- Email: [email protected]
Automated blood pressure monitoring for the assessment of antihypertensive treatment
Automated Blood Pressure Monitoring for th Assessment of Antihypertensive Treatment Michael A. Weber, MD
Whole-day automated monitoring with small portable devices allows the circadian pattern of blood pressure to be assessed conveniently in individual patients. Measurements provided by this technique appear to be more reproducible and physiologically relevant than conventional office readings. Ambulatory monitoring can enhance the evaluation of antihypertensive therapy, especially in clinical trials, in 3 ways. First, it can facilitate the diagnosis of hypertension and avoid potentially inappropriate treatment in patients whose hypertension is erroneously diagnosed by office measurements. Moreover, the monitoring technique appears to prevent placebo responses in therapeutic studies. Second, whole-day monitoring allows clear quantification of treatment effects. Because the standard deviations of treatment-induced blood pressure differences are lower with this method than with conventional readings, valid statistical evaluation of antihypertensive effects can be performed with far fewer patients. Finally, duration of action of drugs can be measured. By dividing the day into sequential periods (typically 12 two-hour periods), it is possible to determine statistical differences between baseline and treatment blood pressure values throughout the dosing intervals of the drugs being tested. This method may be more sensitive than the traditional “peak and trough” approach for assessing duration of action; it is not influenced by unsuspected discrepancies between the pharmacokinetic and pharmacodynamic properties of drugs, and can also provide information on nighttime effects. Thus, automated whole-day blood pressure monitoring appears to be a powerful tool for the evaluation of antihypertensive therapy. (Am J Cardiol 1988;62:97G-1026)
From the Hypertension Center, Veterans Administration Medical Center, Long Beach, California, and the University of California, Irvine, California. Address for reprints: Michael A. Weber, MD, Veterans Administration Medical Center, 5901 East Seventh Street, Long Beach, California 90822.
he most accurate and sensitive method for continuous measurementof blood pressureis by intraarterial monitoring.1,2But for reasonsof safety and convenience,especially when repeated studies are required in the same patient, noninvasive methods are now strongly preferred. The developmentof lightweight, portable devicescapable of repeatedly measuring blood pressure by arm-cuff methods have made whole-day blood pressure monitoring a practical technique. Two differing methodsof measurementare usedby thesedevices:Auscultatory, in which a transducerinterprets Korotkoff sounds,and oscillometric, in which the cuff system is able to perceivefluctuations of the brachial artery and derive values for systolic and diastolic blood pressures. Each of thesemethodshas beenshown to provide values that correlate quite closelywith simultaneouslymeasured intraarterial measurements,and that are virtually identical to readings obtained with a conventional mercury sphygmomanometer. Validity of Monitoring: As a basisfor using this technique for evaluating antihypertensive therapy, it is desirable to demonstrate that the whole-day blood pressure pattern obtainedwith thesedevicesis reproducible from 1 day to the next within individual patients. Studies with the intraarterial system have shown that the 24-hour blood pressurepattern is repeatablewhen studied on consecutive days and also when studied on separate days approximately 6 weeksapart.3 Using noninvasive equipment, we also have found that whole-day blood pressure values obtained under controlled in-hospital conditions4 or in truly ambulatory patients5 are also reproducible during periods of several weeks (Fig. 1). The physiologic relevanceof data provided by automated ambulatory blood pressuremonitoring is supported by 2 additional lines of evidence.First, blood pressure valuesobtained with this technique provide a clear circadian pattern (Fig. 1). Blood pressuretypically is at its highest value during the daytime hours, then decreases steadily to a nadir soon after midnight, and finally increasesin a fairly sharp fashion during the early morning hours near the time of arousal. It seemslikely that this pattern of blood pressureis determined to someextent by activity of the sympatheticnervoussystem,for it hasbeen shown that plasma concentrationsof the neurotransmitter, norepinephrine,havea circadian pattern closelysimilar to that of blood pressure.6This relation may actually be of someclinical and therapeutic importance,becauseit suggeststhat sympathetic activity may be an important factor in producing the rapid early morning increase in blood pressure.Considerableattention has been focused
T
THE AMERICAN JOURNAL OF CARDIOLOGY OCTOBER 5,1988
976
A SYMPOSIUM:
CALCIUM,
THE CARDIOVASCULAR
SYSTEM
on these hours, becauseit has been reported that both cerebrovascularand coronary eventsmay occur with increasedfrequency at this time of day.‘s8Thus, data provided by automated portable equipment on blood pressure values and their responsesto treatment at this time of day may be of considerableimportance. A recent study has focused attention on the ability of certain antihypertensivedrugs to influence adrenergic factors and even /3endorphins during the nocturnal hours in hypertensive patients.9 A further confirmation of the relevanceof blood pressure monitoring comesfrom studies relating blood pressure values to echocardiographically derived measurements of cardiac left ventricular muscle mass. It is assumed that the massof the left ventricle is a consistent and reproducible index of vascular involvement in the hypertensive process. The averaged whole-day blood pressure in normotensive and mildly hypertensive patients correlates far more closely with the calculated left ventricular muscle mass than do conventionally measured clinic or office blood pressures.lOJ1This relation doesnot appear to occur in patients with large hearts at later stagesof hypertensivedisease,presumably because somedegreeof systolic dysfunction in thesepatients may make it difficult for them to sustain their original high levels of blood pressure.12Moreover, at least 1 study has suggestedthat ambulatory monitoring is superior to conventional readingsin predicting prognosisin hypertensive patients.13 ROLE
AND THE KIDNEY
presentedpreviously,14this technique can aid in making the diagnosisof hypertension.Theoretically, this process shouldidentify patients in whom treatment is appropriate and most likely to be effective. Second, blood pressure monitoring can help determine the efficacy of antihypertensivetreatment, allowing valid comparisonsof baseline and treatment blood pressure values in comparatively small patient groups.Third, this technique allows investigators to make assumptionsconcerning duration of action, specifically making it possible to judge whether blood pressureis changedthroughout the dosingintervals of antihypertensive drugs. Diagnostic application: Various clinical trials of antihypertensive therapy have had disappointing results; greater reductions in the incidence of cardiovascular complications have not beenobservedin actively treated patients.15J6However, the Australian Therapeutic Trial in Mild Hypertension,I6 found that almost half of the
r
OF MONITORING
Automated long-term blood pressuremonitoring potentially can contribute to the assessmentof antihypertensive treatment in 3 separateways. Firs!, as has been HYPERTENSIVE
-
AVERAGE
+40
d
SYSTOLIC BLOOD
r
PRESSURE
(MM HG)
r
.
l 110
AVERAGE
o
FIRST
0
SECOND
&HR
.. .
