WO1996032055A1

WO1996032055A1 - Monitoring the occurrence of apneic and hypopneic arousals

Info

Publication number: WO1996032055A1
Application number: PCT/AU1996/000218
Authority: WO
Inventors: Gregory Alan Colla; John William Ernest Brydon
Original assignee: Resmed Limited
Priority date: 1995-04-11
Filing date: 1996-04-11
Publication date: 1996-10-17
Also published as: US6363270B1; US6091973A; AUPN236595A0

Abstract

The occurrence of an arousal in a patient associated with an apneic or hypopneic episode can be determined. Sensors (10) are placed on a patient to obtain signals representative of at least two physiological variables, for example skin conductance, heart rate and blood oxygen concentration. The signals are conditioned by conditioning circuitry (12), then processed by a processor (18) to correlate at least two thereof. A coincident change in at least two of the processed signals is indicative of the occurrence of an arousal, that in turn indicates an apneic or hypopneic episode has occurred. A patient thus can be diagnosed as suffering conditions such as obstructive sleep apnea.

Description

Monitoring the Occurrence of Apneic and Hypopneic Arousals

Field of the Invention

This invention relates to methods and apparatus for the determination or monitoring of arousals that are indicative of an apneic or hypopneic episode. An "A/H episode", as used hereafter, is to be understood as including both obstructive apneas (lack of breathing) or hypopneas (reduction in breathing) occurring during sleep.

Background of the Invention People suffering from Obstructive Sleep Apnea (OS A) and related conditions experience many A/H episodes during sleep. The conventional treatment for OS A is the well known Continuous Positive Airway Pressure (CPAP) treatment. An A/H episode often has an associated arousal, which is a nervous system response to low blood oxygen level and/or high blood carbon dioxide level. The condition of OS A normally is diagnosed by laboratory based polysomnography (PSG). PSG involves the measurement of sleep and respiratory variables including EEG, EOG, chin EMG, ECG, respiratory activity, nasal airflow, chest and abdominal movements, abdominal effort and oxygen saturation. The data gathered leads to a calculation of the Respiratory Disturbance Index (RDI) which is the average number of arousals per hour due to respiratory disturbance. PSG is uncomfortable for a patient due to the placement of numerous electrodes on the patient's head or face and the wearing of a mask or nasal prongs. PSG is an expensive procedure and has the inconvenience of requiring the patient to attend a sleep clinic for a whole night requiring continuous technician attendance. In the clinic, cortical arousals can be detected by measuring and interpreting, either automatically or by eye, changes in the patient's EEG and EMG. Arousals are an important indicator of the amount and quality of sleep obtained by a patient and the degree of severity of the patient's apneas. To obtain a display of the EEG, electrodes must be attached to the patient's head and the signals from these electrodes passed through high gain amplifiers before subsequent display and recording. This is often inconvenient and time consuming; the signals also can become distorted or disappear due to electrode dislodgement or other artefacts such as sweating. To detect an arousal, the EEG, once displayed, requires interpretation either visually by a skilled operator or automatically by a computer-based analysis system. Apart from the difficulties of attaining accurate EEG data, EEG apparatus also is relatively more expensive than much other biomedical apparatus.

EEG data on its own is not an accurate determination of the occurrence of A/H episodes - an arousal can be due to other factors as noted.

Prior art disclosures that form background information to the present invention include (a) R. Ferber, et al, "ASDA Standards of Practice - Portable Recording in the Assessment of Obstructive Sleep Apnea", Sleep, 17(4): 378-392, 1994; (b) L. Ferini- Strambi, et al, "Heart rate variability during sleep in snorers with and without obstructive sleep apnea", Chest, 102(4): 1023-7, Oct. 1992; and (c) C. Guilleminault, et al, "A review of 50 children with obstructive sleep apnea syndrome", Lung. 159:275-287, 1981.

