Accepted Manuscript Title: A new integrated instrumental approach to autonomic nervous system assessment Author: I. Corazza G. Barletta P. Guaraldi A. Cecere G. Calandra-Buonaura E. Altini R. Zannoli P. Cortelli PII: DOI: Reference:

S0169-2607(14)00304-6 http://dx.doi.org/doi:10.1016/j.cmpb.2014.08.002 COMM 3843

To appear in:

Computer Methods and Programs in Biomedicine

Received date: Revised date: Accepted date:

4-3-2014 15-7-2014 5-8-2014

Please cite this article as: I. Corazza, G. Barletta, P. Guaraldi, A. Cecere, G. CalandraBuonaura, E. Altini, R. Zannoli, P. Cortelli, A new integrated instrumental approach to autonomic nervous system assessment., Computer Methods and Programs in Biomedicine (2014), http://dx.doi.org/10.1016/j.cmpb.2014.08.002 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

A new integrated instrumental approach to autonomic nervous system assessment I Corazza1*, G Barletta2,3*, P Guaraldi2,3, A Cecere2,3,G Calandra-Buonaura2,3, E Altini1, R Zannoli1 and P Cortelli2,3 (1) Experimental, Diagnostic and Specialty Medicine Department, University of Bologna, Italy

(3) IRCCS, Institute of Neurological Sciences of Bologna, Bologna, Italy

cr

* These authors contributed equally to this work

us

Corresponding author: Ivan Corazza

an

DIMES – University of Bologna Via Massarenti, 9

Phone: +390516363588

M

40138 Bologna Italy

ip t

(2) Department of Biomedical and Neuromotor Sciences, University of Bologna, Italy

Ac ce pt e

d

Email: [email protected]

Page 1 of 24

Abstract Background and Objectives The autonomic nervous system (ANS) regulates involuntary body functions and is commonly evaluated by measuring reflex responses of systolic and diastolic blood pressure (BP) and heart rate (HR) to physiological and pharmacological stimuli. However, BP and HR values may not sufficient

ip t

be to explain specific ANS events and other parameters like the electrocardiogram (ECG), BP waves, the respiratory rate and the electroencephalogram (EEG) are mandatory. Although ANS

cr

behaviour and its response to stimuli are well-known, their clinical evaluation is often based on individual medical training and experience. As a result, ANS laboratories have been customized,

us

making it impossible to standardize procedures and share results with colleagues.

The aim of our study was to build a powerful versatile instrument easy-to-use in clinical practice to

an

standardize procedures and allow a cross-analysis of all the parameters of interest for ANS evaluation. Methods

M

The new ANScovery System developed by neurologists and technicians is a two-step device: 1) integrating physiological information from different already existing commercial modules, making

d

it possible to cross-analyse, store and share data; 2) standardizing procedures by an innovative tutor

Ac ce pt e

monitor able to guide the patient throughout ANS testing. Results and Conclusions

The daily use of the new ANScovery System in clinical practice has proved it is a versatile easy to use instrument. Standardization of the manoeuvres and step-by-step guidance throughout the procedure avoid repetitions and allow intra and inter-patient data comparison.

Keywords: autonomic nervous system, cardiovascular reflexes, integrated solution, signal analysis

Page 2 of 24

1. Introduction The ANS consists of several interconnected areas distributed throughout the neuraxis and known as the central autonomic network (CAN). The CAN is involved in tonic, reflex and adaptive control of autonomic functions receiving visceroceptive, humoral and exteroceptive information and modulating sympathetic and parasympathetic neuronal outputs of the CAN. [1]

ip t

Because the anatomical location of the ANS renders it inaccessible to simple and direct physiological testing, a series of clinical tests assessing autonomic function and dysfunction have

cr

been developed to overcome this problem by measuring the end-organ responses to various physiological perturbations. [2-5]

us

In 1982, Ewing and Clarke advocated a battery of five tests suitable for bedside autonomic function testing. These tests provided an assessment of both sympathetic and parasympathetic nervous

an

system function. With some modifications, this test battery still forms the core of the cardiovascular autonomic evaluation performed by many autonomic laboratories. [6]

The background physiology of these cardiovascular tests includes parasympathetic cardiac control

