Skip Navigation


NDT Advance Access originally published online on June 7, 2007
Nephrology Dialysis Transplantation 2007 22(10):2950-2961; doi:10.1093/ndt/gfm288
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
22/10/2950    most recent
gfm288v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Lodi, C. A.
Right arrow Articles by Ghidini, C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Lodi, C. A.
Right arrow Articles by Ghidini, C.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author [2007]. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org


In vitro and in vivo evaluation of an oscillometric device for monitoring blood pressure in dialysis patients

Carlo Alberto Lodi1, Charles Estridge2 and Carlo Ghidini3

1R&D, Gambro Dasco S.p.A, Medolla (MO), Italy, 2Department of Nephrology, Toronto General Hospital, Toronto, Canada and 3Lean S.r.l., Medolla (MO), Italy

Correspondence and offprint requests to: Carlo Alberto Lodi, R&D, Gambro Dasco S.p.A., Via Modenese, 66, I – 41036 Medolla, (MO), Italy. Email: carlo.lodi{at}gambro.com



   Abstract
Top
 Abstract
Introduction
 Subjects and methods
 Results
 Discussion
 Acknowledgements
 References
 
Background. Nowadays, an increasing number of dialysis machines lodge a blood pressure (BP) measuring device, whose accuracy has a clear implication for the patients’ clinical management.

Methods. An automated oscillometric sphygmomanometer (HD-BPMTM by Gambro Dasco) used during haemodialysis was submitted to both in vitro and in vivo tests, in order to evaluate some modifications aimed at improving measurement accuracy and consistency. The results were compared with those obtained by another oscillometric monitor (BX-100TM by Colin).

Three steps of evaluation were followed. First, the maintenance of the overall accuracy requirements prescribed by ANSI/AAMI SP-10 standard was verified. Then, an in vitro validation was carried out by using a test simulator. Finally, during a multi-centre field trial, 392 BP measurement sessions on 53 dialysis patients were collected. Every session consisted of two consecutive intra-dialysis measurements by the oscillometric monitors, each one performed simultaneously to an auscultatory measurement. A comparison with the intra-arterial method was performed as well.

Results. When compared with an in vivo data set previously collected, the HD-BPM accuracy complied with required limits. Second, the internal repeatability with respect to the simulator was satisfactory (SD of the differences between device and simulator readings: HD-BPM: systolic = 5.7, diastolic = 4.2; BX-100: systolic = 4.2, diastolic = 5.5 mmHg). Moreover, the comparison between oscillometric and auscultatory methods during in vivo trial gave similar results for the two monitors, even if systolic pressure SD exceeds the limit recommended by ANSI/AAMI SP-10 (mean value of the differences ± SD: HD-BPM: systolic = 0.5 ± 9.0, diastolic = 1.5 ± 6.9; BX-100: systolic = 3.1 ± 8.2, diastolic = –2.0 ± 7.6 mmHg).

Conclusions. These data underline the importance of performing accuracy evaluations for BP monitors in the conditions where they normally work, by using well-accepted protocols.

Keywords: arrhythmias; blood pressure; dialysis patients; mercury sphygmomanometer; oscillometric device; validation



   Introduction
Top
 Abstract
 Introduction
Subjects and methods
 Results
 Discussion
 Acknowledgements
 References
 
Blood pressure (BP) measurement in haemodialysis (HD) patients is of paramount importance for monitoring and preventing hypertensive or hypotensive events. It is well known, in fact, that hypertension and consequent cardiovascular disease in chronic HD patients contributes significantly to their morbidity and mortality [1,2]. Second, arterial BP hypotension during and after treatment represents a major complication of HD [3,4]. However, the relationship between BP and mortality in HD patients remains a major unresolved question [5,6].

BP in clinical settings is usually measured non-invasively by sphygmomanometers, which can be either manually operated or automatic. Among the latter, oscillometric devices are becoming more and more popular in dialysis units, due to their feasibility and ease of use [7,8]. Most of them are fully integrated into modern dialysis equipment, in such a way that the operation of pressure measurement has become a completely automated action into the task of nursing personnel.

An accurate BP measurement is difficult to be obtained in any subject for several reasons ([9] for a review), but further complicating factors intervene during a dialysis treatment (e.g. [10,11]). First of all, there are elements strictly linked to the HD patient conditions, such as the large BP variations caused by plasma ultrafiltration and extra-cellular fluid volume displacement. Secondly, some drawbacks of the oscillometric method become more serious in HD setting. In fact, it is known that the increased rigidity in the arterial wall due to aging and diabetes significantly affects oscillometric measure (e.g. the mathematical study by Ursino and Cristalli [12] and the clinical study by van Popele et al. [13]); obesity causes the oscillometric pulses to get weaker and weaker; finally, cardiac arrhythmias, in particular atrial fibrillation, afflict both auscultatory and oscillometric device accuracy [14]. All the above cited pathologies are more and more common in the dialysis population. Thirdly, there are problems linked to the correct handling of the operator. Often, according to our experience as well, accepted guidelines on the correct methodology of BP measurement [15] become difficult to follow in a busy dialysis unit. Hence, it can happen that an incorrect cuff size is used [16]; sometimes the cuff is wrapped in a wrong position on the patient's limb or the clothes between the skin and cuff are too thick. Moreover, single and not averaged BP measures are usually assessed. Furthermore, it is not uncommon that BP reading is negatively affected by artefact movement and talking during measurement, or by a recent meal, ingestion of caffeine or smoking. In this regard, Rahman et al. [17] compared BP readings obtained during usual HD treatment with those obtained according to standard guidelines for BP measurement [15] and found a poor range of agreement (i.e. –20 to +49 mmHg), which is much greater than what would be acceptable for accuracy in the clinical setting.

