Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
CA 02519041 2005-09-13
WO 2004/080516 PCT/CA2004/000382
TITLE OF INVENTION
IMPROVED SYNCHRONY BETWEEN END OF VENTILATOR CYCLES
AND END OF PATIENT EFFORTS DURING ASSISTED VENTILATION
REFERENCE TO RELATED APPLICATIONS
[0001] This application claims priority under the provisions of 35 USC 119(e)
from US Provisional Patent application No. 60/454,533 filed March 14, 2003 and
under
35 USC 119(a) from International Patent Application No. PCT/CA03/00976 filed
June
27, 2003.
BACKGROUND TO THE INVENTION
[0002] In assisted ventilation, ventilator cycles are triggered by patient
inspiratory efforts. There is no mechanism, however, to insure that ventilator
cycles
terminate at, or near, the end of inspiratory effort. Because the duration of
patient
inspiratory efforts (neural TI) varies over a wide range (0.5 to.2.5 seconds),
the lack of a
link between end of ventilator and patient inspiratory cycles often results in
ventilator
cycles extending well beyond the inspiratory effort (delayed cycling off) or
terminating
before the end of inspiratory effort, forcing exhalation when the patient is
still trying to
inhale. The delayed cycling off in particular is often severe with the
ventilator cycle
extending throughout the patient's expiratory phase (Figure 1). Because such
delayed
cycling off interferes with lung emptying during the patient's expiratory
phase, the next
breath usually begins before lung volwne has return ed to the neutral level.
This delays
ventilator triggering and often causes many patient cycles to be ineffective
in triggering
the ventilator (ineffective efforts, Figure 1).
(0003] Non-synchrony between patient and ventilator is extremely common.
Leung et al found that, on average, 28% of patient's efforts are ineffective
(Leung P,
Jubran A, Tobin MJ (1997), Comparison of assisted ventilator modes on
triggering,
patient effort, and dyspnea. Am J Respir Crit Care Med 155:1940-1948).
Considering
that ineffective efforts are the extreme manifestation of non-synchrony, less
severe, yet
substantial, delays must occur even more frequently. Non-synchrony is believed
to cause
distress, leading to excessive sedation and sleep disruption, as well as
errors is clinical
assessment of patients since the respiratory rate of the ventilator can be
quite different
from that of the patient (e.g. Figure 1).
CA 02519041 2005-09-13
WO 2004/080516 PCT/CA2004/000382
2
[0004] Cycling-off errors result from the fact that, except with Proportional
Assist Ventilation, current ventilator modes do not include any provision that
links the
end of ventilator cycle to end of patient's inspiratory effort. In the most
common form of
assisted ventilation, volume-cycled ventilation, the user sets the duration of
the inflation
cycle without knowledge of the duration of patient's inspiratory effort. Thus,
any
agreement between the ends of ventilator and patient inspiratory phases is
coincidental.
With the second most common form, pressure support ventilation, the inflation
phase
ends when inspiratory flow decreases below a specified value. Although the
time at
which this threshold is reached is, to some extent, related to patient effort,
it is to the
largest extent related to the values of passive resistance and elastance of
the patient. In
patients in whom the product [resistance/elastance], otherwise known as
respiratory time
constant, is high, the ventilator cycle may extend well beyond patient effort,
while in
those with a low time constant the cycle may end before the end of patient's
effort
(~ounes M (1993) Patient-.ventilator interaction with pressure-assisted
modalities of
ventilatory support. Seminars in Respiratory Medicine 14:299-322; Yamada Y9 Du
HL
(2000) Analysis of the mechanisms of expiratory asynchrony in pressure support
ventilation: a mathematical approach. J Appl Physiol ~~:2143-2150). The
present
invention concerns methods and devices to insure that the end of the
ventilator cycle
does not deviate substantially from the end of patient's effort. This is
achieved by
insuring that the duration of the ventilator's inflation phase is a
physi~logic fraction
(0.25-0.50) of the patient's respiratory cycle duration (patient TTOT). In
this fashion
enough time is available for lung emptying during the patient's expiratory
phase. By
extension, this also reduces dynamic hyperinflation at the onset of patient
efforts, thereby
also minimizing trigger delays and further improving synchrony.
[0005] In PCT/CA03/00976, filed June 27, 2003, (WO 2004/002561), from
which this application claims priority, I described an approach to generate a
semi-
quantitative estimate of the pressure waveform generated by the patient's
respiratory
muscles. This waveform can be used to identify the onset and end of patient's
efforts.
According to the aforementioned invention, the end of patient's inspiratory
effort,
detected by said invention, can be used to cycle off the ventilator, thereby
insuring
synchrony between the ends of ventilator and patient's inspiratory phases.
There is,
however, one potential complication to this approach. At times, end of patient
effort
CA 02519041 2005-09-13
WO 2004/080516 PCT/CA2004/000382
3
occurs soon after ventilator triggering. This is because inspiratory muscle
activity can be
inhibited if inspiratory flow is high, and the ventilator frequently delivers
excessive flow
soon after triggering. Thus, this approach may result in medically
unacceptable inflation
times. It was recommended that a back-up procedure be included to insure that
the
duration of inflation phase is physiologically appropriate. A number of
approaches to
insure a physiologically appropriate duration of the inflation phase were
proposed. These
were in pau derived from a separate application concerned specifically with
methods to
synchronize end of ventilator cycle with end of patient effort that do not
require
knowledge of when said patient efforts end (US provisional Application
60/454,533,
March 14, 2003, from which this application claims priority). The current
application
describes rationale and implementation of said methods in detail and,
additionally,
introduces other approaches described in the March 14, 2003 US Provisional
application
(60/454,533) and not referred to in PCT/CA03/00976. The following is the
rationale and
method for ensuring that the duration of the inflation phase remains within
physiologic
limits.