M

and sympathetic cardiovascular control. The preganglionic parasympathetic nerves are located in the nucleus ambiguus and the dorsal motor nucleus of the vagus in the medulla oblongata of the

d

brain stem and project to the heart via vagus nerves terminating in cardiac ganglia that innervate the

Ac ce pt e

sino-atrial node, right and left atria, and atrio-ventricular node, where the neurotransmitter acetylcholine binds to muscarinic receptors to decrease heart rate (HR), atrial contractility, and A-V nodal conduction. Inputs from the arterial baroreflex, arterial chemoreflex, and cardiopulmonary receptors mediate cardiovascular reflexes that alter cardiovagal nerve activity. The activity of the sympathetic nervous system influences mean arterial pressure through many mechanisms, including effects on total peripheral vascular resistance, HR, myocardial contractility, venous vascular tone, and blood volume. The tonic excitatory drive to preganglionic sympathetic neurons located in the intermediolateral column of the thoracolumbar spinal cord originates primarily in premotor neurons

Page 3 of 24

in the rostral ventrolateral medulla, with other central nervous system sites such as the hypothalamic paraventricular nucleus contributing to variable degrees depending on the physiological and pathological state. Preganglionic sympathetic neurons project to sympathetic ganglia where they release acetylcholine as the primary neurotransmitter and they provide widespread innervation of arterial and venous systems, visceral organs and adrenal glands. The arterial baroreflex, arterial chemoreflex, and cardiopulmonary receptors mediate cardiovascular

ip t

reflexes that also alter blood pressure. In addition, neural activity originating within the CAN associated with various behaviours (stress, anxiety, and/or fear) may trigger changes in sympathetic

cr

output that influence cardiovascular control. [7, 8]

Autonomic function in health and disease is now widely assessed myriad disciplines (neurology,

us

cardiology, psychology, psychophysiology, obstetrics, anaesthesiology, and psychiatry) and ANS laboratories are increasingly required to disclose autonomic dysfunction, determine the anatomic

an

and physiologic distribution of the deficit, grade its severity, and monitor the patient’s response to therapy. [2]

According to the well-known physiology of cardiovascular reflex tests the BP responses to head-up

M

tilt test (HUTT), Valsalva manoeuvre (VM), and isometric exercise (IE) are considered a good estimate of sympathetic function while the HR response to deep breathing (DB), and VM are

d

considered a good estimate of cardiovagal function. [2-5]

Ac ce pt e

In order to understand and analyse the beat-to-beat changes in BP and HR induced by the manoeuvres and to check the correctness of the procedure ECG and pressures waves must be acquired in real time [9] with instrumentation such as the Finometer (FMS, Amsterdam, The Netherlands) or Nexfin System (Edwards Lifesciences, Irvine, CA, USA; NIBP100D, Biopac, Goleta, CA, USA). [10]

However, measuring only BP and HR without real-time monitoring and simultaneous measurement of the respiratory rate and EEG signals with video recordings may lead to the wrong diagnosis (i.e.

Page 4 of 24

mistaking a seizure or the effects of hyperventilation for syncope). EEG monitoring provides dynamic information on brain function, disclosing early changes in neurologic status and a possible correlation between the central nervous system and the ANS [11-13]. Respiratory rate assessment is fundamental to evaluate physiological sinus arrhythmia and the possible effects of hyperventilation, apnoeas or breath-holding on cardiovascular parameters [6]. Video recording is crucial during the post-processing phase to relate the patient’s physiological responses to the accuracy of manoeuvres,

ip t

evaluating any extraneous body movements and artifacts. Unfortunately, few ANS laboratories use standard procedures and new acquisition modules (AM)

cr

are usually added to those already present without the possibility to interconnect; other laboratories

adopt all-in-one solutions, allowing the integration and cross-analysis of some but not all

us

parameters depending on the manufacturer. [10]

The wide variety of instrumentation and the lack of standardization in today’s ANS laboratories are

an

critical shortcomings for diagnosis, results comparison and clinical advances. Since the cost of equipment is high and health system budgets are limited, it is unrealistic to think that standardization could be achieved with a powerful but expensive all-in-one system. For this reason,

M

a good solution to the problem could be a modular system able to interface the AM already present in an ANS laboratory and open to improvement with new devices.

d

The aim of our study was to build a powerful versatile instrument to guide the patient through test