It is worth noting that a device deputed to measure BP in dialysis patients should necessarily cope with all the listed issues. To verify the efficacy of this device, it should be important to perform an adequate analysis of the accuracy, consistency and repeatability of the BP measure according to well accepted and standardized protocols (e.g. [18,19]).

The aim of this work is to describe the validation results of the oscillometric sphygmomanometer Gambro Dasco HD-BPMTM (software revision 5.x; Gambro Dasco S.p.A., Medolla, Italy) after the computation algorithm was recently revised in order to improve its performance in a dialysis unit.

The validation process applied can be subdivided into three steps. First of all, the compliance to the requirements of overall accuracy prescribed by ANSI/AAMI SP-10 standard [18] was verified. Second, an in vitro validation session was performed: a wide variety of BP waveforms generated by a non-invasive BP (NIBP) test simulator [20] was applied to HD-BPM, according to a protocol proposed in literature [21]. Third, a multi-centre field trial was organized and the readings of BP measurements in patients during dialysis were collected. Moreover, during the last two phases, the results of HD-BPM were compared with the ones obtained in the same experimental and clinical conditions by another commercially available device, Colin BX-100TM (Colin Corporation, Aichi, Japan), which is declared to comply with ANSI/AAMI SP-10 requirements as well. This same model or analogous ones are currently used both as stand-alone and as accessory in some dialysis machines, such as Gambro AK 200TM (Gambro AB, Lund, Sweden), Fresenius 4008HTM (Fresenius Medical Care, Bad Homburg, Germany) and B. Braun DialogTM (B. Braun Medical Inc., Bethlehem, USA).



   Subjects and methods
Top
 Abstract
 Introduction
 Subjects and methods
Results
 Discussion
 Acknowledgements
 References
 
The automatic NIBP monitor device HD-BPM is an optional item of dialysis equipment Hospal IntegraTM and Gambro PhoenixTM. The HD-BPM estimates systolic (SBP) and diastolic blood pressures (DBP) together with pulse rate by using the oscillometric method. For a detailed description of the HD-BPM device, refer to [22]. Recently, also due to multiple requests coming from the field, the computation algorithm of HD-BPM was revised in terms of identification of the oscillometric curve and robustness against noise artefact.

HD-BPM was analysed together with another oscillometric NIBP monitor device, namely BX-100 by Colin. The stand-alone version equipped with its own user interface was used for this study. This evaluation was performed according to the following three consecutive steps.

Step I: overall accuracy according to ANSI/AAMI standard
An initial version of HD-BPM was positively validated according to ANSI/AAMI SP-10 standard [18], by comparing HD-BPM readings in 92 subjects with BP values found by two independent observers using mercury sphygmomanometer (for the detailed description of the test protocol and results, refer to [22]). Four subjects among them underwent chronic HD. For each subject, three consecutive determinations were performed, in order to have 276 (92 x 3) measurements to be statistically analysed.

Instead of repeating the in vivo assessment, a sort of off-line ANSI/AAMI validation was performed for the new revision of HD-BPM, due to the fact that the 276 cuff pressure waveforms of the preceding validation were stored into an electronic database. The same waveforms were then given as an input to the new version of HD-BPM algorithm running on a PC. This operation was possible since the modifications applied on HD-BPM affected only the parts of the algorithm dealing with the data elaboration. Hence, the results obtained were compared with the readings of the preceding auscultatory measurements, in order to determine the new accuracy values.

Step II: in vitro validation using an NIBP test simulator
For in vitro and in vivo tests a stand-alone version of HD-BPM was used: it includes the same electronic board currently integrated into the dialysis equipment and a dedicated PC software program of user interface. Before the tests and at regular time basis, HD-BPM was statically calibrated against a reference pressure transducer (90DX WesternTM Meter by Mesa Laboratories, Inc., Lakewood, USA).

The functional performances of HD-BPM and BX-100 monitors were first assessed in vitro by using the NIBP test simulator Bio-Tek BP-PumpTM (Bio-Tek Instruments Inc., Winooski, USA). An evaluation protocol first proposed in [21] was adopted: in particular, the part dealing with simulations of adult pressure waveforms was used. The protocol (Table 1) covered a wide range of simulated pressures, pulse rates, pulse strengths, motion and tremor artefacts and specific subject conditions (e.g. weak pulse, geriatric and obese subject, cardiac disorders).