[0006] In sp.~ntaneously breathing subjects and patients, the duration of the
inspiratory phase (TI) ranges between 25% and 50% of respiratory cycle
duration (TTOT)~
In studies by the inventor using proportional assist ventilation (PAV), with
which the
duration of the ventilator's inflation phase mirrors the patient's own TI, the
ratio of TI to
TTOT (Ti/TTOT ratio) was also found to be between 0.25 and 0.5. Therefore, one
approach
to insure that the duration of the inflation phase is within the physiologic
range is to
constrain the duration of the inflation phase to be between 0.25 and 0.50 of
the total
cycle duration of patient's own efforts (to be distinguished from duration of
ventilator
cycles). Implementation of this procedure requires knowledge of the true
respiratory rate
of the patient (as opposed to ventilator rate). The inventor, in association
with his
students and teclmicians, described a method for visually determining true
patient rate by
identifying visually distinctive patterns in the waveforms of respiratory flow
and airway
pressure (Giannouli et al, American Journal of Respiratory and Critical Care
Medicine,
vol 159, pages 1716-1725, 1999). According to this approach, true patient rate
is the sum
of ventilator rate, the number of ineffective efforts occurring during the
ventilator's
exhalation phase (arrows marked "c", Figure 1) and the number of additional
efforts
occurring during inflations triggered by an earlier effort (arrows marlced
"b", Figure 1).
CA 02519041 2005-09-13
WO 2004/080516 PCT/CA2004/000382
4
In PCT/CA03/00976 ventilator cycles triggered by patient (arrows "a", Figure
1) as well
as ineffective efforts occurring during exhalation (arrows "b", Figure 1) are
to be
automatically detected from the new composite signal generated from the flow,
Paw and
volume signals. In the present invention, I describe another approach for
identifying
ineffective efforts. An approach was described in PCT/CA03/00976 to identify
additional efforts occurring during the inflation phase (arrows "c", Figure
1). This
approach is retained here with minor modifications.
[0007] As indicated in US Provisional application 60/454,533, and also in
PCT/CA03/00976, once the true respiratory rate of patient is known, it becomes
possible
to calculate the real duration of respiratory cycles of the patient (TTOT =
60/respiratory
rate) and determine the range of inflation times consistent with a physiologic
TI/TTOT~
For example, if patient's rate is 30/min, TTOT 1S 2.0 seconds and the
physiological range
for the inflation phase is 0.5-1.0 second reflecting a TI/TTOT range of 0.25
to 0.50. The
desirable duration of the ventilator's inflation phase is then determined by
multiplying
patient TTOT bY a user selected physiologic TI/TTOT ratio or a suitable
default value (e.g.
0.4). The ventilator's inflation phase can then be made to cycle off after
said desirable
duration.
[0008] There are a number of ways by which the duration of the ventilator's
inflation phase can be made to correspond to desirable TI. Qne approach,
discussed in
US Provisional application 60/454,533 and also proposed in PCT/CA03/00976, is
to
terminate the inflation phase at the specified desirable duration following
onset of
inspiratory effort or following the time of ventilator triggering. With this
approach
ventilator inflation varies strictly with average respiratory rate discerned
from a number
of elapsed breaths. There is no provision, therefore, for accommodating breath-
by-breath
changes in duration of inspiratory effort since the desirable duration is
predetermined
before the effort begins (based on an average result obtained from a number of
elapsed
breaths). Another approach, particularly suited for pressure support
ventilation, is to
retain the usual criterion for terminating the inflation phase, namely when
inspiratory
flow reaches a specified threshold, but flow threshold is adjusted to produce
the desired
TI. This would permit breath-by-breath changes in patient's TI to influence
ventilator TI
in current breaths but ventilator TI would, on average, correspond to
desirable TI. This
general approach was proposed in US Provisional application 60/454,533. In
CA 02519041 2005-09-13
WO 2004/080516 PCT/CA2004/000382
PCT/CA03/00976 I proposed that this general approach, be implemented by
measuring
the flow occurring at the desirable TI in a number of elapsed breaths. This
would then
become the flow threshold for terminating the inflation phase in prospective
(i.e. current)
breaths. An alternative approach proposed in US Provisional application
60/454,533 (but
not in PCT/CA03/00976) is to measure actual ventilator TI in a number of
elapsed
breaths. This actual value is compared with desirable TI with the difference
(i.e. actual TI
- desirable TI) representing an error signal that can be used for closed-loop
control of he
flow threshold for cycling off, using any of a number of closed-loop control
approaches.
Alternatively, tile error signal can be the difference between actual TI/TTOT-
(i.e. actual
TI/patient TTOT) and desirable TI/TTOT~
[0009] In my experience, patient's respiratory rate often changes
substantially
from time to time. An essential feature of this invention is, therefore, the
provision for
automatic means to monitor patient respiratory rate and to update the relevant
values
(e.g. desirable TI, actual TI, TI error ..etc) at frequent intervals.
[0010] With current methods of assisted ventilation tidal volume is directly
related to the duration of the inflation phase. Changes in the duration of the
inflation
phase produced by the methods of the current invention are, therefore,
expected to result
in corresponding changes in tidal volume. In another aspect of the current
invention
provision is made to partially or completely offset the resulting changes in
tidal volume
by concomitantly increasing inspiratory flow (in the case of vohune-cycled
ventilation)
or the support pressure (in the case of pressure support or assist/pressure
control
ventilation) when the duration of the inflation phase is decreased, and vice
versa.
