EP1150733A1

EP1150733A1 - Gas supply for sleep apnea

Info

Publication number: EP1150733A1
Application number: EP00905108A
Authority: EP
Inventors: Hossein Nadjafizadeh; Pascal Nicolazzi; Véronique GRILLER LANOIR
Original assignee: Mallinckrodt Developpement France SAS
Current assignee: Mallinckrodt Developpement France SAS
Priority date: 1999-02-12
Filing date: 2000-02-10
Publication date: 2001-11-07
Also published as: US7370650B2; JP2002540820A; FR2789593A1; JP4776077B2; US6814074B1; US7992557B2; FR2789593B1; US20080196724A1; WO2000047262A1; US20050081854A1; CA2361631C; CA2361631A1

Abstract

The invention concerns methods for controlling an apparatus supplying air pressure to a patient suffering from sleep disorders such as apnea, the patient wearing a mask (MVA) through which pressurised air is supplied to his upper anatomical airways by the apparatus, which consist in: measuring the air pressure in the mask (MVA) and the air flow rate supplied to the mask (MVA); determining from the measured variables, the occurrence or not of signs denoting sleep disorders.

Description

GAS SUPPLY DEVICE FOR SLEEP APNEA

The invention relates to a method for controlling an apparatus for supplying air pressure to a patient suffering from sleep disorders.

The invention also relates to an apparatus for supplying air pressure to a patient suffering from sleep disorders.

These sleep disorders are respiratory and tend to awaken the patient.

These are for example apneas, hypopneas, acoustic vibrations or snoring, the limitation of the respiratory flow due to a tightening of the upper airways of the patient.

US-A-5,458,137 describes a method and device for controlling breathing in sleep disorders, which use multiple and variable pressure levels.

A pressure source supplies compressed respiratory gas at a relatively low pressure to the user's airways.

Pressure sensors monitor pressures and convert them into electrical signals.

Electrical signals are filtered and processed to extract specific characteristics such as duration and energy levels.

If these characteristics exceed selected thresholds of duration and energy level beyond a minimum period of time, the microprocessor indicates the presence of a respiratory sleep disorder. If a selected number of these events appear for a selected period of time, the microprocessor adjusts the pressure supplied by the source.

Document US-A-5 490 502 describes a method and an apparatus for optimizing the positive pressure controlled in order to minimize the air flow coming from a generator while ensuring that the limitation of flow in the patient's airways does not not produce.

It is planned to detect the flow limitation by analyzing a respiratory flow wave.

As soon as the presence of a flow limitation has been analyzed, the system determines an action to be carried out for adjusting the positive pressure commanded.

The pressure is increased, decreased or maintained depending on whether the flow limitation has been detected and depending on the previous actions implemented by the system.

Mention may also be made of documents US-A-5,335,654, EP-A-661,071 and EP-A-651,971.

The invention aims to improve the methods and devices of the state of the art, to automatically and continuously adapt the pressure delivered to the state of the patient and to prevent and prevent the appearance of disorders.

A first object of the invention is a method of controlling an apparatus for supplying air pressure to a patient suffering from sleep disorders such as apneas.

A second object of the invention is an apparatus for supplying air pressure to a patient suffering from sleep disorders such as apnea, implementing the method of supply.

The patient wears a mask through which pressurized air is supplied to his upper airways by the device. According to the invention, there is provided a control algorithm using a device output rate signal for the detection of apnea, hypopnea, rate limitation events, leaks, and using the analysis. pressure information to determine the presence of snoring, also called acoustic vibrations.

The pressure supplied to the patient's upper airway by the device can be kept constant, increased or decreased depending on the determination of the event that was performed by the control algorithm.

Thus, if no breathing is detected by the control algorithm in a predetermined minimum time dependent on a calculated average breathing time, the presence of apnea is determined.

This predetermined minimum apnea detection time is for example equal to a time constant, for example 10 s, added to a proportionality factor multiplied by the calculated average breathing time, this factor being for example equal to 5/8.

For each apnea, the output flow signal is amplified and filtered to determine the presence or absence of heart oscillations.

If cardiac oscillations were detected during the last elapsed time interval, for example equal to 5 s, then the apnea is classified as central and no command occurs in the algorithm.

If no cardiac oscillation has been detected in this time interval, the apnea is classified as obstructive, and the pressure is increased by a predetermined value once and, during the same apnea, twice more regularly, by example every 15 s.

The control algorithm compares variations in peak-to-peak flow during the patient's last breath compared to a predetermined number of previous breaths, for example equal to 8.

After each breath, a classification is made into: - normal breathing, if the last peak-to-peak flow value is within a determined range compared to the average value over the previous 8 breaths, for example from 40% to 150% or 140% of these;

- hypopneic respiration, if the last flow value is below this range; - hyperpneic breathing, if the last flow value is beyond this range.

A hypopnea determination is carried out if the detection of hypopneic respiration has occurred for at least a determined time, for example 10 s, and ends after a determined number of normal or hyperpneic breaths, for example equal to 2.

A determination of hypopnea leads to a determined increase in pressure, for example of 1 cm H2O first, then, during the same hypopnea, to an increase in pressure of another determined value and this regularly, for example of 0, 5 cm H2O every two hypopneic breaths.