View this table:
[in this window]
[in a new window]

  		Table 1. Test protocol of in vitro validation

 
Generally, five successive BP measures were performed for each type of test at 2-min intervals. Two quantities were computed to evaluate accuracy and repeatability of the NIBP monitors: the mean value (bias, Formula ) and the SD of the differences between the monitor's reading and the target pressure set on the simulator.

Step III: multi-centre field trial
Between November 2003 and August 2004, we studied 53 end-stage renal disease (ESRD) patients treated with thrice-a-week HD. Four dialysis centres were involved: Danville Dialysis Service in Danville, IL, USA (four patients); Hôpital Régional de Sion-Hérens-Conthey in Sion, Switzerland (seven patients); Toronto General Hospital in Toronto, Ontario, Canada (28 patients); Hôpital St-Joseph in Trois-Rivières, Québec, Canada (14 patients). Table 2 reports patient (23) characteristics for each centre. An arterio-venous fistula was used as vascular access in all patients. Twenty-six patients were hypertensive and six hypotensive. Subjects with atrial fibrillation were excluded.


View this table:
[in this window]
[in a new window]

  		Table 2. Patient characteristics

 
During the field trial, HD-BPM, BX-100 and a mercury sphygmomanometer were connected with the same pneumatic tube, through a couple of T connectors. By properly clamping the tube, it was possible to choose which of the two oscillometric monitors were to be used together with the mercury sphygmomanometer. At this proposal, it is worth noting that both HD-BPM and BX-100 deflate the cuff linearly and with a deflation rate compatible with the limits suggested by ANSI/AAMI SP-10 for simultaneous manual auscultation (i.e. 2–4 mmHg per heartbeat [18]). The same cuff was used for all measurements on any particular patient with both NIBP devices. A total of 392 measurement sessions were executed during dialysis treatments. Each session was composed of two pressure measurements: the first one through the stand-alone HD-BPM monitor simultaneously with an auscultatory measurement, the second by means of BX-100 monitor simultaneously with another auscultatory measurement. The operator performing the auscultatory measure was not the same for every session. The arm was placed at heart level during all measurements and care was taken to use an appropriate cuff size, by previously measuring the upper arm circumference [15]. Generally, four measurement sessions were performed every dialysis treatment, over a time interval of 1 h, starting about 30 min from the beginning of the treatment. Statistical evaluation of the differences between measurements by either of the two oscillometric devices and the respective auscultatory measurement was performed by using the Bland–Altman procedure [23]. The differences between automatic measurements (with either device) and observers’ measurements were plotted against the mean BP by the two methods (oscillometry and auscultatory). The SD of the differences between the two simultaneous readings provided a measure of the intra-method variability.

The in vivo validation was extended in December 2006 by comparing BP readings of the two oscillatory devices with invasive intra-arterial measurements performed in an intensive care unit (ICU). In particular, 11 patients admitted at ICU of Malpighi Hospital, Bologna, Italy, were studied (all males, age: 65 ± 17 years). Sixteen measurement sessions were performed: in each of them, the invasive measure was executed through a catheter inserted into the brachial artery (InfinityTM by Draeger Medical Systems Inc., Danvers, MA, USA) simultaneously to the oscillometric one on the same limb. In five patients, the session was repeated twice.



   Results
Top
 Abstract
 Introduction
 Subjects and methods
 Results
Discussion
 Acknowledgements
 References
 
Step I: overall accuracy according to ANSI/AAMI standard
The mean and the SD of the differences between the auscultatory measurements collected during ANSI/AAMI validation and HD-BPM SW rev. 5.x pressure estimates resulted in 0.9 ± 7.6 mmHg for SBP and –0.2 ± 5.4 mmHg for DBP. Fifty-seven percent of the differences in SBP and 75% of the differences in DBP were <5 mmHg; 28% of SBP and 20% of DBP were between 5 and 10 mmHg; 12% of SBP and 4% of DBP were between 10 and 15 mmHg; and 4% of SBP and 1% of DBP were >15 mmHg. The agreement between auscultatory and oscillometric measurements were studied also according to the Bland-Altman statistical method: 5% of SBP and 5% of DBP measurements fell outside the limits of agreements, defined as Formula ± 2 x SD (Figure 1). The accuracy of the measurement by the oscillometric method was not affected by the level of either SBP or DBP.


Figure 1
View larger version (19K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 1. Bland–Altman scatter plots of the differences between systolic (SBP, left) and diastolic (DBP, right) blood pressure measured by oscillometric device HD-BPM SW rev. 5.x and the corresponding ones by auscultatory method measured during preceding ANSI/AAMI validation study. Bias (Figure 1) and SD are mean value and standard deviation of the differences between methods. Limits of agreement: Figure 1 (dotted line); Figure 1 ± SD (dash–dotted line); Figure 1 ± 2SD (dashed line).

