Continuous non-invasive arterial pressure shows high

accuracy in comparison to invasive intra-arterial blood

pressure measurement

I. INTRODUCTION

Continuous blood pressure （BP） monitoring is required in

a multitude of clinical settings, especially in perioperative

care. For inpatient surgeries, the American Society

of Anesthesiologists （ASA） requires continuous perioperative

blood pressure monitoring at least for patients with

severe systemic disease; this necessitates the invasive

placement of an intra-arterial catheter. In all other cases

intermittent non-invasive blood pressure monitoring （NBP）

is the standard of care. Therefore, the patients` blood

pressure may not be monitored at all times.

A recent representative survey（1） among Austrian and

German physicians （N=198） provides evidence that, in

82% of inpatient surgeries, non-invasive blood pressure

monitoring is used. However, in 25% of these cases,

especially in surgeries where hemodynamic instabilities

can be expected or where aggressive management of

blood pressure might be required （e.g. in urologic, extended

laparoscopic, orthopaedic or vascular surgeries, in

surgeries in gynecology and obstetrics, in medium to extended

intestinal surgery and elective or urgent trauma

surgery（2））, anesthetists would prefer a non-invasive continuous

blood pressure monitoring to have better control

over the patient’s hemodynamics. In the remaining 18%

of inpatient surgeries, BP is measured continuously using

invasive catheters （IBP）, mainly in patients where cardiovascular

instability is expected and thus ASA guidelines

specifically require continuous BP measurement and/or

where repeated blood gas analysis is needed. Note that,

in 26% of these cases the invasive catheter is inserted

only to enable continuous blood pressure monitoring.

However, this is a time-consuming and cost-intensive

procedure, causing pain for the patient and including

the risk of infection, and thus should be replaced by a

non-invasive method if possible.

There are a number of studies stressing the importance

of continuous perioperative blood pressure monitoring:

e.g., more than 20% of all hypotensive episodes during

surgeries may be missed by intermittent upper-arm

blood pressure readings and another 20% may be detected

with a delay（3）. This in turn may prevent immediate

treatment or even lead to missing complete hypotensive

episodes. It has been shown that intraoperative

hypotension preceeds 56% of perioperative cardiac

arrests（4） and is associated with a significant increase of

the 1-year post surgical mortality rate（5）, indicating that

intermittent NBP monitoring can be insufficient.

Consequently, there seems to be a discrepancy between

the number of cases where continuous blood

pressure monitoring is needed and those cases where

it is actually used: Due to its invasive nature and associated

risks, intra-arterial catheters can only be justified in

a limited number of patients whereas anesthetists would

like to perform risk-free continuous BP monitoring in a greater

number of cases. For exactly these situations CNAP™

has recently become available（6, 7, 8, 9, 10, 11）. CNAP™ is designed

for anesthetists who look for more control in situations

when continuous blood pressure is desirable, but

the risks and burden of an arterial line are not justified.

CNAP™ provides continuous, non-invasive and risk-free

beat-to-beat blood pressure measurement.

The aim of the present report was to evaluate the accuracy

of CNAP™ in a real-life perioperative setting by

comparing simultaneous measurements of CNAP™ to

continuous intra-arterial pressure monitoring.

II. Methods

Data recording

The measurements were conducted in a perioperative

setting at the Department of Anesthesiology at Landeskrankenhaus

Bruck an der Mur （Austria）. In all patients

included in this report, continuous BP monitoring was

indicated by clinical safety standards. Arterial pressure

was measured simultaneously with an invasive catheter

（Edwards Life Sciences™ Pressure Monitoring Set, Irvine,

USA, connected to Datex Ohmeda S/5 monitor, GE, Helsinki,

Finland） and the CNAP™ Monitor 500i （CNSystems

Medizintechnik AG, Graz, Austria） in fifteen patients undergoing

orthopedic, cardiac and vascular surgeries

（seven female and eight male patients, mean age of

71 years, range 33 to 82 years, ASA classifications I-III: I in

1 case, II in 12 cases, III in 2 cases）. The arterial catheter

was placed ipsi-laterally （n=5） or contra-laterally （n=10）

to the CNAP™ finger cuff in the A. radialis or A. brachialis,

depending on indication and requirements. The surgery

durations averaged 1h39min with a minimum of 44min

and a maximum of 3h01min, the total duration of recordings

obtained was approx. 25 hours.

Data processing

From the IBP as well as from the CNAP™ signal, systolic,

diastolic and mean pressure values were derived for

each second. If one of the signals was missing （e.g. due

to transmission faults or artifacts） for one data point, all

other measurements for that data point were consequently

discarded. Otherwise, no further data processing

was performed and a total of 75,485 data points

were included into the statistical comparison.

