjects entered the sleep chamber, electrodes were attached for the following uninterrupted electrophysiological recordings: two electro-encephalograms (EEG leads F3-A2 and C3-A2); two electro-oculograms (right and left from the outer canthus with reference to the left mastoid); one electromyogram of the mentalis and one electrocardiogram. Sleep stages were scored for each 30 s period of the night according to established criteria (11 ). Subjects went to bed at 22:30 h. The light was switched off at 23:00 hand patients were awakened at 08:00 h. Throughout the experiment, the patients were observed with closed circuit television.
Procedure
Blood sampling and plasma measurements
During the investigation, the control subjects and the hypertensive patients were on their usual Na diet. Na excretion in the 24-h urine samples collected the day before the experiment was between 150 and 212 meq/24h. The normal subjects underwent one control night. In the hypertensive patients, all drugs including hypotensive agents were withdrawn for 30 days, during which they received a placebo. They were studied during one night (NI) between the 18th and the 21 st day of placebo. Mean blood pressure in the morning was 152 ± 4 mmHg for systolic and 99 ± 2 mmHg for diastolic. Four of the patients were studied during a second night (N2) after the first administration of 4 mg
An intravenous catheter was inserted, under local anesthesia, into an antecubital vein at 21 :00 hand kept patent with heparinized saline solution (500 I.U. heparin · rnl' of 0.9 g NaCI . 100 ml'). Blood was collected continuously from 23:00 to 08:00 h in an adjoining room. Samples were collected every 10 min in plastic tubes containing either EDTA (1 mg . mt' blood) for plasma renin activity, or heparin (50 J.lg . rnl' blood)for active renin and ACE activity measurements. They were immediatelycentrifuged at 4 C and the plasma stored at -25 C. A maximum of 200 ml blood was removed during each night. Plasma renin activity was measured by radio-immunoassay of the angiotensin-I generated after plas-
560
PRA oscillations in hypertensive patients
ma incubation (12); the intra-assay coefficient of variation (CV) for duplicate samples was 4.0% for levels comprised between 10-20 ngAI . mr' . rr' : 6.0% for levels between 2-10 ngAI . ml' . rr' : 10.0% for levels between 1-2 ngAI . rnr' . h"; for levels less than 1 ngAI . rnl" . h-1 , it was 30%. Active renin level was measured with a radioimmunometric method using commercially available kits (Diagnostics Pasteur, Lyon, France) . This method uses two monoclonal antirenin antibodies; the intra-assay coefficient of variation for duplicate samples was 10%. ACE activity was measured with the fluorimet ric method of Unger et al. (13); the intra-assay CV for duplicate samples was 6%.The detection limit for PRA was 0.18 ngAI . rnl' . h', and for act ive rening, 10 ng . 1-'.
PRA curves , illustrated in Figures 1-4, were smoothed using the moving averages method over a 3-point span.
Data analysis The data were subjected to spectral analysis to determine the periods of the nocturnal oscillations. First, the data were smoothed using a three-point moving average. Then, a difference filter, defined as Xd1 = Xl - X(t -1) was used to remove low-frequency components. The Fourier transformation of the autocorrelation function, and a Blackman and Tukey window were used to determine the spectral density function (14). The bandwidth of the spectral window was 0.01 cycle ' rnirr' . and the frequency spacing was 0.0019 cycle . rnlrr' . Frequencies with the highest spectral densities were identified. Their significance was tested by comparing their variance with the residual variance with the residual variance in the remaining frequency bands of the spectrum. All values are expressed as means ± SE. Individual
RESULTS
i", l~ .
L~~'~,"OCI ..
!
0
1 )
~ ,
1I n lin n 1r11L rfl[
~ ~LJ~ ~ILJ I'-:
1t:[ ~ ~~~
~
~
~
TIME
~~~
, "'wM
__
~ ~ ! /'j~;
,
V
,~
Nocturnal oscillations in plasma renin activity during sleep in hypertensive patients: the influence of perindopril G. Brandenberger*, J.L. Imbs**, J.P. Libert*, J. Ehrhart*, C. Simon*, J.Ph. Santoni***, and M. Follenius* * Laboratoire de Physiologie et de Psychologie Environnementales UMR 32, CNRS/INRS, 67087 Strasbourg Cedex, ** Service d'Hypertension Arterielle et des Maladies Vasculaires du CHR de Strasbourg, 67000 Strasbourg, *** Institut de Recherches Internationales Servier 22, 92201 Neuilly sur Seine, France.