 
Hence, it can be concluded that the new version of the HD-BPM satisfies the requirements prescribed by ANSI/AAMI SP-10 standard (in particular, bias ± SD not >5 ± 8 mmHg).

Step II: in vitro validation using a NIBP test simulator
Figures 2–4GoGo depict most of the results obtained during the performance assessment of HD-BPM and BX-100 monitors by using Bio-Tek NIBP test simulator.


Figure 2
View larger version (21K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 2. (A) Mean value (points) and SD (bars) of the differences between oscillometric device and non-invasive blood pressure (NIBP) test simulator readings for different couples of SBP and DBP values (refer to ‘Pressure range’ line in Table 1). (B) Mean value (points) and SD (bars) of the differences between oscillometric device and NIBP test simulator SBP and DBP readings for different pulse rate values (refer to ‘Pulse rate range’ line in Table 1). Left column: HD-BPM SW rev. 5.x monitor; right column: BX-100 monitor. Bias is mean value of the differences between monitor and test simulator readings.

 

Figure 3
View larger version (30K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 3. (A) Bias (histograms) and SD (bars) of the differences between oscillometric device and NIBP test simulator SBP and DBP readings for ‘Various physio-pathological situations’ (refer to corresponding line in Table 1 for the characteristics of the tests used). (B) SD (histograms) of the differences between oscillometric device and NIBP test simulator SBP, DBP and pulse rate readings for different ‘Cardiac disorders and respiratory artefacts’ (refer to corresponding line in Table 1 for the characteristics of the tests used). Left column: HD-BPM SW rev. 5.x monitor; right column: BX-100 monitor.

 

Figure 4
View larger version (16K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 4. SD (histograms) of the differences between oscillometric device and NIBP test simulator SBP and DBP readings for different ‘Oscillometric pulse strengths’ (refer to corresponding line in Table 1 for the characteristics of the tests used). Left column: HD-BPM monitor; right column: BX-100 monitor.

 
Figure 2 reports bias and SD of the differences for SBP and DBP readings over a range of pressures from 60/30 to 255/195 mmHg (Figure 2A) and for pulse rates from 40 to 200 bpm (Figure 2B). Bias related to different pressures was worse for HD-BPM than for BX-100, especially for high BP values, while bias related to different pulse rates were the same for the two devices; SD of HD-BPM is always lower than BX-100 (mean values of SD were 0.6 and 1.6 mmHg, respectively).

In Figure 3A, the performance of the monitors for waveforms simulating particular physio-pathological situations is shown. While bias of HD-BPM was generally worse than that of BX-100 (bias mean value: 3.6 and –0.2 mmHg), SD was better (SD mean value: 0.5 and 1.6 mmHg, respectively). Figure 3B deals with cardiac disorders and respiratory artefacts: due to the fact that for these waveforms BP generated by test simulator exhibits a beat-to-beat variation, only SD was plotted. Also in this case, SD of HD-BPM was better than that of BX-100, especially for the condition of premature ventricular contraction.

Figure 4 deals with the ability of the monitors to cope with different amplitudes of pulse. Pulse strength is varied from 100 to 10% of the nominal 2 mmHg oscillometric pulse. HD-BPM had difficulties in measuring BP with a 25% of pulse reduction, while BX-100 succeeded; when the reduction is 10%, HD-BPM failed, but BX-100 commits significant errors.

Finally, Table 3 summarizes the results obtained by executing the NIBP simulator tests listed in Table 1: mean value, SD, maximum and minimum of the differences between the readings of the analysed devices and simulator are reported. While SBP and DBP biases of HD-BPM were worse than those of BX-100, the SD is of the same amount. The NIBP test simulator makes it possible to evaluate the pulse rate estimation as well: the results were not significantly different for the two oscillometric devices.


View this table:
[in this window]
[in a new window]

  		Table 3. Summary of the results with simulator Bio-Tek by applying the test protocol reported in Table 1 (309 measurements)

 
Step III: multi-centre field trial
Based on the auscultatory determinations among observed HD patients, SBP and DBP ranged, respectively, from 68 to 220 mmHg and from 38 and 120 mmHg. Mean values of the first and second auscultatory measure were SBP = 132.2 ± 30.1; DBP = 71.0 ± 15.0 and SBP = 132.0 ± 30.6; DBP = 71.0 ± 15.1, respectively: based on Student's t-test for paired samples, no significant differences existed between the two measurement values (P > 0.70). The mean SBP of the two auscultatory measures was >180 mmHg in 8% of cases and <100 mmHg in 13% of cases. The mean DBP of the two auscultatory measures was >100 mmHg in 4% of cases and <60 mmHg in 18% of cases.

Table 4 summarizes the results obtained during the multi-centre field trial. As it can be seen from this table, HD-BPM comparison and BX-100 comparison failed in five and seven sessions, respectively out of 392 because either auscultatory measure or oscillometric measure could not be taken.