SUMMARY OF INVENTION
[0011] In summary, this invention concerns a novel approach for cycling off
ventilators in which the duration of the inspiratory phase is constrained to
be a
physiological fraction (0.25 to 0.50) of the duration of patient breathing
cycles (patient
TTOT)~ One aspect of the invention is the provision of means for ongoing
automatic
determination of patient TTOT from the flow and/or airway pressure signals. In
another
aspect, control of cycling off time in pressure support ventilation is
effected by
measuring the difference between actual and desirable TI and using this error
signal
(difference between actual and desirable TI) to determine the flow threshold
for cycling
off using closed loop control methods. In still another aspect of the
invention, the level of
CA 02519041 2005-09-13
WO 2004/080516 PCT/CA2004/000382
6
delivered flow or pressure during the inflation phase is altered in concert
with changes in
ventilator TI so as to partially or completely offset the changes in tidal
volume that would
result from uncompensated changes in the dtuation of the inflation phase.
[0012] In accordance with one aspect of the present invention, there is
provided a
method for automatic ongoing adjustment of the cycling-off time of ventilator
inflation
phase during assisted ventilation in accordance with true respiratory rate of
a patient,
comprising generating electrical signals) corresponding to rate of gas flow
exchanged
between patient and ventilator (flow) and/or ~to airway pressure (PaW),
determining true
respiratory rate of patient (patient RR) on an ongoing basis from the flow
and/or PaW
signals, estimating current average cycle duration of patient respiratory
efforts (current
patient T~oT) from the patient RR, calculating a current desirable duration of
the
inhalation phase (desirable TI) from the product of current patient TTOT arid
a TI/TTOT
ratio chosen to be in the physiological range (usually 0.25 to 0.50) and
causing ventilator
inflation phase to terminate in accordance with the desirable TI.
[0013] In accordance with a further aspect of the present invention, there is
provided a device for automatic ongoing adjustment of the cycling-off time of
ventilator
inflation phase during assisted ventilation in accordance with true
respiratory rate of the
patient, comprising circuitry for generating electrical signals) corresponding
to the flow
exchanged between patient and ventilator (flow) and/or to airway pressure
(P~~,,), digital
or analog circuitry means for determining true respiratory rate of patient
(patient RR) on
an ongoing basis from the flow and/or PaW signals, digital or analog circuitry
means for
estimating current average cycle duration of patient respiratory efforts
(current patient
TTOT) from the patient RR, digital or analog circuitry means for calculating a
current
desirable duration of the inhalation phase (desirable TI) from the product of
current
patient TTOT and a TI/TTOT ratio chosen to be in the physiological range
(usually 0.25 to
0.50), and means to cycle off ventilator inflation phase in accordance with
the desirable
Ti.
BRIEF DESCRIPTION OF DRAWINGS
[0014] Figure 1 contains tracings showing flow and airway pressure in a
ventilated patient, along with diaphragm pressure to indicate patient's own
efforts.
Arrows marlced "a" denote efforts that triggered ventilator cycles. Arrows
marked "b"
indicate efforts that occurred in the exhalation phase but failed to trigger
the ventilator
CA 02519041 2005-09-13
WO 2004/080516 PCT/CA2004/000382
(ineffective efforts). Arrows marked "c" denote extra efforts that occurred
during the
same ventilator inflation phase triggered by an earlier effort (additional
efforts).
[0015] Figure 2 contains tracings showing a method of detecting ineffective
(IE)
and additional (AE) efforts from the derivative of the flow signal (~flow/Ot).
[0016] Figure 3 is a block diagram of the preferred embodiment of digital
implementation of the present invention.
[0017] Figures 4 to 13 are flow charts of the different functions listed in
the
block diagram of Figure 3.
DESCRIPTION OF THE PREFERRED EMBODIMENT OF THE INVENTION
[0018] A digital implementation of a preferred embodiment of the invention
will
be described here (Figure 3) because the method of the invention is primarily
intended
for incorporation in microprocessor-based ventilators. As such, the method can
be
installed in a free-standing microprocessor that interacts with the
ventilator's control
circuitry or may be fully incorporated in the ventilator's resident computer.
It is
recognized, however, that most of the functions described here can be
implemented
using standard analog circuits.
[0019] The basic hardware. requirements (microprocessor, 1) are a Central
Processing Unit (CPU, 2), Random Access Memory (RAM, 3) and Read Only Memory
(ROM, (4))0
A. INPUTS
[0020] It is assumed here that inputs are in digital form. If some or all are
available only in analog form, an analog to digital converter (not shown) must
be
installed upstream from the CPU to receive and digitize the analog inputs.
[0021] Inputs may vary depending on user preference and independent
availability, within the host ventilator, of signals required for
implementation of the
present invention.. In the preferred embodiment illustrated in Figure 3 (5),
it is assumed
that the device of the present invention will be responsible for determining
patient's
respiratory rate but that signals corresponding to onset and end of ventilator
breaths are
already available from the host ventilator. Modifications to this arrangement
will be
described at the appropriate locations below.