The control algorithm analyzes and compares, breath by breath, the waveform of the respiratory flow with a sinusoidal waveform of the same period and the same slope.

After the comparison based on two criteria of form of flow, each breath is first classified as normal, intermediate or limited flow.

A final classification based on the combination of the classification of flow and the occurrence of snoring, changes the classification of breaths from normal to intermediate, respectively from intermediate to limited flow breathing.

A treatment is decided when a certain number, for example 2, successive breaths at a limited rate or a certain number, for example 5, of successive intermediate breaths take place after for example two normal breaths.

This treatment causes a determined increase in pressure, repeated regularly a certain number of times, for example 0.3 cm H2O three times every two breaths. For each breath, the pressure signal is amplified and filtered to detect the presence or absence of acoustic vibrations or snoring.

A valid snoring determination is made by the control algorithm, if the detected acoustic vibration has occurred at least for a certain time, for example 7% of the average duration of the last three breaths, and with a shorter period to a factor proportional to this average time, for example 120% of it.

In the case of a valid snoring, the algorithm increases the pressure by a determined value, for example by 1 cm H2O, if the last command due to a snoring occurred for more than a determined time, for example 1 minute.

An average leak is determined to be equal to the average flow during breathing.

The control algorithm continuously compares the current leak to a leak limit, which limit can be adjusted from the pressure.

If the current leak exceeds the limit, all pressure increase commands generated following event detections are blocked.

After detecting an apnea or a snoring event or a hypopnea command or a treatment decision, the algorithm will decrease the pressure by a determined value, for example by 0.5 cm H2O, in a first step after a determined time, for example 5 minutes, and regularly for the following decreases, for example every minute.

A determined holding pressure, for example 8 cm H2O is supplied by the device if no breath has been detected for a determined time, for example two minutes, or if the pressure supplied has been greater than or equal to a value determined for a determined time, for example at 17 cm H2O for 10 or 30 minutes. An advantage of the method is an automatic adaptation of the detection criteria to the respiratory characteristics of the patient.

Thus, any modification of the respiratory rhythm is taken into account by the algorithm to carry out the detection.

The fact of using an average value of respiratory cycle time over a certain number of previous respiratory cycles has the effect of regularly monitoring variations in the respiratory cycle and amplitude and better detection.

The invention will be better understood on reading the description which follows, made with reference to the figures.

Figure 1 is a diagram showing the apparatus for supplying air pressure to the patient.

FIG. 2 represents a decision-making algorithm for a first pressure increase command.

FIG. 3 represents an algorithm for indicating the appearance of disorders.

Figure 4 shows a breathing qualification algorithm.

FIG. 5 represents an algorithm for detecting central and obstructive apnea and for controlling pressure as a function of the result of these detections, as well as an algorithm for reducing pressure according to the previous appearance or not of events representative of disorders of the sleep.

FIG. 6 represents an algorithm for qualifying a normal ventilation cycle, hyperventilation or hypoventilation.

FIG. 7 represents an algorithm for detecting hypopneic respiration.

FIG. 8 represents an algorithm for detecting hyperpneic respiration. Figure 9 shows a normal breathing detection algorithm.

Figure 10 shows a high pressure detection algorithm.

FIG. 11 represents an algorithm for detecting mask leakage.

FIG. 12 represents an algorithm for detecting acoustic vibrations.

FIG. 13 represents an algorithm for reducing pressure in the event of detection of acoustic vibrations.

In FIG. 1, the apparatus for supplying air pressure to a patient comprises a central pressure treatment and control unit U, a controlled MPD module for supplying pressure, an MVA mask for the patient's upper airways , a CF duct for supplying air pressure from the MPD module to the MVA mask.

The air flow supplied to the patient and the air pressure prevailing in the MVA mask are measured by a CDAF sensor of the supplied air flow, connected to the central unit U and by a CPM pressure sensor in the mask. MVA, connected to the central unit U.

It is determined from the measured variables, whether events representative of sleep disorders appear or not.

The algorithms of the method according to the invention are implemented by software integrated into the central unit U.

In FIG. 2, it is determined from the measured variables whether the patient's current respiratory cycle corresponds to a predetermined valid respiratory cycle.

A BLN indicator of disturbance appearance is placed in a first ON state of appearance of disorders, if the appearance of one or more of the events representative of sleep disorders is determined. The BLN indicator is put in a second OFF state of absence of disorders, if the appearance of the events representative of sleep disorders is not determined.

There is a first CCAR number of valid respiratory cycles determined since the last pressure command.

There is a second number CCON of valid respiratory cycles determined since the last passage of the BLN indicator to the first ON state.

There is a third number RC of successive passages from the indicator BLN from the second state OFF to the first state ON.

When the BLN indicator is in the first ON state, command C1 commands a first determined increase in the supplied air pressure, when, at the same time:

• the current respiratory cycle has been determined to be valid;

• the first CCAR number is greater than a first predetermined whole RP number; • the second number CCON corresponds to another or more second predetermined whole numbers N;

• the third RC number is greater than or equal to a third predetermined integer X.