STUDY
2’kHR STUDY
.:.. . ..... $ipt 0:
I
.
DIASTOLIC BLOOD h
. f:. *ii* .. .... .* .-. 0: ::::e. l 0.:’ . z:.. .**
l .**
PRESSURE
HG)
-40i
l
HYPERTENSIVE 0 PM ,,ME
NORMOTENSIVE
0 AR
OF DAY
L FIGURE 1. Average systolic and diastolic blood pressure values of 56 normotensive subjects during 2 separate days (approximately 6 weeks apart) of whole-day ambulatory blood pressure monitoring. Points shown are the mean f standard deviation of the readings obtained during each of 12 two-hour periods of the day.
986
NORhlOTENSlVE
THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 62
FIGURE 2. Differences in patients between conventionally measured office blood pressures and whole-day blood pressure averages. Subjects with values above the zero line are those whose whole-day averages were higher than the office measurements. (Reprinted with permission from Weber MA, Drayer JIM. Role of blood pressure monitoring in the diagnosis of hypertension. J Hypertens 1986;4:suppl5:S325-S327.)
placebo-treated patients had clear decreasesin blood pressureduring the study, in many casesto completely normal levels. Although there may be various explanations for this surprising effect, it seemsreasonable to assumethat many of these patients with strong placebo responseswere not truly hypertensive.Obviously, the inclusion of a large number of such patients in a trial can dilute its power to show a benefit of treatment. The phenomenonof “office hypertension” or “white coat syndrome” has been reported by several investigators.t7-I9 It has been argued that the artifactually high blood pressuremeasuredby the physician is not simply a reflection of nervousnessor anxiety by the patient, but instead is a true conditioned responsethat can be reinforced during the first several visits to the physician’s office.19Differences between blood pressuresmeasured in the office and during whole-day ambulatory monitoring are shownin Figure 2. It is evident that there is a wide variation among persons;most subjectshave office blood pressuresthat are higher than the whole-day values, in some substantially so, whereas other patients appear to havefalsely low readingsin the office. It is calculated that between 20 and 30% of patients diagnosed as having hypertension by conventional means do not have high blood pressure when studied for the entire day.t7m20 As discussedlater, the inclusion of “non-confirmed” hypertensivepatients in a trial can have a meaningful im-
FIGURE 3. Ambulatory whole-day systolic and diastolic blood pressure at baseline and during treatment with diltiazem (240 to 360 mg/day) in 15 hypertensive patients. Values shown are the means (for all patients together) for each of the 12 twohour periods comprising the full day. (Reprinted with permission from JAMA?)
pact on the blood pressureresponsesto antihypertensive therapy. Efficacy studies: A key difference between clinical trials using conventional blood pressure measurements and those using automated blood pressuremonitoring is in the likelihood of placeboresponse.As discussedearlier, conventional clinical trials have observedsignificant placeboeffectson blood pressure.15J6 In contrast, the strong reproducibility of the whole-day blood pressurefrom 1 day of monitoring to another5appearsto prevent placebo responses.Indeed, in clinical studies that use 24-hour blood pressuremonitoring there appearsto be a complete absenceof placebo effect on blood pressurevalues.21-23 This may be important in allowing thesetechniquesto discriminate meaningful antihypertensive effectsin relatively small patient groups.22A recent report, confirming the absenceof placebo effects with ambulatory blood pressure monitoring,24claimed that the standard deviation of the whole-dayblood pressuredifferencesbetween treatmentsin hypertensivesubjectswas sufficiently small to enable conclusions to be reached using one-third as many patients aswould be required if conventional blood pressure readings were used. This finding could be of considerablevalue in facilitating clinical trials of antihypertensivetreatment. In the sameway that automated monitoring appears to improve the accuracy of the initial diagnosis, it may
FIGURE 4. Ambulatory whole-day systolic and diastolic blood pressure at baseline and during treatment with diltiazem (240 to 360 mg/day) in 9 hypertensive patients. Values shown are the means (for all patients together) for each of the 12 twohour periods comprising the full day. THE AMERICAN JOURNAL OF CARDIOLOGY OCTOBER 5,1988
99G
A SYMPOSIUM:
CALCIUM,
THE CARDIOVASCULAR
SYSTEM AND THE KIDNEY
also add to the accuracy of efficacy determinations. A given just oncedaily. 3oBecausethe entire daily dosewas recent study has demonstrated that even carefully ob- given as a single administration in the once-daily group, tained clinical measuresof blood pressureduring antihy- there was a greater reduction in blood pressureafter the pertensivetrials correlate poorly with blood pressureval- morning dosing than after the twice-daily administraues provided by automated monitoring; of interest, this tions, It was noteworthy, however,that the greater blood study noted that blood pressuremeasurementsobtained pressure decrement produced by the single high dose by patients themselvesat home agreedmore closely with disappearedduring the nighttime hours, and there wasan the monitoring data than did the clinic measurements.25 “escape” during the final few hours beforethe next day’s Another report, based on patients undergoing routine morning dose.Clinicai measurementsof blood pressure managementin a clinic, also discoveredthat several pa- during the standard daytime hours during which patients tients with apparent resistanceto antihypertensive thera- typically cometo an office or clinic could havemistakenly py had acceptablecontrol of their blood pressurewhen indicated that the once-daily regimen of the standard studied by ambulatory monitoring.26 formulation was effective. Duration of effects: By comparing the whole-day There have been some interesting studies with the blood pressure pattern during placebo treatment with converting enzyme inhibitor captopril. An earlier study that obtained during antihypertensive therapy, it is possi- using whole-day blood pressure monitoring confirmed ble to determine whether blood pressure is decreased that this agent given twice daily significantly decreased during the entire dosinginterval. Studies performed with blood pressurewhile maintaining its circadian pattern at this monitoring method, therefore, have allowed investi- values lower than the baseline values.31A more recent gators to reach conclusions concerning the duration of study, however,has consideredthe duration of effect of action of antihypertensive drugs when administered in the samedrug given in a single daily dose.32Interestingly, differing regimens.23,27 the investigators in this latter study argued that duration This technique is also useful when differing regimens of effect should be studied as a separateissuefrom antiof the same treatment are compared with each other hypertensiveefficacy, and thus they eliminated from conrather than with placebo.For example,such studieshave sideration thosepatients (10 of 31) whoseblood pressures been able to determine whether a sustained releasefor- did not appearto respondadequately to captopril. When mulation of an antihypertensive agent is as effective throughout the full 24-hour period aswhen the samedrug in a shorter-acting formulation is administered 3 times daily.28,29We have used the monitoring technique in a study in which the standard formulation of prazosingiven twice daily was compared with the same formulation
CLINICAL VALUES 110 r
2 g 1 R
l
105
l
I l
100 :.