Summary of the Invention It is an object of the present invention to overcome or at least ameliorate one or more of the problems with the prior art. In one preferred form the invention seeks to provide a method and apparatus for monitoring the occurrence of arousals due to A/H episodes. The invention in another broad form has as an object the determination of an index of sleep quality that can be used as a diagnostic indicator of a person suffering from conditions such as OSA. Yet further, the invention has as an object the provision of diagnostic apparatus for the recording the occurrence of A/H episodes. A preferred embodiment particularly contemplates detecting sympathetic arousals indicative of A/H episodes by signals derived from a patient which are easier to detect than by EEG, and by use of apparatus that is less costly than that for EEG.

It has been determined that the coincident occurrence of change in the sympathetic physiological variable of sweating and another physiological variable, or of a change in at least two sympathetic physiological variables not necessarily including sweating, is indicative of an arousal associated with an apneic or hypopneic episode.

Therefore, the invention discloses apparatus for monitoring the occurrence of arousals associated with apneic or hypopneic episodes, the apparatus comprising: sensor means for obtaining two or more signals from a patient, one signal representative of skin conductance and the other signals each representative of a physiological variable; and means for correlating said skin conductance signal and at least one other said signal, a coincident change in at least said skin conductance signal and said at least one other said signal being indicative of the occurrence of an arousal.

The correlation can be other than a correlation in the strict mathematical sense, including coincident visual inspection, logical ANDing, indicator voting or weighted sums relative to a threshold.

Preferably, the sensor means comprises electrodes attachable to a patient's body. The signals derived from the electrodes can be signal-conditioned. Furthermore, the means for correlating includes data processing means. Each said signal can have an associated confidence level so that a weighting can be applied in the correlation.

The invention further discloses a method for monitoring the occurrence of arousals associated with apneic or hypopneic episodes, the method comprising the steps of: obtaining two or more signals from a patient, one signal representative of skin conductance and the other signals each representative of a physiological variable; correlating the skin conductance signal and at least one other said signal; and determining whether a coincident change in at least said skin conductance signal and said at least one other signal occurs as an indication of the occurrence of an arousal. The changes in physiological states can include reduction in blood oxygen saturation, change in heart rate, change in skin conductance, change in skin temperature, increase in blood pressure and changes in breathing sounds.

The invention further discloses apparatus for the diagnosis of apneic or hypopneic episodes resulting in arousals, the apparatus comprising: sensor means for obtaining from a patient two or more signals, each signal representative of a physiological variable and including skin conductance; and means for correlating said two or more signals, one or more coincident changes in at least said skin conductance signal and one other said signal being indicative of an arousal associated with an apneic or hypopneic episode.

The invention yet further discloses a method for the diagnosis of apneic or hypopneic episodes resulting in arousals, the method comprising die steps of: obtaining two or more signals from a patient, each signal representative of a physiological variable and including skin conductance; storing said signals as a time sequence; following completion of obtaining said signals, correlating said two or more signals; and determining whether a coincident change in at least said skin conductance signal and one other said signal occurs as an indication of the occurrence of an arousal associated with an apneic or hypopneic episode.

The invention yet further discloses apparatus for the acquisition of data from a patient from which the diagnosis of apneic or hypopneic episodes resulting in arousals can be made, die apparatus comprising: sensor means for obtaining from the patient two or more signals, each signal representative of a physiological variable and including skin conductance; and a memory for storing said signals in a time sequence for subsequent correlation. The invention yet further discloses a method for the determination for a patient of an index of sleep quality, me method comprising the steps of: obtaining two or more signals from the patient, each signal representative of a physiological variable; correlating said two or more signals; and recording instances where there is a coincident change in said correlated signals, the number of instances of coincidence being proportional to said index.

The invention yet further discloses apparatus for monitoring the occurrence of sympathetic arousals associated with apneic or hypopneic episodes, the apparatus comprising: sensor means for obtaining two or more signals from a patient, each signal representative of a sympathetic physiological variable; and means for correlating said two or more signals, a coincident change in at least two of said signals being indicative of the occurrence of a said arousal.