When the BLN indicator goes from the second OFF state to the first ON state, the first determined increase in supplied air pressure is controlled by command C1 when, only and at the same time:

• the current respiratory cycle has been determined to be valid;

• the first CCAR number is greater than a first predetermined whole RP number;

In one embodiment, the second integers N are between 1 and 300. In another embodiment, the second integers N are the first three multiples of a determined integer No.

In another embodiment, the second integers N are 2, 4 and 6 respectively, No being equal to 2.

In another embodiment, the first predetermined integer RP is between 1 and 255.

In another embodiment, the first predetermined integer RP is equal to 10.

In another embodiment, the third predetermined integer X is between 1 and 100.

In another embodiment, the third predetermined integer X is equal to 1.

In another embodiment, the first determined command C1 for increasing the pressure is less than + 10 mbar.

In another embodiment, the first determined command C1 to increase pressure is substantially equal to + 0.3 mbar.

The first and third CCAR numbers are reset to 0; CR of valid respiratory cycles counted and passages counted, after the second count CCON counted of valid cycles has reached the greater of the second predetermined integers N.

The second number counted CCON is reset to 0 when the BLN indicator goes from the second OFF state to the first ON state.

The predetermined valid respiratory cycle corresponds to a maximum respiratory flow greater than a predetermined flow value such as 50 ml / s, to a inspiratory volume greater than a predetermined volume value such as 0.05 I and an absence of saturation during flow detection.

In FIG. 3, to give the ON or OFF state to the BLN indicator of occurrence of troubles, - an ER state variable is initialized when the device is switched on to a third state d absence of NIR processing and the BLN indicator in the second OFF state.

Then, sequentially, - on the basis of the variables measured, the respiratory cycles are qualified as belonging to different categories such as limited flow cycle, intermediate cycle, normal cycle and invalid cycle, each corresponding respectively to weightings RSV0, REVO; RSV1, REV1; RSV2, REV2; 0, 0;

- the weights of the category of the cycle currently qualified are assigned to first and second SV accumulators; Weighting EV;

- if the qualified cycle belongs to the invalid cycle category, the ER state variable is returned to the third NIR state and the BLN indicator to the second OFF state and a first FLC counter is initialized to a predetermined value.

If the state variable ER corresponds to the third NIR state:

• if the value of a first accumulator SV is less than a first comparative value, the counter FLC is reset to its predetermined value;

• if the value of the first accumulator SV is substantially equal to its first comparative value, action is dispensed with and one passes to the next test; • if the value of the first accumulator SV is greater than its first comparative value, the state variable ER is passed to a fourth state PR of processing possibility and the indicator BLN is set to the second state OFF.

If the state of the state variable ER corresponds to the fourth state PR and • if the value of the first accumulator SV is less than its first comparative value, the first counter FLC is reset to its predetermined value, the variable d 'is reset ER state and the BLN indicator respectively in the third and second NIR states; OFF; • if the value of the first accumulator SV is substantially equal to its first comparative value, action is dispensed with and one passes to the next test;

• if the value of the first accumulator SV is greater than its first comparative value, the first FLC counter is made to take its previous value added with the value of the first accumulator SV and if then the value of the first FLC counter is greater than or equal to a stop high predetermined RMS:

• a second NC counter is reset to a predetermined value;

• the state variable ER is passed to a fifth state IR for processing; and

• the BLN indicator is set to the first ON state.

If the state of the ER state variable corresponds to the fifth IR processing state:

If the value of the second EV accumulator is greater than a second comparative value, the second counter NC is made to take its previous value added to the value of the second EV accumulator and

^• if then the value of the second counter NC is greater than or equal to a lower stop RME, it resets the state variable ER and the indicator BLN to the third and second NIR states respectively; OFF and the first and second FLC counters are reset; NC to their respective predetermined values;

^• or otherwise, the BLN indicator is set to its first ON state;

• if the value of the second accumulator EV is less than its second comparative value, the second counter NC is reset to its respective predetermined value and the indicator BLN is set to the first state ON

• if the value of the second EV accumulator is substantially equal to its second comparative value, action is dispensed with.

In one embodiment, the weightings RSV2, REV2; RSV1, REV1; RSVO, REVO; 0, 0 corresponding to the normal cycle, intermediate cycle, limited flow cycle and invalid cycle, are respectively substantially equal to -1; 1; 5 and 0 for the first SV accumulator and are respectively substantially equal to 1; -1; -1 and 0 for the second EV accumulator.

The first and second comparative values and the predetermined initialization values of the first and second FLC counters; NC are each substantially equal to 0. RMS high and low stops; RME are respectively substantially equal to 10 and 2.

In FIG. 4, the respiratory cycles measured are qualified.

The predetermined valid respiratory cycle corresponds to a maximum inspiratory flow greater than a predetermined flow value such as 50 ml / s, to an inspiratory volume greater than a predetermined volume value such as 0.05 I, to a lack of saturation during flow detection, at a measured inspiratory time comprised in a predetermined interval such as from 0.5 s to 6 s and at a measured duration of respiratory cycles comprised in another predetermined interval such as from 1.5 s to 20 s .