s 6 3 0 P :: g EY
24 HOUR AVERAGES
a%
95-
l
90 -
l
85 -
z
SOL
(-DENOTES HYPERTENSION CRITERION)
‘$7 :
FIGURE 5. Conventional office blood pressure values and whole-day ambulatory averaged blood pressure values at baseline in 15 patients treated with diltiazem. Arrows indicate levels at which hypertension is diagnosed. (Arrow is at lower level for whole-day average since this value includes the lower nocturnal readings that typically make the whole-day average 5 mm Hg less than daytime readings.) (Reprinted with permission from JAMA.23)
100G
THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 62
FIGURE 6. Ambulatory whole-day systolic and diastolic blood pressure at baseline and during treatment with diltiazem (240 to 360 mg/day) In 6 normotensive (“non-confirmed” hypertensive) patients. Values shown are the means (for all patients together) for each of the 12 two-hour periods comprising the full day. (Reprinted with permission from JAMA.23)
the remaining 21 responding patients were now considered by themselves,it was found that the treatment was not only efficacious but that it now produced an antihypertensiveeffect during the entire 24-hour period. Interpreting data: As can be seen from the data in Figure 1, it is most convenient to present data obtained with the noninvasive monitoring technique as the averagesof the 12 two-hour periods that comprisethe whole day. Typically, we obtain between4 and 8 readingswithin each 2-hour period for each patient so as to provide a meaningful blood pressureaverage for that period. It is then possibleto compareeachof the IZhour periodswith the respectiveperiods during 2 separatedays of monitoring. This comparison can be made using an analysis of variance. It is necessary,however,to ensurethat data are available for all of the 2-hour periods so as to make the comparisonvalid. If, for technical or other reasons,there are 1 or more periodsin which data are missing, it is then necessaryto useformulas that can extrapolate valuesfor these periods. Techniques such as split-plot analysis of variance make it possibleto determine not only whether there is a quantitative difference between the values in two 24-hour blood pressurecurves, but also whether the curves are parallel to each other. If parallelism is maintained during the 24-hour period, then it is reasonableto assumethat a treatment has full-day efficacy. A lack of parallelism raisesthe possibility that there may have been a lossof efficacy during 1 or more of the 2-hour periodsor that the effects of treatment were inconsistent. Another obvious benefit of the analysis by 2-hour periods during the day is that it allows the efficacy of 1 drug to be compared with another. For each drug, it is possiblewithin each of the 12 two-hour periods to calculate the difference between the baseline and the treatment valuesfor the group of patients asa whole. Thus, for each drug being compared, it is possible to calculate a “delta” value for each of the 12 two-hour periods, and then to comparethe deltas for the 2 drugs throughout the day. Other methods for analyzing data from whole-day blood pressure monitoring include simply using the whole-day blood pressureaveragesand doing direct comparisons between placebo and treatment values, or between 2 different treatments.Another approachwould be ,to calculate the percentageof values throughout a day that are abovean arbitrary level (for example,90 mm Hg for diastolic blood pressure),and determine if treatment produces a significant change in the frequency of such readings.Other investigatorshave regardedthe degreeof blood pressure variability during the day as being as important as the actual blood pressurevalues, and have performed tests on treatment-induced changes in variability.33As yet, there is no generalagreementon the best method for analyzing thesedata. In many instances(Fig. 3 and 4), efficacy and duration of antihypertensivetherapy seemreadily apparent on visual inspectionof the baseline and treatment circadian patterns. An illustrative experience: Whole-day ambulatory blood pressuremonitoring can facilitate the diagnosisof hypertension, and assessthe efficacy and duration of
action of antihypertensivetreatment. Each of theseattributes wasof value in a recent study of a calcium-channel antagonist,diltiazem, in patients with essentialhypertension.23Placebotreatment, as discussedearlier, produced no changesin blood pressure,but a group of 15 patients who received diltiazem (in its standard tablet formulation) twice daily had significant and sustaineddecreases in blood pressure(Fig. 3). Thesepatients were all entered into the active treatment phaseof the study on the basisof conventionaldiastolic blood pressurevaluesof 95 mm Hg or greater. However, on subsequentanalysis of the baseline 24-hour blood pressurevalues, it becameapparent that 6 of these 15patients did not satisfy the usual criterion for diagnosisof hypertension (Fig. 5). If the effectsof the drug treatment were now evaluatedseparatelyin the 6 patients with “non-confirmed” hypertension (Fig. 6) and in the 9 patients with establishedhypertension (Fig. 4), it is evident that the drug had clearly different effects within the 2 subgroups.Diltiazem produced no significant changein blood pressurein patients with “normal” baselinevalues,whereasit had powerful antihypertensive efficacy in thosewho wereclearly hypertensive.Thus, the technique of whole-day blood pressuremonitoring in this study made it possibleto define patients in whom treatment appearedto be most appropriate as well as providing a clear quantification of the antihypertensiveproperties of the drug.