View this table:
[in this window]
[in a new window]

  		Table 4. Summary of the results of the field trial: auscultation as a reference

 
Biases were better for HD-BPM than for BX-100; regarding SD, BX-100 was better for systolic, and the opposite was valid for diastolic pressure. High correlation (r2 > 0.9, P < 0.001) existed between systolic pressures of oscillometric and auscultatory techniques, while correlation was a little weaker for diastolic pressures (r2 {cong} 0.8, P < 0.001). According to Student's t-test for paired samples, the oscillometric readings were not significantly different with respect to the corresponding auscultatory ones only for HD-BPM SBP (P > 0.35); in all the other cases, a statistical difference existed (P < 0.001). Table 4 also reports the same data regarding the comparison between the two consecutive auscultatory measures (i.e. AUS1 and AUS2), which gave a measure of the intra-subject BP short-term variability. Bias is very low both for SBP and for DBP; SD for SBP was comparable to that of the oscillometric devices, while the one for DBP was lower.

Figure 5 shows the distribution of the differences between the two methods: in brief, as regards HD-BPM, the percentages of the differences within 10 mmHg were 80% for systolic and 89% for diastolic; regarding BX-100, these percentages were 83 and 84%, respectively.


Figure 5
View larger version (21K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 5. Distribution of absolute differences between oscillometric devices and auscultatory BP determinations.

 
The degree of agreement between the oscillometric and auscultatory measures is represented in Figure 6. The limits of agreement, calculated according to Bland and Altman [23] as Formula ± 2 x SD, and the percentages of the differences falling outside them were the following: –17.6 and 18.4 mmHg (HD-BPM SBP, 4.9% outside); –13.1 and 16.1 mmHg (HD-BPM DBP, 2.8% outside); –11.9 and 18.0 mmHg (BX-100 SBP, 7.8% outside); and –17.5 and 13.6 mmHg (BX-100 DBP, 5.2% outside). No clear trends existed for the errors with respect to the averages of BP measures.


Figure 6
View larger version (38K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 6. Bland–Altman scatter plots of the differences between blood pressure measured by oscillometric devices (SBP, upper panels and DBP, lower panels) and the corresponding ones measured by auscultatory method, during field trial. Left column: HD-BPM monitor; right column: BX-100 monitor. Limits of agreement: Figure 6 (dotted line); Figure 6 ± SD (dash–dotted line); Figure 6 ± 2SD (dashed line).

 
Table 5 reports the results of the comparison between oscillometric and intra-arterial measures. Intra-arterial readings were corrected with a constant offset to have consistency with auscultatory measures, according to [18]; their ranges were 82–153 mmHg for SBP and 49–88 mmHg for DBP; mean values of the first and second measure were SBP = 119.9 ± 24.1; 118.8 ± 23.6 and DBP = 64.1 ± 11.4; 64.1 ± 12.1, respectively. Regarding SBP differences between the two methods, while biases of both NIBP devices are limited, SDs overcome the ANSI/AAMI limit with HD-BPM being the worse. The opposite happens for DBP: biases are greater, SD of HD-BPM is better than BX-100 and well below 8 mmHg.


View this table:
[in this window]
[in a new window]

  		Table 5. Summary of the results of the field trial: intra-arterial (INV) as a reference

 


   Discussion
Top
 Abstract
 Introduction
 Subjects and methods
 Results
 Discussion
Acknowledgements
 References
 
BP measurement is important during HD sessions to monitor the status of the cardiovascular system and to prevent acute hypotension or hypertensive events. Hence, the accuracy of the instruments measuring BP has a large impact for the clinical management of the dialysis patients. Modern dialysis machines often include an integrated device, which measures BP and pulse rate in an automatic way, typically by using the oscillometric principle [7]. To assure the accuracy of such devices, they would need to be validated according to their intended use.

This article presented the results obtained applying an in vitro and in vivo validation protocol aimed at functionally characterizing an automated oscillometric NIBP monitor (HD-BPM by Gambro Dasco), whose algorithm was recently modified to have more accurate and reliable estimates during the dialysis session. The readings obtained with this oscillometric monitor were always compared with simultaneous measurements taken first with mercury sphygmomanometer, which is still considered the gold standard of BP monitoring, not only in dialysis centres [9,18,24], and then with invasive intra-arterial measurements. In order to have a further term of comparison, another automated BP device based on the same principle (BX-100 by Colin) was analysed.

Many different examples of BP monitoring device validation and/or comparison can be found in literature, but only few works deal with the dialysis patient population. Keavney et al. [24] evaluated a 24-h ambulatory BP device using contemporary auscultatory and oscillometric methods (TM2421TM by A&D Co., Tokyo, Japan). The study was composed by an initial validation according to ANSI/AAMI SP-10 standard, followed by an extensive field trial with healthy and essential hypertensive subjects. Fagugli et al. [25] studied the same device on a population of 44 HD patients. Peixoto et al. [26] validated SpaceLabs 90207TM ambulatory BP monitor (Spacelabs Medical, Issaquah, USA) collecting 255 readings from 85 HD patients during the treatment.