If the measured respiratory cycle is determined to be valid, then

- We calculate an equivalent sinusoidal curve respecting predetermined characteristics compared to the inspiratory curve of the inspiratory cycle measured;

- a surface area criterion CS is calculated proportional to the ratio of the area delimited by the inspiratory curve to the area delimited by the equivalent sinusoidal curve, each taken over the same time interval, included in the inspiratory phase of the measured respiratory cycle ;

- a correlation criterion CC is calculated between the inspiratory curve of the measured inspiratory cycle and the equivalent sinusoidal curve;

- if the calculated correlation criterion CC is greater than or equal to a first predetermined normal limit LN, and if the calculated surface criterion CS is greater than a second predetermined surface limit LS, the measured respiratory cycle is qualified as normal and otherwise, it is called a limited flow cycle.

If the measured respiratory cycle has been qualified as a limited flow cycle, • if the calculated surface criterion CS is greater than a third predetermined expert LE limit, the measured respiratory cycle is normalized, • or otherwise, If the surface area criterion CS calculated is greater than a fourth predetermined limit LD of flow rate, the measured respiratory cycle of intermediate is requalified,

• and if not, it is called a limited flow cycle.

The second area limit LS, the fourth rate limit LD and the third expert limit LE being predetermined in ascending order.

The predetermined characteristics of the equivalent sinusoidal curve include a half-period substantially equal to the inspiratory time measured and a slope at the origin substantially equal to that of the inspiratory curve when it reaches substantially one third of its maximum amplitude.

In one embodiment, the calculated surface criterion CS is substantially equal to a hundred times the ratio of the areas taken each from substantially a quarter to three quarter of the duration of the inspiratory phase of the respiratory cycle measured.

The calculated correlation criterion CC is substantially equal to the maximum of a hundred times the correlation coefficients between the inspiratory curve and the equivalent sinusoidal curve taken respectively on the second half of the inspiratory phase and on the whole of the latter.

The first, second, fourth and third LN limits; LS; LD; LE being respectively between 45 and 100; 0 and 100; 0 and 100; 0 and 100 and being for example substantially equal to 87; 40; 60 and 90 respectively.

In FIG. 5, obstructive apneas and central apneas are detected.

The algorithm represented in FIG. 5 is carried out during each of several (NINT) consecutive predetermined time intervals TACG ^' ).

The predetermined consecutive TACG time intervals are those included in a predetermined apnea detection PDAC period. In this algorithm, for example, by hardware means such as analog or digital filters, the oscillations of the measured flow curve are detected, which are of frequencies included in a frequency range P2.

Then it is detected whether the amplitude of the detected oscillations of the measured flow curve passes successively above and then below a first predetermined SAC threshold for central apnea or if this amplitude remains below the first SAC threshold for central apnea , as shown diagrammatically on the right of FIG. 5 by:

- the appearance of an obstructive apnea flow curve (curve constantly below the first SAC threshold);

- the shape of a flow curve in central apnea (curve passing several times successively above and then below the first SAC threshold).

In the presence of at least one detection of a passage above and then below the first SAC threshold, there is a CAC (D) detection of central apnea.

Then, at each PDAC apnea detection period,

• the sum SIG of the numbers CAC (i) of central apnea detections counted, successively over the (D + 1) last periods of apnea detection, is made, • C2 is commanded a second predetermined increase in delivered air pressure if the sum SIG of the CAC (i) numbers of detections counted is less than or equal to a second predetermined SQAC threshold for qualification of central apnea;

• a maintenance of the delivered air pressure is commanded, if the sum SIG of the CAC (i) numbers of detections counted is greater than the second SQAC threshold.

In one embodiment, the second threshold SQAC for qualification of central apnea is between 0 and 50, and is for example substantially equal to 10.

The predetermined consecutive TACG) time intervals correspond to ten (NINT) consecutive time intervals of each substantially 100 ms, the PDAC apnea detection period corresponding substantially to 1 s.

The second C2 pressure increase command is between 1 and 10 mbar and is for example substantially equal to + 1 mbar. The number (D + 1) of PDAC apnea detection periods, over which the CAC (i) numbers counted for central apnea detection are calculated is substantially equal to 5.

The second range P2 of oscillation frequency is between substantially 2.5 and 47 Hz.

The CAC (i) counted numbers of central apnea detections are reset to 0 when the device is switched on.

There is also shown in FIG. 5 an algorithm for reducing pressure according to the previous appearance or not of events representative of sleep disorders.

According to this algorithm, represented at the bottom of FIG. 5, the pressure P measured is compared with a predetermined pressure value MPL.

After determining the occurrence of one or more of the events,

If the pressure P measured is lower than the predetermined MPL value, a third command C3 of predetermined pressure reduction is performed,

If the pressure P measured is greater than or equal to the predetermined MPL value, a fourth predetermined command C4 of pressure reduction is carried out,

• then, if no occurrence of an event has been detected after one or more of the C3 commands; C4 pressure reduction, the fourth predetermined command C4 pressure reduction is performed.

The fourth pressure reduction command C4 is such that it causes a greater pressure reduction per unit of time than that caused by the third pressure command C3.

In one embodiment, the fourth command C4 for pressure reduction is substantially -0.5 mbar / 1 minute and the third command C3 for pressure reduction is substantially -0.5 mbar / 5 minutes, the comparative MPL value of pressure is between 4 and 19 mbar and is for example substantially equal to 17 mbar.

