| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Istituto Scientifico Ospedale
S. Luca, ABSTRACT Omboni, Stefano, Gianfranco Parati, Antonella Groppelli,
Luisa Ulian, and Giuseppe Mancia. Performance of the AM-5600 blood
pressure monitor: comparison with ambulatory intra-arterial pressure.
J. Appl. Physiol. 82(2): 698-703, 1997.
24-hour ambulatory blood pressure monitoring; noninvasive blood
pressure measurement; intra-arterial blood pressure measurement
INTRODUCTION SEVERAL STUDIES have shown that blood pressure values
obtained by noninvasive ambulatory blood pressure monitoring may have limited accuracy. To overcome this problem, standardized validation criteria have been proposed by different national committees (2, 14,
15, 19). However, with the partial exception of the Association for the
Advancement of Medical Instrumentation (AAMI) protocol (2, 19), these
criteria are based on the comparison between automatic and conventional
sphygmomanometric values obtained at rest, and no reference is made to
the performance of ambulatory blood pressure-monitoring devices in the
conditions for which they are designed, i.e., during daily life (12,
18). This is a serious omission, because even if accurate at rest, a
device may be inaccurate in monitoring ambulatory intra-arterial blood pressure (5, 10, 11, 16).
AM-5600 (9) is a novel noninvasive ambulatory blood pressure monitoring
device characterized by three potential advantages: 1) blood pressure recordings are
based on microphonic criteria that can be replaced by oscillometric
criteria if the microphone fails to identify reliable Korotkoff sounds;
2) the reliability of each blood
pressure measurement can be checked by plotting the recorded Korotkoff
sounds in parallel with the pressure changes within the arm cuff; and
3) blood pressure monitoring is
obtained simultaneously with ECG monitoring from two standard leads,
electrocardiogram (ECG) ST segment depression having the
ability to trigger blood pressure measurements.
The aim of the present study was to evaluate the accuracy of the
AM-5600 device, both at rest and in daily life conditions, by
comparison with simultaneous intra-arterial blood pressure recording.
METHODS
The AM-5600 is a new device that simultaneously monitors
electrocardiogram (ECG) and noninvasive blood pressure (BP) over a 24-h
period. BP readings (Korotkoff sounds and cuff air pressure) are stored into the recorder, allowing the removal of BP artifacts after a visual
check. In 12 subjects with essential hypertension, we compared BP
values simultaneously provided by the AM-5600 and intra-arterial
recordings. At rest, noninvasive systolic BP (SBP) values were lower
(5.4 ± 4.9 mmHg) and diastolic BP (DBP) values were higher (7.3 ± 7.3 mmHg) than were intra-arterial values. In ambulatory
conditions (9 subjects), between-method discrepancies were +0.8 ± 6.1 and +12.2 ± 7.4 mmHg for 24-h SBP and DBP, respectively. AM-5600 underestimated 24-h intra-arterial SBP and DBP SD, but it
accurately tracked intra-arterial SBP and DBP changes. Editing removed
22.1% of total readings, slightly reducing between-method discrepancies. Thus the AM-5600 provides an accurate average estimate of resting and ambulatory SBP and, for DBP, a less accurate estimate that is slightly improved by editing. The AM-5600 allows accurate description of SBP and DBP profiles and thus may be suitable to describe the abrupt BP changes accompanying a number of clinical events.
3 dB at 8-10 Hz; its
signal was linear between 50 and 250 mmHg when calibrated against a
mercury column before and after the 24-h recording period, with no
substantial drift of the zero signal. Further technical details are
reported in previous studies (3).
In each patient, the ambulatory blood pressure recording started at 1 PM, i.e., after collection of data at
rest. During the recording, the patient was free to move within the
hospital buildings and gardens and to engage in the social activities
of the hospital inpatients (watching TV, visiting with relatives,
playing cards). Each patient was asked only to comply with the hospital
meal and bedtimes and to remain with his/her arm still at the time of
each automatic blood pressure measurement. The patients were also asked to report in a diary the kind of activity they were engaged in at the
time of each measurement. An investigator visited the patient during
the recording period to check for possible problems with either the
noninvasive or the intra-arterial devices.
Data analysis of resting values.