Sackl-Pietsch E., Department of Anesthesiology, Landeskrankenhaus Bruck an der Mur, Austria

Data comparison

For a comprehensive evaluation of CNAP™, its underlying

mechanisms have to be considered: CNAP™ is an

integrated solution where relative BP changes are measured

at the finger sensor which are turned into absolute

values based on initial readings from its integrated

NBP-unit. This fact needs to be taken into consideration

when comparing the blood pressure readings recorded

by CNAP™ and IBP.

Since three measurement positions are combined in this

comparison （CNAP™ finger sensor, CNAP™ NBP-unit and

IBP catheter）, some physiological facts have to be taken

into account: namely, transformations of BP amplitudes

and waveforms as illustrated in figure 1. This implies that

a systematic offset between CNAP™ and IBP can be expected.

Thus, it is not surprising that even the AAMI-SP10 standard

recommended by the FDA reports substantial differences

between indirect NBP and direct intra-arterial

measurements（12）. A meta-analysis with the results of nine

studies totaling 330 patients was performed which quantifies

this systematic offset: The average differences between

arterial and NBP-cuff systolic BP ranged from 0.8 to

13.4 mmHg with standard deviations （SD） ranging from 0

to 13.0 mmHg. Diastolic BP showed average differences

from 0.8 to 18.0 mmHg with SDs ranging from 0.0 to 10.2

mmHg.

This offset may be even magnified when IBP and NBP

recordings are taken on contra-lateral arms. Note that in

10 out of the 15 patients reported on here, CNAP™ and

IBP were placed on contra-lateral arms.

Therefore, the following differences between CNAP™

and IBP can be expected:

（i） Differences between the two BP waveforms.

（ii） The characteristic offset between the absolute values

of systolic, diastolic and mean pressure.

Figure 1: Different blood pressure waveforms and amplitudes in

the （1） A. brachialis, （2） A. radialis and （3） A. digitalis, resulting

in different systolic and diastolic values

III. RESULTS

Waveform comparison

Figure 2 shows blood pressure waveforms recorded by

CNAP™ compared directly to intra-arterial blood pressure

waveforms. The upper graph shows a short episode

of stable blood pressure. The bottom-up arrows indicate

rising and the top-down arrows indicate falling BP ramps

considered as results of volume status, the Frank-Starling

mechanism and autonomic regulation（13）. The lower graph

shows BP changes caused by perioperative treatment

or patient movement. Due to the fact that data

was recorded in the clinical routine, no further information

about the patient`s treatment at this special time

slice is available. Nevertheless, the good accordance of

waveforms indicates that CNAP™ can follow fast blood

pressure variations changes as well as IBP.

Hemodynamic changes

For clinical application it is important to ensure that

CNAP™ is able to monitor fast hemodynamic changes.

In figure 3 an example is displayed where short-term hemodynamic

variability during 25 minutes of orthopedic

surgery can be observed clearly: CNAP™ and IBP display

a parallel hemodynamic trend with the typical offset

between indirect and direct measurement methods.

Sackl-Pietsch E., Department of Anesthesiology, Landeskrankenhaus Bruck an der Mur, Austria 2

Continuous non-invasive arterial pressure shows high accuracy in comparison to invasive intra-arterial

blood pressure measurement

Figure 2: Blood pressure tracings showing the agreement of

CNAP™ （solid line） with IBP （dotted line） during anesthesia.

Boxplots for all patients’ data sets

Figure 4 shows boxplots for all 15 data sets, for mean

BP values. This graph illustrates that most of the patients

show a characteristic offset between CNAP™ and IBP.

Bland-Altman-plots for the complete data set

The differences of CNAP™ and IBP data points were

computed for every data point （n = 75,485） and plotted

vs. their average, resulting in the Bland-Altman-plot of

Figure 5. No distinct trend of blood pressure difference

in relation to the absolute mean values of pressure can

be detected, i.e. the diffe difference between the two

recording methods is the same over the whole range of

values.

Furthermore, table 1 shows mean values and standard

deviations of differences of CNAP™ to IBP for systolic,

mean and diastolic pressure for each patient separay

as well as for the whole sample.

Figure 3: Comparison of short-term trends of systolic, diastolic

and mean blood pressure measurements from CNAP™ （solid

lines） and from IBP （dotted lines） during 25 min of anesthesia.

Figure 4: Boxplots of differences between CNAP™ and IBP

values for all 15 patients （mean BP [mmHg]）. The boxes contain

the middle 50% of the data, the horizontal lines show the

median. The upper and lower edges of the boxes indicate the

75th and 25th percentiles, respectively. The 5-95% range of the

data is indicated by the ends of the vertical lines.