This pressure reduction algorithm as a function of the appearance or not of events is implemented after that of detection of central and obstructive apneas as represented in FIG. 5 but is also implemented, in embodiments not shown, after other algorithms such as:

- that of treatment decision-making, when the BLN indicator has gone from the first ON state to the second OFF state; - that of hypopneic respiration detection described below;

- that of acoustic vibration detection described below.

In FIGS. 6 to 9, the respiratory cycles are qualified as hyperventilated, hypoventilated or cycles with normal ventilation and pressure commands are generated according to the qualifications performed.

At each end of the measured respiratory cycle, the average amplitude AM is calculated over a fourth predetermined number Y4 of previous respiratory cycles.

As shown in Figure 7, if the measured amplitude of the last respiratory cycle is less than the calculated average amplitude AM multiplied by a first predetermined hypopnea factor FHO, then the duration TC is added to a CTHO counter in hypopnea time of the last respiratory cycle measured,

• if the current value of the time counter in hypopnea CTHO is greater than or equal to a minimum hypopnea time TMHO, it is controlled by a command

C5 a fifth predetermined increase in pressure,

• after the end of a fifth predetermined number Y5 of respiratory cycles following the fifth command C5 to increase pressure, C6 commands a sixth predetermined increase in pressure; • after the end of a sixth predetermined number Y6 of respiratory cycles, greater than the fifth number Y5, according to the fifth command C5 of increase in pressure, a seventh increase in pressure is commanded by a command C7. The CTHO hypopnea time counter is initialized to 0 when the device is switched on.

In one embodiment, the fourth determined number Y4 of respiratory cycles for calculating average amplitude is substantially equal to 8.

The first predetermined factor FHO for hypopnea is between 1 and 100% and is for example substantially equal to 40%.

The minimum TMHO hypopnea time is between 1 s and 25 s and is for example substantially equal to 10 s.

The fifth and sixth predetermined numbers Y5; Y6 of respiratory cycles are substantially equal to 2 and 4 respectively.

The fifth predetermined pressure increase C5 is between 0.1 mbar and 10 mbar and is for example substantially equal to + 1 mbar.

The sixth and seventh increases C6; C7 predetermined pressure are each less than the fifth command C5 and are for example each substantially equal to half of the fifth increase C5 pressure.

As shown in Figures 8 and 9, if the measured amplitude of the last respiratory cycle is greater than or equal to the calculated average amplitude AM multiplied by the first hypopnea factor FHO, then the TCM time of average respiratory cycles is calculated over a seventh predetermined number Y7 of previous cycles.

If the measured duration TC of the last cycle is greater than an eighth predetermined number Y8 multiplied by the average respiratory cycle time calculated TCM, the measured duration TC of the last cycle, multiplied by a second factor F2, is added to the time counter in hypopnea CTHO hypopnea.

If the measured amplitude of the last respiratory cycle measured is greater than a third factor F3 of hyperventilation, greater than the first factor FHO of hypopnea, multiplied by the calculated average amplitude AM, we qualify the last hyperventilated cycle, we increment by one a counter of hyperventilated cycles CCH, we reset to 0 a CCN counter of cycles with normal ventilation and

• if the value of the CCH counter for hyperventilated cycles is greater than or equal to a ninth predetermined number Y9,

• if the duration of the last TC cycle is greater than or equal to the eighth number Y8 multiplied by the calculated average cycle time TCM, the second factor F2 multiplied by the duration of the last respiratory cycle TC is added to the CTHO counter for time in hypopnee, ^" and if not, the CTHO counter in hypopnea is reset to 0; then a counter CCHO of hypoventilated cycles is reset to 0 and the mean amplitude AM of respiratory cycle is calculated on the predetermined number Y4 of previous respiratory cycles.

If the amplitude measured from the last respiratory cycle measured is less than or equal to the third factor F3 multiplied by the average amplitude calculated AM, the last cycle cycle with normal ventilation is qualified, the CCH counter for hyperventilated cycles is reset to 0 and increments the CCN counter of cycles with normal ventilation by one, and • if the value of the CCN counter of cycles with normal ventilation is greater than or equal to a tenth predetermined number Y10,

• if the duration of the last TC cycle is greater than or equal to the eighth number Y8 multiplied by the calculated average cycle time TCM, the second factor F2 multiplied by the duration of the last TC cycle is assigned to the CTHO counter in time in hypopnee and reset at 0 the CCN cycle counter with normal ventilation,

• and if not, the CTHO counter in hypopnea is reset to 0; then the CCHO counter of hypoventilated cycles is reset to 0 and the average amplitude of the respiratory cycle is calculated on the predetermined number Y4 of respiratory cycles.

In one embodiment, the second factor F2 is substantially equal to 5/8.

The third hyperventilation factor F3 is between 100% and 200% and is for example substantially equal to 140%. The seventh, eighth, ninth and tenth predetermined numbers Y7; Y8; Y9; Y10 are respectively substantially equal to 3; 2; 2; and 2.

In FIG. 10, it is detected if the pressure is too high.

If the measured pressure P is less than a predetermined high pressure value PH, a time counter in high pressure TPH is reset to 0.