The blood pressure measurements and the continuous ECG tracing stored
in the memory of the noninvasive recorder were sent to an AST Bravo
486/25 personal computer (AST Research, Irvine, CA) via a parallel
interface. As shown in Fig. 1,
top, for each blood pressure
measurement, Korotkoff sounds and cuff air pressure were simultaneously
displayed in parallel with an ECG lead. Because the display of these
noninvasive signals with the concomitant intra-arterial blood pressure
signal was not possible (because AM-5600 does not provide the real-time
output of Korotkoff sounds and cuff air pressure), the coupling of each
noninvasive blood pressure measurement with the simultaneous
intra-arterial value was made as follows.
1) The AM-5600 monitor was
programmed to inflate the cuff up to 200 mmHg, and the deflation rate
was set at 4 mmHg/s down to a pressure of 40 mmHg, at which point the air pump valve was fully opened to allow quick deflation to 0 mmHg.
2) The beginning and the end of the
cuff deflation were identified on the chart by an event marker. The
cuff air pressures corresponding to the first and the fifth Korotkoff
sounds were identified as systolic and diastolic blood pressure values
and compared with the corresponding intra-arterial values. The latter values were selected by 1) counting
the number of QRS complexes in the ECG tracing from the time when cuff
deflation started (event marker) to the time of appearance of Korotkoff
first and fifth sounds, respectively;
2) counting the corresponding number
of intra-arterial blood pressure waveforms from the initial event marker; and 3) averaging the
intra-arterial systolic and diastolic blood pressure values of the beat
selected with the values derived from the preceding and the following
beats.
The agreement between the AM-5600 values and the intra-arterial values was tested by calculating the between-method difference in each subject and by averaging mean individual differences for the group as a whole (4). Furthermore, we calculated the SD of the individual mean differences, the average of the individual SD for the group as a whole, and the 95% confidence limits of the mean between-method difference when all measurements obtained in the 12 subjects (n = 122) were pooled. Data analysis in ambulatory conditions. The 24-h intra-arterial signal recorded on the Oxford tape was replayed at 60 × the real time and displayed on an oscilloscope to check its quality and exclude possible artifacts (dampening of the signal, ectopic beats, and so on). The edited signal was sent to a Digital PDP 11/23 computer (Digital Equipment, Maynard, MA), sampled at 165 Hz, digitized on 12 bits, and stored on a magnetic disk. Systolic and diastolic blood pressures were computed over consecutive 3-s segments. These segments were used to calculate 24-h average values, hourly average values, and 24-h SD, the last being taken as an index of blood pressure variability. These values were compared with the corresponding values provided by the AM-5600 recorder. The comparison did not involve individual measurements, because during ambulatory monitoring it was not possible to synchronize the occurrence of each noninvasive measurement with the corresponding intra-arterial value, as done at rest. It should be emphasized that, even though we included in our analysis 9 rather than 15 subjects, as recommended by AAMI standards, comparisons of AM-5600 blood pressure values with intra-arterial blood pressure values over the 24 h in ambulatory conditions provided a number of data and a range of conditions greater than those required by the AAMI and other guidelines (2, 14, 15, 19) on validation of ambulatory blood pressure-monitoring devices. In addition, each blood pressure value provided by the AM-5600 recorder was displayed on the computer screen along with the corresponding Korotkoff sound and cuff air pressure. This allowed us to identify several blood pressure readings originating from artifactual Korotkoff sounds (Fig. 1, middle) that had not been recognized and automatically replaced by oscillometric criteria (Fig. 1, bottom). These readings (22.1 ± 9.5% of total 24-h readings, 20.1 ± 9.0% of daytime readings, and 2.0 ± 1.8% of nighttime readings) were removed, and comparison with the 24-h and hourly average intra-arterial values was thus made also after full editing of noninvasive values. Ambulatory blood pressure data from individual subjects were averaged to obtain mean values for the group as a whole. The statistical significance of the differences between the AM-5600 and intra-arterial data was assessed by a paired Student's t-test (with Bonferroni correction when multiple comparisons were performed). The Pearson
2 test of independence was used
to assess the degree of concordance or discordance in the hour-to-hour
blood pressure changes detected by the intra-arterial and noninvasive
methods. These hour-to-hour changes were quantified by computing the
difference between the average systolic and diastolic blood pressure
values of 1 h of the recording and those derived from the preceding
one. The relationship between the directional changes of the values
provided by the two methods was quantified by computing the
2 (2) and the contingency (C)
coefficients. The 2 test is based on the assumption that
the sets of data obtained by two methods are independent, a
statistically significant 2 rejecting this assumption
and thus proving dependency. C is taken as an index of correlation
between data obtained by the two methods (11). Results were always
shown as mean values ± SD, unless differently indicated.