CA 02519041 2005-09-13
WO 2004/080516 PCT/CA2004/000382
8
A 1 Iraputs corr~es~ondin~ to flow 6~ andlof~ airway~ressure ~P~"~:
[0022] Virtually all modern ventilators monitor air flow within the ventilator
circuit and generate a signal corresponding to the rate. of gas flow exchanged
between
patient and ventilator. Fuuthermore, airway pressure (PaW) is routinely
monitored. These
resident signals can be used as inputs to the microprocessor implementing the
current
invention. Alternatively, if the present invention is incorporated in an
external device,
flow and PaW signals can be generated independently by standard techniques
(for
example, as described in PCT/CA03/00976).
[0023] Flow (6) and PaW (7) signals are used to determine a) patient's
respiratory
rate and b) to implement a specific method (custom change function) of closed
loop
control of "cycling-off' flow threshold in the pressure support mode (see
8.2.2.4,
below). In the event patient's respiratory rate is determined from only one of
these
signals (for example, flow only or PaW only) and a different method of closed
loop
control of "cycling-off' flow threshold is used, the other input can be
omitted.
A.2 Mode (8):
[0024] The present invention is primarily intended for use when the ventilator
is
in the pressure support mode (PSV). It can, however, also be used in the
assist/control
modes (A/C). Because implementation of this invention varies with the mode
used (see
FIJ1~TCTIC~I~IS, below), an input reflecting the mode being used is
recommended. ~Jhen
the current invention is incorporated within the ventilator, this input can be
obtained
directly from the ventilator's control system. Alternatively, if the invention
is
incorporated in an external device (for example, as described in
PCT/CA03/00976), the
mode of ventilation is entered by the user.
A.3 Z'on~:
[0025] This is a signal that indicates either the time of onset of patient
inspiratory
effort or the time of onset of a ventilator inflation cycle depending on which
is available.
In practice, if the present invention is implemented, dynamic hyperinflation
is minimized
and there should be little difference between the two times (i.e. trigger
delay should be
minimal). In all current ventilators, the ventilator control system generates
a triggering
signal that initiates a ventilator cycle. This signal can be used as To" (9).
Ventilator
cycles can be triggered either in response to patient effort (patient-
triggered cycles) or by
the ventilator itself if a triggering effort did not occur within a time
specified by a user-
CA 02519041 2005-09-13
WO 2004/080516 PCT/CA2004/000382
9
selected baclc-up rate (ventilator-triggered cycles). Ventilators can
distinguish between
these two types of triggers. If the ventilator's trigger signal is to be used
as To", only the
signals related to patient-triggered cycles are communicated to the
microprocessor
implementing the current invention. If the current invention is implemented in
an
external device (i.e. the ventilator's trigger signals are not available), the
onset of
ventilator cycle can be identified externally from the pressure and/or flow
signal using
any of a number of obvious techniques (for example, as described in
PCT/CA03/00976).
[0026] According to recent developments (PCT/CA03/00976), onset of patient
inspiratory effort can be identified non-invasively. Such devices/methods may
be
incorporated in future ventilators or be used as external devices. In either
case, the onset
of patient inspiratory effort identified by such device/method, or by other
means
available to the ventilator, can be used as To" for the sake of the current
invention.
A.4 T~ 10
[0027] This is a signal that indicates the end of the ventilator's inflation
phase. It
can be obtained directly from the. ventilator's control system (cycling-off
signal) or be
derived independently from the flow (6) and/or PaW (7) signal using any of a
number of
obvious methods (for example, as described in PCT/CA03/00976).
A.5 1?esired T,lTTOT r atio 11
[002] This ratio is preferably a user-selected input that would normally range
between 0.25 and 0.50. Alternatively, it can be replaced by a default value. A
default
value of 0.4 would be appropriate, but other values preferred by manufacturers
may be
used instead. One embodiment is to link the default TI/TTOT to patient
respiratory rate
with high default ratios being used when patient rate is high, and vice versa.
This would
preclude having very short TI when patient rate is very high and very long TI
when rate is
slow. A suggested relation is to use a default ratio of 0.5 when rate is 50
miri 1 and a ratio
of 0.3 when rate is 10 miri 1, with intermediate values for intermediate
rates.
B. FUNCTIONS:
B.1 Real-time unctions:
[0029] The timed interrupt request process (Timed IRQ process, 12) is executed
at suitable intervals (for example, every 5 msec). This collects data from
various inputs
(see Figure 3 for inputs), calculates the time derivative of flow and stores
collected and
CA 02519041 2005-09-13
WO 2004/080516 PCT/CA2004/000382
derived data in memory. This also checks for the times at which To" and Toy,
occur and
stores them in memory.
B.2 Non Real-time functions:
B.2.1 Powef° ON start-up Routivce (13):
[0030] The power on start-up routine clears the memory and enables the
Interrupt
Request (IRQ) Process.
B.2.2 FunctiofZS to Determine Patient. Respirator°y Rate:
[0031] As indicated in the Background section above, patient's respiratory
rate
may be quite different from ventilator's rate (for example, Figure 1). It is
the patient's
rate that needs to be kno~cm in order to set the ventilator's inflation time
to be in the
physiological range. Furthermore, since patient's rate may vary considerably
from time
to time, it is necessary to monitor patient's rate on an ongoing basis. In
this preferred
embodiment (1), I have developed an automatic continuous digital approach
based on the
visual (identified by eye) approach described by Giarmouli et al (Am. J.
Respir. Crit.