If the value of the time counter for high pressure TPH is greater than a maximum time for high pressure TMPH and

• if the maximum set pressure value Pmaxi is less than a predetermined safety pressure value PSEC, the pressure P is controlled at this maximum set pressure value Pmaxi; • if the minimum set pressure value Pmini is greater than a predetermined safety pressure PSEC value, the pressure P is controlled at this minimum set pressure value Pmini;

• if the two previous conditions are not fulfilled, the pressure P is controlled at the safety pressure value PSEC, then

• the TPH high pressure time counter is reset to 0.

In one embodiment, the high pressure value PH is between 10 mbar and 25 mbar and is for example substantially equal to 17 mbar.

The maximum high pressure time TMPH is between 1 and 100 minutes and is for example substantially equal to 10 minutes or 30 minutes.

The PSEC safety pressure value is approximately 8 mbar.

In FIG. 11, an air leak is measured, substantially equal to the average flow during the patient's breathing. If the measured air leakage is greater than a predetermined NFM leakage level, the pressure increase commands are invalidated.

In one realization, NFM = A x Pfiltrée + B.

According to this formula, the predetermined level of leakage NFM is substantially equal to a coefficient of leakage multiplied by a filtered air pressure in the mask, added to a coefficient B additive of leakage, the coefficient A of leakage being between 0 and 10 l / minute.mbar and for example being substantially equal to 2.5 l / minute.mbar.

The additive leakage coefficient B is between 0 and 100 l / min and is for example substantially equal to 50 l / min.

In FIG. 12, it is detected whether the measured pressure curve has oscillations, such as acoustic vibrations, included in a frequency range P1.

This detection is carried out for example by hardware means such as analog or digital filters.

The time RF1 of presence of oscillations detected between two successive absences of detected oscillations is measured and the time RF0 of absence of oscillations detected between two successive presences of detected oscillations;

If the sum of the measured times of absence and presence of oscillations detected RF0; RF1 is within a prescribed BIP time range; BSP.

If the time of presence of oscillations RF1 measured is greater than or equal to a minimum time of oscillations TMRH and if the value of a counter CTAR of time elapsed since the penultimate time that the preceding time conditions have been fulfilled, is greater than a prescribed waiting time TAR, C8 commands an eighth predetermined increase in pressure and resets the elapsed time counter CTAR to 0. The acoustic vibration detection and control algorithms in the event of acoustic vibrations are implemented at prescribed time intervals, in particular regularly and for example every 100 ms.

At the start of the acoustic vibration detection algorithm represented in FIG. 12, if the value of the CTAR elapsed time counter is less than the prescribed waiting time TAR, this counter of the time interval is incremented (INC CTAR) prescribed mentioned above.

If the sum of the measured times of presence and absence of oscillations detected RFO; RF1 is below the prescribed BIP time range; BSP or if the measured presence time of detected oscillations RF1 is less than the minimum oscillation time TMRH,

• the measured time RFO of absence of detected oscillations is replaced by the sum of the measured times of absence and presence of detected oscillations RFO; RF1, then

• the measured time of presence of detected oscillations RF1 is reset to 0.

If the sum of the measured times of absence and presence of oscillations detected RFO; RF1 is beyond the predetermined time range BIP; BSP or a predetermined maximum time TCMax, each of the measured times of absence and presence of oscillations detected RFO is reset to 0; RF1.

If the two conditions mentioned above on the sum of the presence and absence times RF1, RFO and on the presence time RF1 are not fulfilled, each of the measured absence and presence times of 0 is reset to 0 oscillations detected RFO; RF1.

In one embodiment, the predetermined maximum time TCMax is substantially equal to twice the time TCM of average respiratory cycle over the last three measured cycles.

The prescribed BIP time range; BSP is substantially between 10% and 120% of the calculated average cycle time TCM. The minimum oscillation time TMRH is substantially equal to 7% of the calculated average cycle time TCM.

The prescribed waiting time TAR is between 1 and 30 minutes and is for example substantially equal to 1 minute.

The eighth command C8 for increasing the pressure is between 0.1 mbar and 10 mbar and is for example substantially equal to 1 mbar.

The range P1 of oscillation detection frequency is between substantially 30 and 300 Hz.

The chronology of the detected events is memorized and the memorized chronology is noted, for example after one night.

To this end, the central unit U of the device includes a memory, not shown, which can be written and read with the chronology of the events detected.

This chronology can be viewed for example on a monitor by reading the contents of the memory, via a computer not shown.

Claims

1. Method for controlling an apparatus for supplying air pressure to a patient suffering from sleep disorders such as apnea, the patient wearing a mask by which pressurized air is supplied to his upper airways by the device, in which:

- the air pressure in the mask and the air flow supplied to the mask are measured;

- it is determined from the measured variables whether events representative of sleep disorders appear or not, characterized in that

- the pressure (P) measured is compared to a predetermined pressure value (MPL);

- after determining the appearance of one or more of the events,

• if the pressure (P) measured is lower than the predetermined value (MPL), a third command (C3) is made to decrease the pressure, • if the pressure (P) measured is greater than or equal to the value (MPL) predetermined, a fourth predetermined command (C4) for reducing pressure is carried out,

• then, if no occurrence of an event has been detected after one or more of the pressure reduction commands (C3; C4), the fourth predetermined pressure reduction command (C4) is carried out.