Figure 5: Bland-Altman-plot of differences vs. average of all

data points （CNAP™ vs. IBP values, n=75,485） for mean BP

[mmHg].

Systolic BP Mean BP Diastolic BP

patient mean SD mean SD mean SD

1 -10,03 13,83 4,29 9,87 8,80 6,80

2 2,56 7,54 16,09 5,82 19,24 5,88

3 -2,81 7,17 6,99 6,51 12,27 7,21

4 -7,82 12,06 1,88 12,62 9,94 14,11

5 1,31 6,63 14,41 5,88 20,25 4,70

6 -16,43 5,11 -9,44 4,15 -3,99 4,38

7 -1,33 8,00 5,44 6,15 14,46 5,07

8 -10,77 5,69 1,91 3,71 7,34 2,86

9 -11,20 7,78 -0,81 6,71 3,75 6,91

10 -9,93 7,82 1,93 3,86 7,16 3,22

11 -25,82 8,37 -7,48 4,62 0,22 3,89

12 -1,45 6,95 6,52 7,73 10,95 6,41

13 0,24 11,62 6,81 7,91 11,09 7,34

14 33,55 4,59 32,00 7,10 37,77 5,77

15 2,89 10,49 13,58 5,84 19,69 4,99

Total -2,96 13,81 6,66 11,23 12,36 10,91

TABLE 1: Means and standard deviations （SD） of differences

between CNAPTM and IBP [mmHg].

Sackl-Pietsch E., Department of Anesthesiology, Landeskrankenhaus Bruck an der Mur, Austria 3

Continuous non-invasive arterial pressure shows high accuracy in comparison to invasive intra-arterial

blood pressure measurement

IV. DISCUSSION

Within an every day clinical setting, CNAP™ and IBP readings

were recorded simultaneously during inpatient surgeries.

The results of this perioperative comparison indicate

that CNAP™ has a high usability during anesthetic

care: the overall statistical analyses of systolic, mean and

diastolic blood pressure show small differences and standard

deviations between the two methods. The graphical

comparison of BP waveforms and short-term trends

during anesthesia indicates that CNAP™ can follow hemodynamic

variability as fast as IBP. These results give

strong support to a high accuracy of the non-invasive

CNAP™ device in comparison to the invasive measurement.

The waveforms of CNAP™ and IBP shown in figure 2 comply

well with the physiological expectations （see section “Methods”）.

As can be seen, CNAP™ corresponds to the IBP

signal both in resting conditions as well as in movement.

For perioperative usability of the CNAP™ system, it is essential

to show that CNAP™ can deal with hemodynamic

changes as well as IBP: The trends of systolic, diastolic

and mean BP depicted in figure 3 show excellent visual

accordance between the two devices.

To illustrate the overall agreement between CNAP™ and

IBP, figures 4 and 5 sum up the results for all 15 patients.

The validation of CNAP™ with a total observation duration

of about 25 hours and 75,485 data points is very acceptable:

The mean values and standard deviations of

differences to the intra-arterial recordings comply with

the results of the meta-analysis recommended by the

FDA.

As can be seen in figure 4, all patients have their own

characteristic offset between CNAP™ and IBP. Only patient

no. 14 seems to slightly deviate from the rest with a

higher pressure difference which may be explained by

the patients’ arteries: in patient no. 14 the vessels were

described by the clinician as ‘stiff’ and the IBP readings

as ‘dependent on bedding’, thus making the arterial reference

less reliable and the results surprisingly good. On

the other hand, not even in the case where a patient’s

peripheral perfusion was described by the physician

as “poor” （patient no. 11） did the CNAP™ system fail to

quickly find a suitable BP waveform and the results compared

to IBP are very satisfactory.

The individual, physiologically-determined offset can

also be seen clearly in the cluster of data points of each

patient in figure 5 （e.g., note patient no. 14 in the upper

right-hand corner）. Nevertheless, the Bland-Altman-plot

between CNAP™ and IBP shows no distinct trend of mean

pressure difference in relation to the average values of

pressure, i.e. the difference between the two recording

methods is the same over the whole range of values. This

indicates that CNAP™ measurement is reliable in normal,

hypotensive and hypertensive episodes.

The mean values and standard deviations of differences

between CNAP™ and IBP reported in table 1 confirm the

findings of the meta-analysis in the current ANSI standard.

These results are very satisfactory considering the patient

sample included in this report. Note that data was recorded

in patients with severe systemic disease or during

higher-risk surgeries where the placement of an invasive

catheter was motivated by safety considerations.