The invention yet further discloses a metiiod for monitoring the occurrence of sympathetic arousals associated with apneic or hypopneic episodes, the method comprising the steps of obtaining two or more signals from a patient, each signal representative of a sympathetic physiological variable; and correlating said two or more signals whereby a coincident change in at least two of said signals is indicative of the occurrence of a said arousal. Embodiments of the invention have the advantage of being non-invasive. This is of particular importance when diagnosis of conditions such as OSA must be undertaken on a baby or small child who may be disturbed by die conventional PSG apparatus.

One embodiment, relating to apparatus for the acquisition of data from which the diagnosis of apneic or hypopneic episodes, and d us the likelihood of the presence of conditions including OSA exist in a patient, can, in the first instance, substitute for a conventional sleep study and allow a candidate to collect data when in the home in a non-invasive manner. The data then can be analysed by a medical practitioner and a determination made of whether further detailed investigations need to be conducted.

Brief Description of the Drawings

Fig. 1 is a schematic block diagram of diagnostic apparatus embodying die invention; Fig. 2 is a schematic block diagram of a circuit for measuring skin conductance;

Fig. 3 is a plot of skin conductance with time;

Fig. 4 is a plot of skin conductance together with airflow rate, heart rate and Saθ₂; and

Fig. 5 is a plot of skin conductance measured periodically both before and during CPAP treatment is implemented.

Description of Preferred Embodiments The sequence of events leading to an arousal due to an apnea or hypopnea are as follows:

1. The patient's upper airway obstructs as they try to inspire, or the central nervous system drive to breathe is inhibited and d e patient makes no effort to breathe. 2. The oxygen level in the patient's arterial blood begins to fall and die carbon dioxide level rises.

3. After some time, which may exceed a minute, blood gas sensitive receptors in the carotid and/or aortic arteries force an arousal causing the patient to partially wake and recommence breathing. This sequence is associated wim a number of physiological changes (apart from those of the EEG referred to above) some of which are driven by me sympathetic nervous system which can be monitored:

1. A fall in blood oxygen saturation (SaO₂) with subsequent post-apneic rise. Saθ2 can be measured by pulse oximetry. 2. A decrease in heart rate as die patient's hypercapnia increases, followed by an increase when they arouse and restart breathing. Heart rate can be measured by ECG techniques. 3. A change in skin conductance due to sweating, measured by application of an electrical potential to measure conductance/impedance.

4. Increase in blood pressure. Blood pressure can be measured by continuous sphygomomanometry. 5. Change in constriction of blood vessels in die skin leading to a change in skin temperature and a change in tissue volume, particularly that in the fingers. This can be measured by plethysmography. 6. A reduction in breadiing sounds as the apnea occurs followed by a snore or snort when me apnea is broken. Taken separately, each of the above changes is only ambiguously related to an arousal after an obstructive event. Each also is subject to artefacts in die signal which are a function of how it is measured. However, by combining several of die above measurements, a much more artefact-free and specific diagnostic indicator of an arousal due to an A/H episode can be obtained. Fig. 1 shows a plurality of sensors 10j-10_n from which signals of the type described above are obtained from the patient. The sensors typically will comprise a sensing portion to detect physiological functions such as blood oxygen saturation, heart rate, skin conductance, blood pressure, skin temperature and audible noises, coupled wim amplifying/buffering devices to derive an electrical signal. The electrical signals each are passed to a corresponding preprocessing section

12]-12_n. The conditioning typically can include averaging, thresholding linearisation and differentiation for iose signals which exhibit a marked change. Each of die n signals has a characteristic pattern or range of patterns associated with apneic and hypopneic events. Each matching pattern serves as an indicator for an apneic or hypopneic event. A characteristic pattern for each signal is as follows.

1. The blood oxygen saturation signal (Saθ₂> is low-pass filtered, resulting in a baseline value. Signal values which differ from die baseline by a predetermined ti reshold, typically 3%, are positive indicators. 2. The heart rate signal is low-pass filtered to remove high frequency components, including d e sinus rhythm. This signal is dien differentiated, so diat changes in die heart rate are evident. A fall in the heart rate followed by a marked increase in heart rate is an indicator. When the difference between die slowest and fastest heart-rate is greater dian a specified minimum, typically 20 beats per minute, me indicator scores positively.