REFERENCES 1. Bevan AT, Honour AJ, Scott FH. Portable recorder for continuous arterial pressure measurement in man. J Physioi (LA&) 1966;38:186-190. 2. Millar-Craig MW, Hawes B, Whittington J. A new system for recording ambulatory blood pressure in man. Med Biol Eng Comput 1978;16:727-731. 3. Gould BA, Mann S, Davies AB, Altman DG, Raftery EB. Can placebo therapy influence arten’al blood pressure? Clin Sci 1981,61:478S-482X 4. Weber MA, Drayer JIM, Wyle FA, DeYoung JL. Reproducibility of the whole-day blood pressure pattern in essential hypertension. Clin Exp Hypertens 1982;A4:1377-1390, 5. Weber MA, Drayer JIM, Nakamura DK, Wyle FA. The circadian blood pressure pattern in ambulatory normal subjects. Am J Cardioll984:54:115-119. 6. Linsell CR, Lightman SL, Mullen PE, Brown MJ, Causon RC. Circadian rhythms of epinephrine and norepinephrine in man. J Clin Endocrinol Metab 1985,60:1210-1215. 7. Agnoli A, Manfredi M, Mossuto L, Piccinelli A. Rapport entre les rythmes hemeronyctaux de lo tension arterielle et sa pathogenie de i’insuffisance oosculake cerebrale. Rev Neural 1975;131:597-606. 6. Meyers A, Dewar HA. Circumstances attending 100 sudden deaths from coronary artery disease with coroners’ necrotxies. Br Heart J 1975;37:11331143. 9. Stump KO, Thwbald R, Kolloch R, Kraft K. Nocturnal changes in blood pressure ofyoungpaiients with mild essential hypertension are related to changes in plasma beta-endorphin concentrations. J Hypertens 1977;S:suppl5:~87-~89. 10. Rowlands DB, Ireland MA, Glover DR, McLeau RAB, Stallard TJ, Littler WA. The relationship between ambulatory blood pressure and echocardiographically assessed left ventricular hypertrophy. Clin Sci 1981;6l:suppl lOlS-
103s. II. Drayer JIM, Weber MA, DeYoung JL. BP as a determinant of cardiac left ventricular muscle mass. Arch Intern Med 1983:143:90-92, 12. Drayer JIM, Gardin JM, Brewer DD, We&r MA. Disparate relationships between blood pressure and left ventricular mass in patients with and without left ventricular hypertrophy. Hypertension 1987,9:suppl lI:Il-61 -II-64. 13. Perloff D, Sokolow M, Cowan R. The prognostic value of ambulatory blood pressure. JAMA 1983;249:2793-2798. 14. W&r MA. Whole-day blood pressure. Hypertension 1988;11:288-289. 15. Medical ResearchCouncil Working Party. MRC trial of treatment of mild hypertension principal results. Br Med J 1985;291:97-104. 16. Report by the ManagementCommittee. The Australian Therapeutic Trial in Mild Hypertension. Lancet 1980;1:1261-1267.
THE AMERICAN JOURNAL OF CARDIOLOGY OCTOBER 5.1988
1OlG
A SYMPOSIUM: CALCIUM, THE CARDIOVASCULAR SYSTEM AND THE KIDNEY 17. Drayer JIM, Weber MA, Nakamura DK. Automated ambulatory blood pressure monitoring: a study in age-matchednormotensiue and hypertensive men. Am Heart J 1985;109:1334-1338.
18. Mancia G, Bertinieri G, Parati G, PomidossiG, Ferrari A, Gregorini L, Zanchetti A. Effects of blood pressure measurement by the doctor on patients’ blood pressure and heart rate. Lancet 1983;2:495-697.
19. Pickering TG, JamesGD, Boddie C, Harshfield GA, Blank S, Laragh JH. How common is white coat hypertension? JAMA 1988;259:225-228. 20. White WB. Assessment of patients with office hypertension by 24-hour noninvasive ambulatory blood pressure monitoring. Arch Int Med 1986;146: 2196-2199. 21. Gould BA, Davies AB, Mann S, Altman DG, Raftery EB. Does placebo lower blood pressure? Lancer 198l;ii:l377-1381. 22. Drayer JIM, Weber MA, DeYoung JL, Brewer DD. Long-term blood pressure monitoring in the evaluation of antihypertensive therapy. Arch Int Med 1983:143:898-901. 23. Weber MA, Cheung DG, Graettinger WF, Lipson JL. Characterization of antihypertensiue therapy by whole-day BP monitoring. JAMA 1988;259:32813285. 24. Conway J, JohnstonJ, Coats A, SomersV, Sleight P. The use of ambulatory blood pressure monitoring to improve the accuracy and reduce the numbers of subjects in clinical trials of antihypertensive agents. J Hypertens 1988,6:111116. 25. Gould BA, Hornung RS, Kieso H, CashmanPMM, Raftery EB. An eoaluation of self-recorded blood pressure during drug trials. Hypertension 1986; 8:267-271.
1026
THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 62
26. Porchet M, BussienJP, Waeber B, NussbergerJ, Brunner HR. Unpredictability of blood pressures recorded outside the clinic in the treated hypertensive patient. J Cardiovasc Pharmacol 1986;8:332-335. 27. White WB, Smith VE, McCabe EJ, Meeran MK. Effects of chronicnitrendipine on casual (office) and 24-hour ambulatory blood pressure. Clin Pharmacol Ther 1985;38:60-64.
28. Zachariah PK. ShepsS, Schirger A, Spiekman RE, O’Brien PC, Simpson KK. Verapamil and 24-hour ambulatory blood pressure monitoring in essential hypertension. Am J Cardiol 1986:57:74D-790.
29. Nissinen A, Koistinen A, Tuomilehto J, SundbergS, Gordin A. Sustained release verapamil in hypertension-rest&from a noninvasiue ambulatory blood pressure monitoring and a clinical study. Eur J Clin Phnrmacol 1986;31:255260. 30. Weber MA, Tonkon MJ, Klein RC. Bloodpressure monitoringfor assessing the duration of action of antihypertensive treatment. J Clin Pharmacol 1987; 27:751-755. 31. Meijer JL, Ardesch HG, van Rooijen JC, De Bruijn JHB. Low dose captopril twice daily lowers blood pressure without disturbance of the normal circadian rhythm. Postgrad Med J 1986,62:101-105.
32. Mancia G, Parati G, PomidossiG, Colombo A, Cuspidi C, Lattuada S, Antivalle M, Rindi M, Libretti A, Gianfranco B, Zanchetti A. Eualuation of the antihypertensive effect of once-a-day captoprii by 24-hour ambulatory blood pressure monitoring. J Hypertens 1987;S:suppl 5:S591 -S593. 33. de Faire U, Forslund L, OdenA. Variability of bloodpressure in ambulatory hypertensive patients: effects of verapamil on twice and thrice daily dose regimens. Acta Med Stand 1986;220:411-418.