ABSTRACT. In previous studies, we established a strong concordance between nocturnal oscillations in plasma renin activity (PRA) and REMNREM sleep cycles. To determine whether this relation persists in the case of moderate essential hypertension and if it is influenced by antihypertensive therapies affecting renin release, six normal subjects and six hypertensive patients were studied. The normal subjects underwent one control night. The hypertensive patients were studied during a first night when a ptacebo was given. Four of them underwent a second night following a single dose of an angiotensin-converting enzyme (ACE) inhibitor, perindopril; and a third night, 45 days later, with the antihypertensive treatment. In addition, two of the patients underwent two nightstudies, after a single and repeated doses of a beta-blocker, atenolol, to see whether preventing renin release modified the sleep structure. The relationship between the nocturnal PRA oscillations and the sleep stage patterns persisted in
hypertensive patients receiving placebo. In patients who had low PRA levels, the increases associated with NREM sleep were small. However, the mean relative amplitude of the oscillations, expressed as a percentage of the nocturnal mean, was about 600/0, which was similar to that in normotensive subjects. Active renin and PRA oscillations were closely coupled. ACE activity profiles displayed damped fluctuations and no systematic relationship with sleep stages. Perindopril, in single or repeated doses led to striking increases in PRA and amplified the nocturnal oscillations without disturbing their relationship to specific sleep stages. Atenolol almost supressed PRA fluctuations, while regular REM-NREM sleep cycles persisted. These results indicate that the relation between PRA oscillations and sleep stage alternation persists in moderate essential hypertension, and is preserved during perindopril therapy which increases the oscillation amplitude.
INTRODUCTION
sleep phases invariably coincide with increasing PRA levels, and REM sleep phases always occur when PRA levels are decreasing. Spontaneous and provoked awakenings blunt the rise in PRA normally associated with NREM sleep. So, PRA curves exactly reflect the patterns of sleep stage distribution: when sleep cycles are regular, PRA levels display a strong ultradian rhythm. For irregular sleep cycles, PRA curves reflect all disturbances in the sleep structure (4). The renin-angiotensin system plays an important role in the control of arterial blood pressure, so, in recent years great attention has been paid to its inhibition, with different drugs, in an attempt to normalize blood pressure (5-9). This study was designed to determine whether the previously described relation between nocturnal
Sleep is organized in cycles of different stages, grossly divided into Rapid Eye Movement (REM) sleep and Non-REM (NREM) sleep, each characterized by specific electrophysiological, autonomic and motor changes. Extending the results ot Mullen (1), we have reported similar cycles averaging 100 min both for REM-NREM sleep alternation and for nocturnal levels in plasma renin activity (PRA) (2, 3). Both phenomena are strongly linked. NREM Key-words: Renin activity, sleep, ultradian rhythm, hypertension, betablocker, angiotensin-converting-enzyme inhibitor. Correspondence: Dr. G. Brandenberger, Laboratoire de Physiologie et de Psychologie Environnementales, 21 rue Becquerel, 67087 Strasbourg Cedex, France.
Received October 6, 1989; accepted April 19, 1990.
559
G. Brandenberger, J.L. Imbs, J.P. Libert, et al.
PRA levels and the sleep stage patterns persists in moderate essential hypertension before and after the administration of a new ACE inhibitor, perindopril, well-known to increase renin release by removing feedback from angiotensin II.The nocturnal profiles, obtained from blood collected at 10-min intervals, were compared to the concomitant sleep stage patterns in untreated patients, after the first and after repeated doses of the antihypertensive drug. In addition, as it has been shown that disturbing sleep modifies renin release from the kidneys (4), a beta-blocker, atenolol was given to see whether, conversely, any alteration in renin release modifies the sleep structure. MATERIALS AND METHODS Subjects
Studies were performed in 6 male patients with moderate essential hypertension and in six normal subjects matched for age (43.2 ± 3.8 yr vs 41.6 ± 3.5 yr) and weight (82.7 ± 5.1 kg vs 81.1 ± 4.8 kg). The control subjects were normal volunteers, in good health, with a mean systolic pressure of 119 ± 9 mmHg and a mean diastolic pressure of 78 ± 4 mmHg. The hypertensive patients had a documented evidence of hypertension for about 8 years prior to the study. At the end of a 4-week placebo period, their mean blood pressure was 152 ± 4 mmHg for systolic and 99 ± 1 .5 mmHg for diastolic (± SE). Cases of heart or lung disease, as well as neurological, renal, endocrine or metabolic disorders, and secondary forms of hypertension were excluded. They all gave their informed consent. The experiments were approved by the Strasbourg Hospital Ethics Committee.