3. The skin conductance is low-pass filtered, resulting in a baseline value, men is high-pass filtered to accentuate changes. When die conductance increases above a specific ratio of die baseline, typically 100%, the indicator scores positively.

4. The blood pressure signal (both systolic and diastolic) is low-pass filtered to produce a baseline value. The elevation of d e blood pressure by more man a predetermined d reshold, typically 25% , results in a positive indicator. 5. The skin temperature is low-pass filtered to produce a baseline value.

If me skin temperature falls below a specific threshold from d e baseline, typically between 0.5 and 1.0°C, d e indicator is positive.

6. Breathing sounds are band-pass filtered, typically between 30 and

300 Hz. The energy from this signal is estimated by squaring or rectifying die filtered signal, and is used as a measure of snore intensity. A significant fall in me snore intensity, typically by 6 dB, followed by a sharp increase, is a positive indicator.

The conditioned signals 14-ι-14_n, together widi a quality signal 16ι-16_n giving a measure of the quality (confidence) of the signal, are passed to a multi-input processing device 18. The multiple input signals are processed algorithmically, widi ones of die signals 14 having die higher quality factor being given more weight. The detection of an isolated matching pattern does not signify an apneic or hypopneic event. However, die coincidence of two or more of these indicators increases me confidence of event detection. Various algorithms can be used:

1. Indicator ANDing, whereby each of me indicators must be positive to indicate an arousal.

2. Indicator voting, whereby most of the indicators must be positive to indicate an arousal.

3. Weighted sum, whereby each of the indicators has a predetermined weight. The products of each indicator and its weight are accumulated. The sum indicates an arousal if me sum is above a predetermined direshold.

The processing device 18 dius makes a determination of whether an arousal associated with an A/H episode has occurred, and further determines die percentage chance that this 'diagnosis' is correct, and yet further provides a determination of die duration of the A/H episode diat preceded me arousal. These determined functions are logged to a memory device 20 for subsequent analysis, for example by plotting wi i time and observing coincidences, and also can be indicated on a display device 22 in real time.

From mis information an index of arousals can be compiled, typically as arousals per hour, providing an alternative indicator to ROI, as previously discussed, to serve as a useful indicator of sleep quality.

In another embodiment die signals obtained from the sensors can be directiy stored in die memory device 20 in a time sequence for subsequent correlative processing by the processing device 18, or by a processor independent of die data acquisition module. This embodiment can serve as a data acquisition device for home usage, with the correlative processing occurring at a medical practitioner's surgery (for example) at a later time.

Particular embodiments now will be described, relating to clinical measurements performed on three patients, and for which d e physiological variables measured include skin conductance and heart rate (particularly cardiac rate, both of which are sympathetic variables).

Heart rate is measured by a conventional technique, namely three lead ECG. It is preferred diat the electrodes are located at d e patient's body's extremities to avoid apprehension and reduce d e likelihood of unwanted artifacts due to gross body movement.

The skin conductance measurement similarly is performed at me body's extremities, in one case being on me sole of a foot where sweat glands are concentrated. An alternate site is die forehead. Skin conductance is to be understood as the reciprocal of skin resistance, which is affected by sweating. In die complex number domain, conductance is die real part of skin admittance. As follows, skin also exhibits a reactive property (essentially a capacitance), die reciprocal of which is the susceptance.

Fig. 2 shows a schematic block diagram of a skin conductance measurement system 30 representing one form of a sensor 10 and preprocessing section 12. An oscillator 32 applies a 1 kHz signal of approximately 0.6 V_pp signal between two electrodes placed apart on die sole of the patient's foot, thereby applying an excitation to die patient's skin 34. The current passing dirough die electrodes is dependent upon d e conductance and susceptance of die skin and interconnecting tissue. A sense resistor 36 converts me current to a signal voltage cs, which is amplified by an amplifier 38. A demodulator circuit 40,42 removes d e effects of the skin susceptance (i.e. due to the skin's capacitance) from the amplified signal C_$, so that the resulting signal, SCL, is proportional to the skin conductance (level). The baseline of die SCL is removed by a high pass filter -44 and amplified (not shown) so mat responses to arousals are accentuated and more easily monitored. The resultant signal, SCR, is indicative of die skin conductance response.