Finally, Amoore and colleagues in vitro tested [27,28] various oscillometric devices for ambulatory BP monitoring by using NIBP test simulators according to the protocol proposed in [21].

The validation protocol here presented is well-extended since it comprised both in vivo validation with healthy and HD subjects and in vitro validation with a large extent of situations.

The first step of this validation consisted in verifying the compliance of HD-BPM device to ANSI/AAMI SP-10 standard requirements regarding overall accuracy (Section 4.4.5 of [18]). This verification was executed in a simulation setting, by using an off-line BP data set recorded during a preceding in vivo validation [22]. The results showed that with the modifications introduced, HD-BPM still satisfied ANSI/AAMI requirements.

As a second step, the estimates of the two NIBP monitors HD-BPM and BX-100 were collected by giving as input different BP waveforms generated by an NIBP test simulator [20]. Although this validation method is not fully accepted, it was possible to cover a wide range of simulated pressures, pulse rates, pulse strengths and artefact levels, such as may be encountered in clinical practice. In particular, an already published evaluation protocol was followed [21], in order to offer the most comparable results.

Since the simulated waveform is different from the physiological one, as demonstrated in [8,29], NIBP test simulators are not used to reveal systematic differences between the tested monitors, but to evaluate the internal repeatability and the robustness against artefacts. Therefore, for this step, SD are more significant than bias (Table 3). The repeatability of the pressure waveforms generated by NIBP test simulators is claimed to be within 1 mmHg [27]. Hence, Amoore [28] and van Montfrans [8] proposed, as limit of good repeatability, an intra-device SD <2 mmHg, calculated on 15 determinations of 120/80 mmHg BP waveform, without and with artefact, generated by NIBP test simulator Bio-Tek. Both HD-BPM and BX-100 monitors complied with this criteria for SBP and DBP without artefact (Figure 2A) and with motion/tremor artefact level 1 (mean values of SD: HD-BPM = 1.6, BX-100 = 2.1 mmHg), while they failed for the other levels (data not shown).

The so-called NIBP playback system [30] can partially solve the limitations of current NIBP test simulator. In this system, in fact, pressure pulses of a stored data set previously recorded from actual patients are selected and summed up to the pressure signal during the deflation of a cuff wrapped around a cylinder: in this way, the monitor under test is reading an oscillation envelope similar to the original data from a real patient. If the data set is comprehensive of the situations present in a dialysis clinical setting, this system could potentially cover Step I and II of this validation process.

The third and major step of the validation consisted of the field trial, along which the oscillometric measures by HD-BPM were compared with simultaneous auscultatory measures in 53 ESRD patients during dialysis treatment in four different centres. The requirements prescribed in ANSI/AAMI SP-10 for BP comparison were not fully addressed, especially because the main purpose of this study was to verify the accuracy of the oscillometric devices during usual treatment in a dialysis unit, instead of more quiet settings like home or ambulatory. In particular, no consecutive measurements, at least with the same device, were taken and only one trained observer, instead of two, made the auscultatory recording, not being the same person during the various sessions. Moreover, the number of the patients enrolled for the study was lower than the recommended one (85), even if the total amount of the BP measurements can be considered significant (392, while ANSI/AAMI prescribes a minimum number of 255 measurements). Furthermore, in this case, the comparison was enriched by using another oscillometric device, in particular by adding a consecutive comparative session.

Notwithstanding these limitations, the results were analysed using the ANSI/AAMI criteria and the requirements regarding accuracy resulted in them not being completely fulfilled. In fact, SBP SD of HD-BPM exceeded 1 mmHg of the maximum limit; SBP SD of BX-100 is slightly better, but not lower than 8 mmHg. On the contrary, biases and DBP SD were comprised into the limits, with the values from HD-BPM being lower. It is worth noting that oscillometric SBP SD were similar to the one obtained comparing the two auscultatory measures: hence, the dispersion of the oscillometric method is comparable with the short-term BP variability recorded by a mercury sphygmomanometer.

One may argue on the unbalanced design of the multi-centre field trial (from Table 2 it appears, in fact, that Danville and Sion centres provide 60% of measures obtained in 20% of the patients). Indeed, by randomly choosing only two consecutive measures for each patient, the differences of the readings between NIBP devices and auscultation result in not being statistically different with respect to the ones obtained from all the 392 sessions (P > 0.10 according to Student's t-test for unpaired samples).

Regarding validation of other devices also, Keavney et al. [24] found that SBP SD between oscillometric method and observer did not satisfy ANSI/AAMI criteria in some cases. Moreover, the results here obtained were not so far from the ones reported by Peixoto et al. [26] in patients during HD: the differences between auscultatory and oscillometric readings were 0.5 ± 7.5 mmHg for SBP and –0.2 ± 5.2 mmHg for DBP. Additionally, the authors observed that the device was less accurate in extreme ranges of SBP.