Care Ieiled. 159: 1716-1725, 1999). This approach is described here only to
illustrate that
patient rate can be monitored automatically on an ongoing basis using simple
processing
of universally available signals (flow and/or PaW). There are a number of
other
approaches that can be employed to achieve the same end. For example, a signal
combining flow, volume and P~~,,, can be generated from which true respiratory
rate can
be estimated, as described in PCT/CA03/00976. Alternatively, although such
methods
have not yet been specifically described, it may be possible to obtain
patient's rate by
spectral analysis of the flow and/or PaW signal (loolcing for the frequency of
significant
power peaks in the respiratory rate range (10 to 50 miri 1)) or by other
mathematical
analyses of these signals. Furthermore, it is theoretically possible to
estimate patient rate
from signals other than flow and/or Pav,,, for example from changes in
electrical
impedance or inductance of the chest wall, from strain gauges placed on the
chest wall,
or from monitoring electrical activity of respiratory muscles. Other methods
may be
developed in the future to estimate patient's rate. The specific way by which
patient rate
is continuously monitored is not the subject of the current patent
application. Where
methods other than the one described here are used to continuously monitor
patient's rate
the result of such determination can be inputted directly in the
microprocessor of the
present invention.
CA 02519041 2005-09-13
WO 2004/080516 PCT/CA2004/000382
11
[0032] In the approach used in the preferred embodiment (1), patient's rate is
estimated from the sum of a) patient triggered ventilator cycles ("a" arrows,
Figure 1), b)
respiratory efforts occurring during the ventilator's exhalation phase that
did not trigger
ventilator cycles (ineffective efforts, "b" arrows, Figure 1), and c)
additional inspiratory
efforts occurring during the ventilator's inflation phase (additional efforts,
"c" arrows,
Figure 1). Patient triggered ventilator cycles (To") are identified by the IRQ
process (12)
and stored in memory. Separate functions are included for detection of
ineffective efforts
(16) and additional efforts (18). A fourth function (20) sums the 3 results
over specified
elapsed intervals to obtain patient's rate/minute and average patient cycle
duration
(patient TTOT). Before implementing these functions it is desirable to
determine the time
of occurrence of peals inspiratory and expiratory flow.
B.2.2.1 Peak inspiratory flow function (14):
[0033] This function determines the magnitude and time of occurrence of
maxunum flow during the inflation phase in each elapsed ventilator cycle. It
searches the
flow signal between T~" and To~° of the preceding inflation phase
looking for the highest
value and stores the actual flow value and its time.
B.2.2.2 Peals expiratory flow function (15):
[0034] This function determines the magnitude and time of occurrence of peak
expiratory flow during the expiratory phase in each elapsed ventilator cycle.
It searches
the flow signal between To~ and the next To" looking for the lowest value and
stores the
actual flow value and its time.
8.2.2.3 Ineffective efforts function (16~: '
[0035] This function searches the flow signal of each elapsed exhalation phase
in
the interval between peals expiratory flow (15) and next To" (9) for evidence
of efforts
that did not trigger a ventilator cycle. Figure 2 shows the principle of the
preferred
approach described herein. Figure 2 shows tracings of airway pressure (PaW),
flow, rate
of change in flow (~flow/~t) and diaphragm pressure. An ineffective effort
occurred at
the arrow (arrow$ marked IE). In the passive state, once expiratory flow
reaches its peak
value, it should progressively decrease (i.e. flow becomes less negative)
until the next
ventilator cycle. This should result in a continuously positive Oflow/Ot
signal. When an
inspiratory effort occurs, expiratory flow initially moves toward zero at a
faster rate (and
flow may become transiently. positive, Figure 2). If the effort ends without
triggering the
CA 02519041 2005-09-13
WO 2004/080516 PCT/CA2004/000382
12
ventilator, as illustrated in figure 2 (arrow marked IE), expiratory flow
increases again
(i.e. flow becomes more negative). After reaching a maximum value, expiratory
flow
begins decreasing again toward zero and continues to do so witil the next
inspiratory
effort. This sequence results in a characteristic pattern in the dflow/~t
signal. The signal
rises at a faster rate than before with the onset of effort. Then the signal
declines
transiently into the negative range (point "a", Figure 2) and finally crosses
zero again
into positive range (point "b", Figure 2). During passive expiration, ~flow/~t
should not
become negative except very briefly in association with transient noise, .such
as
secretions or tube vibrations. Such artifactual negative transients have much
shorter
durations than ineffective efforts (see arrows marked "noise" in Figure 2).
Accordingly,
identification of ineffective efforts in this preferred embodiment of the
invention is based
on detecting negative transients in the Oflow/Ot signal having a duration that
is greater
than that of the usual noise. From experience, I found that a negative
transient duration
of approximately 0.15 second provides a good separation between noise and
ineffective
efforts. Two other optional conditions are implemented in the preferred
embodiment that,
based on experience, minimize false identification of ineffective efforts: a)
requiring that
flow at the onset of the negative transient in Oflow/~t (point "a", Figure 2)
be higher than
flow at the end of the transient (point "b", Figure 2) by a specified amount.
In Figure 2,
the difference in flow between the two points was 0.4~ 1/second. A minimum
difference of
0.075 1/second is recommended. I~Tote that a second negative transient related
to noise did
not meet this criterion, (b) requiring that PaW at the onset of the negative
transient in
aflow/~t (point "a", Figure 2) be lower than PaW at the end of the transient
(point "b",
Figure 2). When an ineffective effou is identified, its time is stored in
memory. From
this, the number of ineffective efforts per minute can be calculated and
displayed (17).