Ac ce pt e

procedures and cross-analyse all the cardiovascular parameters of interest for ANS evaluation. Our solution was designed and built in the ANS laboratory of the University of Bologna (Bologna, Italy) based on the clinical and research experience of ANS specialists.  [14] Our approach was to collect signals from commercial AM by means of a specific analogue-to-digital (AD) converter and dedicated open software to visualize, store and analyse the clinical information and generate a final report with the results. This guarantees standardization of the procedures, simultaneous acquisition of all the parameters required, the addition of any new AM to upgrade the system, and storage of all patient data in only one repository for statistical analysis and research purposes. Standard protocols

Page 5 of 24

were followed for each clinical procedure. [2-4, 6, 15] An example is reported of the results obtained by applying a standard battery of tests to two groups of healthy and pathological subjects.

2. Material and methods The core of our proposal was the possibility to integrate physiological information from different commercial modules during ANS evaluation in order to cross-analyse and store data. The solution

ip t

was based on an analogue-digital approach and comprises four main blocks: 1) Commercial AM with analogue outputs;

cr

2) Multipurpose AD converter connected with a medical grade personal computer (PC) to integrate, analyse and store all the signals of interest;

us

3) Video recorder directly interfaced to the PC; 4) Centralized web-based database to store data.

an

To complete the proposal, an external device (tutor monitor, TM) was developed and interfaced to the system to guide the patient through test procedures, standardizing execution of the manoeuvres.

2.1 Commercial AM

M

A schematic view of our system (ANScovery) is shown in figure 1.

For ANS evaluation acquiring all the parameters of interest (BP, ECG, EEG, nasal and abdominal

d

breathing), the following commercial AM were used:

Ac ce pt e

Model 15LT (GRASS TECHNOLOGIES, Quincy, MA) polygraph amplifiers for: - ECG: Lead II with foam, wet, gel, ECG Electrodes ( Kendall™); - BREATH: Abdominal or chest respiration motions with piezo-crystal respiration effort sensor (SleepSense) and respiratory airflow with thermal flow sensor (SleepSense);

- PERIPHERAL VASOMOTOR TONE: with finger photoelectric transducer (51P35H GRASS TECHNOLOGIES):

Page 6 of 24

- EEG: two or four bipolar leads derived from Ag/AgCl disk electrodes positioned on the scalp according to the10/20 international system; - Finometer MIDI (Finapres Medical Systems, Amsterdam) for continuous beat-to-beat BP measurement.

ch

ECG

Model 15 LT 15A54

3

EEG

Model 15 LT 15A54

2

BREATH

Model 15 LT 15A94

2

Model 15 LT 15A12

1

VASOMOTOR TONE Finometer Midi

1

M

BLOOD PRESSURE

an

PERIPHERAL

cr

Monitoring device

us

Parameter

ip t

Table 1 shows the technical number of channels for each parameter.

Table 1. Clinical parameters, acquisition module and number of channels (ch) for each device.

d

2.2 Data acquisition (multipurpose AD converter)

Ac ce pt e

The AD unit was designed and built in collaboration with SparkBio Srl (Bologna, Italy) and acquires up to 16 analogical signals at a sample rate of 500 Hz with a resolution of 12 bits. The sampling rate and number of channels are easily set by the user through the software interface of the ANScovery System. The only requirement is that the product of these two factors remains constant. Moreover, the instrument allows the input ranges to be set according to the external monitor set-up, being the optimal solution to integrate products with different technical specifications. The AD converter is serial connected with a Microsoft Windows-based PC with dedicated software for visualization, storage and analysis. The software is built with Microsoft Visual Studio based on

Page 7 of 24

NET Framework SDK and will combine signals from the AD converter with additional signals directly acquired from the PC (i.e. video recordings). HR is calculated in real time and a tachogram channel is added and visualized together with the other signals. Systolic (SBP) and diastolic BP (DBP) values are shown throughout the procedure.