Even if restricted to a small number of subjects and sessions, the comparison with intra-arterial measurement method provided another test bench for assessing together the two NIBP devices, especially where vascular pathologies are concerned. Their behaviour is complementary this time as well: HD-BPM provided better SBP bias and DBP SD, while BX-100 is better for SBP SD and DBP bias. The result is in accordance with the observation that it is difficult to find a sphygmomanometer based on indirect methods yielding both SBP and DBP values that are within ± 5 mmHg of intra-arterial values [18].

In summary, we believe that these three validation approaches complete each other and provide a sufficiently wide test bench for assessing NIBP devices. In fact, the first step was designed to perform an initial validation on previously recorded waveforms of mainly normal subjects, albeit fully according to ANSI/AAMI SP-10; the second step evaluated BP reading repeatability, simulating a large range of measurement conditions, such as cardiac arrhythmia; the third step finally allowed NIBP devices to be validated in several dialysis centres. To strengthen this evaluation, two reference methods were separately used, that is auscultatory and intra-arterial measurement. The difficulties found during field trial can most probably be explained by the multiplicity of factors specific to the dialysis population, which complicate BP measurement. Among them, the not uncommon presence of other pathologies in the ESRD patient, observed in the present study as well (Table 2), can alter the BP measuring device performance (refer to Introduction section).

Beside the unremovable drawbacks of BP estimation in dialysis patients, according to our experience in this validation test, other issues exist, whose influence can be, however, significantly reduced by adhering more carefully to an agreed BP measurement protocol [9,10,17]. These aspects are mainly linked to the operator use and maintenance of the BP measuring device, attention to the patient's correct behaviour during measurement and the periodical verification and calibration of the instrument.

Hence, this work highlights the importance of assessing BP measurement accuracy and feasibility during working conditions, by offering an example of validation protocol to evaluate BP monitor devices for dialysis patients.



   Acknowledgements
Top
 Abstract
 Introduction
 Subjects and methods
 Results
 Discussion
 Acknowledgements
References
 
We would like to thank the nursing and technical staff of the hospitals in Danville (IL, US), Sion (Switzerland), Toronto, Trois-Rivières (Canada) and Bologna (Italy) for their kind collaboration in collecting the data and retrieving patients’ characteristics.

Conflict of interest statement. C.A.L. currently works as an employee in Gambro Dasco S.p.A., which is the manufacturer of the product HD-BPM, which is the subject of this study. C.E. does not have any involvements that might raise the question of bias in the work reported or in the conclusions, implications or opinions stated and C.G. currently works as an employee in Lean S.r.l., a company which supplies the product HD-BPM to Gambro Dasco S.p.A.



   References
Top
 Abstract
 Introduction
 Subjects and methods
 Results
 Discussion
 Acknowledgements
 References
 