Whole-day automated monitoring with small portable devices allows the circadian pattern of blood pressure to be assessed conveniently in individual patients. Measurements provided by this technique appear to be more reproducible and physiologically relevant than conventional office readings. Ambulatory monitoring can enhance the evaluation of antihypertensive therapy, especially in clinical trials, in 3 ways. First, it can facilitate the diagnosis of hypertension and avoid potentially inappropriate treatment in patients whose hypertension is erroneously diagnosed by office measurements. Moreover, the monitoring technique appears to prevent placebo responses in therapeutic studies. Second, whole-day monitoring allows clear quantification of treatment effects. Because the standard deviations of treatment-induced blood pressure differences are lower with this method than with conventional readings, valid statistical evaluation of antihypertensive effects can be performed with far fewer patients. Finally, duration of action of drugs can be measured. By dividing the day into sequential periods (typically 12 two-hour periods), it is possible to determine statistical differences between baseline and treatment blood pressure values throughout the dosing intervals of the drugs being tested. This method may be more sensitive than the traditional “peak and trough” approach for assessing duration of action; it is not influenced by unsuspected discrepancies between the pharmacokinetic and pharmacodynamic properties of drugs, and can also provide information on nighttime effects. Thus, automated whole-day blood pressure monitoring appears to be a powerful tool for the evaluation of antihypertensive therapy. (Am J Cardiol 1988;62:97G-1026)
From the Hypertension Center, Veterans Administration Medical Center, Long Beach, California, and the University of California, Irvine, California. Address for reprints: Michael A. Weber, MD, Veterans Administration Medical Center, 5901 East Seventh Street, Long Beach, California 90822.
he most accurate and sensitive method for continuous measurementof blood pressureis by intraarterial monitoring.1,2But for reasonsof safety and convenience,especially when repeated studies are required in the same patient, noninvasive methods are now strongly preferred. The developmentof lightweight, portable devicescapable of repeatedly measuring blood pressure by arm-cuff methods have made whole-day blood pressure monitoring a practical technique. Two differing methodsof measurementare usedby thesedevices:Auscultatory, in which a transducerinterprets Korotkoff sounds,and oscillometric, in which the cuff system is able to perceivefluctuations of the brachial artery and derive values for systolic and diastolic blood pressures. Each of thesemethodshas beenshown to provide values that correlate quite closelywith simultaneouslymeasured intraarterial measurements,and that are virtually identical to readings obtained with a conventional mercury sphygmomanometer. Validity of Monitoring: As a basisfor using this technique for evaluating antihypertensive therapy, it is desirable to demonstrate that the whole-day blood pressure pattern obtainedwith thesedevicesis reproducible from 1 day to the next within individual patients. Studies with the intraarterial system have shown that the 24-hour blood pressurepattern is repeatablewhen studied on consecutive days and also when studied on separate days approximately 6 weeksapart.3 Using noninvasive equipment, we also have found that whole-day blood pressure values obtained under controlled in-hospital conditions4 or in truly ambulatory patients5 are also reproducible during periods of several weeks (Fig. 1). The physiologic relevanceof data provided by automated ambulatory blood pressuremonitoring is supported by 2 additional lines of evidence.First, blood pressure valuesobtained with this technique provide a clear circadian pattern (Fig. 1). Blood pressuretypically is at its highest value during the daytime hours, then decreases steadily to a nadir soon after midnight, and finally increasesin a fairly sharp fashion during the early morning hours near the time of arousal. It seemslikely that this pattern of blood pressureis determined to someextent by activity of the sympatheticnervoussystem,for it hasbeen shown that plasma concentrationsof the neurotransmitter, norepinephrine,havea circadian pattern closelysimilar to that of blood pressure.6This relation may actually be of someclinical and therapeutic importance,becauseit suggeststhat sympathetic activity may be an important factor in producing the rapid early morning increase in blood pressure.Considerableattention has been focused
T
THE AMERICAN JOURNAL OF CARDIOLOGY OCTOBER 5,1988
976
A SYMPOSIUM:
CALCIUM,
THE CARDIOVASCULAR
SYSTEM
on these hours, becauseit has been reported that both cerebrovascularand coronary eventsmay occur with increasedfrequency at this time of day.‘s8Thus, data provided by automated portable equipment on blood pressure values and their responsesto treatment at this time of day may be of considerableimportance. A recent study has focused attention on the ability of certain antihypertensivedrugs to influence adrenergic factors and even /3endorphins during the nocturnal hours in hypertensive patients.9 A further confirmation of the relevanceof blood pressure monitoring comesfrom studies relating blood pressure values to echocardiographically derived measurements of cardiac left ventricular muscle mass. It is assumed that the massof the left ventricle is a consistent and reproducible index of vascular involvement in the hypertensive process. The averaged whole-day blood pressure in normotensive and mildly hypertensive patients correlates far more closely with the calculated left ventricular muscle mass than do conventionally measured clinic or office blood pressures.lOJ1This relation doesnot appear to occur in patients with large hearts at later stagesof hypertensivedisease,presumably because somedegreeof systolic dysfunction in thesepatients may make it difficult for them to sustain their original high levels of blood pressure.12Moreover, at least 1 study has suggestedthat ambulatory monitoring is superior to conventional readingsin predicting prognosisin hypertensive patients.13 ROLE
AND THE KIDNEY
presentedpreviously,14this technique can aid in making the diagnosisof hypertension.Theoretically, this process shouldidentify patients in whom treatment is appropriate and most likely to be effective. Second, blood pressure monitoring can help determine the efficacy of antihypertensivetreatment, allowing valid comparisonsof baseline and treatment blood pressure values in comparatively small patient groups.Third, this technique allows investigators to make assumptionsconcerning duration of action, specifically making it possible to judge whether blood pressureis changedthroughout the dosingintervals of antihypertensive drugs. Diagnostic application: Various clinical trials of antihypertensive therapy have had disappointing results; greater reductions in the incidence of cardiovascular complications have not beenobservedin actively treated patients.15J6However, the Australian Therapeutic Trial in Mild Hypertension,I6 found that almost half of the
r
OF MONITORING
Automated long-term blood pressuremonitoring potentially can contribute to the assessmentof antihypertensive treatment in 3 separateways. Firs!, as has been HYPERTENSIVE
-
AVERAGE
+40
d
SYSTOLIC BLOOD
r
PRESSURE
(MM HG)
r
.
l 110
AVERAGE
o
FIRST
0
SECOND
&HR
.. .