P < 0.05 was set as the minimum
level of statistical significance. All computations were carried out by a commercially available statistical package (SPSS/PC+, SPSS, Cary, NC;
Ref. 13).
RESULTS
5.4 and +7.9 mmHg for systolic and diastolic blood
pressure, respectively (P < 0.01 for
both). The means of the corresponding individual SD were 3.5 and 3.6 mmHg, respectively.
|
|||||||||||||||||||||||||||||||||||||||||||||||||||
) between AM-5600 and
intra-arterial systolic blood pressure (SBP,
left) and diastolic blood pressure
(DBP, right) values at rest. Points
refer to the between-method discrepancy for each blood pressure
measurement in 12 subjects (total, n = 122). Dashed and continuous lines refer to average discrepancies and to
their 95% confidence intervals, respectively. Individual discrepancies
are plotted vs. corresponding intra-arterial blood pressures.
Average 24-h and hourly blood pressures. As shown in Table 2, before editing of artifacts, the average 24-h blood pressure values obtained by AM-5600 and intra-arterially showed variable discrepancies between subjects. The mean discrepancy in the group as a whole was 0.8 mmHg (not significant) for systolic and 12.2 mmHg (P < 0.01) for diastolic blood pressure. Thus, in ambulatory conditions, the overall systolic blood pressure discrepancy was less than at rest, whereas the average diastolic blood pressure discrepancy was greater than at rest. After removal of artifacts, the 24-h between-method discrepancies were reduced, but the reductions were small and not statistically significant.
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
) and
noninvasively (
). Noninvasive hourly averages were computed after editing of blood pressure artifacts. Data are shown as hourly averages ± SE from 9 subjects. In each panel, larger symbols on right refer to 24-h mean values.
** Statistically significant difference between 24-h mean values;
P < 0.01.
Blood pressure variability. As shown in Fig. 4, the individual 24-h SD of systolic and diastolic blood pressure were significantly lower when estimated by the AM-5600 than intra-arterially (P < 0.01). A further nonsignificant reduction was observed after editing. Thus, noninvasive ambulatory blood pressure recording underestimates the magnitude of overall 24-h blood pressure variations. However, when blood pressure variations were assessed as changes between hourly averages, the results obtained by the intra-arterial method and by the AM-5600 (edited data) were similar (Fig. 5).
,
IA) and noninvasive (, AM-5600) recordings. The latter were obtained
before and after editing of artifacts. 
, Average data for group as a
whole. ** Statistically significant difference between different
sets of data; P < 0.01.
)
and noninvasive () SBP and DBP values. Noninvasive data are
shown after removal of blood pressure artifacts.
Bottom: results of
2 test.
DISCUSSION
Our study shows that the AM-5600 device underestimates at-rest intra-arterial systolic blood pressure and overestimates intra-arterial diastolic blood pressure by ~5 and 7 mmHg, respectively. These under- and overestimations may be due to a limited accuracy of the AM-5600 device. However, similar discrepancies have been reported when resting blood pressures taken by a mercury sphygmomanometer have been compared with intra-arterial values (8). Thus, it is likely that the results we obtained are accounted for by inherent differences between blood pressures measured indirectly and directly and that the overall performance of the AM-5600 with subjects at rest (i.e., measuring device plus algorithm) is at least as satisfactory as that provided by sphygmomanometry.
The main result of our study, however, is that the performance of the AM-5600 was different in ambulatory conditions compared with at rest. The difference was that, although with noticeable individual variations, the average discrepancy between AM-5600 and intra-arterial systolic blood pressure was less during the 24-h period than at rest. On the contrary, the average discrepancy between AM-5600 and intra-arterial diastolic blood pressure was greater during the 24-h period than at rest. This provides further evidence that because of inadequacy of the measuring device and/or algorithms used to identify blood pressure values (5, 10, 11, 16), testing a noninvasive blood pressure-monitoring device at rest may not predict the performance of the device in daily life conditions. The adequacy of the device and algorithms needs to be addressed by comparison with simultaneous intra-arterial blood pressure monitoring, which thus is a necessary step in noninvasive ambulatory blood pressure-monitoring device testing. All the above observations hold true for average group values. It should be emphasized, however, that the between-method discrepancies in systolic and diastolic blood pressure were characterized by a large interindividual variability, which requires conclusions on the accuracy of the AM-5600 device in any individual subject to be drawn with caution.