  1. Mitra S, Chandna SM, Farrington K. What is hypertension in chronic haemodialysis? The role of interdialytic blood pressure monitoring. Nephrol Dial Transplant (1999) 14:2915–2921.[Abstract/Free Full Text]
  2. London GM. Controversy on optimal blood pressure on haemodialysis: lower is not always better. Nephrol Dial Transplant (2001) 16:475–478.[Free Full Text]
  3. Shoji T, Tsubakihara Y, Fujii M, Imai E. Hemodialysis-associated hypotension as an independent risk factor for two-year mortality in hemodialysis patients. Kidney Int (2004) 66:1212–1220.[CrossRef][Web of Science][Medline]
  4. Kovesdy CP, Trivedi BK, Kalantar-Zadeh K, Anderson JE. Association of low blood pressure with increased mortality in patients with moderate to severe chronic kidney disease. Nephrol Dial Transplant (2006) 21:1257–1262.[Abstract/Free Full Text]
  5. Zager PG, Nikolic J, Brown RH, et al. ‘U’ curve association of blood pressure and mortality in hemodialysis patients. Medical Directors of Dialysis Clinic, Inc. Kidney Int (1998) 54:561–569.[CrossRef][Web of Science][Medline]
  6. Stidley CA, Hunt WC, Tentori F, et al. Changing relationship of blood pressure with mortality over time among hemodialysis patients. J Am Soc Nephrol (2006) 17:513–520.[Abstract/Free Full Text]
  7. Drzewiecki G, Hood R, Apple H. Theory of the oscillometric maximum and the systolic and diastolic detection ratios. Ann Biomed Eng (1994) 22:88–96.[CrossRef][Web of Science][Medline]
  8. van Montfrans GA. Oscillometric blood pressure measurement: progress and problems. Blood Press Monit (2001) 6:287–290.[CrossRef][Web of Science][Medline]
  9. Tholl U, Forstner K, Anlauf M. Measuring blood pressure: pitfalls and recommendations. Nephrol Dial Transplant (2004) 19:766–770.[Free Full Text]
  10. Lazar AE, Smith MC, Rahman M. Blood pressure measurement in hemodialysis patients. Semin Dial (2004) 17:250–254.[CrossRef][Web of Science][Medline]
  11. Sankaranarayanan N, Santos SF, Peixoto AJ. Blood pressure measurement in dialysis patients. Adv Chronic Kidney Dis (2004) 11:134–142.[CrossRef][Web of Science][Medline]
  12. Ursino M, Cristalli C. A mathematical study of some biomechanical factors affecting the oscillometric blood pressure measurement. IEEE Trans Biomed Eng (1996) 43:761–778.[CrossRef][Web of Science][Medline]
  13. van Popele N, Bos WJW, de Beer NAM, et al. Arterial stiffness as underlying mechanism of disagreement between an oscillometric blood pressure monitor and a sphygmomanometer. Hypertension (2000) 36:484–488.[Abstract/Free Full Text]
  14. White WB, Mansoor GA. Ambulatory blood pressure monitoring. Curr Opin Nephrol Hypertens (1993) 2:928–934.[CrossRef][Medline]
  15. Perloff D, Grim C, Flack J, et al. (American Heart Association) Human blood pressure determination by sphygmomanometry. Circulation (1993) 88:2460–2470AHA.[Free Full Text]
  16. Fonseca-Reyes S, de Alba-Garcia JG, Parra-Carrillo JZ, Paczka-Zapata JA. Effect of standard cuff on blood pressure readings in patients with obese arms. How frequent are arms of a ‘large circumference’? Blood Press Monit (2003) 8:101–106.[Web of Science][Medline]
  17. Rahman M, Griffin V, Kumar A, et al. A comparison of standardized vs ‘usual’ blood pressure measurements in hemodialysis patients. Am J Kidney Dis (2002) 39:1226–1230.[CrossRef][Web of Science][Medline]
  18. Association for the Advancement of Medical Instrumentation. American National Standard. Manual, electronic, or automated sphygmomanometers. In: ANSI/AAMI SP-10 (1992 and 2002) Arlington, VA: AAMI.
  19. O’Brien E, Petrie J, Littler W, et al. The British Hypertension Society protocol for the evaluation of blood pressure measuring devices. J Hypertens (1993) 11:S43–S62.[Web of Science]
  20. Ng KG, Small CF. Review of methods & simulators for evaluation of non-invasive blood pressure monitors. J Clin Eng (1992) 17:469–479.[Medline]
  21. Geake WB, Amoore JN, Scott DH. An automated system for the functional evaluation of oscillometric non-invasive blood pressure monitors. J Med Eng Technol (1995) 19:162–176.[Web of Science][Medline]
  22. Cavalcanti S, Marchesi G, Ghidini C. Validation of automated oscillometric sphygmomanometer (HDBPM) for arterial pressure measurement during haemodialysis. Med Biol Eng Comput (2000) 38:98–101.[CrossRef][Web of Science][Medline]
  23. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet (1986) 8:307–310.
  24. Keavney B, Bird R, Caiazza A, Casadei B, Conway J. Measurement of blood pressure using the auscultatory and oscillometric methods in the same cuff deflation: validation and field trial of the A&D TM2421 monitor. J Hum Hypertens (2000) 14:573–579.[CrossRef][Web of Science][Medline]
  25. Fagugli RM, Vecchi L, Valente F, Santirosi P, Laviola MM. Comparison between oscillometric and auscultatory methods of ambulatory blood pressure measurement in hemodialysis patients. Clin Nephrol (2002) 57:283–288.[Web of Science][Medline]
  26. Peixoto AJ, Gray TA, Crowley ST. Validation of the SpaceLabs 90207 ambulatory blood pressure device for hemodialysis patients. Blood Press Monit (1999) 4:217–221.[Medline]
  27. Amoore JN, Geake WB, Scott DH. Oscillometric non-invasive blood pressure measurements: the influence of the make of instrument on readings? Med Biol Eng Comput (1997) 35:131–134.[CrossRef][Web of Science][Medline]
  28. Amoore JN. A simulation study of the consistency of oscillometric blood pressure measurements with and without artifacts. Blood Press Monit (2000) 5:69–79.[Web of Science][Medline]
  29. Amoore JN, Scott DH. Can simulators evaluate systematic differences between oscillometric non-invasive blood-pressure monitors? Blood Press Monit (2000) 5:81–89.[Web of Science][Medline]
  30. Medero R, Booth J. NIBP Playback system. Johnson & Johnson Medical. Patent No. EP0748610. 12-18-1996. US.
Received for publication: 22. 8.06
Accepted in revised form: 17. 4.07


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?



This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
22/10/2950    most recent
gfm288v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Lodi, C. A.
Right arrow Articles by Ghidini, C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Lodi, C. A.
Right arrow Articles by Ghidini, C.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?