STUDY
2’kHR STUDY
.:.. . ..... $ipt 0:
I
.
DIASTOLIC BLOOD h
. f:. *ii* .. .... .* .-. 0: ::::e. l 0.:’ . z:.. .**
l .**
PRESSURE
HG)
-40i
l
HYPERTENSIVE 0 PM ,,ME
NORMOTENSIVE
0 AR
OF DAY
L FIGURE 1. Average systolic and diastolic blood pressure values of 56 normotensive subjects during 2 separate days (approximately 6 weeks apart) of whole-day ambulatory blood pressure monitoring. Points shown are the mean f standard deviation of the readings obtained during each of 12 two-hour periods of the day.
986
NORhlOTENSlVE
THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 62
FIGURE 2. Differences in patients between conventionally measured office blood pressures and whole-day blood pressure averages. Subjects with values above the zero line are those whose whole-day averages were higher than the office measurements. (Reprinted with permission from Weber MA, Drayer JIM. Role of blood pressure monitoring in the diagnosis of hypertension. J Hypertens 1986;4:suppl5:S325-S327.)
placebo-treated patients had clear decreasesin blood pressureduring the study, in many casesto completely normal levels. Although there may be various explanations for this surprising effect, it seemsreasonable to assumethat many of these patients with strong placebo responseswere not truly hypertensive.Obviously, the inclusion of a large number of such patients in a trial can dilute its power to show a benefit of treatment. The phenomenonof “office hypertension” or “white coat syndrome” has been reported by several investigators.t7-I9 It has been argued that the artifactually high blood pressuremeasuredby the physician is not simply a reflection of nervousnessor anxiety by the patient, but instead is a true conditioned responsethat can be reinforced during the first several visits to the physician’s office.19Differences between blood pressuresmeasured in the office and during whole-day ambulatory monitoring are shownin Figure 2. It is evident that there is a wide variation among persons;most subjectshave office blood pressuresthat are higher than the whole-day values, in some substantially so, whereas other patients appear to havefalsely low readingsin the office. It is calculated that between 20 and 30% of patients diagnosed as having hypertension by conventional means do not have high blood pressure when studied for the entire day.t7m20 As discussedlater, the inclusion of “non-confirmed” hypertensivepatients in a trial can have a meaningful im-
FIGURE 3. Ambulatory whole-day systolic and diastolic blood pressure at baseline and during treatment with diltiazem (240 to 360 mg/day) in 15 hypertensive patients. Values shown are the means (for all patients together) for each of the 12 twohour periods comprising the full day. (Reprinted with permission from JAMA?)
pact on the blood pressureresponsesto antihypertensive therapy. Efficacy studies: A key difference between clinical trials using conventional blood pressure measurements and those using automated blood pressuremonitoring is in the likelihood of placeboresponse.As discussedearlier, conventional clinical trials have observedsignificant placeboeffectson blood pressure.15J6 In contrast, the strong reproducibility of the whole-day blood pressurefrom 1 day of monitoring to another5appearsto prevent placebo responses.Indeed, in clinical studies that use 24-hour blood pressuremonitoring there appearsto be a complete absenceof placebo effect on blood pressurevalues.21-23 This may be important in allowing thesetechniquesto discriminate meaningful antihypertensive effectsin relatively small patient groups.22A recent report, confirming the absenceof placebo effects with ambulatory blood pressure monitoring,24claimed that the standard deviation of the whole-dayblood pressuredifferencesbetween treatmentsin hypertensivesubjectswas sufficiently small to enable conclusions to be reached using one-third as many patients aswould be required if conventional blood pressure readings were used. This finding could be of considerablevalue in facilitating clinical trials of antihypertensivetreatment. In the sameway that automated monitoring appears to improve the accuracy of the initial diagnosis, it may
FIGURE 4. Ambulatory whole-day systolic and diastolic blood pressure at baseline and during treatment with diltiazem (240 to 360 mg/day) in 9 hypertensive patients. Values shown are the means (for all patients together) for each of the 12 twohour periods comprising the full day. THE AMERICAN JOURNAL OF CARDIOLOGY OCTOBER 5,1988
99G
A SYMPOSIUM:
CALCIUM,
THE CARDIOVASCULAR
SYSTEM AND THE KIDNEY
also add to the accuracy of efficacy determinations. A given just oncedaily. 3oBecausethe entire daily dosewas recent study has demonstrated that even carefully ob- given as a single administration in the once-daily group, tained clinical measuresof blood pressureduring antihy- there was a greater reduction in blood pressureafter the pertensivetrials correlate poorly with blood pressureval- morning dosing than after the twice-daily administraues provided by automated monitoring; of interest, this tions, It was noteworthy, however,that the greater blood study noted that blood pressuremeasurementsobtained pressure decrement produced by the single high dose by patients themselvesat home agreedmore closely with disappearedduring the nighttime hours, and there wasan the monitoring data than did the clinic measurements.25 “escape” during the final few hours beforethe next day’s Another report, based on patients undergoing routine morning dose.Clinicai measurementsof blood pressure managementin a clinic, also discoveredthat several pa- during the standard daytime hours during which patients tients with apparent resistanceto antihypertensive thera- typically cometo an office or clinic could havemistakenly py had acceptablecontrol of their blood pressurewhen indicated that the once-daily regimen of the standard studied by ambulatory monitoring.26 formulation was effective. Duration of effects: By comparing the whole-day There have been some interesting studies with the blood pressure pattern during placebo treatment with converting enzyme inhibitor captopril. An earlier study that obtained during antihypertensive therapy, it is possi- using whole-day blood pressure monitoring confirmed ble to determine whether blood pressure is decreased that this agent given twice daily significantly decreased during the entire dosinginterval. Studies performed with blood pressurewhile maintaining its circadian pattern at this monitoring method, therefore, have allowed investi- values lower than the baseline values.31A more recent gators to reach conclusions concerning the duration of study, however,has consideredthe duration of effect of action of antihypertensive drugs when administered in the samedrug given in a single daily dose.32Interestingly, differing regimens.23,27 the investigators in this latter study argued that duration This technique is also useful when differing regimens of effect should be studied as a separateissuefrom antiof the same treatment are compared with each other hypertensiveefficacy, and thus they eliminated from conrather than with placebo.For example,such studieshave sideration thosepatients (10 of 31) whoseblood pressures been able to determine whether a sustained releasefor- did not appearto respondadequately to captopril. When mulation of an antihypertensive agent is as effective throughout the full 24-hour period aswhen the samedrug in a shorter-acting formulation is administered 3 times daily.28,29We have used the monitoring technique in a study in which the standard formulation of prazosingiven twice daily was compared with the same formulation
CLINICAL VALUES 110 r
2 g 1 R
l
105
l
I l
100 :.