Our observations do not allow us to reach a definite conclusion as to whether the AM-5600 is better or worse than other noninvasive devices, because this would require comparing the devices in the same subjects. However, several favorable features of the AM-5600 should be emphasized. 1) Compared with other devices we tested, AM-5600 provides an estimate of 24-h average systolic blood pressure close to the intra-arterial value (5, 11). 2) A close correspondence between AM-5600 and intra-arterial systolic blood pressure was evident not only for the 24-h average but also for almost each hour of the 24-h time. 3) The hour-to-hour systolic and diastolic blood pressure changes provided by AM-5600 were nearly always qualitatively and quantitatively similar to the changes occurring intra-arterially. Finally, the AM-5600 editing procedure made it possible to reduce, although to only a limited extent, the discrepancy with the intra-arterial values.
Despite the ability of AM-5600 to reflect accurately hour-to-hour blood pressure changes, the estimate of overall 24-h blood pressure variability by AM-5600 was different from the real variability value derived from intra-arterial recording. For both systolic and diastolic blood pressure, the difference consisted in an underestimation of the real variability phenomena. The underestimation was a substantial one, because the AM-5600 systolic and diastolic 24-h SD were ~20-30% less than the corresponding intra-arterial ones. This expands our previous results (7) that showed 1) actual blood pressure variability may be inaccurately quantified by intermittent blood pressure sampling, and 2) this is particularly the case if the between-sample interval is too long, presumably because a long interval without sampling prevents short-term blood pressure changes from being detected. These observations strengthen the conclusion that estimates of blood pressure variability obtained noninvasively should be interpreted with caution.
Our data permit three further considerations. 1) In our study, intra-arterial blood pressure was taken as the gold standard, both at rest and in ambulatory conditions. However, some error in intra-arterial blood pressure values might have occurred, particularly in ambulatory conditions because of movement artifacts, dampening of the intra-arterial signal caused by blood clotting, small air bubbles in the catheter-transducing system, and drift of the zero signal with time. Thus, although the accuracy of intra-arterial blood pressure measurements was always carefully checked, it is not possible to exclude the possibility that, at least to some degree, the between-method discrepancy was not accounted for by the AM-5600. 2) Our conclusion regarding AM-5600 applies to the type of patients we studied, and the performance of the device in other categories of patients (e.g., subjects with severe hypertension or hypotension, elderly subjects, and obese subjects) has to be separately tested. This need for expanded testing is supported by our observations that the AM-5600 intra-arterial systolic blood pressure discrepancy at rest showed a tendency to change with the increasing baseline blood pressure value.
Finally, the ability of AM-5600 to closely reflect changes in hourly systolic and diastolic blood pressure implies that this device is capable of accurately detecting the blood pressure increases and reductions that may occur during a 24-h time period and that can be associated with important clinical events, such as angina pectoris and silent myocardial ischemia (2, 6, 17).
FOOTNOTES
Address for reprint requests: S. Omboni, Centro di Fisiologia Clinica e Ipertensione, Ospedale Maggiore, Via F. Sforza 35, 20122 Milano, Italy (E-mail: DARKSIDE{at}ZEROCITY.IT).
Received 29 March 1996; accepted in final form 26 September 1996.
REFERENCES
| 1. | Asmar, R., A. Benetos, B. Pannier, E. Agnes, J. Topouchian, B. Laloux, and M. Safar. Prevalence and circadian variations of ST-segment depression and its concomitant blood pressure changes in asymptomatic systemic hypertension. Am. J. Cardiol. 77: 384-390, 1996. [Medline] |
| 2. | Association for the Advancement of Medical Instrumentation American National Standard for Electronic or Automated Sphygmomanometers: ANSI/AAMI SP 10-1992. Arlington, VA: Association for the Advancement of Medical Instrumentation, 1993, p. 40. |
| 3. | Bevan, A. T., A. J. Honour, and F. D. Stott. Direct arterial pressure recording in unrestricted man. Clin. Sci. 36: 329-344, 1969. [Medline] |
| 4. | Bland, J. M., and D. G. Altman. Statistical methods for assessing agreement between two methods of clinical measurements. Lancet 1: 307-310, 1986. [Medline] |
| 5. | Casadei, R., G. Parati, G. Pomidossi, A. Groppelli, S. Trazzi, M. Di Rienzo, and G. Mancia. 24-hour blood pressure monitoring: evaluation of Spacelabs 5300 monitor by comparison with intra-arterial blood pressure recording in ambulant subjects. J. Hypertens. 6: 797-803, 1988. [Medline] |
| 6. |
Davies, A. B.,
V. Bala Subramanian,
P. M. M. Cashman,
and
E. B. Raftery.
Simultaneous recording of continuous arterial pressure, heart rate, and ST segment in ambulant patients with stable angina pectoris.