Fig. 3 shows a time window of one half an hour during a sleep study performed on a patient known to be a sufferer of Obstructive Sleep Apnea. The upper trace represents die skin conductance SCL, and die lower trace the skin conductance response SCR. The sensed response at approximately t = 267.5 minutes corresponded with an apneic arousal o ierwise determined by PSG. Skin conductance d us is a reliable partial indicator of an arousal associated widi an A/H episode. Fig. 4 shows an excerpt from a sleep study conducted on a different patient showing a relevant 30 minute window. The patient is a known sufferer of OSA. The uppermost trace is heart rate filtered over one minute to remove sinus rhydim. The units are beats per minute, and me measurement is made wid a pulse oximeter. The second trace (in descending order) is blood oxygen saturation (Saθ₂) also measured with a pulse oximeter. The scale is 0-100%. The third trace is d e SCR phase reference signal and the fourth trace the SCL base reference signal as measures of skin conductance. The units of conductance are relative only. The lowermost trace is of respiratory airflow measured widi nasal cannulae and a pressure transducer.

Fig. 4 shows a continuous decrease in nasal airflow until approximately t = 11 minutes corresponding to a hypopneic event. At diis point nasal airflow increased wid die taking of a sudden deep breath. The SaU2 trace shows a corresponding fall of approximately 5%, which, because of heavy filtering, appears after die resumption of normal breaming, aldiough in fact was coincident widi d e sudden deep breath. The chain of events was shallow breadiing (hyponea), followed by a fall in Saθ₂, followed by a resumption of normal breathing. Associated wid die arousal causing die resumption of normal breathing is tachycardia, widi the heart rate increasing by about 20 bpm. The arousal is coincident in time with the sudden increase in skin conductance and heart rate. Thus diese two quantities can be utilised in concert as an indicator of arousals associated wi i A/H episodes. Interestingly, at approximately t = 20 minutes, mere is a sudden interruption in respiratory airflow as a result of the patient sighing, however diere is no corresponding indicator in either skin conductance or heart rate to suggest iat the cessation of flow was due to an apnea. The data indicates diat the correlation of skin conductance widi one or more other physiological variables (whedier associated widi the sympadietic nervous system or not) or any two or more sympathetic physiological variables (for example blood pressure and heart rate), is an accurate determinant of me occurrence of an arousal associated widi an A/H episode. Further, it is believed diat die correlation of two sympathetic physiological variables, not necessarily including skin conductance, and for example heart rate and blood pressure, also is an accurate determinant.

Fig. 5 shows a plot of die variance of skin conductance for die diird patient during a sleep study in which between t = 20:30 hours and t = 1 :15 hours no CPAP treatment was in place, and between t = 1 : 15 hours and t = 5:30 hours CPAP treatment was in place. The sampling points are at five minute epochs. The 10 and 90 centiles have been shown. The plot indicates die reduction in variance of skin conductance bodi before and after die commencement of CPAP treatment as an indication of die reduction in die number of arousals due to A/H episodes and dius die reliability of skin conductance as a partial indicator of A/H episodes.

In summary, embodiments of d e invention can serve to monitor me occurrence of apneic or hypopneic episodes, to diagnose die occurrence of A/H episodes resulting in arousals, to acquire data from which die diagnosis of A/H episodes resulting from arousals can be made, to determine an index of sleep quality, and to monitor the sympathetic nervous system.

Claims

CLAIMS:

1. Apparatus for monitoring me occurrence of arousals associated wid apneic or hypopneic episodes, d e apparatus comprising: sensor means for obtaining two or more signals from a patient, one signal representative of skin conductance and die odier signals each representative of a physiological variable; and means for correlating said skin conductance signal and at least one odier said signal, a coincident change in at least said skin conductance signal and said at least one odier said signal being indicative of me occurrence of an arousal.

2. Apparatus as claimed in claim 1, wherein said means for correlating includes a data processor for receiving and analysing said signals.