s 6 3 0 P :: g EY
24 HOUR AVERAGES
a%
95-
l
90 -
l
85 -
z
SOL
(-DENOTES HYPERTENSION CRITERION)
‘$7 :
FIGURE 5. Conventional office blood pressure values and whole-day ambulatory averaged blood pressure values at baseline in 15 patients treated with diltiazem. Arrows indicate levels at which hypertension is diagnosed. (Arrow is at lower level for whole-day average since this value includes the lower nocturnal readings that typically make the whole-day average 5 mm Hg less than daytime readings.) (Reprinted with permission from JAMA.23)
100G
THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 62
FIGURE 6. Ambulatory whole-day systolic and diastolic blood pressure at baseline and during treatment with diltiazem (240 to 360 mg/day) In 6 normotensive (“non-confirmed” hypertensive) patients. Values shown are the means (for all patients together) for each of the 12 two-hour periods comprising the full day. (Reprinted with permission from JAMA.23)
the remaining 21 responding patients were now considered by themselves,it was found that the treatment was not only efficacious but that it now produced an antihypertensiveeffect during the entire 24-hour period. Interpreting data: As can be seen from the data in Figure 1, it is most convenient to present data obtained with the noninvasive monitoring technique as the averagesof the 12 two-hour periods that comprisethe whole day. Typically, we obtain between4 and 8 readingswithin each 2-hour period for each patient so as to provide a meaningful blood pressureaverage for that period. It is then possibleto compareeachof the IZhour periodswith the respectiveperiods during 2 separatedays of monitoring. This comparison can be made using an analysis of variance. It is necessary,however,to ensurethat data are available for all of the 2-hour periods so as to make the comparisonvalid. If, for technical or other reasons,there are 1 or more periodsin which data are missing, it is then necessaryto useformulas that can extrapolate valuesfor these periods. Techniques such as split-plot analysis of variance make it possibleto determine not only whether there is a quantitative difference between the values in two 24-hour blood pressurecurves, but also whether the curves are parallel to each other. If parallelism is maintained during the 24-hour period, then it is reasonableto assumethat a treatment has full-day efficacy. A lack of parallelism raisesthe possibility that there may have been a lossof efficacy during 1 or more of the 2-hour periodsor that the effects of treatment were inconsistent. Another obvious benefit of the analysis by 2-hour periods during the day is that it allows the efficacy of 1 drug to be compared with another. For each drug, it is possiblewithin each of the 12 two-hour periods to calculate the difference between the baseline and the treatment valuesfor the group of patients asa whole. Thus, for each drug being compared, it is possible to calculate a “delta” value for each of the 12 two-hour periods, and then to comparethe deltas for the 2 drugs throughout the day. Other methods for analyzing data from whole-day blood pressure monitoring include simply using the whole-day blood pressureaveragesand doing direct comparisons between placebo and treatment values, or between 2 different treatments.Another approachwould be ,to calculate the percentageof values throughout a day that are abovean arbitrary level (for example,90 mm Hg for diastolic blood pressure),and determine if treatment produces a significant change in the frequency of such readings.Other investigatorshave regardedthe degreeof blood pressure variability during the day as being as important as the actual blood pressurevalues, and have performed tests on treatment-induced changes in variability.33As yet, there is no generalagreementon the best method for analyzing thesedata. In many instances(Fig. 3 and 4), efficacy and duration of antihypertensivetherapy seemreadily apparent on visual inspectionof the baseline and treatment circadian patterns. An illustrative experience: Whole-day ambulatory blood pressuremonitoring can facilitate the diagnosisof hypertension, and assessthe efficacy and duration of
action of antihypertensivetreatment. Each of theseattributes wasof value in a recent study of a calcium-channel antagonist,diltiazem, in patients with essentialhypertension.23Placebotreatment, as discussedearlier, produced no changesin blood pressure,but a group of 15 patients who received diltiazem (in its standard tablet formulation) twice daily had significant and sustaineddecreases in blood pressure(Fig. 3). Thesepatients were all entered into the active treatment phaseof the study on the basisof conventionaldiastolic blood pressurevaluesof 95 mm Hg or greater. However, on subsequentanalysis of the baseline 24-hour blood pressurevalues, it becameapparent that 6 of these 15patients did not satisfy the usual criterion for diagnosisof hypertension (Fig. 5). If the effectsof the drug treatment were now evaluatedseparatelyin the 6 patients with “non-confirmed” hypertension (Fig. 6) and in the 9 patients with establishedhypertension (Fig. 4), it is evident that the drug had clearly different effects within the 2 subgroups.Diltiazem produced no significant changein blood pressurein patients with “normal” baselinevalues,whereasit had powerful antihypertensive efficacy in thosewho wereclearly hypertensive.Thus, the technique of whole-day blood pressuremonitoring in this study made it possibleto define patients in whom treatment appearedto be most appropriate as well as providing a clear quantification of the antihypertensiveproperties of the drug.