Br. Heart. J.
50:
85-91,
1983.
|
| 7. |
Di Rienzo, M.,
G. Grassi,
A. Pedotti,
and
G. Mancia.
Continuous vs intermittent blood pressure measurements in estimating 24-hour average blood pressure.
Hypertension
5:
264-269,
1983.
|
| 8. | Geddes, L. A. The Direct and Indirect Measurement of Blood Pressure. Chicago, IL: Yearbook Medical Publishers, 1970. |
| 9. | Germano, G., A. Caparra, S. Valentino, F. Coia, L. Federico, and A. Santucci. A validation of the data obtained with the simultaneous recording of blood pressure and the 24 hour electrocardiogram. (In Italian.) Cardiologia 38: 383-392, 1993. [Medline] |
| 10. | Gould, B. A., R. S. Hornung, P. M. M. Cashman, D. Altman, and E. B. Raftery. An evaluation of the Avionic's Pressurometer III 1978 at home and in the hospital. Clin. Cardiol. 9: 335-343, 1986. [Medline] |
| 11. |
Groppelli, A.,
S. Omboni,
G. Parati,
and
G. Mancia.
Evaluation of noninvasive blood pressure monitoring devices Spacelabs 90202 and 90207 versus resting and ambulatory 24-hour intra-arterial blood pressure.
Hypertension
20:
227-232,
1992.
|
| 12. | Mancia, G., and G. Parati. Commentary on the revised British Hypertension Society protocol for evaluation of blood pressure measuring devices: a critique of aspects related to 24-hour ambulatory blood pressure measurement. J. Hypertens. 11: 595-597, 1993. [Medline] |
| 13. | Norusis, M. J. SPSS/PC+ v2.0 Base Manual. Cary, NC: SPSS, 1990. |
| 14. | O'Brien, E., J. Petrie, W. Littler, M. De Swiet, P. L. Padfield, D. G. Altman, M. Bland, A. Coats, and N. Atkins. Short report: an outline of the revised British Hypertension Society protocol for the evaluation of blood pressure measuring devices. J. Hypertens. 11: 677-679, 1993. [Medline] |
| 15. | O'Brien, E., J. Petrie, W. Littler, M. De Swiet, P. L. Padfield, K. O'Malley, M. Jamieson, D. Altman, M. Bland, and N. Atkins. The British Hypertension Society protocol for the evaluation of automated and semi-automated blood pressure measuring devices with special reference to ambulatory systems. J. Hypertens. 8: 607-619, 1990. [Medline] |
| 16. | Parati, G., A. Groppelli, S. Omboni, S. Trazzi, A. Frattola, and G. Mancia. Testing the accuracy of blood pressure monitoring devices in ambulatory conditions. J. Hypertens. 9, Suppl. 8: S7-S11, 1991. |
| 17. | Trenkwalder, P., R. Dobrindt, M. Plaschke, and M. Lydtin. Usefulness of simultaneous ambulatory electrocardiographic and blood pressure monitoring in detecting myocardial ischemia in patients >70 years of age with systemic hypertension. Am. J. Cardiol. 72: 927-931, 1993. [Medline] |
| 18. | White, W. B. The assessment of portable, non-invasive blood pressure recorders. J. Hypertens. 8: 591-593, 1990. [Medline] |
| 19. |
White, W. B.,
A. S. Berson,
C. Robbins,
M. J. Jamieson,
L. M. Prisant,
E. Roccella,
and
S. G. Sheps.
National standard for measurement of resting and ambulatory blood pressures with automated sphygmomanometers.
Hypertension
21:
504-509,
1993.
|
This article has been cited by other articles:
|
|
 |
|
  E. Umana, W. Ahmed, M. A. Fraley, and M. A. Alpert Comparison of Oscillometric and Intraarterial Systolic and Diastolic Blood Pressures in Lean, Overweight, and Obese Patients Angiology, January 1, 2006; 57(1): 41 - 45. [Abstract] [PDF]
|
|
|
|
 |
|
 E. O'Brien, B. Waeber, G. Parati, J. Staessen, and M. G Myers Blood pressure measuring devices: recommendations of the European Society of Hypertension BMJ, March 3, 2001; 322(7285): 531 - 536. [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
ABSTRACT