3. Apparatus as claimed in claim 2, wherein said data processing means analyses said signals by any one or more operations of mathematical correlation, logical

ANDing, indicator voting, and weighted sums relative to a direshold.

4. Apparatus as claimed in claim 3, further comprising a memory for storing said signals for analysis and/or for output to a reproduction device.

5. Apparatus as claimed in any one of claims 1 to 4, wherein said odier signals representative of a physiological variable include blood oxygen saturation, heart rate, skin temperature, blood pressure and breadiing sounds.

6. Apparatus as claimed in any one of claims 1 to 4, wherein said sensor means obtain signals representative at least of die patient's skin conductance and cardiac rate.

7. Apparatus as claimed in claim 4, further comprising an indicator coupled to said data processing means, and giving a visual indication of die occurrence of an arousal.

8. A method for monitoring the occurrence of arousals associated widi apneic or hypopneic episodes, die method comprising the steps of: obtaining two or more signals from a patient, one signal representative of skin conductance and die other signals each representative of a physiological variable; correlating die skin conductance signal and at least one odier said signal; and determining whether a coincident change in at least said skin conductance signal and said at least one other signal occurs as an indication of the occurrence of an arousal.

9. A method as claimed in claim 8, whereby said step of correlating includes any one or more of die processes of mathematical correlation, logical ANDing, indicator voting, and a weighted sum relative to a threshold.

10. A method as claimed in claim 8 or claim 9, whereby said odier signals representative of a physiological variable include blood oxygen saturation, heart rate, skin temperature, blood pressure and breadiing sounds.

11. Apparatus for die diagnosis of apneic or hypopneic episodes resulting in arousals, die apparatus comprising: sensor means for obtaining two or more signals from a patient, each signal representative of a physiological variable and including skin conductance; and means for correlating said two or more signals, one or more coincident changes in at least said skin conductance signal and one odier said signal being indicative of an arousal associated with an apneic or hypopneic episode.

12. Apparatus as claimed in claim 11, further comprising a memory for storing said signals in a time sequence.

13. A method for the diagnosis of apneic or hypopneic episodes resulting in arousals, die method comprising the steps of: obtaining two or more signals from a patient, each signal representative of a physiological variable and including skin conductance; storing said signals as a time sequence; following completion of obtaining said signals, correlating said two or more signals; and determining whed er a coincident change in at least said skin conductance signal and one other said signal occurs as an indication of die occurrence of an arousal associated widi an apneic or hypopneic episode.

14. Apparatus for the acquisition of data from a patient from which the diagnosis of apneic or hypopneic episodes resulting in arousals can be made, die apparatus comprising: sensor means for obtaining from the patient two or more signals, each signal representative of a physiological variable and including skin conductance; and a memory for storing said signals in a time sequence for subsequent correlation.

15. A method for die determination for a patient of an index of sleep quality, the method comprising d e steps of: obtaining two or more signals from die patient, each signal representative of a physiological variable; correlating said two or more signals; and recording instances where diere is a coincident change in said correlated signals, the number of instances of coincidence being proportional to said index.

16. A method as claimed in claim 15, comprising die further step of storing said signals as a time sequence, and performing said correlation only upon completion of said signal obtaining step.

17. A method as claimed in eimer one of claims 15 or 16, wherein one said signal is representative of skin conductance and said step of correlating includes at least said skin conductance signal.

18. A med od as claimed in eidier one of claims 15 or 16, wherein said physiological variables are associated only widi the sympathetic nervous system.

19. Apparatus for monitoring the occurrence of sympathetic arousals associated wid apneic or hypopneic episodes, die apparatus comprising: sensor means for obtaining two or more signals from a patient, each signal representative of a sympametic physiological variable; and means for correlating said two or more signals, a coincident change in at least two of said signals being indicative of me occurrence of a said arousal.

20. A mediod for monitoring the occurrence of sympathetic arousals associated wid apneic or hypopneic episodes, me method comprising the steps of obtaining two or more signals from a patient, each signal representative of a sympathetic physiological variable; and correlating said two or more signals whereby a coincident change in at least two of said signals is indicative of die occurrence of a said arousal.