[0036] During the exhalation phase PaW is a fwction of expiratory flow and
resistance of the exhalation tube/valve combination. Accordingly, changes in
flow
produce corresponding changes in PaW; when expiratory flow decreases (i.e.
becomes less
negative) PaW also decreases (i.e. becomes less positive) (for example, note
that PaW
during exhalation is a mirror image of flow, Figure 2). For this reason,
detection of
ineffective efforts can be made from the PaW signal using a similar approach
to that
described above for flow. PaW is differentiated (OPaW/Ot). A positive
transient in ~PaW/Ot
of sufficient width, and associated with a threshold increase in Pa,~,, would
indicate an
CA 02519041 2005-09-13
WO 2004/080516 PCT/CA2004/000382
13
ineffective effort. In my experience, however, use of flow signal is
preferable since
changes in PaW associated with ineffective efforts can be quite subtle,
particularly when
exhalation line resistance is low.
B.2.2.4 Additional efforts function (18):
[0037] This function detects additional efforts occurring during the inflation
phase and is applicable only in pressure-cycled modes (for example, PSV,
pressure
control). The principles employed are similar to those for ineffective effort
detection
(16). In pressure-cycled modes, once inspiratory flow reaches its peak value
it should
progressively decline towards zero. A secondary increase of sufficient
duration occurring
during the same inflation invariably indicates an additional effort (Giannouli
et al, Am. J.
Respir. Crit. Care Med. 159: 1716-1725, 1999). This pattern results in a
characteristic
change in Oflow/~t (Figure 2). Oflow/~t is negative early in the inflation
phase, beyond
peak flow, as expected. However, instead of remaining negative until the end
of the
phase, it becomes positive (point "c", Figure 2) for a while before becoming
negative
again (point "d", Figure 2). In the preferred embodiment, additional efforts
axe identified
if there are positive transients in ~flow/~t between the time of peak
inspiratory flow and
Toff~ To eliminate artifactual positive transients related to non-specific
noise, two
additional optional requirements are specified: a) Flow at the end of the
positive
~flow/~t transient (point "d", Figure 2) should be higher than flow at the
onset of the
transient (point "°999 Figure 2) by a specified amount. In the
illustrated example it was
higher by 0.2 1/second. A minimum required value of 0.05 1/sec is suggested.
b) The
difference between instantaneous flow beyond point "c" and flow at point "c"
is
integrated between point "c" and Toy (shaded area). The integral should exceed
a
specified value. A value of 0.03 1 is suggested. When an additional effort is
identified its
time is stored in memory. From this, the number of additional efforts per
minute can be
calculated and displayed (19).
B.2.2.5 Patient rate/TTOT function (20):
[0038] This function simply adds all events identified as To" (by the IRQ
process
(12)), ineffective efforts (identified by ineffective efforts function (16))
and additional
efforts (identified by additional efforts function (18)) stored in memory over
a specified
period. The specified period may be 1.0 minute, or any other interval selected
by
manufacturer or user. A preferred approach (20) is to count all events
identified over an
CA 02519041 2005-09-13
WO 2004/080516 PCT/CA2004/000382
14
interval corresponding to a specified number of ventilator cycles (e.g. 10
ventilator
cycles). From this, average patient rate is calculated from [(number of events
*60)/(period covered by specified number of ventilator cycles)]. Patient TTO~r
is then
calculated from [60/patient rate]. Patient rate and average TTOT values are
updated at
suitable intervals, preferably after each elapsed ventilator cycle. Patient
rate and/or
average patient TTOT may be displayed (21).
[0039] At times in the assist/control modes the patient is apneic, there being
no
respiratory efforts, effective or not. Likewise, at times in the pressure
support mode, the
patient develops recurrent periods of central apnea during which there are no
efforts.
Inclusion of these apneic periods in the calculation of average patient
rate/TTOT would
result in substantial underestimation in the respiratory rate of the patient
when he/she is
making respiratory efforts. By extension, this error would result in
overestimation of
patient TTOT when patient is making respiratory efforts. A number of
approaches can be
implemented to avoid this error. For example, periods in excess of a specified
duration
(for example, 10 seconds) during which there were no efforts of any kind (i.e.
effective,
ineffective or additional) are excluded from analysis. In another approach,
the intervals
between successive efforts (whether they triggered the ventilator (as
indicated by To"),
were ineffective or additional) are tabulated. Intervals exceeding the normal
variance of
this variable (for example, >mean + 2 standard deviations) are excluded from
analysis of
patient respiratory rate/TTOT.
B.2.3 T_Functions to calculate ventilator TT_error:
B.2.3.1 Desirable TI function (22):
[0040] This calculates the ventilator inflation phase duration that would
result in
a physiologically desirable TI/TTOT. Desirable TI is . estimated from patient
current
average TTOT value (20) and the desirable TI/TTOT ratio, with the latter being
either a user
selected value (11) or a default value (see A.5). Because patient current
average TTOT
value (20) is updated continuously at suitable intervals, desirable TI is
automatically
updated at the same intervals, preferably after each elapsed ventilator cycle.
Desirable TI
is communicated to the ventilator (23) for use to adjust ventilator TI or for
display to the
user. °It is recommended that a minimum (e.g. 0.5 second) and a maximum
(e.g. 2.5
second) be assigned to this value.
CA 02519041 2005-09-13
WO 2004/080516 PCT/CA2004/000382
B.2.3.2 Actual TI function (24):
[0041] This calculates the average duration of inflation phase of a suitable
number of elapsed ventilator cycles. Ventilator TI is calculated for each
elapsed breath
from the difference between To" and Toy of that breath. Results of individual
breaths are
stored in a buffer. Actual TI is the average of such values over a suitable
number of
elapsed breaths. This number should ideally be the same as the number used to
estimate
average patient TTOT (20). In the preferred embodiment, the number is 10
breaths. Actual
TI is updated at suitable intervals, preferably after each elapsed ventilator
cycle.