2.3 Video recorder

ip t

The video and audio recording system is integrated into the PC and comprises an internal interface (MORPHIS MATROX EVO), an external camera (PANASONIC VIDEOCAMERA mod. WV-

cr

CS580/G for acquisition of high-resolution video clips at 25fps, and joystick Videotec mod. DCJ for the remote control of the video camera) and a high sensitivity microphone to record ambient

us

sounds. A secondary monitor has been added to the PC for better video display. The video interface is directly driven by the AD converter to guarantee a simultaneous start-up.

an

Since the timing of the electronic components used for video and signals acquisition is intrinsically variable, the ANScovery system also implemented a software synchronization protocol: after a fixed time, depending on the settled sampling rate, the numbers of buffered samples is compared to

M

the acquired frames. In case of incongruity, a correction is made. Given that the AD converter has an internal clock at 20kHz and the typical sampling rates used for signal acquisition are 500/100Hz,

d

the timing errors of the analogical signals due to clock variability is 20/40 times less than the

Ac ce pt e

sampling rate. When the timing discrepancy between the frame rate and the sampled data becomes equal to the sampling period, a sample is added or erased from the buffered data. Since the video frame rate is lower than the AD sampling rate, errors in video acquisition are rare. In case of undersampling, each lost frame is compensated by an additional one equal to the last frame correctly acquired, thereby always maintaining the frame rate equal to 25fps. Thus, the signals and video are realigned in real time and any time lags kept shorter than one sampling period, irrespective of the procedure duration.

Page 8 of 24

2.4 Tutor monitor A PC-based tutor monitor (TM) was designed and built to guide and support patients during execution of the ANS tests by means of visual and acoustic feedback (time, event mark, test parameters). The beginning and end of each manoeuvre are automatically marked and a marker channel is added and acquired together with the other parameters. This channel is important for automatic

ip t

recognition of the manoeuvre and for software analysis. A dedicated pressure module (composed of two pressure sensors) is connected to the monitor to acquire the pressure generated by the patient

cr

during VM and IE. A mouthpiece is connected to one pressure sensor and an armcuff is connected with an air hose to the other. TM shows these pressure signals in real time, providing positive

us

feedback to patients and helping them perform the manoeuvres. For example, during the VM the patient’s expiratory pressure is shown as an incremental green bar. A threshold mark is set by the

an

operator (red mark) and is clearly visible on the screen. The manoeuvre starts when the green bar crosses the threshold and stops after 15s.

The start and stop time instants are stored and automatically sent to the AD unit as marker signals.

M

Correct manoeuvre identification is fundamental for procedure analysis and clinical evaluation. For

manoeuvres.

d

each manoeuvre, the TM shows different frames (figure 2) guiding the patient through the

Ac ce pt e

The clinical set-up in the ANS laboratory is completed by a tilting bed (MFTT tilt-bed) to change the patient’s position from clinostatism to orthostatism depending on the procedure [2, 3].

2.5 Clinical protocol for ANS evaluation

The standard clinical protocol for ANS analysis comprises a sequence of cardiovascular tests that can be customized depending on the patient’s condition and the correct execution of manoeuvres. The ANScovery System can manage the following tests: HUTT, VM, DB, CF, CP, right and left

Page 9 of 24

sinus massage, standing, IE (or hand grip), mental stress and hyperventilation. The standard sequence of tests performed in the University of Bologna ANS laboratory [14] includes: 1) Head-up tilt test (HUTT); 2) Three repetitions of the Valsalva manoeuvre (VM); 3) Deep breathing (DB); 4) Cold face (CF);

cr

Tests are executed and evaluated following standard procedures [2-4, 6, 9, 15].

ip t

5) Isometric exercise (IE) or hand grip.

2.6 Software analysis

us

ANScovery software manages the visualization and storage of real-time signals and video recordings. Visualization can be customized by the user and some basic analysis procedures can be

time and amplitude measurements, marking events).

an

applied both in real time and in the post-processing phase (i.e. digital filtering, tachogram creation,

To date, the ANScovery System has fully analysed and correlated HR and BP during the different

• Detection of events;

M

tests [3, 5]. At the end of acquisition, the analysis software follows four steps:

d

• Association between event and manoeuvre according to standard protocols;

Ac ce pt e

• Manoeuvre analysis with subsequent visualization of the sequences or identified points (SBP, DBP, HR) following the criteria explained below;

• Final report (Microsoft Word format) including the results of the parameters evaluated for each manoeuvre and concurrent data dispatch to the web database.