perindopril. Since oral perindopril doses act within 4h (1 0), the drug was given exceptionally at 2200 h. Mean blood pressure in the morning was 139 ± 4 mmHg for systolic and 88 ± 2 mmHg for diastolic. The patients underwent a third night (N3) between the 43rd and 47th days of the antihypertensive treatment. The patients received daily 4 mg perindopril taken, as usual, in the morning. Blood pressure was then 137 ± 3 mmHg for systolic and 81 ± 4 mmHg for diastolic. In addition, two of the patients were treated with a beta-blocker atenolol (ICI-Pharma, Cergy, France) at doses of 50 mg once a day. They underwenttwo night-studies after a single dose (50 mg at 22:00 h) and repeated 50 mg morning doses, 45 days later. Blood pressure was then normalized (140 and 138 mmHg for systolic and 95 and 90 mmHg for diastolic). The studies were performed in a sound-proof, airconditioned sleep chamber. Thesubjects were kept ina thermoneutral environment(20 C). A standard meal (900 kcal; carbohydrate: 45%; fat: 34%; protein: 21 %) was served at 19:15 h. Before the subjects entered the sleep chamber, electrodes were attached for the following uninterrupted electrophysiological recordings: two electro-encephalograms (EEG leads F3-A2 and C3-A2); two electro-oculograms (right and left from the outer canthus with reference to the left mastoid); one electromyogram of the mentalis and one electrocardiogram. Sleep stages were scored for each 30 s period of the night according to established criteria (11 ). Subjects went to bed at 22:30 h. The light was switched off at 23:00 hand patients were awakened at 08:00 h. Throughout the experiment, the patients were observed with closed circuit television.
Procedure
Blood sampling and plasma measurements
During the investigation, the control subjects and the hypertensive patients were on their usual Na diet. Na excretion in the 24-h urine samples collected the day before the experiment was between 150 and 212 meq/24h. The normal subjects underwent one control night. In the hypertensive patients, all drugs including hypotensive agents were withdrawn for 30 days, during which they received a placebo. They were studied during one night (NI) between the 18th and the 21 st day of placebo. Mean blood pressure in the morning was 152 ± 4 mmHg for systolic and 99 ± 2 mmHg for diastolic. Four of the patients were studied during a second night (N2) after the first administration of 4 mg
An intravenous catheter was inserted, under local anesthesia, into an antecubital vein at 21 :00 hand kept patent with heparinized saline solution (500 I.U. heparin · rnl' of 0.9 g NaCI . 100 ml'). Blood was collected continuously from 23:00 to 08:00 h in an adjoining room. Samples were collected every 10 min in plastic tubes containing either EDTA (1 mg . mt' blood) for plasma renin activity, or heparin (50 J.lg . rnl' blood)for active renin and ACE activity measurements. They were immediatelycentrifuged at 4 C and the plasma stored at -25 C. A maximum of 200 ml blood was removed during each night. Plasma renin activity was measured by radio-immunoassay of the angiotensin-I generated after plas-
560
PRA oscillations in hypertensive patients
ma incubation (12); the intra-assay coefficient of variation (CV) for duplicate samples was 4.0% for levels comprised between 10-20 ngAI . mr' . rr' : 6.0% for levels between 2-10 ngAI . ml' . rr' : 10.0% for levels between 1-2 ngAI . rnr' . h"; for levels less than 1 ngAI . rnl" . h-1 , it was 30%. Active renin level was measured with a radioimmunometric method using commercially available kits (Diagnostics Pasteur, Lyon, France) . This method uses two monoclonal antirenin antibodies; the intra-assay coefficient of variation for duplicate samples was 10%. ACE activity was measured with the fluorimet ric method of Unger et al. (13); the intra-assay CV for duplicate samples was 6%.The detection limit for PRA was 0.18 ngAI . rnl' . h', and for act ive rening, 10 ng . 1-'.
PRA curves , illustrated in Figures 1-4, were smoothed using the moving averages method over a 3-point span.
Data analysis The data were subjected to spectral analysis to determine the periods of the nocturnal oscillations. First, the data were smoothed using a three-point moving average. Then, a difference filter, defined as Xd1 = Xl - X(t -1) was used to remove low-frequency components. The Fourier transformation of the autocorrelation function, and a Blackman and Tukey window were used to determine the spectral density function (14). The bandwidth of the spectral window was 0.01 cycle ' rnirr' . and the frequency spacing was 0.0019 cycle . rnlrr' . Frequencies with the highest spectral densities were identified. Their significance was tested by comparing their variance with the residual variance with the residual variance in the remaining frequency bands of the spectrum. All values are expressed as means ± SE. Individual
RESULTS
i", l~ .
L~~'~,"OCI ..
!
0
1 )
~ ,
1I n lin n 1r11L rfl[
~ ~LJ~ ~ILJ I'-:
1t:[ ~ ~~~
~
~
~
TIME
~~~
, "'wM
__
~ ~ ! /'j~;
,
V
,~