B.2.3.3 TI error function (25):
[0042] This function calculates the average difference between actual (24) and
desirable (22) TI. An alternate format is to calculate the difference between
actual and
desired TI/TTOT with the former calculated from actual TI (24)/patient TTOT
(20). TI error
is updated at suitable intervals, preferably after each elapsed ventilator
cycle. TI error is
c~mmunicated to the ventilator (26) for use to adjust ventilator TI or for
display to the
user.
B 2 4 Functions for the control of ventilator cycling-off time:
[0043] There are a number of ways by which the output of the aforementioned
functions can be used to continuously adjust ventilator cycling in order to
obtain a
desirable TI~TTOT~ The choice would clearly be up to the ventilator
manufacturer. ~nly a
few examples of possible approaches are discussed here.
[0044] In the assist/control modes, the desirable TI output (23) can be used
to
continuously update the programmed ventilator cycle duration (ventilator TI)
in the
ventilator's control circuitry. In this fashion, ventilator TI changes
continuously and
appropriately in response to spontaneous changes in patient rate, which are
quite
frequent. Because in the volume cycled mode tidal volume is directly related
to
ventilator TI, changes in the latter induced by the current invention will
necessarily result
in similar changes in delivered tidal volume. This may be desirable in some
cases in that
an increase in patient rate, with a consequent reduction in ventilator TI,
would result in a
reduction in tidal volume, maintaining ventilation approximately the same and
avoiding
over-ventilation. Some users, however, may prefer to partially or completely
avoid
changes in tidal volume in response to changes in respiratory rate. In this
case, an option
may be provided whereby a decrease in ventilator TI is automatically offset by
CA 02519041 2005-09-13
WO 2004/080516 PCT/CA2004/000382
16
appropriate and simultaneous increase in flow rate, and vice versa. The
adjustment in
flow rate can be designed to completely or only partially offset the change in
tidal
volume resulting from the change in ventilator TI. For example, the
simultaneous
change in flow rate can be set to be a fraction of the % change in ventilator
TI with said
fraction being user-specified or a default value (for example, 50%, 60% etc).
[0045] Frequently, the duration of patient inspiratory effort varies from
breath to
breath. In this case, the use of a ventilator TI, corresponding to an
estimated average
patient TI, may result in the ventilator cycling off before the end of
inspiratory effort in '
some breaths. In another aspect of this invention, the implemented ventilator
TI
corresponds to the desirable TI generated by the current invention (23) plus a
specified
amount (for example, 0.2 sec) or specified fraction (for example, 10% etc). In
this
fashion, the frequency of cycles in which the ventilator inflation phase
terminates before
patient effort is reduced. The increase, over desirable TI, to be implemented
may be a
user input or a default value. It is readily possible to identify cycles in
which the
ventilator breath terminated prematurely from observing the flow pattern on
the
ventilator screen. The adjustable incremental amount to be used can,
accordingly, be set
by the user to minimize the occurrence of such events.
[0046] The same approach can be used to cycle off the ventilator in the
pressure
support mode. Thus, instead of the conventional flow-based cycling-off
mechanism, the
ventilator can be made to cycle off at the desirable TI identified by the
present invention
(23). An option to alter the pressure level simultaneously, as desirable TI
changes, may
also be provided to partially or completely offset the changes in tidal volume
resulting
from the different TI. Additionally, an option to increase the desirable TI
(23) by a
specified amount or percent (as in the case of volume cycled ventilation
described above)
before implementation can be provided to minimize instances of ventilator
cycle
terminating before patient effort.
[0047] An alternative and preferred approach, however, is to retain the flow-
based cycling-off mechanism and utilize the results of the current invention
to
continuously adjust the flow threshold for cycling off. By retaining the flow-
based
cycling-off mechanism, spontaneous changes in duration of patient inspiratory
effort
continue to influence ventilator TI, since a longer patient TI will delay the
point at which
CA 02519041 2005-09-13
WO 2004/080516 PCT/CA2004/000382
17
the flow.threshold is reached, and vice versa. However, the present invention
can provide
the flow threshold that will result in a desirable TI/TTOT~
[0048] There are several approaches by which the results of the current
invention
can be used to continuously adjust the cycling-off flow threshold in order to
obtain a
desirable TI/TTOT ratio (closed loop control of flow threshold). Four such
functions are
described here: All functions utilize the TI error signal (26) to effect
changes in flow
threshold. The ventilator manufacturer may select one of them or utilize some
other
control paradigm of his choice. It is recognized that because of the
spontaneous breath-
by-breath differences in patient TTOT and TI, feedback should not operate on a
breath-by-
breath basis. Rather, the average error calculated over a number of elapsed
breaths
should be used, as done here (25). Furthermore, because it is not clinically
critical to
rapidly adjust the flow threshold, feedback with slow response is preferred to
avoid
instability.
8.2.4.1 Fixed change function (27):
[0049] In this approach, a fixed increment or decrement in flow threshold is
implemented depending on magnitude and polarity of the TI error signal (25).