REFERENCES 1. Bevan AT, Honour AJ, Scott FH. Portable recorder for continuous arterial pressure measurement in man. J Physioi (LA&) 1966;38:186-190. 2. Millar-Craig MW, Hawes B, Whittington J. A new system for recording ambulatory blood pressure in man. Med Biol Eng Comput 1978;16:727-731. 3. Gould BA, Mann S, Davies AB, Altman DG, Raftery EB. Can placebo therapy influence arten’al blood pressure? Clin Sci 1981,61:478S-482X 4. Weber MA, Drayer JIM, Wyle FA, DeYoung JL. Reproducibility of the whole-day blood pressure pattern in essential hypertension. Clin Exp Hypertens 1982;A4:1377-1390, 5. Weber MA, Drayer JIM, Nakamura DK, Wyle FA. The circadian blood pressure pattern in ambulatory normal subjects. Am J Cardioll984:54:115-119. 6. Linsell CR, Lightman SL, Mullen PE, Brown MJ, Causon RC. Circadian rhythms of epinephrine and norepinephrine in man. J Clin Endocrinol Metab 1985,60:1210-1215. 7. Agnoli A, Manfredi M, Mossuto L, Piccinelli A. Rapport entre les rythmes hemeronyctaux de lo tension arterielle et sa pathogenie de i’insuffisance oosculake cerebrale. Rev Neural 1975;131:597-606. 6. Meyers A, Dewar HA. Circumstances attending 100 sudden deaths from coronary artery disease with coroners’ necrotxies. Br Heart J 1975;37:11331143. 9. Stump KO, Thwbald R, Kolloch R, Kraft K. Nocturnal changes in blood pressure ofyoungpaiients with mild essential hypertension are related to changes in plasma beta-endorphin concentrations. J Hypertens 1977;S:suppl5:~87-~89. 10. Rowlands DB, Ireland MA, Glover DR, McLeau RAB, Stallard TJ, Littler WA. The relationship between ambulatory blood pressure and echocardiographically assessed left ventricular hypertrophy. Clin Sci 1981;6l:suppl lOlS-
103s. II. Drayer JIM, Weber MA, DeYoung JL. BP as a determinant of cardiac left ventricular muscle mass. Arch Intern Med 1983:143:90-92, 12. Drayer JIM, Gardin JM, Brewer DD, We&r MA. Disparate relationships between blood pressure and left ventricular mass in patients with and without left ventricular hypertrophy. Hypertension 1987,9:suppl lI:Il-61 -II-64. 13. Perloff D, Sokolow M, Cowan R. The prognostic value of ambulatory blood pressure. JAMA 1983;249:2793-2798. 14. W&r MA. Whole-day blood pressure. Hypertension 1988;11:288-289. 15. Medical ResearchCouncil Working Party. MRC trial of treatment of mild hypertension principal results. Br Med J 1985;291:97-104. 16. Report by the ManagementCommittee. The Australian Therapeutic Trial in Mild Hypertension. Lancet 1980;1:1261-1267.
THE AMERICAN JOURNAL OF CARDIOLOGY OCTOBER 5.1988
1OlG
A SYMPOSIUM: CALCIUM, THE CARDIOVASCULAR SYSTEM AND THE KIDNEY 17. Drayer JIM, Weber MA, Nakamura DK. Automated ambulatory blood pressure monitoring: a study in age-matchednormotensiue and hypertensive men. Am Heart J 1985;109:1334-1338.
18. Mancia G, Bertinieri G, Parati G, PomidossiG, Ferrari A, Gregorini L, Zanchetti A. Effects of blood pressure measurement by the doctor on patients’ blood pressure and heart rate. Lancet 1983;2:495-697.
19. Pickering TG, JamesGD, Boddie C, Harshfield GA, Blank S, Laragh JH. How common is white coat hypertension? JAMA 1988;259:225-228. 20. White WB. Assessment of patients with office hypertension by 24-hour noninvasive ambulatory blood pressure monitoring. Arch Int Med 1986;146: 2196-2199. 21. Gould BA, Davies AB, Mann S, Altman DG, Raftery EB. Does placebo lower blood pressure? Lancer 198l;ii:l377-1381. 22. Drayer JIM, Weber MA, DeYoung JL, Brewer DD. Long-term blood pressure monitoring in the evaluation of antihypertensive therapy. Arch Int Med 1983:143:898-901. 23. Weber MA, Cheung DG, Graettinger WF, Lipson JL. Characterization of antihypertensiue therapy by whole-day BP monitoring. JAMA 1988;259:32813285. 24. Conway J, JohnstonJ, Coats A, SomersV, Sleight P. The use of ambulatory blood pressure monitoring to improve the accuracy and reduce the numbers of subjects in clinical trials of antihypertensive agents. J Hypertens 1988,6:111116. 25. Gould BA, Hornung RS, Kieso H, CashmanPMM, Raftery EB. An eoaluation of self-recorded blood pressure during drug trials. Hypertension 1986; 8:267-271.
1026
THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 62
26. Porchet M, BussienJP, Waeber B, NussbergerJ, Brunner HR. Unpredictability of blood pressures recorded outside the clinic in the treated hypertensive patient. J Cardiovasc Pharmacol 1986;8:332-335. 27. White WB, Smith VE, McCabe EJ, Meeran MK. Effects of chronicnitrendipine on casual (office) and 24-hour ambulatory blood pressure. Clin Pharmacol Ther 1985;38:60-64.
28. Zachariah PK. ShepsS, Schirger A, Spiekman RE, O’Brien PC, Simpson KK. Verapamil and 24-hour ambulatory blood pressure monitoring in essential hypertension. Am J Cardiol 1986:57:74D-790.
29. Nissinen A, Koistinen A, Tuomilehto J, SundbergS, Gordin A. Sustained release verapamil in hypertension-rest&from a noninvasiue ambulatory blood pressure monitoring and a clinical study. Eur J Clin Phnrmacol 1986;31:255260. 30. Weber MA, Tonkon MJ, Klein RC. Bloodpressure monitoringfor assessing the duration of action of antihypertensive treatment. J Clin Pharmacol 1987; 27:751-755. 31. Meijer JL, Ardesch HG, van Rooijen JC, De Bruijn JHB. Low dose captopril twice daily lowers blood pressure without disturbance of the normal circadian rhythm. Postgrad Med J 1986,62:101-105.
32. Mancia G, Parati G, PomidossiG, Colombo A, Cuspidi C, Lattuada S, Antivalle M, Rindi M, Libretti A, Gianfranco B, Zanchetti A. Eualuation of the antihypertensive effect of once-a-day captoprii by 24-hour ambulatory blood pressure monitoring. J Hypertens 1987;S:suppl 5:S591 -S593. 33. de Faire U, Forslund L, OdenA. Variability of bloodpressure in ambulatory hypertensive patients: effects of verapamil on twice and thrice daily dose regimens. Acta Med Stand 1986;220:411-418.