2. Method according to claim 1, characterized in that the fourth control (C4) pressure reduction is such that it causes a greater pressure reduction per unit of time than that caused by the third control (C3).

3. Method according to any one of claims 1 and 2, characterized in that the fourth control (C4) pressure reduction is substantially -0.5 mbar / 1 minute and the third control (C3) pressure reduction is approximately 0.5 mbar / 5 minutes, the comparative pressure value (MPL) is between 4 and 19 mbar and is, for example, substantially equal to 17 mbar.

4. Method according to any one of claims 1 to 3, characterized in that:

- at the end of each measured respiratory cycle, the average amplitude (AM) is calculated over a fourth predetermined number (Y4) of previous respiratory cycles, - if the measured amplitude of the last respiratory cycle is less than the calculated average amplitude (AM) multiplied by a first predetermined hypopnea factor (FHO), then the duration (hypopnee time) is added to the counter (CTHO) TC) of the last measured respiratory cycle, • if the current value of the hypopnea time counter (CTHO) is greater than or equal to a minimum hypopnea time (TMHO), a fifth predetermined pressure increase is commanded (C5),

• after the end of a fifth predetermined number (Y5) of respiratory cycles following the fifth command (C5) to increase pressure, a sixth predetermined pressure increase is commanded (C6);

• after the end of a sixth predetermined number (Y6) of respiratory cycles, greater than the fifth number (Y5), following the fifth command (C5) to increase pressure, a seventh pressure increase is commanded (C7), - the hypopnea time counter (CTHO) being initialized to 0 when the device is switched on.

5. Method according to claim 4, characterized in that:

the fourth determined number (Y4) of respiratory cycles for calculating average amplitude is substantially equal to 8, the first predetermined factor (FHO) for hypopnea is between 1 and 100% and is for example substantially equal to 40%, the minimum time of hypopnea (TMHO) is between 1 s and 25 s and is for example substantially equal to 10 s, the fifth and sixth predetermined numbers (Y5; Y6) of respiratory cycles are substantially equal to 2 and 4 respectively, the fifth predetermined increase (C5) in pressure is between 0.1 mbar and 10 mbar and is for example substantially equal to + 1 mbar, the sixth and seventh increases (C6; C7) predetermined pressure each being less than the fifth command (C5) and for example each being substantially equal to half of the fifth increase (C5) in pressure.

6. Method according to claim 5, characterized in that: - if the measured amplitude of the last respiratory cycle is greater than or equal to the calculated average amplitude (AM) multiplied by the first hypopnea factor (FHO), then the time (TCM) of average respiratory cycles is calculated over one seventh predetermined number (Y7) of previous cycles; • if the measured duration (TC) of the last cycle is greater than an eighth predetermined number (Y8) multiplied by the calculated average respiratory cycle time (TCM), the duration (TC) is added to the hypopnea time counter (CTHO) measured from the last cycle, multiplied by a second factor (F2) of hypopnea,

• if the measured amplitude of the last respiratory cycle measured is greater than a third factor (F3) of hyperventilation, greater than the first factor (FHO) of hypopnea, multiplied by the calculated average amplitude (AM), we qualify the last hyperventilated cycle, a hyperventilated cycle counter (CCH) is incremented by one, a counter (CCN) of cycles with normal ventilation is reset to 0 and

• if the counter value (CCH) of hyperventilated cycles is greater than or equal to a ninth predetermined number (Y9),

• if the duration of the last cycle (TC) is greater than or equal to the eighth number (Y8) multiplied by the calculated average cycle time (TCM), we add to the counter (CTHO) of time in hypopnee the second factor (F2) multiplied by the duration of the last respiratory cycle (TC), - and if not, the counter (CTHO) of time in hypopnea is reset to 0; then a counter (CCHO) of hypoventilated cycles is reset to 0 and the average amplitude (AM) of respiratory cycle is calculated on the predetermined number (Y4) of previous respiratory cycles,

• if the measured amplitude of the last measured respiratory cycle is less than or equal to the third factor (F3) multiplied by the calculated average amplitude (AM), the last cycle cycle with normal ventilation is qualified, the counter is reset to 0 CCH) of hyperventilated cycles and the counter (CCN) of cycles with normal ventilation is incremented by one, and

• if the value of the counter (CCN) of cycles with normal ventilation is greater than or equal to a tenth predetermined number (Y10),

• if the duration of the last cycle (TC) is greater than or equal to the eighth number (Y8) multiplied by the calculated average cycle time (TCM), the second factor (F2) multiplied is assigned to the hypopnee time counter (CTHO) by the duration of the last cycle (TC) and the counter (CCN) of cycle with normal ventilation is reset to 0, • and if not, the counter (CTHO) of time in hypopnee is reset to 0; then the counter (CCHO) of hypoventilated cycles is reset to 0 and the average amplitude of the respiratory cycle is calculated on the predetermined number (Y4) of respiratory cycles.