For
example, if -0.1<TI error <0.1, no change is implemented. If TI error>0.1
second, flow
threshold is increased by a fixed amount and if it is <-0.1 second, threshold
is decreased
by a fixed amount. A value of 0.05 1/second is used in the illustrated
embodiment (27)
but other values can obviously be used. The larger the step change, the faster
the
response but the more likely it is for the system to overcorrect and have an
oscillatory
response. Because the full effect of the implemented change on TI error will
not become
apparent until a number of breaths have elapsed (since the TI error (25) is
based on
average of a number of breaths) step changes in flow threshold are computed
only every
"n" breaths, where "n" is the number of breaths used in calculating TI error
(25) (see
Main Program Loop Function (33). In the preferred embodiment, I have used n=10
(33).
B.2.4.2 Custom change function (28):
[0050] Here the recommended change in flow threshold is based on the average
rate of change in flow in the terminal part of the inflation phase in a
suitable number of
elapsed breaths. To determine the flow VS time slope, flow at Toy and at a
suitable
interval before Tuff is measured in a suitable number of elapsed breaths. The
difference
between average flow at the two points divided by the interval between the two
points of
CA 02519041 2005-09-13
WO 2004/080516 PCT/CA2004/000382
18
measurement provides the relevant slope. In the preferred embodiment (28), I
have used
an interval of 0.2 second and 10 elapsed breaths. The recommended change in
flow
threshold is then calculated from TI error * calculated slope. Because the
flow VS time
relation in pressure support ventilation is usually not linear, the calculated
slope over an
arbitrarily selected interval may not be representative of the slope before or
after this
interval. For this reason, it is prudent not to apply the recommended change
(from TI
error * calculated slope) all at once. In the preferred embodiment (28), the
recommended
change is multiplied by an attenuation factor (e.g. 0.5). This will slow the
correction
somewhat but will improve stability. Other attenuation factors may be used
depending on
manufacturer preference. An alternative approach (not illustrated) is to fit
the flow VS
time relation in elapsed breaths with a non-linear function and calculate the
required
change in flow threshold from TI error (25) and said non-linear function.
[0051] As in the case of the fixed change fiulction (27), because the full
effect of
the implemented change on TI error will not become apparent until a number of
breaths
have elapsed (since the TI error (25) is based on average of a number of
breaths) custom
changes in flow threshold are computed only every ."n" breaths where "n" is
the number
of breaths used in calculating TI error (25). (see Main Program Loop Function
(33).
8.2.4.3 Hybrid change function (297:
[0052] Although the custom change function (28) should, on average, result in
faster correction than the fixed change function (27), at times flow is quite
flat over short
intervals near the end of the inflation phase. In this case, the custom
function (28) would
result in very small recommended changes. In the hybrid function, the changes
recommended by both the fixed (27) and custom (28) functions are calculated
and the
one with the higher absolute value is used. The hybrid function is executed at
the same
time the other two functions are executed (i.e. every "n" breaths). The hybrid
function is
preferred for general use and its output is the one utilized in the preferred
embodiment
(see Timed IRQ process (12)). However, the result of the fixed (27) or custom
(28)
functions may be the preferred output under some circumstances.
[0053] Based on manufacturer preference, the recommended change in flow
threshold, as derived from any of the above three functions (27-29), can be
outputted
(30) in 1/second or as % peals inspiratory flow. For the salve of the latter
expression, the
recommended change is divided by peals inspiratory flow obtained from the
peals
CA 02519041 2005-09-13
WO 2004/080516 PCT/CA2004/000382
19
inspiratory flow function (14). The recommended change can then be added to or
subtracted from the flow threshold value currently stored in the ventilator's
control
system. Alternatively, the recommended change can be added to the average
value of
flow at T°~ (obtained from the custom change function (28)), which is
an approximation
of the current flow threshold, and the result is expressed (30) as recommended
flow
threshold, as opposed to recommended change in flow threshold. This, again,
can be
expressed either in 1/second or as % peak flow according to manufacturer
preference.
B 2 4 4 Proportional/Inte~rallDerivative (PID) function (31):
[0054] Here, rather than computing a recommended change in flow threshold, as
done by the above three functions (27 to 29), flow threshold is directly
controlled by the
TI error function (25) using the standard PID approach. A composite error
signal with
three components is generated (31) from the TI error signal (26). One
component is
proportional to TI error (26), the other is proportional to the integral of TI
error (26) and
the third is propot-tional to the derivative of TI error (26). The gains of
the individual
components are adjusted for optimal performance. This composite error signal
(32) is
then used for ongoing adjustment of flow threshold for cycling off.
[0055] With all methods described above for closed-loop control of fl~w
threshold (27,28,29,31) it is recommended that the ventilator manufacturer
place a limit
on how low the flow threshold for cycling off can go. This will prevent
instances of
cycling off being prevented because of positive offsets in the ventilator's
flow signal.
The magnitude of this set minimum value will clearly depend on the quality of
the
ventilator's flow signal (i.e. tendency to drift ...etc).
B 2 5 Main~rogram loop function (33):
[0056] This function is initiated with every T°ff. It then executes the
various
functions in the appropriate order at a fixed point in the breath cycle to
guarantee the
availability of all necessary variables.
[0057] Detailed flowcharts of the various functions used in the preferred
embodiment are shown in Figures 4 to 13.
SUMMARY OF DISCLOSURE
[0058] In surmnary of this disclosure, the present invention provides method
and
apparatus for automatic ongoing adjustment of the cycling-off time of
ventilator inflation
CA 02519041 2005-09-13
WO 2004/080516 PCT/CA2004/000382
phase during assisted ventilation in accordance with true respiratory rate of
the patient.
Modifications are possible within the scope of this invention.