The user is able to check and correct each step of the analysis. Some more analysis algorithms have been implemented such as frequency analysis (Fast Fourier Transform and autoregression), HR variability, baroreceptor sensitivity and other simple calculations like signal derivative, integrals, and digital filtration (low-pass, high-pass, band-pass, notch).

Page 10 of 24

To guarantee the compatibility between ANScovery software and other commercial stand-alone devices able to acquire at least pressure and ECG (FMS Finometer, Portapres and Finapress) data and beat-to-beat files exported from the different machines can be read allowing the same ANS analysis.

2.7 Database

ip t

ANScovery software can be connected to a web-server database (ANS Web-application) to store the parameters of interest of all patients. The patient’s case history and clinical data (diagnosis,

cr

drugs, etc.) must be inserted directly by the physician, whereas physiological values related to ANS

evaluation are directly stored in the database by the ANScovery software. All data are encrypted

us

and stored in a safe server according to current regulations concerning privacy and data protection. Figure 6 (a) shows two database screenshots related to general patient information (a) and VM

an

analysis (b).

The whole software package, including the acquisition and analysis modules, TM and remote database, was developed following the requirements of the IEC 62304. As a final step, during

M

clinical and technical validation, the ANScovery System was CE-marked according to the medical

d

devices EU directives (93/42/CEE and 2007/47).

Ac ce pt e

In order to demonstrate the feasibility of the ANScovery System, we extracted from our database two groups of subjects who underwent the standard battery of tests performed in our laboratories: 37 healthy subjects (age: 63±8 years, 24 males and 13 females) and 33 patients (age: 64±7 years, 27 males and 6 females) with a diagnosis of autonomic neuropathy. Each subject was submitted to a standard battery of tests as reported in paragraph 2.5. The duration of tilt test was 10 minutes. Table 2 shows the parameters evaluated for each manoeuvre [3, 5] with its description.

Page 11 of 24

Test

Description response to HUTT as the difference (∆) between values at 3 min and basal values of SBP as the mean value of the last 5 min of supine rest preceding HUTT

∆DBP (mmHg)

response to HUTT as the difference (∆) between values at 3 min and basal values of DBP as the mean value of the last 5 min of supine rest preceding HUTT

∆HR (bpm)

response to HUTT as the difference (∆) between values at 3 min and basal values of HR as the mean value of the last 5 min of supine rest preceding HUTT

VR

Valsalva ratio as the ratio between the highest HR reached in phase II and the lowest HR of phase IV reflex bradycardia

∆BP IV (mmHg)

∆ BP IV as difference (∆) between max SBP phase IV and basal values of SBP as the mean value of the last 10 beats preceding Valsalva manoeuvre

Deep breathing

∆IE (bpm)

Sinus arrhythmia during DB as average of the 10 shortest R-R intervals during inspiration – average of the 10 longest R-R during expiration

Cold face

∆SBP (mmHg)

Response to CF as the difference (∆) between the highest value reached during test and basal values of SBP as the mean value of the last 20 beats preceding the test

∆DBP (mmHg)

Response to CF as the difference (∆) between the highest value reached during test and basal values of DBP as the mean value of the last 20 beats preceding the test

∆HR (bpm)

Response to CF as the difference (∆) between the lowest value reached during test and basal values of HR as the mean value of the last 20 beats preceding the test

∆SBP (mmHg)

Response to HG as the difference (∆) between values after 5 min of isometric effort and basal values of SBP as the mean value of the last 3 min preceding isometric effort

∆DBP (mmHg)

Response to HG as the difference (∆) between values after 5 min of isometric effort and basal values of DBP as the mean value of the last 3 min preceding isometric effort

∆HR (bpm)

Response to HG as the difference (∆) between values after 5 min of isometric effort and basal values of HR as the mean value of the last 3 min preceding isometric effort

cr

us

an

Ac ce pt e

Handgrip

M

Valsalva manoeuvre

ip t

∆SBP (mmHg)

d

Head-up tilt test

Evaluated parameter

The Kolgomorov-Smirnov test was used to verify the normal distribution of data; then a parametric Student t-test was performed to compare the results of the two groups. A p value

A new integrated instrumental approach to autonomic nervous system assessment.

The autonomic nervous system (ANS) regulates involuntary body functions and is commonly evaluated by measuring reflex responses of systolic and diasto...
403KB Sizes 6 Downloads 10 Views