Although the results of this report indicate a high clinical

usability of CNAP™, some remarks have to be made

about the comparison to IBP measurements: There is

common agreement that “true” blood pressure is best

determined directly using a reliable, calibrated transducer

in an artery. Nevertheless, there is also consensus

that the direct intra-arterial measurement is fraught with

its inherent set of issues, including variability with radial

position, vasoconstriction, the effects of flow-velocity

changes and the frequency response of amplifier and

transducer. Taking this into account, the results of this

present report are even more remarkable.

V. Conclusion

On the whole, the reported results provide clear evidence

of an excellent clinical feasibility and high accuracy

of the non-invasive BP measurement device CNAP™

in comparison to IBP.

With intermittent measurement of oscillometric sphygmomanometers

（NBP）, short-term but clinically relevant

hemodynamic changes during anesthesia are not satisfactorily

detectable. Therefore, the demand from

anesthetists for a system providing non-invasive, continuous

beat-to-beat BP is increasing.

CNAP™ provides patient comfort and usability similar to

a standard upper-arm NBP and clinical data shows that

its accuracy is comparable to IBP. Thus, CNAP™ is the

convenient solution for anesthetists who want to have

comprehensive hemodynamic control to ensure highest

patient safety.

Sackl-Pietsch E., Department of Anesthesiology, Landeskrankenhaus Bruck an der Mur, Austria 4

Continuous non-invasive arterial pressure shows high accuracy in comparison to invasive intra-arterial

blood pressure measurement

VI. REFERENCES

1. von Skerst B: Market survey, N=198 physicians in Germany

and Austria, Dec.2007 - Mar 2008, InnoTech Consult

GmbH, Germany

2. Ezekiel MR. Handbook of Anesthesiology. Current Clinical

Strategies Publishing. 2003

3. Dueck R, Jameson LC. Reliability of hypotension detection

with noninvasive radial artery beat-to-beat versus upper

arm cuff BP monitoring. Anesth Analg 2006, 102 Suppl: S10

4. Sprung J, Warner ME, Contreras ME et al. Predictors of

Survival following Cardiac Arrest in Patients Undergoing

Noncardiac Surgery. Anesthesiology 2003; 99:259–69

5. Monk TG, Saini V, Weldon BC, Sigl JC. Anesthetic management

and one-year mortality after noncardiac surgery.

Anesth Analg. 2005 Jan;100（1）:4-10.

6. Fortin J, Gratze G, Wach P, Skrabal: Automated noninvasive

assessment of cardiovascular function, spectra

analysis and baroreceptor sensitivity for the diagnosis

of syncopes. World Congress on Medical Physics and

Biomedical Engineering. Med & Biol Eng & Comput, 35,

Supplement I, 466 （1997）.

7. Gratze G, Fortin J, Holler A, Grasenick K, Pfurtscheller G,

Wach P, Kotanko P, Skrabal F: A software package for

non-invasive, real time beat to beat monitoring of stroke

volume, blood pressure, total peripheral resistance and for

assessment of autonomic function. Comp in Bio & Medicine;

28, 121-142 （1998）.

8. Fortin J, Habenbacher W, Gruellenberger R, Wach P, Skrabal

F: Real-time Monitor for hemodynamic beat-to-beat

parameters and power spectra analysis of the biosignals.

Proc. of the 20th Annual International Conference of the

IEEE Eng in Medicine and Biology Society, 20, 1 （1998）.

9. Fortin J, Marte W, Grüllenberger R, Hacker A, Habenbacher

W, Heller A, Wagner Ch, Wach P, Skrabal F: Continuous

non-invasive blood pressure monitoring using concentrically

interlocking control loops. Computers in Biology

and Medicine 36 （2006） 941–957

10. Fortin J, Alkan S, Wrede C E, Sackl-Pietsch E and Wach P:

Continuous Non-invasive Arterial Pressure （CNAP™） – An

Innovative Approach of the Vascular Unloading Technique.

Submitted for publication in Blood Pressure April

2008.

11. Fortin J: Continuous Non-invasive Measurements of Cardiovascular

Function. PhD-thesis, Institute of Biomedical Engineering,

University of Technology Graz, 2007, pp. 103-21.

12. Association for the Advancement of Medical Instrumentation.

American National Standard. Manual, electronic or

automated sphygmomanometers ANSI/AAMI SP10-2002/

A1. 3330 Washington Boulevard, Suite 400, Arlington, VA

22201-4598, USA: AAMI; 2003

13. Parati G, Omboni St, Frattola A, Di Rienzo M, Zanchetti A,

Mancia G. Dynamic evaluation of the baroreflex in ambulant

subject. In: Blood pressure and heart rate variability,

edited by di Rienzo et al. IOS Press, 1992, pp. 123-137.

Sackl-Pietsch E., Department of Anesthesiology, Landeskrankenhaus Bruck an der Mur, Austria 5