7. Method according to claim 6, characterized in that the second factor (F2) is substantially equal to 5/8, the third hyperventilation factor (F3) is between 100% and 200% and is for example substantially equal to 140%, the seventh, eighth, ninth and tenth predetermined numbers (Y7; Y8; Y9; Y10) are respectively substantially equal to 3; 2; 2; and 2.

8. Method according to any one of claims 1 to 7, characterized in that: - if the measured pressure (P) is less than a predetermined high pressure value (PH), a pressure time counter is reset to 0 high (TPH); - if the value of the high pressure time counter (TPH) is greater than a maximum high pressure time (TMPH) and

• if the maximum set pressure value (Pmaxi) is less than a predetermined safety pressure value (PSEC), the pressure (P) is controlled at this maximum set pressure value (Pmaxi);

• if the minimum set pressure value (Pmini) is greater than a predetermined safety pressure value (PSEC), the pressure (P) is controlled at this minimum set pressure value (Pmini); • if the two previous conditions are not met, the pressure (P) is controlled to the safety pressure value (PSEC); then

• the high pressure time counter (TPH) is reset to 0.

9. Method according to claim 8, characterized in that the high pressure value (PH) is between 10 mbar and 25 mbar and is for example substantially equal to 17 mbar, the maximum high pressure time (TMPH) is between 1 and 100 minutes and is for example substantially equal to 10 minutes or 30 minutes, the safety pressure value (PSEC) is approximately equal to 8 mbar.

10. Method according to any one of claims 1 to 9, characterized in that: an air leak is measured, substantially equal to the average flow during the patient's breathing,

- if the measured air leakage is greater than a predetermined leakage level (NFM), the pressure increase commands are invalidated.

11. Method according to claim 10, characterized in that the predetermined level of leakage (NFM) is substantially equal to a coefficient (A) of leakage multiplied by a filtered air pressure in the mask, added to a coefficient (B) leakage additive, the leakage coefficient (A) being between 0 and 10 l / minute.mbar and for example being substantially equal to 2.5 l / minute.mbar, and the leakage additive coefficient (B) is between 0 and 100 l / min and is for example substantially equal to 50 l / min.

12. Method according to any one of claims 1 to 11, characterized in that:

- it is detected whether the measured pressure curve has oscillations, such as acoustic vibrations, included in a frequency range (P1), - the time (RF1) of presence of oscillations detected between two successive absences of oscillations is measured detected and the time (RFO) of absence of oscillations detected between two successive presences of detected oscillations;

- if the sum of the measured times of absence and presence of detected oscillations (RFO; RF1) is within a prescribed time range (BIP; BSP), - if the measured time of presence of oscillations (RF1) is greater or equal to a minimum oscillation time (TMRH) and

- if the value of a counter (CTAR) of time elapsed since the penultimate time that the previous time conditions have been met, is greater than a prescribed waiting time (TAR), an eighth is ordered (C8) predetermined pressure increase and the elapsed time counter (CTAR) is reset to 0.

13. Method according to claim 12, characterized in that:

- if the sum of the measured times of presence and absence of detected oscillations (RFO; RF1) is below the prescribed time range (BIP; BSP) or if the time of measured presence of detected oscillations (RF1) is less than the minimum oscillation time (TMRH),

• the measured time (RFO) of absence of detected oscillations is replaced by the sum of the measured times of absence and presence of detected oscillation (RFO; RF1), then

• the measured time of presence of detected oscillations (RF1) is reset to 0, and

- if the sum of the measured times of absence and presence of detected oscillations (RFO; RF1) is beyond the predetermined time range (BIP; BSP) or a predetermined maximum time (TCMax), we reset to 0 each of the measured times of absence and presence of detected oscillations (RFO; RF1); and otherwise

- each of the measured times of absence and presence of detected oscillations is reset to 0 (RFO; RF1).

14. Method according to any one of claims 12 and 13, characterized in that: the predetermined maximum time is substantially equal to twice the time (TCM) of average respiratory cycle over the last three measured cycles;

- the prescribed time range (BIP; BSP) is substantially between 10% and 120% of the calculated average cycle time (TCM); - the minimum oscillation time (TMRH) is substantially equal to 7% of the calculated average cycle time (TCM);

- the prescribed waiting time (TAR) is between 1 and 30 minutes and is for example substantially equal to 1 minute;

the eighth command (C8) for increasing the pressure is between 0.1 mbar and 10 mbar and is for example substantially equal to 1 mbar;

- The range (P1) of oscillation detection frequency is between substantially 30 and 300 Hz.

15. Method according to any one of claims 1 to 14, characterized in that the chronology of the events detected is memorized and the memorized chronology is noted, for example after one night.

16. Apparatus for supplying air pressure to a patient suffering from sleep disorders such as apnea, implementing a method according to any one of Claims 1 to 15, characterized in that it comprises a central pressure processing and control unit (U), a controlled pressure supply module (MPD), a mask (MVA) for the patient's upper airways, a duct (CF) for supplying air pressure from the module (MPD) to the mask (MVA), a sensor (CDAF) for supplied air flow, connected to the central unit (U) and a sensor (CPM) for pressure in the mask (MVA), connected to the central unit (U).

17. Apparatus according to claim 16, characterized in that it comprises a memory of the chronology of the detected events, the content of which is capable of being recorded, for example after one night.